Behaviour of electric and magnetic dipole transitions of Eu³⁺, ⁵D₀ → ⁷F₀ and Eu–O charge transfer band in Li⁺ co-doped YPO₄:Eu³⁺

Abstract

The effect of Li⁺ co-doping on the photoluminescence properties of YPO₄:Eu is discussed. Interesting behaviours, such as the presence of intermediate bands, shifting of the Eu–O charge transfer band (Eu–O CTB) to a lower wavelength, variation in intensities of magnetic (⁵D₀ → ⁷F₁) and electric dipole (⁵D₀ → ⁷F₂) transitions of Eu³⁺ and shift of ⁵D₀ → ⁷F₀ to higher energy with increasing excitation wavelengths are observed. The Eu³⁺ ion does not have an absorption band in the range 340–350 nm, but after excitation at these wavelengths, a broad emission band (370–570 nm), as well as sharp peaks of Eu³⁺, could be observed. This is due to strong energy transfer from the intermediate band of the host to the Eu³⁺ ion. X-ray photoelectron spectroscopy (XPS) study also confirms that intermediate band emission is not due to Eu²⁺ ion emission. The blue shifting of Eu–O CTB is because of the increase in the optical electronegativity of the Eu³⁺ ion on Li⁺ co-doping. The variation in intensities of the ⁵D₀ → ⁷F₂ and ⁵D₀ → ⁷F₁ dipole transitions is related to (i) overlapping interaction parameters within the ground and excited states, (ii) exchange interaction among atoms/ions and (iii) density of the incoming photons. Shift of ⁵D₀ → ⁷F₀ to a higher energy with increasing excitation wavelengths is because of change in the second order crystal field parameter B₂₀ with excitation wavelength. The significant enhancement of luminescence intensity is found with Li⁺ co-doping due to the increase in crystallinity.

---

1. Introduction

Lanthanide ion (Ln³⁺) doped materials have been employed in a variety of advances and challenges in light emitting diodes (LED). Eu³⁺ doped YPO₄ (wide band gap of E_g = 8.6 eV) is useful for optoelectronic device applications.¹ The failure of Judd–Ofelt theory in the ⁵D₀ → ⁷F₀ transition of Eu³⁺ is well known. It is due to the mixing of two different states with different J values (J mixing).² In some systems, it is observed that the emission intensity of ⁵D₀ → ⁷F₁ is linearly dependent on the ⁵D₀ → ⁷F₀ excitation energy.³ However, changes of the ⁵D₀ → ⁷F₀ peak position and line-width to the lower wavelength value with increasing site selective excitation wavelengths from (467–577 nm) have been reported.^2a Similar observations were reported.^2b On the other hand, the opposite effect was not reported, to the best of the authors' knowledge. Variation of this peak intensity with excitation wavelength has not been discussed much in the literature. Even at room temperature, the ⁷F₁ level can be thermally populated in part and also lower energy photons via thermally excited Eu³⁺ ions can populate electrons at the ⁵D₀ level using a laser as an excitation source.³ It was reported that single crystal YPO₄:Eu³⁺ did not show the ⁵D₀ → ⁷F₀ transition, but its intensity appeared in the sintered powder sample due to the strain developed in the lattice.⁴

Eu³⁺ ions are used in spectroscopy to determine the environment around the Eu³⁺ ion in the host as compared to other Ln³⁺ ions, since its electric dipole transition is hypersensitive to its local symmetry. If the ratio of the electric to magnetic dipole transitions varies significantly with excitation wavelength (i.e. more or less than 1), it is difficult to the determine the environment of Eu³⁺ in the same host. This has not been discussed much in the literature.

Since Eu³⁺ has a 4f⁶ configuration, it needs to gain one more electron to achieve the half-filled 4f⁷ configuration, which is relatively stable compared to partially filled configurations. When Eu³⁺ is linked to the O²⁻ ligand, there is a chance of electron transfer from O²⁻ to Eu³⁺ to form Eu²⁺–O⁻ (simply Eu–O). During this, there is a broad absorption band at 240–270 nm, depending on the host. This is known as the Eu–O charge transfer band (CTB), and this band changes depending on the particle size of the host and environment of Eu³⁺. Could it happen on Li⁺ co-doped YPO₄:Eu³⁺? Will Li⁺ ions go to the lattice sites or interstitial sites on Li⁺ co-doped YPO₄:Eu³⁺? There is controversy regarding the Li⁺ co-dopant system, where authors claim that Y³⁺ is substituted by Li⁺ in YPO₄. But upon expansion of their X-ray diffraction patterns, it was found that the positions of the peaks shift to a lower 2 theta angle upon Li⁺ doping.⁵ In the case of substitution, the opposite effect should occur, since the ionic size of Li⁺ (∼0.76 Å) is much less than that of Y³⁺ (∼0.9 Å).⁶ In some reports Li⁺ co-doping causes the peaks to shift to a higher 2θ angle in the XRD patterns.⁷ In the cases of alkali metal (Li, Na, K) doping to many luminescent materials, there were changes in luminescence intensity, but there are no conclusive reasons so far.^7–10 Sometimes, it was found that the energy transfer rate from Eu–O or W–O to Eu³⁺ is very poor, and thus it is difficult to distinguish variations in luminescence intensity.⁸ In recent work on Li⁺ co-doped YPO₄:Eu, 10 at% Li⁺ could give optimum luminescence intensity due to improved crystallinity and increased particle size.¹¹ Here we used the polyol synthesis route (ethylene glycol as both solvent and capping), due to the smaller number of hydrocarbons present in its formula, low cost and effective control of particle sizes. Recently, we have successively synthesized highly crystalline nanoparticles of YPO₄ and other hosts such as Eu³⁺/Tb³⁺doped CaMoO₄ nanoparticles using this method at low temperature.^12,13

In this article, we analyse the luminescence properties of a 5 at% Eu³⁺ doped YPO₄ sample with or without Li⁺ co-dopant under different excitation wavelengths (240–399 nm). 5 at% Eu³⁺ doped YPO₄ is referred to as YPO₄:Eu in subsequent paragraphs. We observe that the magnetic dipole transition intensity is dominant over the electric dipole transition at the excitation wavelengths (240, 250, 290, 300, 340, 350, 399 nm) whereas the reverse occurs at 260–280 nm excitations. Eu–O CTB shows a blue shift with Li⁺ ion concentration. Interestingly, the forbidden transition (⁵D₀ → ⁷F₀) appears distinctly on Li⁺ co-doping. The variations in peak position, line-width and intensity are found with the different excitation wavelengths. These interesting observations are discussed here.

2. Experimental section

2.1. Synthesis

The samples of Eu³⁺ (5 at%) doped YPO₄ (i.e., YPO₄:Eu) and Li⁺ (Li⁺ = 3, 5, 7 and 10 at%) co-doped YPO₄:Eu were prepared using polyethylene glycol route at low temperature (∼100–120 °C for 1 h). The detailed preparation method is reported elsewhere.¹⁴ The as-prepared sample was annealed at 900 °C in ambient atmosphere at a rate of 2 °C min⁻¹ for 4 h in order to remove the dangling bonds on the surface of the nanoparticles. Samples annealed at 900 °C are studied here because they have high crystallinity.

2.2. Characterization

Detailed characterization of the sample was given in ref. 14. Chemical bonding in the sample was measured using X-ray photoelectron spectroscopy (XPS) SPECS, Germany (Mg-Kα X-ray source, hv = 1253.6 eV). The photoluminescence (PL) spectra of these powder phosphors were recorded using a Hitachi F-4500 spectrometer with a 150 W Xe lamp as a source at a spectral resolution of 3 nm. All the measurements were carried out at room temperature. PL decay was recorded with an Edinburgh instrument F920 equipped with Nd-YAG laser pumped optical parameters (OPO), with a pulse width of 10 ns and a repetition frequency of 10 Hz as the excitation source. Excitation and emission wavelengths are fixed at 468 and 615 nm, respectively.

3. Results and discussion

3.1 XPS study

We have analysed the X-ray Photoelectron Spectra (XPS) of 0 and 10 at% Li⁺ co-doped YPO₄:Eu. The binding energy (BE) of each Y³⁺, P⁵⁺ and O²⁻ ion increases by 0.4 eV when Li⁺ (10 at%) is co-doped to YPO₄:Eu (Fig. 1). This is related to the improved crystallinity on annealing. Their corresponding FWHM decreases on Li⁺ co-doping and their values are shown in the respective figures. Here, we discuss the peak positions/binding energies of ions for 10 at% Li⁺ co-doped YPO₄:Eu. As shown in Fig. 1(a), a broad peak with BE ∼ 45.5 eV is assigned to Li(1s). The BE of Eu²⁺(4d_3/2) was reported at ∼133.1 eV and P(2p) was also in the same band.¹⁵ In our study, there is a peak centered at 133.1 eV and thus we could not determine the exact positions of Eu²⁺(4d_3/2) and P(2p). However, it was reported that there was a peak at ∼127.1 eV, which can determine Eu²⁺ (4d_5/2),¹⁶ but we could not achieve it. Now, we are looking for Eu³⁺, which can have peaks at 135 eV for 4d_5/2 and 141 eV for 4d_3/2.¹⁵ After expansion in the 126–144 eV range, a small peak at ∼141.2 eV is found, which matches with Eu³⁺(4d_3/2), and the peak for Eu³⁺(4d_5/2) will be assigned to the peak at 133 eV. So, the peak at 133 eV will be from P(2p). The presence of Eu³⁺ is also supported by luminescence study (discussed later). A broad peak at ∼158.8 eV is assigned to Y(3d_3/2) (Fig. 1(b)). This confirms the high probability of Eu³⁺ present in the samples (Fig. 1(c)). A broad peak at ∼301 eV is assigned to Y(2p_1/2) and its FWHM is ∼2.9 eV (Fig. 1(d)). Fig. 1(e) shows the BE peak at ∼531.06 eV having a FWHM of ∼1.8 eV, which corresponds to O(1s). The peak is symmetrical, indicating the presence of one kind of O, which is in a tetragonal structure (i.e. highly crystalline). If there is O defect in the lattice, there will be a peak/hump at ∼529 eV.¹⁵

Fig. 1 XPS spectra of 0 and 10 at% Li⁺ co-doped YPO₄:Eu. Peaks corresponding to the core binding energies of individual elements are shown in (a)–(e). The binding energy of Eu²⁺ 4d_5/2 is missing, whereas that of Eu³⁺ 4d_5/2 is observed after the expansion (c).

3.2 Luminescence study

3.2.1 Excitation and emission studies

Fig. 2(a) shows the excitation spectrum of 5 at% Li⁺ co-doped YPO₄:Eu by monitoring the emission wavelength at 594 nm. The excitation spectrum consists of a strong absorption band between 200–275 nm being centered at ∼237.6 nm with FWHM ∼20 nm, which can be assigned to the Eu–O CT. The sharp lines in the wavelength range 275 to 500 nm centered at 324, 367, 387, 398, and 468 nm correspond to ⁷F_0,1 → ⁵H_3,6, ⁷F_0,1 → ⁵D₀, ⁷F_0,1 → ⁵G₁, ⁵L₇, ⁷F₀ → ⁵L₆ and ⁷F₀ → ⁵D₂ of Eu³⁺, respectively.^17,18 The absorption intensity of the ⁷F₀ → ⁵L₆ transition at 398 nm (FWHM ∼ 6 nm) is 1.4 times stronger than the Eu–O CT absorption, indicating a weak energy transfer from the Eu–O CT band to Eu³⁺.^19,20 Fig. S1† (ESI†) shows the excitation spectra of (0–10) at% Li⁺ co-doped YPO₄:Eu by monitoring the emission wavelength at 594 nm. The wavelength corresponding to the Eu–O CT band decreases from 238.6 to 235.5 nm and the corresponding FWHM decreases from 21 to 18 nm, with Li⁺ ion concentration up to 10 at% (Fig. 2(b)). The blue shift in the CT band with Li⁺ ion concentration suggests the increase in the ionicity of the Eu–O bond.

Fig. 2 (a) Excitation spectra (monitoring emission at 594 nm) of 5 at% Li⁺ co-doped YPO₄:Eu; the inset shows its excitation spectra on monitoring emissions at 430 and 450 nm. (b) Change in peak position of Eu–O CT band (left) and corresponding FWHM with Li⁺ ion concentration up to 10 at%. (c) Difference in electronegativity (χ(O²⁻) − χ(Eu³⁺)) vs. Li⁺ ion concentration on Li⁺ co-doped YPO₄:Eu.

To understand the shifting of the Eu–O CT band, the optical electronegativity of the ligand ion is studied using the Jørgensen empirical equation of charge transfer, which is expressed as

Δ_Eu–O = [χ(ligand) − χ(Eu³⁺ ion)]3 × 10⁴ cm⁻¹ (1)

where Δ_Eu–O is the peak position of the Eu–O CT band and χ(ligand) and χ(Eu³⁺) are the optical electronegativity values of the ligand ion (O²⁻) and the Eu³⁺ ion, respectively.²¹ However, the optical electronegativity of the ion changes in different crystal field environments. From Fig. 2(b), the peak position values (Δ_Eu–O) of the Eu–O CT band are 238.5 nm (41 929 cm⁻¹), 238.2 nm (41 980 cm⁻¹), 237.5 nm (42 100 cm⁻¹), 237.3 nm (41 143 cm⁻¹) and 235.6 nm (42 441 cm⁻¹) for Li⁺ = 0, 3, 5, 7 and 10 at%, respectively. Then, we have calculated the difference in electronegativity (χ(O²⁻) − χ(Eu³⁺)), and the values are shown in Fig. 2(c). The difference in electronegativity increases with increasing Li⁺, indicating an increase in ionic strength. Y³⁺ is surrounded by 8 O²⁻ ions to form dodecahedron (YO₈) in a tetragonal YPO₄ structure and its site symmetry of Y³⁺ or Eu³⁺ is D_2d. The ionic size of Eu³⁺ in 8 co-ordination number (CN) is 1.07 Å, whereas Li⁺ has 0.59 Å (4CN) or 0.76 Å (6CN). Due to differences in ionic sizes, co-ordination numbers and charges of the Eu³⁺ and Li⁺ ions arise; the environment of Eu³⁺ is disturbed after the incorporation of Li⁺. It is expected to a create more positive charge in the Li⁺ co-doped system because Li⁺ ions occupy the interstitial sites of the YPO₄ structure, which is confirmed by XRD study. This will induce more ionicity in the Eu–O CT band.

Fig. 3(a–c) show the emission spectra of Li⁺ co-doped YPO₄:Eu (Li⁺ = 0, 5 and 10 at%) under different excitation wavelengths. A broad band emission in 370–570 nm (centered at ∼430 nm) with a FWHM of ∼75 nm is observed, and this is related to intermediate band emission within the band gap of the YPO₄ host. To check the intermediate band arising from the sample, we have carried out excitation spectra of 5 at% Li⁺ co-doped YPO₄:Eu on monitoring emission at 430 and 450 nm (inset of Fig. 2(a)). It is found that there is an increase in absorption intensity from 300 to 400 nm. This is related to intermediate bands present within a large band gap of YPO₄ (E_g = 8.6 eV). Notably, there is no absorption band in 340–350 nm in the case of the pure Eu³⁺ ion.²² There was a report on intermediate band observations (absorption and emission) in LaPO₄:Eu.²³

Fig. 3 Emission spectra of (a) Li⁺ = 0, (b) Li⁺ = 5 and (c) Li⁺ = 10 at% co-doped YPO₄:Eu samples under different excitation wavelengths. (d) The ratio of the integrated area of the total magnetic and electric dipole transitions of Eu³⁺ (⁵D₀ → ⁷F₁ and ⁵D₀ → ⁷F₂), (I₀₁ + I₀₂) to intermediate host emission (I_H) vs. Li⁺ ion concentrations. Inset of (c) shows the digital photograph of the sample under 266 nm laser excitation.

The intensity of the intermediate band emission peak increases with excitation wavelength from 240 to 399 nm. Actually, the range of intermediate band emission is 360–570 nm, which can be seen after excitation at 340 nm, where there is no absorption peak for Eu³⁺.²² This range covers the main absorption peaks of Eu³⁺ (399 nm, ⁷F₀ → ⁵L₆ and 468 nm, ⁷F₀ → ⁵D₂). The emission spectrum of each sample consists of two sharp peaks at ∼594 and 615 nm, which correspond to the magnetic dipole (⁵D₀ → ⁷F₁) and electric dipole transitions (⁵D₀ → ⁷F₂), respectively. There is a strong resonance energy transfer from the intermediate band to Eu³⁺ and thereby the population of electrons/photons in the excited state of Eu³⁺ is high. This energy transfer process is non-radiative through a dipole–dipole interaction (known as Förster or Dexter-type mechanism).²⁴ The inset of Fig. 3(c) shows the digital photograph of 10 at% Li⁺ co-doped YPO₄:Eu under 266 nm laser excitation. A bright red colour is observed.

Fig. 3(d) shows the ratio of the integrated areas of electric and magnetic dipole transitions of Eu³⁺ to intermediate host emission with Li⁺ ion concentrations. The ratio increases slowly up to 5 at% Li⁺ and then increases steeply with Li⁺ ion concentration. Li⁺ co-doping decreases the intermediate band emission intensity and thereby enhances the Eu³⁺ emission by energy transfer from the intermediate band to Eu³⁺. This is due to improved crystallinity, which enhances the radiative rate of Eu³⁺ transitions on Li⁺ co-doped YPO₄:Eu. This is also supported by XRD and XPS results.

Fig. 4(a) shows the expanded emission spectra of Li⁺ (0, 3, 5, 7 and 10 at%) co-doped YPO₄:Eu in 550–650 nm under an excitation wavelength of 240 nm (Eu–O CT band excitation). The magnetic dipole transition intensity (⁵D₀ → ⁷F₁) is dominant over the electric dipole transition (⁵D₀ → ⁷F₂), suggesting a higher occupancy of Eu³⁺ in a symmetric environment. The emission spectrum excited at 250 nm shows an almost similar trend to that at 240 nm excitation (Fig. S2(a), ESI†). The emission spectrum monitored under 260 nm excitation shows different behavior as compared to both the 240 and 250 nm excitations (Fig. 4(b)). It is found that the intensity of the electric dipole transition is more than the magnetic dipole transition. The electrical-dipole transition is a hypersensitive transition, which is allowed only on the condition that Eu³⁺ ion occupies a site without an inversion center and is very sensitive to the local environment around the Eu³⁺ ion.^19,25 To understand the different behaviors of the electric and magnetic dipole transition intensities, emission spectra are recorded at excitation wavelengths of 270, 280, 290, 300, 340, 350 and 399 nm. We have used a 375 nm cut-off filter for recording emission spectra for all excitations except 399 nm (no filter is used). Excitations at 270 and 280 nm show a similar nature to that at 260 nm excitation (Fig. S2(b) and Fig. S3(a), ESI†). Excitations at 290 to 399 nm (Fig. 4(c), Fig. 5) show a similar nature to those at 240 and 250 nm excitations (Fig. S3(b), ESI†). The asymmetric ratio (A₂₁), which is defined as the ratio of integrated intensities of the electric dipole transition to the magnetic dipole transition is used to understand the luminescence behaviour of the Eu³⁺ ion under different excitation wavelengths (Table 1). There are changes in ratio from less than one to more than one on different excitations or doping concentrations of Li⁺. The above luminescence behaviour causes confusion for analysis of the nature of the Eu³⁺ environment. The reported A₂₁ values are close to ∼1 in Eu³⁺ doped YPO₄ and ∼4–10 in Eu³⁺ doped CaMoO₄.^16,17

Fig. 4 Emission spectra of Li⁺ co-doped YPO₄:Eu (Li⁺ = 0, 3, 5, 7 and 10 at%) in 550–650 nm after different excitations at (a) 240, (b) 260 and (c) 290 nm.

Fig. 5 Emission spectra of Li⁺ co-doped YPO₄:Eu under (a) 340, (b) 350 and (c) 399 nm excitations.

Table 1 A ₂₁ values of Li⁺ co-doped YPO₄:Eu (Li⁺ = 0, 3, 5, 7 and 10 at%) under 240, 250, 260, 270, 280, 290, 300, 340, 350 and 399 nm

Li^+ (at%)	A 21 calculated from different excitation wavelengths (nm)
	240	250	260	270	280	290	300	340	350	399
0	0.56	0.68	0.98	1.26	1.57	0.95	0.77	1.81	1.11	0.53
3	0.49	0.54	0.69	0.83	0.99	0.53	0.50	0.90	0.82	0.66
5	0.47	0.57	0.75	0.85	1.05	0.58	0.49	1.06	0.79	0.45
7	0.45	0.52	0.75	0.75	1.09	0.49	0.43	1.00	0.73	0.49
10	0.47	0.90	1.84	1.74	1.44	0.77	0.60	0.74	0.72	0.47

---

When we look into the crystal symmetry of YO₈ or EuO₈, it does not have an inversion symmetry element (i). The environment of Y³⁺ or Eu³⁺ is asymmetric. According to Judd–Ofelt theory, the electric dipole transition intensity should be more than the magnetic dipole transition in such an asymmetric environment. The variation of magnetic and electric dipole transition intensities with incoming excitation wavelength can come from the following possibilities:²⁶

1. Overlapping interaction parameters within the ground and excited states.

2. Exchange interaction among atoms/ions or groups of atoms (e.g. PO₄).

3. Density of the incoming photons.

The first point indicates the ground state and the excited state of the Eu³⁺ and usually the gap between them is fixed because of the f–f transitions. The second point includes the neighbouring atoms or groups around the Y/EuO₈. In this tetragonal structure, Y/EuO₈ is connected with the PO₄ group. The PO₄ has S₄ symmetry, which has a high polarizing ability because the ionic size of P⁵⁺ is too small (∼0.17 Å, 4CN). This influences the electric and magnetic dipole transition probability of Eu³⁺. The third point refers to the density of the oscillator state around the energy of incoming photons, which influences the transition probability of Eu³⁺ with the wavelength of the incoming light/photon. This also suggests that if the incoming light falls in the VUV or UV or visible regions, the electric and magnetic dipole transitions of Eu³⁺ will change. In VUV excitation, the YPO₄ host absorption will occur and the absorbed energy will be transferred to the excited state of Eu³⁺. In the UV excitation, Eu–O CT or intermediate band absorption occur and this absorbed energy can be transferred to the excited state of Eu³⁺. In the case of visible absorptions, direct Eu³⁺ excitation occurs and the possibility of energy transfer among the Eu³⁺ ions takes place through the diffusion or non radiative process. It is to be noted that the variation of electric and magnetic dipole transitions depends on the intensity of the incoming light at the same wavelength. The population of higher excited levels of Eu³⁺ (⁵D_j=0–3) can be achieved in part via excited Eu³⁺, using a laser as an excitation source with high power.^2,3,27 With increasing intensity of the incoming light, 2 or 3 photon absorption can occur in some systems.²⁷

The intensity of the ⁵D₀ → ⁷F₀ transition is significant in the case of highly Li⁺ co-doped YPO₄:Eu (Li⁺ = 10 at%). Interestingly, the peak position and the intensity vary with excitation wavelength. This observation is unusual to our knowledge. Below 250 nm or above 300 nm excitation, its intensity is very weak (Fig. 6a). After base line correction (Fig. 6b), the peak positions are found to be 579.3 and 577.3 nm at 250 and 300 nm respectively, suggesting a blue shift in the peak position with increasing excitation wavelength (Fig. 6c). The FWHM (Fig. 6c) and integrated intensity (Fig. 6d) increase with increasing excitation wavelength up to 290 nm. Observation of this peak under high Li⁺ co-doping indicates that the forbidden transition ⁵D₀ → ⁷F₀ can be relaxed by distortion of the Eu³⁺ environment (i.e. variation in the crystal field) and j = 0–0 is no longer pure. Instead, j–j mixing occurs. Such j–j mixing is dependent on excitation wavelength as well as crystal field. The crystal field of Eu³⁺ is influenced by the Li⁺ ions, which occupy interstitial sites of YPO₄. Thus, this transition is observed for a high Li⁺ co-doping system. The peak shifting, as well as the intensity variation, can be explained by the above possibilities (1–3). In addition, the thermal population of electrons in the j = 1, 2, 3 levels of ⁷F_j (ground state) may be one factor in such a transition at 577–579 nm.

Fig. 6 Expansion of luminescence spectra of 10 at% Li⁺ co-doped YPO₄:Eu sample between 570–585 nm under different excitation wavelengths (a) before and (b) after baseline correction. (c) Variations of peak ⁵D₀ → ⁷F₀ position (left) and FWHM (right) with excitation wavelength (250–300 nm) and (d) integrated intensity of the ⁵D₀ → ⁷F₀ transition vs. excitation wavelength from 250 to 300 nm.

The intensity of the ⁵D₀ → ⁷F₀ (∼576 nm) transition is very weak compared to both the ⁵D₀ → ⁷F₁ and ⁵D₀ → ⁷F₂ transitions. It is found that the intensity of the ⁵D₀ → ⁷F₀ transition significantly increases with Li⁺ ion concentration as well as change of excitation wavelength from 250 to 300 nm. Due to the non-availability of the crystal field splitting at the ⁵D₀ and/or ⁷F₀ levels (i.e. non-degenerate), this transition can be effectively used for the bond environment around the Eu³⁺ ion. The shift in peak position (blue shift) of the ⁵D₀ → ⁷F₀ transition with increase in excitation wavelength indicates the formation of the (Y/Eu)³⁺–O²⁻–Li⁺ type structure in the host lattice. The FWHM of the ⁵D₀ → ⁷F₀ transition peak increases with increase of excitation wavelength up to 280 nm and is almost constant with a further increase in excitation wavelength up to 300 nm. The FWHM of the ⁵D₀ → ⁷F₀ transition peak is broad and its intensity is very weak at excitation wavelengths of 240 and above 340 nm; thus it is difficult to calculate the exact peak position. Excitations at 250 to 300 nm are used for analysis. Does the excitation wavelength influence the intensity of the ⁵D₀ → ⁷F₀ transition? To the best of our knowledge this type of question has not been properly answered in the literature. There were reports in the red shift of the ⁵D₀ → ⁷F₀ transition peak position, their corresponding FWHM and intensity with increasing site selective excitation wavelengths (576 to 577.5 nm or 461.8 to 467.1 nm) from a laser source in a glass system,^28,29 which are opposite to our study. In TiO₂–SiO₂ glasses, the position of the ⁵D₀ → ⁷F₀ transition does not change with excitation wavelengths, and an even lower photon energy was used to excite Eu³⁺ for emitted light production of higher energy (⁵D₀ → ⁷F₀) partly due to population at the ⁵D₀ level.³ In ref. 2, 3, 28 and 29, the excitation source is a laser and thus site selective excitation can be performed. The differences from other reported cases in this study are that only particular excitation wavelengths can show significant intensity for the ⁵D₀ → ⁷F₀ transition and the source is a xenon lamp. In this work we focus on this problem originating from lattice distortion from the host, crystal field and/or polarization effect of the PO₄ group.

The ⁵D₀ → ⁷F₀ transition is forbidden according to Judd–Ofelt theory. However, its significant intensity observed in many compounds is due to either parity mixing (e.g. d–f) or j–j mixing at the excited state (⁵D_j=0,1,3) or the ground state (F_j=0,1,2,3,4) due to thermal population. The parity mixing (e.g. d–f) is ruled out because of the high energy gap between them. The j–j mixing at the excited state (⁵D_j=0,1,3) is not feasible because of the significant energy gap. There is a possibility of j–j mixing at the ground state of F_j=0,1,2,3,4 due to thermal population. ⁵D₀ → ⁷F_1,3 magnetic dipole transitions are almost independent of local-crystal field strength. The mixing of the higher j = 4, 6 with j = 0 is lower because the higher energy level of j = 4, 6 from j = 0 and also their emission intensities are weak compared to j = 2. The most probable is from mixing j = 2 with j = 0 at the ground state of F_j, which gives rise to a significant transition probability (i.e. intensity) of the ⁵D₀ → ⁷F₀ transition through the second order term of the local crystal field. The relationship between the intensities of ⁵D₀ → ⁷F₀ and the ⁵D₀ → ⁷F₂ is given by

where B₂₀ is the second order crystal field parameter and Δ₂₀ is the energy separation between the ⁷F₂ and ⁷F₀ levels (cm⁻¹).³⁰ Generally, intensity of ⁵D₀ → ⁷F₀ is related to B₂₀, whereas ⁵D₀ → ⁷F₂ is independent of this. In order to calculate B₂₀, the intensities of ⁵D₀ → ⁷F₀ (570–583 nm), ⁵D₀ → ⁷F₁ magnetic dipole (583–603 nm) and ⁵D₀ → ⁷F₂ electric dipole (603–630 nm) are calculated using the Gaussian function after the baseline correction. Typical fitting to a 10 at% Li⁺ co-doping system is shown in Fig. S4 (ESI†). The excitation wavelength is at 280 nm. The fitting parameters are given in the figure itself. The increase in value of A₀₂ (⁵D₀ → ⁷F₀/⁵D₀ → ⁷F₂) is found (i.e. 0.0345 to 0.0647 at 290 nm excitation for 0 and 10 at% Li⁺ co-doping) with Li⁺ ion concentration indicating the larger J-mixing because the Li⁺ ions present in the interstitial sites influence the crystal field.³¹ The A₀₂ value for 10 at% Li⁺ co-doping varies from 0.0136 to 0.0759 for excitation wavelength 250 to 300 nm.

Ray et al.³⁰ considered the Δ₂₀ value to be constant (900 cm⁻¹) in the case of the Y₂O₃ host system. Theoretically, the ground state energy (⁷F₀) should not change at a particular temperature. However, the shifting of the ⁵D₀ → ⁷F₀ peak position with excitation wavelength in this study suggests that Δ₂₀ will vary for each excitation wavelength from 250 to 300 nm because of variation of the crystal field interaction with incoming light. This was not considered in earlier studies. Fig. 7 shows the variation of second order crystal field parameter B₂₀ (right) and A₀₂ (left) with excitation wavelengths ranging from 250 to 300 nm after the consideration of changes in Δ₂₀ for the 10 at% Li⁺ co-doping system. B₂₀ is purely related to the bond environment around the Eu³⁺ ion. It is found that B₂₀ increases proportionally with exciting wavelength from 250 to 300 nm (4.96–4.13 eV). This implies that the increased excitation wavelength from 250 to 300 nm induces more perturbation from the second nearest bonds to the Eu³⁺ ion. Typical B₂₀ values for 10 at% Li⁺ co-doped YPO₄:Eu are found to be 542 and 1365 cm⁻¹ for the 250 to 300 nm excitations, respectively. In the case of Ca(PO₃)₂:Eu³⁺ glass, the high value of B₂₀ of 1600 cm⁻¹ was reported.³² In our study, the transition probability/intensity of ⁵D₀ → ⁷F₀ increases with Li⁺ concentration at a particular excitation wavelength, indicating involvement of the second order crystal field parameter on the Li⁺ co-doped structure. There were reports on the decrease of B₂₀ from ∼600 to 280 cm⁻¹ with increasing excitation wavelength (nm) in three different phosphate glasses,²⁸ however, the trend is opposite to that in this study. In these phosphate glasses, it was suggested that the larger the value of B₂₀, the closer the coordinating ligands to the Eu³⁺ ion. It was also reported on variation of B₂₀ from 950 to 1350 cm⁻¹ on the same excitation wavelength (488 nm from laser source) with increasing Eu³⁺ concentration in Y₂O₃, but the Δ₂₀ value is fixed at 900 cm⁻¹, while B₂₀ is calculated according to eqn (2).³⁰ In this study, the B₂₀ value is high, indicating that the higher perturbation crystal field can be associated in addition to the second order parameter.²⁸

Fig. 7 Variation of I(⁵D₀ → ⁷F₀)/I(⁵D₀ → ⁷F₂) (left) and crystal field parameter B₂₀ (right) of 10 at% Li⁺ co-doped YPO₄:Eu under different excitations.

3.2.2 Luminescence decay study. The luminescence decay curves of the ⁵D₀ level (594 nm) have been measured for Li⁺ co-doped YPO₄:Eu (Li⁺ = 0, 3 and 10 at%). Fig. 8 shows the decay curves for ⁵D₀ → ⁷F₂ emission of these samples when excited at 468 nm laser excitation. The decay curves are not well fitted by the mono-exponential equation (I = I₀exp(−t/τ), where I and I₀ are the intensity at time t and at 0 s, respectively, and τ is the lifetime). Fig. S5 (ESI†) shows the mono-exponential curve fitting to the decay data of 3 at% Li⁺ doped YPO₄:Eu. The fitted parameters are given in the figure itself with χ² = 3.401. The fitting behaviour can be clearly understood by plotting ln(I) vs. t, which is shown in Fig. S5 (ESI†). The fitting parameters are obtained after mono exponential decay equations are given in Table S1 (ESI†). The fitted straight line does not match with the decay data points. This suggests that the environment of the Eu³⁺ ions in the lattice are not the same in different positions.

Fig. 8 (a) Lifetime decay spectra of Li⁺ doped YPO₄:Eu (Li⁺ = 0, 3 and 10 at%) under 468 nm laser excitation (λ_em = 594 nm) and (b) bi-exponential fittings to luminescence decay data of 3 at% Li⁺ co-doped YPO₄:Eu. Fitting parameters are shown in their figures.

All decay data are fitted by using bi-exponential decay equation, which is expressed as

I = I₁e^{(−τ/τ₁)} + I₂e^{(−τ/τ₂)} (3)

where I₁ and I₂ are the intensities at different time intervals and τ₁ and τ₂ their corresponding lifetimes. The bi-exponential fittings to luminescence decay data of 3 at% Li⁺ co-doped YPO₄:Eu are shown in Fig. 8(b). The fitting parameters are given in the figure itself. The average lifetime can be calculated using the equation³³The fitting parameters are obtained after the bi-exponential decay equations given in Table 2. The average lifetime values for 0, 3 and 10 at% Li⁺ co-doped samples are found to be 1.88, 2.07 and 1.49 ms, respectively. These values are in good agreement with reported values for Eu³⁺ emission (2–3 ms).¹⁸ It is concluded that decay time increases with Li⁺ ion concentration up to 3 at% and then decreases. From all decay curves it is found that ∼60% of the emission intensity drops in ∼0.4–0.7 ms and the remaining ∼40% drops in ∼3–4 ms. This fast decay process (τ₁) may be related to Eu³⁺ ions, which are near the defects or surfaces of particles, whereas the slow decay process (τ₂) is related to the Eu³⁺ ions available at the core of the particle. This is based on the core–shell model. On Li⁺ co-doping up to 3 at%, the amount of τ₁ decreases because the crystallite/particle size increases and thus the non-radiative rate from surface defects decreases; whereas the opposite effect is found in the case of τ₂ because of increased crystallinity. Similar explanations were reported in our previous studies.³⁴ Goldys et al.³⁵ reported the different decay processes (fast and slow) in Eu³⁺ doped Gd₂O₃ using time resolved spectroscopy. The effect of OH/H₂O over the surface of the particle is not considered in this study because it is unavoidable in all samples heated at 900 °C at ambient atmosphere.

Table 2 Decay parameters obtained in bi-exponential fitting

Li^+ (at%)	I 1 (%)	τ 1 (ms)	I 2 (%)	τ 2 (ms)	τ av ^a (ms)	χ ^2 ^b
a b , χ^2 is closeness of fit, wk is the weighting factor for data points (), Xk is the calculated lifetime and Fk is the measured lifetime data.
0	64	0.673	36	4.028	1.88	1.559
3	57	0.564	43	4.073	2.07	1.451
10	59	0.395	41	3.068	1.49	1.553

---

4. Conclusions

Luminescence properties of Li⁺ co-doped YPO₄:Eu are discussed in this article (Li⁺ = 0, 3, 5, 7 and 10 at%). The XPS study confirms the lack of the core binding energy peak at ∼127.1 eV that corresponds to Eu²⁺(4d_5/2). It is concluded that the Eu³⁺ → Eu²⁺ conversion does not occur with Li⁺ co-doping. Eu–O CTB shows a blue shift upon Li⁺ co-doping due to the increase in ionicity. The addition of Li⁺ significantly enhances the luminescence intensity of Eu³⁺ due to improved energy transfer from the intermediate band of the host to Eu³⁺. Moreover, we observe some unusual behavior in electric (⁵D₀ → ⁷F₂) and magnetic (⁵D₀ → ⁷F₁) dipole transitions under excitation wavelengths from 240 to 399 nm. The intensity of the ⁵D₀ → ⁷F₂ transition is more than that of the ⁵D₀ → ⁷F₁ transition for excitations at 260–280 nm, and the reverse occurs for excitations at 240, 250 and 290–399 nm. We observe the ⁵D₀ → ⁷F₀ optical transition having maximum intensity at 290 nm excitation. On increasing the excitation wavelength from 250 to 300 nm, the ⁵D₀ → ⁷F₀ transition shifts to lower wavelengths (blue shift), which is opposite to the reported one. The second order crystal field parameter (B₂₀) is used as a probe to understand the crystal field environment around Eu³⁺. The B₂₀ value is found to be in the range ∼542–1365 cm⁻¹. The average lifetime value (τ_av) is found to be 2.07 ms for the 3 at% Li⁺ co-doped YPO₄:Eu sample.

Acknowledgements

This work has been sponsored by the University Grants Commission (UGC) under the D. S. Kothari Postdoctoral Fellowship Scheme (No.F.4-2/2006(BSR)/13-309/2008(BSR)) to Dr A. K. Parchur. Authors thank Dr T. Mukherjee, Chemistry Group and Dr D. Das, R. K. Vatsa and Dr V. Sudarsan, Chemistry Division, BARC for their support and encouragement during this work and Dr R. K. Sahu, NML Jamshedpur for providing XPS data.

References

1. (a) J. W. Stouwdam, G. A. Hebbink, J. Huskens and F. C. J. M. van Veggel, Chem. Mater., 2003, 15, 4604; (b) A. H. Krumpel, A. J. J. Bos, A. Bessière, E. van der Kolk and P. Dorenbos, Phys. Rev. B: Condens. Matter Mater. Phys., 2009, 80, 085103; (c) N. K. Sahu, R. S. Ningthoujam and D. Bahadur, J. Appl. Phys., 2012, 112, 014306.
2. (a) T. Kushida and M. Tanaka, Phys. Rev. B: Condens. Matter, 2002, 65, 195118; (b) M. Tanaka, G. Nishimura and T. Kushida, Phys. Rev. B: Condens. Matter, 2004, 49, 16917.
3. H. You and M. Nogami, J. Phys. Chem. B, 2004, 108, 12003.
4. C. Brecher, H. Samelson, R. Riley and A. J. Lempicki, Chem. Phys., 1968, 49, 3303.
5. R. Balakrishniah, D. W. Kim, S. S. Yi, S. H. Kim, K. Jang, H. S. Lee, B. K. Moon and J. H. Jeong, NSTI-Nanotech., 2010, 3, 773.
6. R. D. Shannon, Acta Crystallogr., Sect. A: Cryst. Phys., Diffr., Theor. Gen. Crystallogr., 1976, 32, 751.
7. J. Liu, H. Lian and C. Shi, Opt. Mater., 2007, 29, 1591.
8. J. Huang, J. Xu, H. Luo, X. Yu and Y. Li, Inorg. Chem., 2011, 50, 11487.
9. S. Neeraj, N. Kijima and A. K. Cheetham, Chem. Phys. Lett., 2004, 387, 2.
10. C. Guo, S. Wang, T. Chen, L. Luan and Y. Xu, Appl. Phys. A: Mater. Sci. Process., 2009, 94, 365.
11. H. Lai, H. Yang, T. Chunyan and X. Yang, Phys. Status Solidi A, 2007, 204, 1178.
12. (a) A. K. Parchur, A. I. Prasad, S. B. Rai, R. Tiwari, R. K. Sahu, G. S. Okram, R. A. Singh and R. S. Ningthoujam, AIP Adv., 2012, 2, 32119; (b) A. K. Parchur, G. S. Okram, R. A. Singh, R. Tewari, L. Pradhan, R. K. Vatsa and R. S. Ningthoujam, AIP Conf. Proc., 2010, 1313, 391.
13. (a) A. K. Parchur and R. S. Ningthoujam, Dalton Trans., 2011, 40, 7590; (b) A. K. Parchur, A. I. Prasad, A. A. Ansari, S. B. Rai and R. S. Ningthoujam, Dalton Trans., 2012, 41, 11032.
14. A. K. Parchur and R. S. Ningthoujam, RSC Adv., 2012 10.1039/c2ra20763j.
15. R. S. Ningthoujam, V. Sudarsan, R. K. Vatsa, R. M. Kadam, Jagannath and A. Gupta, J. Alloys Compd., 2009, 486, 864.
16. M. N. Luwang, R. S. Ningthoujam, Jagannath, S. K. Srivastava and R. K. Vatsa, J. Am. Chem. Soc., 2010, 132, 2759.
17. A. K. Parchur, R. S. Ningthoujam, S. B. Rai, G. S. Okram, R. A. Singh, M. Tyagi, S. C. Gadkari, R. Tewari and R. K. Vatsa, Dalton Trans., 2011, 40, 7595.
18. M. N. Luwang, R. S. Ningthoujam, S. K. Srivastava and R. K. Vatsa, J. Am. Chem. Soc., 2011, 133, 2998.
19. B. R. Judd, Phys. Rev., 1962, 127, 750.
20. R. X. Yan, X. M. Sun, X. Wang, Q. Peng and Y. D. Li, Chem.–Eur. J., 2005, 11, 2183.
21. C. K. Jørgensen, Modern Aspects of Ligand-Field Theory, North Holland, Amsterdam, 1971.
22. R. S. Ningthoujam, V. Sudarsan and S. K. Kulshreshtha, J. Lumin., 2007, 127, 747.
23. L. Li, Y. Su and G. Li, J. Mater. Chem., 2010, 20, 459.
24. (a) T. Förster, Ann. Phys., 1948, 2, 55; (b) D. L. Dexter, J. Chem. Phys., 1953, 21, 836.
25. G. S. Ofelt, J. Chem. Phys., 1962, 37, 511.
26. D. Curie, Luminescence in Crystals, John Wiley & Sons Inc. London, 1st edn, 1960.
27. F. Auzel, Chem. Rev., 2004, 104, 139.
28. J. A. Capobianco, P. P. Proulx, M. Bettinelli and F. Negrisolo, Phys. Rev. B: Condens. Matter, 1990, 42, 5936.
29. H. Wen, G. Jia, C. K. Duanw and P. A. Tanner, Phys. Chem. Chem. Phys., 2010, 12, 9933.
30. S. Ray, P. Pramanik, A. Singha and A. Roya, J. Appl. Phys., 2005, 97, 094312.
31. J. C. Boyer, F. Vetrone, J. A. Capobianco, A. Speghini and M. Bettinelli, J. Phys. Chem. B, 2004, 108, 20137.
32. G. Nishimura and T. Kushida, Phys. Rev. B, 2005, 37, 9075.
33. V. Sudarsan, F. C. J. M. van Veggel, R. A. Herring and M. Raudseppc, J. Mater. Chem., 2005, 15, 1332.
34. (a) N. Yaiphaba, R. S. Ningthoujam, N. S. Singh, R. K. Vatsa, N. R. Singh, S. Dhara, N. L. Misra and R. Tewari, J. Appl. Phys., 2010, 107, 034301; (b) N. S. Singh, R. S. Ningthoujam, M. N. Luwang, S. D. Singh and R. K. Vatsa, Chem. Phys. Lett., 2009, 480, 237; (c) N. S. Singh, R. S. Ningthoujam, N. Yaiphaba, S. D. Singh and R. K. Vatsa, J. Appl. Phys., 2009, 105, 064303.
35. E. M. Goldys, K. D. Tomsia, S. Jinjun, D. Dosev, I. M. Kennedy, S. Yatsunenko and M. Godlewski, J. Am. Chem. Soc., 2006, 128, 14498.

---

Footnotes

† Electron Supplementary Information (ESI) available. See DOI: 10.1039/c2ra22144f

‡ This paper is dedicated to Professor N. S. Gajbhiye on the occasion of his 60th birthday.

---