Photoluminescence of the Eu³⁺-Activated Y_xLu_1−xNbO₄ (x = 0, 0.25, 0.5, 0.75, 1) Solid-Solution Phosphors

Abstract

Eu³⁺-doped Y_xLu_1−xNbO₄ (x = 0, 0.25, 0.5, 0.75, 1) were prepared by the solid-state reaction method. YNbO₄:Eu³⁺ and LuNbO₄:Eu³⁺ crystallize as beta-Fergusonite (SG no. 15) in 1–10 μm diameter particles. Photoluminescence emission spectra show a slight linear variation of emission energies and intensities with the solid-solution composition in terms of Y/Lu content. The energy difference between Stark sublevels of ⁵D₀→⁷F₁ emission increases, while the asymmetry ratio decreases with the composition. From the dispersion relations of pure YNbO₄ and LuNbO₄, the refractive index values for each concentration and emission wavelength are estimated. The Ω₂ Judd–Ofelt parameter shows a linear increase from 6.75 to 7.48 × 10⁻²⁰ cm² from x = 0 to 1, respectively, and Ω₄ from 2.69 to 2.95 × 10⁻²⁰ cm². The lowest non-radiative deexcitation rate was observed with x = 1, and thus LuNbO₄:Eu³⁺ is more efficient phosphor than YNbO₄:Eu³⁺.

1. Introduction

Phosphors designed with rare earth (RE)-activated compounds are a continuously rising research topic in both basic and applied science. Materials containing the trivalent europium ion (Eu³⁺) are well known for their strong luminescence in the orange/red spectral region, making them suitable for use in artificial lights, display devices, luminescent sensing, and biomedical research, among other applications [1,2,3,4]. In addition, the Eu³⁺ ion is a truly unique spectroscopic probe from a theoretical point of view. Because the 4f shell of Eu³⁺ has an even number of electrons ([Xe]4f⁶), the ion exhibits a non-degenerated ground (⁷F₀) and excited (⁵D₀) energy states, as well as non-overlapping ^2S+1L_J multiplets, resulting in emission spectra that are predictably dependent on the host material site symmetry. The energy level structure, the intensities of the f-f transitions (including the Judd–Ofelt theory), and the decay times of the excited states allow the Eu³⁺ ion to be used as a spectroscopic probe [5,6,7,8].

Compounds with the ABO₄ composition (A = RE; B = P, V, Nb) are suggested to be excellent hosts for luminescent materials in the vast field of phosphors [9,10,11] due to the coupling between rare-earth (RE³⁺) ions and the BO₄³⁻ group. RE-Niobates (RENbO₄) have long been known, but have received far less attention than vanadates or phosphates. They are chemically stable and have good dielectric properties, as well as ion and proton conductivities [12,13,14,15]. They have been prepared as single crystals [16,17], thin films [18], and in crystalline power form [19,20] for a variety of potential applications, primarily as optical hosts. Under UV or x-ray excitation, YNbO₄ is a self-activated phosphor with a broad and strong emission band in the blue spectral region around 400 nm (associated with NbO₄³⁻ groups from the host crystalline lattice) [21,22]. RE-Niobates’ luminescent properties can be altered by doping with various rare-earth ions. Emission has been observed from: Tm³⁺, Tb³⁺, Dy³⁺, Sm³⁺, Er³⁺, Nd³⁺ [23,24,25,26,27,28], and Eu³⁺ [20,29]. Recent applications of RE-doped niobates involved luminescent temperature sensors [27,30].

The RENbO₄ structure type has a more complex profile than the REVO₄ materials, which always occurs in the typical tetragonal zircon structure type. YNbO₄ was originally thought to be the parent material of the RENbO₄ group, like the natural mineral fergusonite, leading to the classification of the material as a tetragonal phase with a scheelite (I41/a) structure [31]. However, there are two major crystalline forms of RENbO₄. One is the low-temperature M-phase isostructure with a monoclinic form of the fergusonite, and the other is the high-temperature T-phase corresponding to the tetragonal scheelite. Between 500 and 850 °C, the reversible transition between two phases, monoclinic and tetragonal, occurs [32]. Subsequent research revealed that the low-temperature phase, monoclinic beta-Fergusonite structure can be described by the space group I-centered (I12/a1) or C-centered (C12/c1), as the I2/a space group is a non-standard setting of C2/c (SG no. 15) [33,34,35]. In both settings, Y is surrounded by 8 oxygens in a large, low-symmetry YO₈ octahedron [34]. When other rare earth ions, such as Eu³⁺, occupy the Y site, the tilting of adjacent NbO₄ and NbO₆ polyhedra is expected to influence the luminescence properties of the dopant.

In this study, we wanted to investigate the effect of different ionic radii of RE³⁺ ions on the photoluminescence of the Eu³⁺-activated niobate host. We prepared a set of five Eu-doped Y_xLu_1−xNbO₄:Eu samples (x = 0, 0.25, 0.5, 0.75, and 1) with a fixed Eu concentration (5%) to investigate the influence of the Y to Lu ratio in the host niobate material on Eu³⁺ luminescence features. In that sense, the crystal field splitting of ⁷F₁ manifold, R intensity ratio and Judd–Ofelt parameters were determined. The application of Judd–Ofelt theory was explained, and the difference in the refractive index of the materials was considered.

2. Materials and Methods

The set of five samples, Eu-doped Y_xLu_1−xNbO₄:Eu³⁺ (x = 0, 0.25, 0.5, 0.75, and 1), were prepared by the solid-state reaction method. In the stoichiometric amounts of starting materials, Yttirium(III) oxide (Y₂O₃ Alfa Aesar, 99.9%), Lutetium(III) oxide (Lu₂O₃, Alfa Aeser 99.9%), and Niobium(V) oxide (Nb₂O₅ Alfa Aesar, 99.5%) were mixed with Europium oxide (Eu₂O₃ Alfa Aesar, 99.9%) added in order for Y or Lu to reach 95% (i.e., Y_0.95Eu_0.05NbO₄). With the addition of sodium sulphate decahydrate (Na₂SO₄ x 10H₂O, Alfa Aesar, 99%) as flux and small amount of ethanol, the mixtures were homogenized in a ball mill (BM500, Anton Paar) for several hours. The dried mixture was pre-sintered at 800 °C for 2 h, then sintered at 1450 °C for 8 h and allowed to cool to room temperature. Such obtained powders or pellets prepared from the powder under a load of 5 × 10⁸ Pa were used for measurements.

X-ray diffraction measurements were performed with a Rigaku SmartLab diffractometer (Tokyo, Japan) using Cu Kα radiation (30 mA, 40 kV) measured in the 2θ range from 10° to 90°. The built-in package software was used to obtain relevant structural analysis results (crystal coherence size, microstrain values, unit cell parameters, and data fit parameters). A Mira3 Tescan field emission scanning electron microscope (FE-SEM) (Brno, Czech Republic) was used for microstructural characterization, operated at 20 keV and 5.00 kx magnification. Photoluminescence excitation spectra were measured by a Horiba Jobin Yvon Fluorolog FL3-221 spectrofluorometer (Palaiseau, France) equipped with a 450 W Xenon lamp, TBX-04-D PMT detector and a double-grating monochromator with 1200 g/mm. The excitations were performed by monitoring the emission maxima at 612 nm. Lifetime measurements were performed using the same instrument equipped with a xenon–mercury pulsed lamp. Photoluminescence emission spectra were recorded by a high-resolution spectrograph (FHR 1000) (Palaiseau, France) equipped with a Horiba Jobin–Yvon Intensified Charge Coupled Device–ICCD detector and selectable diffraction gratings of 300 and 1800 g/mm. Samples were excited by a 365 nm LED (Ocean Optics) and driven by an Ocean Insight LDC-1C controller.

3. Results

3.1. Crystal Structure and Morphology

X-ray diffraction measurements confirmed that YNbO₄:Eu crystallizes in a monoclinic Fergusonite-beta-(Y) structure that can be best fitted with the C2/c space group corresponding to pure YNbO₄ ICDD No. 01-083-1319, as can be seen in Figure 1 [36]. The isostructural incorporation of a bigger Eu³⁺ (r_VIII = 1.066 Å) ion instead of a Y³⁺ (r_VIII = 1.019 Å) ion of YO₈ dodecahedra results in a small maximum shift of approximately 0.03° to the lower 2θ values [37]. X-ray diffraction measurements confirmed that LuNbO₄:Eu³⁺ crystallizes in a monoclinic Fergusonite-beta-(Lu) structure that can be best fitted with the I2/a space group corresponding to pure LuNbO₄ ICDD No. 01-074-6538, as can be seen in Figure 1 [35]. A slight peak position shift of around 0.1° to the lower 2θ values is also observed, resulting from the isostructural incorporation of a larger Eu³⁺ (r_VIII = 1.066 Å) ion into the Lu³⁺ (r_VIII = 0.977 Å) position of LuO₈ dodecahedra [36].

Table 1. Structural parameters obtained by Rietveld refinement of XRD data for YNbO₄:Eu³⁺ according to ICDD 01-083-1319, and for LuNbO₄:Eu³⁺ according to ICDD 01-074-6538.

Ref. Parameters	YNbO4:Eu3+ (C2/c)	LuNbO4:Eu3+ (I2/a)
Crystallite size (Å)	533(6)	479(6)
Strain (%)	0.08(2)	0.11(2)
a (Å)	7.62775(17)	5.23879(13)
b (Å)	10.9636(3)	10.8447(3)
c (Å)	5.30476(15)	5.04724(12)
α (o)	90	90
β (o)	138.4389(11)	94.4267(13)
γ (o)	90	90
Rwp 1	10.28%	10.25%
Rp 2	6.30%	6.55%
Re 3	3.23%	3.32%
GOF 4	3.1833	3.0895

¹ Rwp—regression sum of weighted squared errors of fit; ² Rp—regression sum of relative squared errors of fit; ³ Re—regression sum of relative errors of fit; ⁴ GOF—goodness of fit parameter.

shows the relevant structural characteristics for both final compositions of Y_xLu_1−xNbO₄:Eu³⁺ (x = 0 and 1), determined by Rietveld refinement of the experimental data derived from Rigaku SmartLab built-in package software. The YNbO₄:Eu³⁺ (x = 1) parameters are refined to the C2/c space group, with a higher value of parameter a of the unit cell around 7.63 Å and a higher β of 138.44°, whereas LuNbO₄:Eu³⁺ (x = 0) parameters are refined to the I2/a space group, with a lower value of parameter a of the unit cell of around 5.24 Å and a lower β of 94.43°. Both samples have crystallite sizes of roughly 50 ± 3 nm and low strain values. The SEM images of Y_xLu_1−xNbO₄:5%Eu (x = 0, 0.5, 1) materials presented in Figure 1 show irregularly shaped particles ranging in size from 1 to 10 microns. The YNbO₄:5%Eu material comprises a higher proportion of smaller, densely packed particles, whereas the particles in the LuNbO₄:5%Eu material are larger and less densely packed.

3.2. Photoluminescence

The luminescence excitation and emission spectra of the Eu³⁺ (4f⁶) ion in Y_xLu_1−xNbO₄ (x = 0, 0.25, 0.5, 0.75, 1) hosts clearly show characteristic f-f transitions, as shown in Figure 2. The excitation spectra of all samples, observed in Figure 2a, as recorded at λem = 612 nm, show several dominant excitation bands centered at around 540 nm (⁷F₁ → ⁵D₁), 529 nm (⁷F₀ → ⁵D₁), 466 nm (⁷F₀ → ⁵D₀), 393 nm (⁷F₀ → 5L₆), and 365 nm (⁷F₀ → ⁵D₄). The emission spectra of all samples recorded with λexc = 365 nm excitation are presented in Figure 2b. Following the ⁷F₀ → ⁵D₄ excitation, the electrons non-radiatively deexcite to the long-lived ⁵D₀ level, from where the radiative relaxation occurs to the ⁷F_J ground multiplet. In the recorded region, five dominant emission bands from the ⁵D₀ level are centered at 595 nm (⁵D₀ → ⁷F₁), 612 nm (⁵D₀ → ⁷F₂), 655 nm (⁵D₀ → ⁷F₃), 710 nm (⁵D₀ → ⁷F₄), and 745 nm (⁵D₀ → ⁷F₅). The possible ⁵D₀ → ⁷F₆ cannot be recorded by the used experimental setup. As is obvious from Figure 2, the luminescences in all the samples show overlapping bands characteristic for the Eu³⁺ ion. The Stark level energies calculated from the recorded excitation and emission spectra are presented in

Table 2. Calculated energy of Stark levels Y_0.5Lu_0.5NbO₄:5%Eu.

Level	Observed Energy [cm−1]
7F0	0
7F1	369
7F2	1200
7F3	1960
7F4	2999
7F5	3799
7F6	N/A
5D0	17,204
5D1	18,921
5D2	21,402
5D3	23,810
5L6	25,253
5L7	26,008
5G2–6	26,420
5L8	27,100
5D4 + 5L9	27,473
5L10	28,531
5H3,4,7	30,722

.

The intensity of the Eu^{3+ 5}D₀→⁷F₂ (ΔJ = 2) forced electric dipole transition is hypersensitive since it is strongly dependent on the coordinating polyhedron local site symmetry. Given that the environment of Eu³⁺ ion is a (Y,Lu)O₈ dodecahedron in all samples, small changes can be expected because of the variation of ion positions governed by the different ionic radii of Y and Lu ions [36].

The intensity of the parity-allowed Eu^{3+ 5}D₀→⁷F₁ (ΔJ = 1) magnetic dipole transition is considered as independent to the local symmetry; however, maximum splitting of the ⁷F₁ manifold of the Eu³⁺ ion is a function of the host [38]. To observe the influence of the host on the ⁷F₁ manifold, additional measurements were performed using a 1800 g/mm diffraction grating monochromator. To precisely determine the positions of all three maxima, a deconvolution is applied, as presented in Figure 2c.

The positions (Peak energies [cm⁻¹]) of the characteristic emission peak maxima of Y_xLu_1−xNbO₄:Eu (x = 0, 0.25, 0.5, 0.75, 1) are presented in Figure 3a. As can be seen, only small differences of around 10–20 cm⁻¹ are determined, given the similarity of the Eu³⁺ ion local environment. Trends of maximum ΔE splitting, the Stark splitting of ⁷F₁ manifold, and the asymmetry ratio R corresponding to the Y/Lu content are presented in Figure 3b. As is evident, ΔE increases linearly with the increase of the x value (increase in Y content). For the same site symmetry and coordination number, the positions of the Stark levels of emissions originating from the ⁵D₀ level of Eu³⁺ depend on the covalency of the metal-ligand bond [39].

The asymmetry ratio R, which is the ratio of the two characteristic transitions, is frequently used to obtain information on the local site symmetry. As the ratio can be influenced by the refractive index (n) of the host, experimental R values can be corrected for n using the following equation [40]:

$$R_{corr}=\left[\frac{9n^2}{{\left(n^2+2\right)}^2}\right]\frac{I\left(5_{{\mathrm{D}}_0}-7_{{\mathrm{F}}_2}\right)}{I\left(5_{{\mathrm{D}}_0}-7_{{\mathrm{F}}_1}\right)}.$$

(1)

Starting n values for LuNbO₄ and YNbO₄ were obtained from Ref. [10]. The refractive index values for the other hosts were then calculated by approximation: n(Y_xLu_1−xNbO₄) = x∙n(YNbO₄) + (1−x)∙n(LuNbO₄); and they are given in Figure 4. Characteristic n values at different wavelengths are clearly marked in Figure 4, and they are used for further Judd–Ofelt calculations.

3.3. Judd–Ofelt Analysis

Eu³⁺ is the only ion with pure magnetic dipole transitions, ⁵D₀→⁷F₁ and ⁵D₁→⁷F₀, allowing for the self-referenced Judd–Ofelt analysis from the single emission spectrum [5,6,41]. As the magnetic dipole transition strengths are independent of the host matrix, the radiative transition probability of the magnetic dipole transition (MD) can be used for calibration of the spectrum [42]. Then, the Judd–Ofelt parameters can be estimated from the single emission spectrum (without the traditionally used fitting procedure [43,44]), directly from the relation [45,46]:

$${\mathsf{\Omega }}_{\lambda }\left[{\mathrm{cm}}^2\right]=\frac{4.135\times {10}^{-23}}{U^{\lambda }}{\left(\frac{{\nu }_{MD}}{{\nu }_{\lambda }}\right)}^3\frac{9n_{MD}^3}{n_{\lambda }{\left(n_{\lambda }^2+2\right)}^2}\frac{I_{\lambda }}{I_{MD}},$$

(2)

where λ = 2,4,6 abbreviates ⁵D₀→⁷F_λ transitions, and U^λ are the squared reduced matrix elements with values of 0.0032, 0.0023, and 0.0002 for λ = 2,4,6, respectively. ν is the emission barycenter energy, n is the refractive index, and I are the integrated emission intensities (the employed MD transition here is ⁵D₀→⁷F₁). Equation (2) is given for intensities measured in counts. In the case where intensities are given in power units, the power to the ratio of barycenter energies should be equal to 4 [47].

From the Judd–Ofelt parameters, many derived quantities can be estimated directly: radiative transition probabilities, branching ratios, radiative lifetime, and emission cross-sections [48]. The radiative transition probabilities for a given λ and MD transition are given by Equations (3) and (4), respectively [49]:

$$A_{\lambda }\left[{\mathrm{s}}^{-1}\right]=8.034\times {10}^9{\nu }_{\lambda }^3{n}_{\lambda }{\left(n_{\lambda }^2+2\right)}^2{U}^{\lambda }{\mathsf{\Omega }}_{\lambda },$$

(3)

$$A_{MD}\left[{\mathrm{s}}^{-1}\right]=3\times {10}^{-12}{\nu }_{MD}^3{n}_{MD}^3,$$

(4)

The radiative lifetime, branching ratios, and emission cross-sections are estimated by Equations (5)–(7), respectively [50,51]:

$${\tau }_{rad}\left[\mathrm{ms}\right]=\frac{{10}^3}{{{\displaystyle \sum }}_{\lambda =2,4,6}{A}_{\lambda }\left[s^{-1}\right]+A_{MD}\left[s^{-1}\right]},$$

(5)

$${\beta }_{\lambda ,MD}={10}^{-3}{A}_{\lambda ,MD}\left[{\mathrm{s}}^{-1}\right]{\tau }_{rad}\left[\mathrm{ms}\right],$$

(6)

$${\sigma }_{\lambda ,MD}\left[{\mathrm{cm}}^2\right]=\frac{1.33\times {10}^{-5}}{{\nu }_{\lambda ,MD}^4{n}_{\lambda ,MD}^2}\frac{\max i_{\lambda ,MD}}{I_{\lambda ,MD}}{A}_{\lambda ,MD},$$

(7)

where max(i) is the intensity at the transition maximum.

The experimentally measured lifetime values (τ_obs) of the emissions from the ⁵D₀ level are given in

Table 3. Judd–Ofelt parameters of Y_xLu_1−xNbO₄:Eu³⁺, branching ratios (β), radiative lifetimes (τ_rad), emission cross-sections (σ), radiative transition probabilities (A), total radiative transition probabilities (A_R), observed lifetime (τ_obs), non-radiative lifetime (τ_NR), and intrinsic quantum yield (η_int). Transitions ⁵D₀→⁷F_1,2,4 are abbreviated with MD, 2, and 4 in subscripts, respectively.

x	0	0.25	0.50	0.75	1
Ω2∙1020 [cm2]	6.75	6.88	7.37	7.41	7.48
Ω4∙1020 [cm2]	2.69	2.91	2.97	3.01	2.95
βMD [%]	14	14	14	14	14
β2 [%]	73	73	73	73	73
β4 [%]	13	14	14	14	13
τrad [ms]	0.782	0.859	0.916	1.027	1.134
σMD∙1021 [cm2]	2.12	1.81	1.64	1.43	1.49
σ2∙1021 [cm2]	9.96	9.40	9.71	9.01	8.87
σ4∙1021 [cm2]	3.63	3.50	3.64	3.13	3.18
AMD [s−1]	179	163	148	133	120
A2 [s−1]	928	838	795	705	628
A4 [s−1]	171	164	148	136	115
AR [s−1]	1279	1164	1092	974	882
τobs [ms]	0.619	0.643	0.650	0.650	0.652
τNR [ms]	2.970	2.557	2.238	1.771	1.534
ηint [%]	79	75	71	63	57

, allowing for the calculation of the non-radiative lifetime (τ_NR) by [52]:

$${\tau }_{NR}\left[\mathrm{ms}\right]=\frac{{\tau }_{rad}{\tau }_{obs}}{{\tau }_{rad}-{\tau }_{obs}}.$$

(8)

The Judd–Ofelt parameters and derived quantities were calculated by the JOES application software and are given in Table 3 [51]. With the increase of Y content in the hosts (increasing x), certain trends can be observed: the Ω₂ parameter of the hypersensitive transition ⁵D₀→⁷F₂ increases with x, indicating an increasing degree of covalency and decrease of Eu³⁺ site symmetry, and the Ω₄ parameter, which relates to the viscosity and rigidity of the host matrix, shows no significant change [53]. The Ω₆ parameter could not be estimated, as the emission ⁵D₀→⁷F₆ was not observed.

According to the changes in the refractive index and Judd–Ofelt parameters, the radiative transition probabilities and cross-sections show a monotonic increase with x, while the radiative lifetime decreases. As the observable lifetime also increases with x, the non-radiative lifetime is at a minimum in the pure YNbO₄, indicating that LuNbO₄ is a more efficient host matrix for the Eu³⁺ emission.

As the Judd–Ofelt parameters depend on the environment of the ion, the parameters obtained by the semi-empirical method are a statistical average of the Judd–Ofelt parameters for each ion. Thus, if there is a mix of non-equivalent sites in the host matrix, the Judd–Ofelt parameters represent an average value for each site, weighted by their contribution [49]. Thus, the Judd–Ofelt parameters of the mixture of two hosts can be predicted from the Judd–Ofelt parameters of the two pure compounds (a,b):

$${\mathsf{\Omega }}_{\lambda }={\mathrm{x}\mathsf{\Omega }}_{\lambda }^{\left(a\right)}+\left(1-\mathrm{x}\right){\mathsf{\Omega }}_{\lambda }^{\left(b\right)}.$$

(9)

The estimation of the Judd–Ofelt parameters from the emission spectra inherently brings an error of about 10%. By fitting the Judd–Ofelt parameters to the linear relations for each x, the Judd–Ofelt parameters can be refined by the values on the fitted curve, as the estimate reduces the error of estimation for each by $\sqrt{k}$, where k is the number of measurements. Thus, in the case of the mixture of YNbO₄ and LuNbO₄ in five ratios, the error of each estimated Judd–Ofelt parameter is reduced about 2.2 times, and the more correct values are on the fitted curve for each concentration (Figure 5).

4. Conclusions

In this study, we demonstrated how the Y-to-Lu ratio in Y_xLu_1−xNbO₄:Eu³⁺ powder material influenced the Eu³⁺ luminescence features. The materials were synthesized using the solid state reaction method. All of the structures crystallized as beta-Fergusonite, in which the Eu ion replaced the Y or Lu ion in a large, low-symmetry octahedron. In all composition hosts, the luminescence excitation and emission spectra of the Eu³⁺ (4f⁶) ion showed characteristic f-f transitions from which the Stark energy levels were calculated.

The specific features and energy positions of the characteristic ⁵D₀→⁷F₁ magnetic dipole transition were determined when measured with a higher resolution and when spectra deconvolution was used. The maximum ΔE splitting of the Stark splitting of ⁷F₁ manifold and the asymmetry ratio R all exhibit Y/Lu content-dependent trends. Calculations based on the Judd–Ofelt theory were used to estimate specific quantities, concluding that LuNbO₄ is a more efficient host matrix for the Eu³⁺ emission.