Synthesis, luminescence properties and electronic structure of Tb³⁺-doped Y_4−xSiAlO₈N:xTb³⁺ – a novel green phosphor with high thermal stability for white LEDs

Abstract

A series of novel green Y_4−xSiAlO₈N:xTb³⁺ phosphors have been prepared by a high temperature solid state reaction. The phase formation and structural properties were analyzed by X-ray powder diffraction. The XRD results and SEM images show that the Y³⁺ can be substituted by Tb³⁺ for the max content of x = 1.5 and the most suitable sintering temperature is about 1500 °C. The PLE spectra of Y_4−xSiAlO₈N:xTb³⁺ phosphors exhibit a wide excitation band ranging from 200 to 500 nm, which matches well with the characteristic emission of n-UV chips. Under excitation of 380 nm, the phosphor shows four intense emission bands with emission peaks at 488 nm, 543 nm, 585 nm and 624 nm, respectively. These emissions were attributed to the characteristic ⁵D₄ → ⁷F_J (J = 6, 5, 4, 3) transitions of Tb³⁺ ions. The optimum doping concentration of Tb³⁺ was found to be x = 2.0 which indicated that the concentration quenching effect in the Y_4−xSiAlO₈N:xTb³⁺ phosphor is very weak. The critical distance for the Tb³⁺ ions calculated by the concentration quenching is 3.64 Å. The detailed nonradiative energy transfer mechanism between Tb³⁺ ions is confirmed to be via a dipole–dipole interaction by the fluorescence decay analysis. Furthermore, with the introduction of Tb³⁺ ions, the reflectance spectra shows an obvious increment of reflectance in the region of 200–250 nm, which is due to the variation of electronic structure in the conductance band derived from Y³⁺ ions. Finally, the excellent thermal stability of the phosphors was demonstrated by the temperature dependence of the PL spectra. Compared to the initial intensity at room temperature (293.0 K), the relative PL intensity maintains high value of 96.5% at 475.2 K. The detailed thermal quenching behavior has also been interpreted.

---

1. Introduction

In recent years, white light-emitting diodes (LEDs) have attracted substantial attention as excellent candidates to replace incandescent light bulbs owing to their extraordinary luminous efficiency, low power consumption, reliability, and environmental friendliness.^1–3 White LEDs are becoming more and more important to the world's energy saving as they significantly reduce global power requirements and the use of fossil fuels.^4–6 Presently, the most efficient way to create white light is to use blue-LED chips in combination with yellow-emitting phosphors, which partially convert the primary photons emitted by blue-LED chips to longer wavelengths.^1,5,7 However, this combination yields only a poor colour rendering index (CRI; Ra = 70 to 80) and a high correlated colour temperature (∼7750 K), namely cool white light, because of the lack of red and green components in its spectrum.^1,2,8 In order to achieve “warmer” white light with high Ra, a multi-phosphor approach is necessary. Accordingly, an alternative method for white LEDs is the combination of n-UV LED chips (350 to ∼420 nm) and tricolour phosphors, which have been greatly developed recently.^2,8,9 There are several benefits in such a device, including high colour rendering index and stable light colour that are almost independent of the changed current.⁸ However, the disadvantages of such a device are low efficiency of the phosphors and the need for complex coating technology.⁴ Hence, finding novel phosphor materials with superior luminescence properties for this type device is one of the most important and urgent challenge to be met by advanced science and high technology.^3,7

As an indispensible component in white LEDs devices, high efficient phosphors must typically possess high conversion efficiency, good thermal stability, and strong absorption in the n-UV to blue region (370–460 nm).² However, these outstanding properties are usually restricted by concentration quenching and thermal quenching effects in most phosphors. For the concentration quenching, due to the energy interaction between doped luminescent centres, the emission intensity will reach maximum and then decrease with the increasing of luminescent centres quantity. In general, the optimum doping concentration of rare-earth luminescent centres are maintained in low values with about 1–20 mol%. On the other hand, the thermal stability of phosphors is one of the most important issues to be considered, which is considerable influence chromaticity and brightness of white light output.^10–13 Recently, Si₃N₄-based chemical compounds such as nitridosilicates and nitridoalumosilicates have been extensively investigated as good candidates for rare-earth doped phosphors because of their excellent optical properties upon excitation of near ultraviolet (n-UV) or blue light.^3,14 The rare-earth ions such as Eu²⁺, Eu³⁺, Ce³⁺, and Tb³⁺ are the most frequently used activators for wavelength converting due to the high efficiency and abundant emissions based on 4f → 4f or 5d → 4f transitions.⁶ In this paper, a novel green phosphor Y_4−xSiAlO₈N:xTb³⁺ with superior luminescence properties and high thermal stability for n-UV white LEDs has been reported and investigated in detail.

2. Experimental

2.1 Synthesis of samples

A series of Tb³⁺ activated Y_4−xSiAlO₈N:xTb³⁺ phosphors were prepared by a high temperature solid-state reaction. Y₂O₃ (99.99%), SiO₂ (99.99%), Al₂O₃ (99.99%), α-Si₃N₄ (99.9%) and Tb₄O₇ (99.995%) were used as starting materials. They were weighed in stoichiometric proportions and mixed in an agate mortar. Then the mixture were placed into alumina crucible and sintered at 1500 °C or 1550 °C for 6 h under a reducing atmosphere (5%H₂/95%N₂ mixture). The products were then obtained by cooling down to room temperature in the furnace, grinding, and pulverizing for further measurements.

2.2 Characterization

The phase purity of obtained samples were checked by X-ray powder diffraction (XRD) analysis using a D2 PHASER X-ray powder diffractometer (Bruker, Germany), with CuKα radiation operated at 30 kV and 10 mA. The 2θ ranges of all data sets were from 10° to 80° with a step size of 0.02°. The scanning electron microscopy (SEM) micrographs were obtained by using a SU8010 field-emission scanning electron microscope (Hitachi, Japan). The photoluminescence (PL) and photoluminescence excitation (PLE) spectra were measured at room temperature by FL3-211-P spectrofluorometer (Horiba Jobin Yvon, USA) equipped with a 450 W Xe light source. All of the measurement procedures of PL and PLE spectra are the same. The PL decay-curves were measured on the same spectrophotometer, which was combined with a Time-Correlated Single-Photon Counting (TCSPC) system. The temperature-dependent emission spectra were recorded using a FL3-211 P spectrofluorometer equipped with a heating apparatus (THMS-600) and optical fibers. The diffuse reflectance spectra of these samples were measured by an ultraviolet-visible-near infrared spectrophotometer (UV3600) using BaSO₄ as a reference in the range of 200–800 nm. The density of states (DOS) were calculated by first-principles calculations and performed in the density functional theory (DFT) framework using the CASTEP (Cambridge serial total Energy package) module of Materials Studio 8.0. The exchange–correlation effects were treated within the generalized gradient approximation (GGA) with the PBE functional.

3. Results and discussion

3.1 Phase identification and crystal structure

Fig. 1 shows X-ray diffraction (XRD) patterns of Y_4−xSiAlO₈N:xTb³⁺ phosphors which were sintered at 1500 °C. The substitution content (x) of Tb³⁺ for Y³⁺ was varied from 0 to 0.5 in Fig. 1(a), and from 1.0 to 2.5 in Fig. 1(b) respectively. As shown in Fig. 1(a), all of the observed XRD peaks match well with the standard card of Y₄SiAlO₈N phase (JCPDS #48-1630). These results indicate that the monophasic Y₄SiAlO₈N-based phosphors have been synthesized successfully. The crystalline phases could not be changed when the substitution content (x) of Tb³⁺ for Y³⁺ was varied from 0 to 0.5. However, with further substitution of Tb³⁺ for Y³⁺ from 1.0 to 2.5, the X-ray diffraction peaks weakened gradually. As shown in Fig. 1(b), when the substitution content (x) increased from 1.0 to 1.5, the diffraction intensity of (201) and (431) lattice planes decreased obviously. Fortunately, there are hardly any other impure diffraction peaks. It means that the crystal structure of Y_2.5SiAlO₈N:1.5Tb³⁺ was almost the same with Y₄SiAlO₈N phase. When the substitution content (x) increased from 2.0 to 2.5, the diffraction peaks appeared complex. Except for Y₄SiAlO₈N phase, there are some other unknown diffraction peaks. These results indicate that the Y³⁺ can be substituted by Tb³⁺ for the max content of x = 1.5 in the Y₄SiAlO₈N host.

Fig. 1 XRD patterns of Y_4−xSiAlO₈N:xTb³⁺ phosphors with (a) 0 ≤ x ≤ 0.5; and (b) 1.0 ≤ x ≤ 2.5.

Fig. 2 shows typical scanning electron microscopy (SEM) images of as-prepared Y₄SiAlO₈N hosts sintered at different temperature. As shown in Fig. 2(a), when sintered at 1500 °C, the surface morphology of obtained powder particles is very smooth and well-dispersed. It indicates that the degree of crystallinity is very well. The particle shapes are rod like with diameters range from 2 μm to 3 μm and length range from 8 μm to 10 μm. This dimension is very suitable for application in white LEDs package. In order to increase the crystallization properties, we attempted to increase the sintering temperature from 1500 °C to 1550 °C. Unfortunately, we haven't obtained the satisfactory results. As shown in Fig. 2(b), the morphology of obtained powder particle is irregular and broken. This is because the powder particles were sintered too hard under higher temperature. Then the powder particles connected with each other. When we grinded the powder, the well crystallized particles were broken by the mortar simultaneously. As a result, the surface of powder particles show lots of crystal defects. These defects will cause lower luminescent efficiency and even much more scattering loss in the white LEDs. Accordingly, the most suitable sintering temperature is about 1500 °C.

Fig. 2 SEM images of Y₄SiAlO₈N host sintered at (a) 1500 °C; and (b) 1550 °C.

Fig. 3 shows crystal structure of Y₄SiAlO₈N unit cell and coordination spheres of Y1, Y2, Y3 and Y4 sites viewed toward [111] axis. The structure of Y₄SiAlO₈N has four formula units in a unit cell with unit cell volume V = 805.83 ų. The detailed description of the crystal structure can refer to our previous studies.¹⁵ For special attention, the crystal structure of Y₄SiAlO₈N host has four different Y³⁺ sites: Y1 is seven-coordinated with O and one-coordinated with N (coordination number CN = 8); Y2 is seven-coordinated with O (CN = 7); Y3 is six-coordinated with O (CN = 6); Y4 is six-coordinated with O and one-coordinated with N (CN = 7). Consequently, the Tb³⁺ will possibly substitute all of four different Y³⁺ sites in the host.

Fig. 3 Crystal structure of Y₄SiAlO₈N unit cell and coordination spheres of Y1, Y2, Y3 and Y4 sites.

According to first-principles calculations, we rebuilt Y₃SiAlO₈N:Tb crystal structures in which the Tb³⁺ substituted Y1, Y2, Y3 and Y4 respectively. As a result of geometry optimization of crystal structures, the final energy of Y₄SiAlO₈N and Y₃SiAlO₈N:Tb systems are shown in Table 1. The final energy of Y₄SiAlO₈N system is −16 357.62 eV. This value is much more than that of Y₃SiAlO₈N:Tb systems. When one Y atom was substituted by one Tb atom, the final energy of Y₃SiAlO₈N:Tb is −30 225.16 eV, −30 224.03 eV, −30 224.45 eV and −30 225.53 eV for the substitution of Y1, Y2, Y3 and Y4 respectively. In the crystal structure, less system energy means much more stabilization.¹⁶ Hence, the Tb³⁺ ion can be well-incorporated into the Y₄SiAlO₈N crystal lattice within a certain doping concentration, and the substitution probability of Tb for different Y sites is Y4 > Y1 > Y3 > Y2.

Table 1 Final energy of Y₄SiAlO₈N and Y₃SiAlO₈N:Tb systems

No.	Substitution	Final energy (eV)
1	None	−16 357.62497248
2	Tb → Y1	−30 225.16128393
3	Tb → Y2	−30 224.02935319
4	Tb → Y3	−30 224.44562182
5	Tb → Y4	−30 225.52576334

---

3.2 Photoluminescence properties

Fig. 4 shows photoluminescence excitation (PLE) and photoluminescence emission (PL) spectra of Y₄SiAlO₈N host without Tb³⁺ co-doped. Upon 290 nm excitation, the PL spectrum exhibits a broad emission range from 350 nm to 700 nm, which consists of two emission bands peaked at 410 nm and 543 nm respectively. The emission intensity of 543 nm is much stronger than that of 410 nm. When monitored at 543 nm, the PLE spectrum shows a narrow intense excitation band at 290 nm and a broad weak excitation band at about 370 nm due to the energy absorption from valence band to conduction band in the host. These two excitation bands compose a broad excitation band range from 250 nm to 400 nm, which match well with the characteristic excitation band of Tb³⁺ ions. Therefore, the energy transfer between host and Tb³⁺ will occur easily and promote the emission of Tb³⁺ ions in the Y₄SiAlO₈N host.

Fig. 4 PLE and PL spectra of Y₄SiAlO₈N host.

As an important effective activator, the Tb³⁺ ion has been investigated in detail in many hosts.^12,17,18 Fig. 5 represents photoluminescence (PL) emission spectra of Y_4−xSiAlO₈N:xTb³⁺ phosphors with varied Tb³⁺ concentrations. Upon 380 nm excitation, the emission spectra of Y_4−xSiAlO₈N:xTb³⁺ consist of four intense emission peaks in the region of green and red, which are ascribed to characteristic ⁵D₄ → ⁷F_J (J = 6, 5, 4, 3) transitions of Tb³⁺ ions.^18–20 The emission peaks were located at 488 nm (⁵D₄ → ⁷F₆), 543 nm (⁵D₄ → ⁷F₅), 585 nm (⁵D₄ → ⁷F₄) and 624 nm (⁵D₄ → ⁷F₃) respectively. Except for ⁵D₄ → ⁷F_J (J = 6, 5, 4, 3) transitions, the inset of Fig. 5 also show the magnified ⁵D₃ → ⁷F_J (J = 5, 4) transitions of Tb³⁺ ions. The emission peaks were located at 418 nm (⁵D₃ → ⁷F₅), 437 nm (⁵D₃ → ⁷F₄) respectively. Relative to the ⁵D₄ level transitions, the intensity of ⁵D₃ level are much weaker. The upper of Fig. 5 indicates the photograph of Y_4−xSiAlO₈N:xTb³⁺ phosphors which were illuminated by 375 nm n-UV light except x = 0.00. The first sample (x = 0.00) was illuminated by white light due to weak light emission upon 375 nm excitation. It can be seen that the phosphors emitted distinct green light with different Tb³⁺ concentration. When x = 1.50 and 2.00, these two samples show the most brightness compared to others. The brightness variation of these samples is consistent with PL spectra.

Fig. 5 PL emission spectra of Y_4−xSiAlO₈N:xTb³⁺ phosphors with varied Tb³⁺ ion concentrations.

Fig. 6 shows dependence of emission intensities of ⁵D₃ (437 nm) and ⁵D₄ (543 nm) levels on Tb³⁺ concentration in Y_4−xSiAlO₈N:xTb³⁺ phosphors. With the increase of Tb³⁺ concentration, the emission intensity of ⁵D₃ level monotonously decreased. This is because the energy difference between ⁵D₃ and ⁵D₄ levels is close to that between those of ⁷F₆ and ⁷F₀, which usually leads to cross relaxation by the resonant energy-transfer process: ⁵D₃ (Tb³⁺) + ⁷F₆ (Tb³⁺) → ⁵D₄ (Tb³⁺) + ⁷F₀ (Tb³⁺).¹⁷ In addition, with increasing of Tb³⁺ concentration, the emission intensity of ⁵D₄ level first increased and then descended after a maximum at x = 2.0. This is a very high doping concentration relative to other Tb³⁺ doped phosphors.^12,17,20 According to XRD analysis of Y_4−xSiAlO₈N:xTb³⁺ phosphors, the max doping concentration of Tb for Y in the Y₄SiAlO₈N host is x = 1.5. When x = 2.0 and x = 2.5, the XRD patterns show some impure phases which cause the decrease of emission intensities of ⁵D₄ (543 nm) levels. It can be concluded that the concentration quenching effect in Y_4−xSiAlO₈N:xTb³⁺ phosphors is very weak. Before the crystal structure of Y₄SiAlO₈N host was destroyed, the emission intensity of ⁵D₄ level monotonously increased with the increase of Tb³⁺ concentration.

Fig. 6 Dependence of emission intensities of ⁵D₃ and ⁵D₄ in Y_4−xSiAlO₈N:xTb³⁺ phosphors on Tb³⁺ concentration.

Fig. 7 represents photoluminescence excitation spectra of the Y_4−xSiAlO₈N:xTb³⁺ phosphors monitored at its characteristic green emission peak (543 nm). The excitation spectra contain many peaks in region of 200–500 nm, which can be separated into three independent bands. The first excitation band (peaked at 484 nm) and the second excitation band (include 377 nm, 358 nm, 352 nm, 335 nm and 317 nm excitation peaks) are corresponding to absorption of 4f → 4f spin-forbidden transitions of Tb³⁺ ion.¹⁷ These excitation region match well with near-UV chips for applications in near-UV white LEDs. The third excitation band refers to a wide band range from 225 nm to 310 nm which is ascribed to 4f → 5d allowed transition of Tb³⁺ ion.¹⁷ With increasing of Tb³⁺ concentration, by coincidence with the variation of emission intensity, the excitation intensity of 4f → 4f transitions first increased and then descended after a maximum at x = 2.0. However, the intensity of the third excitation band reached maximum at x = 1.5. And excitation centres of third bands changed from 268 nm, 276 nm, 281 nm, 287 nm, 290 nm, 294 nm to 298 nm. Furthermore, Fig. 7 shows that the intensity ratio of third band versus second band decreased with the increase of Tb³⁺ concentration. It can be interpreted that the third excitation band is not only result from 4f → 5d allowed transition of Tb³⁺ ion, but also result from 4f → 5d absorption of Y³⁺ ion. With the increase of Tb³⁺ concentration, the content of Y³⁺ decreased and energy transfer between host and Tb³⁺ reduced. It can be supported by photoluminescence spectra of host in Fig. 4 and diffuse reflectance spectra of Y₄SiAlO₈N host in Fig. 10.

Fig. 7 PL excitation spectra of Y_4−xSiAlO₈N:xTb³⁺ phosphors with varied Tb³⁺ ion concentrations.

3.3 Fluorescence dynamics of Tb³⁺ ions

Fig. 8 shows PL decay curves of Y_4−xSiAlO₈N:xTb³⁺ phosphors excited by 380 nm and monitored at 543 nm. The fluorescence lifetime (τ) and fitting degree CHISQ (χ²) are summarized in Table 2. As shown in Fig. 8(a) and (b), when 0.02 ≤ x ≤ 1.5, although Tb³⁺ ions substituted four Y³⁺ ions with different crystal sites in Y₄SiAlO₈N host, the decay curves can be well fitted into a single exponential decay and the fitting degrees of CHISQ (χ²) are 1.49, 1.55, 1.06, 0.96 and 1.05 respectively. The single exponential decay fitting function is as follow:²¹where I(t) represent the luminescence intensity at time t, A and B₁ are constants, τ₁ is the fluorescence lifetime in the single exponential decay, and t is the time. When the Tb³⁺ concentration increased from 0.02, 0.20, 0.50, 1.00 to 1.50, the fluorescence lifetime (τ) decreased from 2.431 ms, 2.326 ms, 2.238 ms, and 2.081 ms to 1.805 ms due to the nonradiative energy transfers between Tb³⁺ ions. Whereas, when the concentration of Tb³⁺ increased to x = 2.00 and 2.50, the decay curves of Tb³⁺ were no longer shown as single exponential decay. In case of single exponential fitting, the fitting degrees of CHISQ (χ²) are 2.27 and 6.11 respectively, which indicate a worse fitting result, as red line shown in Fig. 8(c) and (d). Simultaneously, the emission intensity decreased sharply at x = 2.50. These results reveal that the concentration quenching effect is predominant in this case. The higher Tb³⁺ concentration would lead to greater probability of nonradiative energy transfer between Tb³⁺ ions within a certain distance. Then, the more accurate fluorescence lifetime (τ) could be observed by two exponentials decay fitting using the following equation:^22,23where I(t) represents the luminescence intensity at time t, A, B₁ and B₂ are constants, τ₁ and τ₂ are the two fluorescence lifetimes in the two exponentials decay, and t is the time. Using these parameters, the average decay time (τ_avg) can be determined by the following formula:^22,23

Fig. 8 PL decay curves of Y_4−xSiAlO₈N:xTb³⁺ phosphors under 380 nm excitation.

Table 2 Fluorescence lifetime and fitting degree (χ²) of Y_4−xSiAlO₈N:xTb³⁺ phosphors

x Tb^3+	x = 0.02	x = 0.20	x = 0.50	x = 1.00	x = 1.50	x = 2.00		x = 2.50
Exponential	Single	Single	Single	Single	Single	Single	Two	Single	Two
CHISQ (χ^2)	1.49	1.55	1.06	0.96	1.05	2.27	0.90	6.11	1.00
Lifetime (ms)	2.431	2.326	2.238	2.081	1.805	1.517	1.450	1.253	1.139

---

As shown in Fig. 8(c) and (d) and Table 2, the fitting degrees of two exponentials decay are CHISQ (χ²) = 0.90 and 1.00 respectively. It can be concluded that the decay curves of high Tb³⁺ concentration in the Y_4−xSiAlO₈N:xTb³⁺ phosphors are suitable for two exponentials decay fitting. The average fluorescence lifetimes of Tb³⁺ are determined as 1.450 ms and 1.139 ms for x = 2.00 and x = 2.50 respectively.

The concentration quenching occurs mainly because the nonradiative energy transfers between Tb³⁺ ions within a certain distance.²⁴ And three mechanisms may be responsible for the nonradiative energy transfers, which are exchange interaction, radiation reabsorption, and electric multipolar interaction. According to the critical concentration, the critical distance R_c can be calculated by the following formula:^24,25

where V is the volume of crystallographic unit cell, x_c is the critical concentration, and N is the number of lattice sites in the unit cell that can be occupied by activator ion. In present case, the unit cell volume V is 805.83 ų, and N = 16.¹⁵ The critical concentration is x_c = 2.0 which can be obtained from fluorescence spectra. According to eqn (4), R_c of Tb³⁺ was calculated to be 3.64 Å. Since the exchange interaction requires a forbidden transition and a typical critical distance less than 5 Å,²⁴ it can be inferred that the mechanism of exchange interaction may plays an important role in nonradiative energy transfers between Tb³⁺ ions in the Y_4−xSiAlO₈N:xTb³⁺ phosphors. In addition, owing to only a little overlap between PLE and PL spectra of Tb³⁺ could be observed by them self in the region of 480–490 nm, the mechanism of radiation reabsorption plays an unimportant role. Thus, the nonradiative energy transfers between Tb³⁺ ions may occur via exchange interaction and electric multipolar interaction. On the basis of Dexter's energy transfer theory of exchange and electric multipolar interactions, the following relation can be obtained:^22,26where η₀ and η are the luminescence quantum efficiencies of Tb³⁺ without and with nonradiative energy transfer; C is the doping concentration of the Tb³⁺ ions; eqn (5) corresponds to the exchange interaction and eqn (6) with n = 6, 8, and 10 corresponds to dipole–dipole, dipole–quadrupole, and quadrupole–quadrupole interactions, respectively. The value of η₀/η can be approximately estimated from the related lifetime's ratio (τ₀/τ). Thus, eqn (5) and (6) can be represented by the following equations:²²

The relationships of ln(τ₀/τ) ∝ C and (τ₀/τ) ∝ C^n/3 are illustrated in Fig. 9. Taking into account significant distortion of the crystal structure observed for x_c = 2.0 and 2.5, these two samples were not presented. The fitting parameters of adjusted R-square are 0.96668, 0.97428, 0.94643 and 0.91167 respectively. It can be concluded that the best linear fitting behavior was observed when n = 6, implying that nonradiative energy transfers between Tb³⁺ ions in the Y_4−xSiAlO₈N:xTb³⁺ phosphors occur via a dipole–dipole mechanism.

Fig. 9 (a) Dependence of ln(τ₀/τ) of Tb³⁺ on C and τ₀/τ of Tb³⁺ on (b) C^6/3, (c) C^8/3, and (d) C^10/3.

3.4 Diffuse reflectance properties

To further study the energy absorption of Y_4−xSiAlO₈N:xTb³⁺ phosphors, the diffuse reflectance spectra of the host lattice and Tb³⁺ doped Y_4−xSiAlO₈N:xTb³⁺ is shown in Fig. 10. Apparently, the Y₄SiAlO₈N host material demonstrates a strong absorption band in the wavelength region of 200–250 nm, which is assigned to the host absorption. With the introduction of Tb³⁺ ions in Y_4−xSiAlO₈N:xTb³⁺, an obvious absorption from 250 to 500 nm assigned to 4f → 4f spin-forbidden transitions and 4f → 5d allowed transition of Tb³⁺ ion is observed.¹⁷ Also, the absorption edge gradually extends to longer wavelength, and the absorption gets enhanced with higher Tb³⁺ concentration. Especially, the three absorption peaks which located at 485 nm, 375 nm and 300 nm enhanced more quickly. These results are consistent with PLE spectra (Fig. 7) of Tb³⁺ ions and demonstrate that the broad and strong absorption range of Y_4−xSiAlO₈N:xTb³⁺ match well with near-UV chips.

Fig. 10 Diffuse reflectance spectra of Y_4−xSiAlO₈N:xTb³⁺ phosphors.

Unexpectedly, the host absorption in the 200–250 nm region reduced evidently with Tb³⁺ introduction. When the concentrations of Tb³⁺ ions reach to x = 1.5 or more, the reflectance in 200–250 nm region even exceed 100%. The reason for this phenomenon is that the reference sample we used in reflectance measurement is BaSO₄ (99.999%) powder, which exhibit excellent reflective capability in near-UV and visible light region. These results reveal that the reflectance of Y_4−xSiAlO₈N:xTb³⁺ (1.5 ≤ x ≤ 2.5) phosphors is greater than that of BaSO₄ (99.999%) powder in the region of 200–250 nm.

3.5 Thermal quenching properties of Y₃SiAlO₈N:1.0Tb³⁺ phosphor

For application in high power white LEDs, the thermal stability of phosphors is one of the important issues to be considered, which is considerable influence chromaticity and brightness of white light output.^11–13 Therefore, the temperature dependence of PL spectra for Y₃SiAlO₈N:1.0Tb³⁺ phosphor upon 380 nm excitation is shown in Fig. 11. It can be seen that the PL intensity decrease slowly with the increase of temperature from 293.0 K to 475.2 K. Since the fact that the temperature of an LED package rises by the heat generation of LED itself during the LED operation, the phosphors of white LEDs are required to maintain their conversion efficiency up to 150 °C.²³ In the upper of Fig. 11, the relative intensity (%) of PL spectra is shown depend on the different heating temperature. When the phosphor temperature was heated to 375.2 K, the relative PL intensity of Y₃SiAlO₈N:1.0Tb³⁺ reduced to 98.3% of the initial intensities at room temperature (293.0 K) and then further decreased to 97.5% at 434.7 K. Finally, the relative PL intensity maintains high value of 96.5% at 475.2 K. Comparing to thermal stabilities of other Tb³⁺ in recent reported literatures,^12,27,28 these results indicate that the Y₃SiAlO₈N:1.0Tb³⁺ phosphor possess excellent thermal stability which is benefit to application in the white LEDs.

Fig. 11 The temperature dependence of the PL spectra for Y₃SiAlO₈N:1.0Tb³⁺ phosphor.

In order to interpret such low thermal quenching behavior of Tb³⁺ in Y_4−xSiAlO₈N:xTb³⁺ phosphors, two possible models are proposed as follows.²⁴ One is the well-known nonradiative relaxation model: the thermal relaxation process from the potential curve of the excited state to the potential curve of the ground state through the crossing point in the configuration coordinate diagram. The other is the thermal ionization model: thermal excitation of the excited state electron to conduction band states. If the quenching process only caused by the nonradiative relaxation process, the activation energy can be calculated by the Arrhenius equation:^29–31

here I₀ is the initial emission intensity, I_T is the intensity at different temperatures, ΔE is activation energy of thermal quenching, c is a constant for a certain host, and k is the Boltzmann constant (8.629 × 10⁻⁵ eV K⁻¹). For convenient observation, eqn (9) can be deformed as follow:

By linear fitting, the relationship of [ln(I₀/I_T) − 1] versus 1/kT for the Y₃SiAlO₈N:1.0Tb³⁺ phosphor is plotted in Fig. 12. The fitted slope and intercept are acquired as −0.143 and 0.19787 respectively. According to above equation, the activation energy ΔE was calculated to be 0.143 eV, which is really quite lower than that of some recently reported Tb³⁺ doped phosphors.^27,28 As we known, the potential curve of the excited state level and the ground state level usually cross connection with each other with a point in the configuration coordinate diagram.^24,29 And the activation energy corresponds to the energy required to thermally activate the excited electrons from the lowest position of excited state level to that crossing point.²³ In the nonradiative relaxation process, the emission intensity decrease with temperature increment is attributed to the thermal excitation which provides the energy for ΔE. Hence, in the same or similar crystal structure, the higher the activation energy ΔE is, the better the thermal stability is. Furthermore, the probability of a nonradiative transition per unit time (α) can be expressed as:²⁹

where s is the frequency factor (s⁻¹), k is the Boltzmann constant, and T is the temperature. Because s, k, and ΔE are constant, it is clear that lower activation energy leads to a greater probability of nonradiative transitions. However, the obtained activation energy value of the Y₃SiAlO₈N:1.0Tb³⁺ phosphor is only about 0.143 eV, which is much lower than that of some other Tb³⁺ doped phosphors.^27,28 These results disagree with the excellent thermal stability of Y₃SiAlO₈N:1.0Tb³⁺ phosphor observed in Fig. 11. However, on the basis of Arrhenius equation (eqn (9)), except for activation energy ΔE, the constant C is also an important parameter for the relation between PL intensity at different temperatures I(T) and temperature (T). According to W. Lv et al.²⁹ C is a rate constant for thermally activated escape, which is constant for the same loss. So the different C value will indicates different thermal quenching process. According to eqn (10) and the intercept (0.19787) acquired by linear fitting in the Fig. 12, the value of C was calculated as about 1.22. Thus, the thermal relaxation process should not be responsible for the thermal quenching of Y₃SiAlO₈N:1.0Tb³⁺ phosphor. And the model of thermal excitation of the excited state electron to conduction band states should be considered for Y_4−xSiAlO₈N:xTb³⁺ phosphors. In fact, the total and partial density of states (DOS) calculation (as shown in Fig. 13) will show that the conduction band of Y₄SiAlO₈N host is mainly composed of dominant 4d (Y) orbitals and minor 2p (O) orbitals. With the introduction of Tb³⁺, the DOS intensity of conduction band decrease to a very low level in the Y₃SiAlO₈N:1.0Tb³⁺ phosphor. As a result, the thermal excitation of the excited states electron to conduction band states will hardly occur, i.e., the PL spectra of Y₃SiAlO₈N:1.0Tb³⁺ phosphor will show a low thermal quenching.

Fig. 12 The activation energy of Y₃SiAlO₈N:1.0Tb³⁺ phosphor.

Fig. 13 (a) Total and partial DOS of Y₄SiAlO₈N host; (b) total and partial DOS of Y₃TbSiAlO₈N phosphor.

3.6 Density of states (DOS) calculations

To illustrate the luminescence phenomenon of Y_4−xSiAlO₈N:xTb³⁺ phosphors, the total and partial density of states (DOS) of Y₄SiAlO₈N host and Y₃TbSiAlO₈N phosphors were determined by the ground state GGA-PBE calculation, as shown in Fig. 13(a) and (b). In the Y₃TbSiAlO₈N model, the Tb substitute for Y4 site for the most substitution probability. For clarity, the Fermi level is set to zero. It can be seen that the electronic structures of both compounds are very different to each other, especially for the obvious variation of 4d orbital of Y atom and the appearance of 4f orbital of Tb atom. As shown in Fig. 13(a), the conduction band of Y₄SiAlO₈N host is mainly composed of dominant 4d (Y) orbitals and minor 2p (O) orbitals. And the valence band is mainly composed of dominant 2p (O) orbitals, minor 4d (Y) orbitals, and some 2p (N) orbitals. The 2p orbitals of N atoms have more contribution than that of O atoms to the top of the valence band. The major energy gap between conduction band and valence band is determined as about 4.97 eV, which agrees well with the absorption edge (250 nm) of the Y₄SiAlO₈N host (Fig. 10). However, for the Y₃TbSiAlO₈N model, the intense DOS peak of Y atom at conduction band disappeared. As a result, the host absorption in the 200–250 nm region decrease drastically and the reflectance in the 200–250 nm region even exceed 100% when the concentration of Tb³⁺ is more than 1.5. Additionally, two sharp DOS peaks of 4f (Tb) emerge around the Fermi level. The upper one is 0.83 eV above the Fermi level and the lower one is −0.46 eV below the Fermi level. Then, the electronic structure in the top of valence band is mainly composed of dominant 4f (Tb) orbitals, which are hybridized with minor 2p (O) and minor 2p (N) orbitals. Also, a strong DOS peak of 5p (Tb) orbitals are observed at lower energy (−20 eV) position in the valence band. These preliminary results provide useful information of Y_4−xSiAlO₈N:xTb³⁺ phosphors which help to understand the luminescence phenomenon. Since the PL properties are mainly determined from the top of the valence band and the bottom of the conduction band,⁶ the DOS of host play a significant role in the luminescence phenomenon. As discussed above, with introduction of Tb³⁺ in Y₄SiAlO₈N host, the DOS of 4d (Y) weaken drastically. Accordingly, the DOS of conduction band is shown as a blank space, which is a favourable property for luminescent materials, helping to accommodate both the ground and excited states of luminescent ions within broad energy band.¹⁶ Owing to above reasons, the Y_4−xSiAlO₈N:xTb³⁺ phosphors show a low thermal quenching and high efficiency of green emission. Hence, these phosphors have a great potential to be a high effective green phosphor for white LEDs application packaged with near-UV chips.

4. Conclusions

In summary, we have successfully developed a novel, bright green-emitting Y_4−xSiAlO₈N:xTb³⁺ phosphor in a reduced atmosphere using high temperature solid state reaction method. The XRD and SEM analysis show that the most suitable sintering temperature is about 1500 °C. The PL intensity monotonously increased with the increase of Tb³⁺ concentration until the crystal structure was destroyed. The optimum doping concentration of Tb³⁺ was found for x = 2.0. Compared to other Tb³⁺ doped phosphors, this phosphor possesses higher quenching concentration which cause stronger emission intensity. On the basis of concentration quenching method, the critical distance of Tb³⁺ ions was calculated as 3.64 Å. The nonradiative energy transfer mechanism between Tb³⁺ ions has been confirmed to be via a dipole–dipole interaction. Due to the variation of conductance band in the phosphors, the reflectance in the region of 200–250 nm increased with the increase of Tb³⁺ concentration. The thermal quenching property has been investigated in detail. The temperature dependence of the PL spectra shows the excellent thermal stability of phosphors. According to Arrhenius equation, the corresponding activation energy is calculated as 0.143 eV. Furthermore, the model of thermal excitation of the excited states electron to conduction band states should be considered responsible for the thermal quenching of Y₃SiAlO₈N:1.0Tb³⁺ phosphor, rather than the thermal relaxation process. Considering the merits of n-UV light excitation, intense green emission, and good thermal stability, these materials have a potential application as white LEDs phosphors.

Acknowledgements

This work was financially supported by the National Natural Science Foundation of China (Grant no. 51272243 and 61405185) and the Zhejiang Provincial Natural Science Foundation of China (LY14E020008 and LZ14F050001).

Notes and references

1. D. Durach, L. Neudert, P. J. Schmidt, O. Oeckler and W. Schnick, Chem. Mater., 2015, 27, 4832–4838.
2. G. Li, C. C. Lin, W.-T. Chen, M. S. Molokeev, V. V. Atuchin, C.-Y. Chiang, W. Zhou, C.-W. Wang, W.-H. Li, H.-S. Sheu, T.-S. Chan, C. Ma and R.-S. Liu, Chem. Mater., 2014, 26, 2991–3001.
3. P. Pust, A. S. Wochnik, E. Baumann, P. J. Schmidt, D. Wiechert, C. Scheu and W. Schnick, Chem. Mater., 2014, 26, 3544–3549.
4. C. C. Lin and R. S. Liu, J. Phys. Chem. Lett., 2011, 2, 1268–1277.
5. M. Peng, X. Yin, P. A. Tanner, M. G. Brik and P. Li, Chem. Mater., 2015, 27, 2938–2945.
6. H. Ji, Z. Huang, Z. Xia, M. S. Molokeev, V. V. Atuchin, M. Fang and Y. Liu, J. Phys. Chem. C, 2015, 119, 2038–2045.
7. J. Meyer and F. Tappe, Adv. Opt. Mater., 2015, 3, 424–430.
8. T. M. Tolhurst, T. D. Boyko, P. Pust, N. W. Johnson, W. Schnick and A. Moewes, Adv. Opt. Mater., 2015, 3, 546–550.
9. Y. Chen, Y. Li, J. Wang, M. Wu and C. Wang, J. Phys. Chem. C, 2014, 118, 12494–12499.
10. P. Fulmek, C. Sommer, P. Hartmann, P. Pachler, H. Hoschopf, G. Langer, J. Nicolics and F. P. Wenzl, Adv. Opt. Mater., 2013, 1, 753–762.
11. Y.-C. Fang, S.-Y. Chu, P.-C. Kao, Y.-M. Chuang and Z.-L. Zeng, J. Electrochem. Soc., 2011, 158, J1.
12. Z. Xia and W. Wu, Dalton Trans., 2013, 42, 12989–12997.
13. W. Lv, M. Jiao, Q. Zhao, B. Shao, W. Lu and H. You, Inorg. Chem., 2014, 53, 11007–11014.
14. T.-C. Liu, B.-M. Cheng, S.-F. Hu and R.-S. Liu, Chem. Mater., 2011, 23, 3698–3705.
15. Y. Hua, X. Li, D. Zhang, H. Ma, D. Deng and S. Xu, New J. Chem., 2016, 40, 5458–5466.
16. Y. Jia, H. Qiao, N. Guo, Y. Zheng, M. Yang, Y. Huang and H. You, Opt. Mater., 2011, 33, 1803–1807.
17. K. Li, M. Shang, D. Geng, H. Lian, Y. Zhang, J. Fan and J. Lin, Inorg. Chem., 2014, 53, 6743–6751.
18. K. Y. Jung and H. W. Lee, J. Lumin., 2007, 126, 469–474.
19. X. Ju, X. Li, W. Li, C. Tao and J. Yang, Mater. Lett., 2012, 77, 35–37.
20. Z. Xia and R.-S. Liu, J. Phys. Chem. C, 2012, 116, 15604–15609.
21. Y. Jia, W. Lu, N. Guo, W. Lu, Q. Zhao and H. You, Phys. Chem. Chem. Phys., 2013, 15, 6057–6062.
22. M. Shang, G. Li, D. Geng, D. Yang, X. Kang, Y. Zhang, H. Lian and J. Lin, J. Phys. Chem. C, 2012, 116, 10222–10231.
23. J. Chen, Y. Liu, M. Fang and Z. Huang, Inorg. Chem., 2014, 53, 11396–11403.
24. J. Zhong, W. Zhuang, X. Xing, R. Liu, Y. Li, Y. Liu and Y. Hu, J. Phys. Chem. C, 2015, 119, 5562–5569.
25. Y. Jia, H. Qiao, Y. Zheng, N. Guo and H. You, Phys. Chem. Chem. Phys., 2012, 14, 3537–3542.
26. Y. Jia, Y. Huang, N. Guo, H. Qiao, Y. Zheng, W. Lv, Q. Zhao and H. You, RSC Adv., 2012, 2, 2678.
27. F. Zhang, Y. Wang and Y. Tao, J. Lumin., 2013, 136, 51–56.
28. X. Zhang, L. Zhou, Q. Pang, J. Shi and M. Gong, J. Phys. Chem. C, 2014, 118, 7591–7598.
29. W. Lv, Y. Jia, Q. Zhao, M. Jiao, B. Shao, W. Lü and H. You, Adv. Opt. Mater., 2014, 2, 183–188.
30. Y. Zhang, X. Li, K. Li, H. Lian, M. Shang and J. Lin, J. Mater. Chem. C, 2015, 3, 3294–3303.
31. X. Wang, T. Seto, Z. Zhao, Y. Li, Q. Wu, H. Li and Y. Wang, J. Mater. Chem. C, 2014, 2, 4476–4481.

---