On peculiarities of Eu³⁺ and Eu²⁺ luminescence in Sr₂GeO₄ host

Abstract

Powders of Sr₂GeO₄:Eu were prepared in air or forming gas and were investigated for their luminescent properties in the 20–300 K range of temperatures. Oxidizing conditions produced a material with the orange-red luminescence of Eu³⁺. As many as eight Eu³⁺ symmetry sites active in luminescence were found at low temperatures. Four of them had the expected quite regular characteristics and were assigned to the four Sr sites offered by the host lattice. In addition, another four Eu³⁺ sites producing the ⁵D₀ → ⁷F₀ emission at extraordinarily short wavelengths in the range of 574–577 nm were clearly revealed. Their CT excitation bands peaked at around 325 nm, which is an unusually low energy for the O²⁻ → Eu³⁺ CT transition. This unusual characteristic of Eu³⁺ luminescence was assigned to the presence of an interstitial oxygen not bound to Ge but strongly interacting with Eu³⁺ ions. The effect was justified using the Wybourne and Downer extension of the Judd–Ofelt theory. Reduced samples did not produce any Eu²⁺ emission at room temperature, but below about 70 K the 5d → 4f luminescence of the ion came into view with an intense appearance at 20 K. No indications were found of Eu²⁺ luminescence from a few sites. A possible explanation of this observation as well as mechanism of the Eu²⁺ luminescence quenching are discussed.

---

Introduction

Rare earth-activated Sr₂SiO₄ have been reported to be efficient luminescent materials.^1–5 The violet-blue emission of Ce³⁺ and the green or yellow-orange (depending on the crystal structure) luminescence of Eu²⁺ were even considered useful for white LEDs.^6–13 Thus, it is surprising that there is no literature on the luminescence of chemically quite similar Sr₂GeO₄:Eu. Despite the analogous chemical formulas, however, the germanate crystallizes in a different, orthorhombic crystal structure, space group Pbn2₁, offering as many as four different symmetry sites for Sr²⁺ ions.¹⁴ Both monoclinic and orthorhombic silicates, Sr₂SiO₄, offer just two Sr sites in their lattices.^15,16

The four different coordination spheres of Sr²⁺ sites in orthorhombic Sr₂GeO₄ are depicted in Fig. 1.¹⁴ Two of them are six-coordinated (Sr1 and Sr2) and the other two are eight-coordinated (Sr3 and Sr4). In Table 1 O²⁻–Sr²⁺ distances are listed.¹⁴ At first it is tempting to consider the two sites within each part (6- and 8-coordinated) similar in symmetry. However, closer analysis reveals that all four polyhedral sites are quite different and this should be mirrored by easily measurable differences in the position of their related f–f luminescence lines when substituted by rare earth luminescent ions. What the four sites have in common is their low local symmetries. Consequently, degeneracy of all the ^SL_J states with J ≠ 0 are expected to be lifted and much extended excitation and luminescence spectra with a longer reach are expected. Fig. 1 also presents a perspective view of the rather loosely packed unit cell.

Fig. 1 Coordination sphere of different sites for Sr²⁺ in orthorhombic Sr₂GeO₄. At the bottom a perspective view of the unit cell showing the rather loose packing of the structure is presented.

Table 1 Sr–O distance (Å) in orthorhombic Sr₂GeO₄ (ref. 14)

Sr1_CN6		Sr2_CN6		Sr3_CN8		Sr4_CN8
O3	2.44	O7	2.38	O8	2.50	O2	2.44
O5	2.65	O2	2.56	O6	2.54	O3	2.52
O8	2.65	O1	2.57	O1	2.57	O4	2.56
O1	2.67	O4	2.57	O4	2.58	O8	2.57
O6	2.69	O5	2.73	O7	2.63	O6	2.61
O2	2.77	O6	2.96	O5	2.70	O5	2.64
—	—	—	—	O3	2.76	O7	2.76
—	—	—	—	O2	2.92	O1	2.82

---

As we mentioned above, there is no literature data on the spectroscopic properties of a rare earth-activated Sr₂GeO₄ host. In the present paper we report on the photoluminescent properties of Sr₂GeO₄ doped with Eu³⁺ or Eu²⁺. In the Sr₂GeO₄ lattice the Eu²⁺ ion may readily substitute for Sr²⁺ because the ionic radii of Eu²⁺_CN6 (1.17 Å) and Eu²⁺_CN8 (1.25 Å) are almost identical with the sizes of Sr²⁺_CN6 (1.18 Å) and Sr²⁺_CN8 (1.26 Å), respectively.¹⁷ Yet, the ionic radii of Eu³⁺ ions are considerably (15–20%) smaller than Sr²⁺, reaching 0.95 Å for Eu³⁺_CN6 and 1.07 Å for Eu³⁺_CN8.¹⁷ On the other hand, the size of Eu³⁺ is more than double the size of Ge⁴⁺_CN4 (0.39 Å).¹⁷ Consequently, in practice, only the four Sr sites are supposed to be available for substitution by either Eu²⁺ or Eu³⁺ ions. Tracking this problem by means of photoluminescence spectroscopy is one of the purposes of this paper.

Experimental

Synthesis

Eu-doped Sr₂GeO₄ powders were prepared by the classic ceramic method. Concentrations of europium ions were 0.1, 0.5 and 2.0% mol. The ratios of the reagents were calculated assuming that europium substitutes for strontium in the host lattice. The substrates, SrCO₃ (Alfa Aesar 99.995%), GeO₂ (Alfa Aesar 99.9999%), Eu(NO₃)₃·6H₂O (europium nitrate was prepared by dissolving Eu₂O₃ – Stanford Materials 99.999% purity in HNO₃ – Chempur pure p.a.), were ground in an agate mortar and heated at 1500 °C for 3 h in air (for the Eu³⁺ emission) or in a reducing atmosphere of 25% H₂ + 75% N₂ forming gas also for 3 h to accomplish the Eu³⁺ → Eu²⁺ reduction.

Measurements

Photoluminescence (PL), excitation spectra (PLE) and decay kinetics (DEC) were measured on a FLS980-sm Fluorescence Spectrometer from Edinburgh Instruments Ltd. A 450 W continuous xenon arc lamp as an excitation source for PL and PLE spectra and a 60 W xenon flash lamp and Fianium WhiteLase Supercontinuum Fiber Laser for DEC measurements were used. TMS302-X Single Grating excitation and emission monochromators of 30 cm focal length were used and the luminescence light was recorded by a Hamamatsu R928P high-gain photomultiplier detector. Emission spectra were corrected for the recording system spectral efficiency and excitation spectra were corrected for the incident light intensity. For all samples X-ray diffraction measurements were performed with a D8 Advance X-ray Diffractometer from Bruker in the range of 2θ = 20–50° with a step of 2θ = 0.01608°. Ni-filtered Cu K_α1 radiation (λ = 1.540596 Å) was utilized. A Hitachi S-3400N Scanning Electron Microscope (SEM) was used to image the microstructure of the powders.

Results and discussion

Fig. 2 shows the measured XRD patterns of Sr₂GeO₄ powders with two different concentrations of europium activator – 0.1% and 2.0% – prepared in air or forming gas. The simulated pattern of the orthorhombic Sr₂GeO₄ is also given (ICSD#83345 (ref. 14)). Independent of the Eu concentration and preparation atmosphere, all powders are single phase orthorhombic materials. While the position of the measured and referenced diffraction lines match very well, some discrepancies in their intensity ratios are observed, especially around 2θ ∼ 31 degree. This may result from some preferential orientation of grains/crystallites, which may occur due to their significant aggregation and sintering even after pulverizing in a mortar (see Fig. 3).

Fig. 2 X-ray diffraction patterns of Sr₂GeO₄ containing 0.1% Eu or 2.0% of Eu and prepared in air and in reducing atmosphere of forming gas.

Fig. 3 SEM images of powders of Sr₂GeO₄:0.1% Eu (a) and Sr₂GeO₄:2.0% Eu (b) fabricated in air and Sr₂GeO₄:0.1% Eu (c) and Sr₂GeO₄:2.0% Eu (d) prepared in a reducing atmosphere of 25% H₂ + 75% N₂.

Fig. 3 presents SEM images of samples with different Eu concentrations and prepared in air or the forming gas. Fig. 3a and b depicts the morphology of Sr₂GeO₄ prepared in air atmosphere for low (0.1%) and high (2.0%) concentrations of activator, respectively. Alteration of the dopant content does not affect the morphology. Independent of the preparation atmosphere, aggregation of crystallites and their partial sintering is clearly observed. The size of the crystallites may be estimated at ∼2–3 μm for samples prepared in air. SEM images of samples containing 0.1% and 2% of Eu and prepared in the reducing atmosphere of the forming gas are presented in Fig. 3c and d, respectively.

The average sizes of their crystallites are clearly larger compared to powders prepared in air and reach roughly 4–10 μm. It is thus clear that the reducing atmosphere enhances the mass flow and thus facilitates the grain growth in Sr₂GeO₄, while the Eu content does not affect this process. Since treatment in the forming gas is expected to enhance the population of oxygen-vacancy sites, , in the powder, it may be hypothesized that their appearance stimulates the host atoms' mobility and thus more significant grain growth is observed.¹⁸

Photoluminescence of Sr₂GeO₄:Eu³⁺

Each sample of air-heated Sr₂GeO₄:Eu shows the characteristic red Eu³⁺ emission upon excitation around 250–270 nm. More orange luminescence appears when the excitation moves towards 300–350 nm. Fig. 4 presents typical PLE spectra for Sr₂GeO₄:Eu³⁺ which show characteristic emissions and PL spectra recorded at RT and at 20 K. The RT PLE spectrum of the 588.6 nm luminescence (Fig. 4a, blue line) is composed of an O²⁻ → Eu³⁺ charge transfer (CT) broad band with a maximum at about 260 nm and a set of narrow lines at longer wavelengths related to the f → f transitions of Eu³⁺.^19,20 Thus, this spectrum is quite characteristic of Eu³⁺ in many other oxides. However, monitoring the 574.2 nm emission we get a quite different PLE spectrum with an intense CT band peaking at 325 nm and only weak lines related to the f → f transitions of Eu³⁺ (Fig. 4a, red line). The O²⁻ → Eu³⁺ CT band located at such long wavelengths in oxides is quite peculiar. Yet, an analogous result was reported for α′-Sr₂SiO₄,^21–23 Ba₂SiO₄,²⁴ and La₂Si₂O₇,²⁵ which are not isostructural to Sr₂GeO₄. The corresponding PL spectra (Fig. 4b) cover the orange-red region from 570 nm to 720 nm and consist of a series of sharp lines due to the ⁵D₀ → ⁷F_J (J = 0, 1, 2, 3, 4) transitions of Eu³⁺. Nevertheless, excitation into the two CT absorption bands (260 nm and 325 nm) produces quite different PL emissions. Excitation at 260 nm produces a classic Eu³⁺ luminescence with the main components related to the ⁵D₀ → ⁷F₁ (∼590 nm), ⁵D₀ → ⁷F₂ (∼620 nm) and ⁵D₀ → ⁷F₄ (∼700 nm). The ⁵D₀ → ⁷F₀ transition could not be unambiguously resolved, having a very low intensity. Also, transition to the ⁷F₃ level produces only a very low-intensity emission around 650–660 nm. Under the 325 nm excitation the strongest luminescence line is observed at 574.2 nm, which is unusual for an Eu³⁺ ion. Different characteristics of the two Eu³⁺ emissions are further confirmed by PLE (Fig. 4c) and PL (Fig. 4d) spectra recorded at 20 K. Again, excitation around 325 nm produces Eu³⁺ luminescence with the dominant line positioned at 574.2 nm, while excitation at 257 nm gives quite a conventional spectral distribution for the Eu³⁺ emission. In both cases, the f–f related features in excitations and in emissions are much better resolved at 20 K than at RT, as expected.

Fig. 4 PLE spectra of the luminescence monitored at 574.2 nm and 588.6 nm recorded at RT (a) and at 20 K (c) and the PL spectra under 257 nm and 325 nm excitation recorded at RT (b) and at 20 K (d) for Sr₂GeO₄:0.1% Eu³⁺ prepared in air at 1500 °C for 3 h.

As a matter of fact, one might doubt whether the 574.2 nm emission results from the ⁵D₀ → ⁷F₀ transition. Yet, its decay time of τ = 1.0 ms (Fig. 5, red points) is identical to the decay times of the other emission features located at longer wavelengths and characteristic of the 325 nm excitation. This proves that all the observed emission lines upon 325 nm excitation result from the radiative relaxation of the same ⁵D₀ level of the Eu³⁺ ion. The decay time of the characteristic luminescence features upon excitation around 260 nm reaches τ = 2.0 ms, as is seen in Fig. 5 (blue trace).

Fig. 5 20 K DEC traces of Sr₂GeO₄:0.1% Eu³⁺ prepared in air at 1500 °C. Decay traces were measured for two different lines: λ_exc = 325 nm λ_em = 574.2 nm (red dots) and λ_exc = 257 nm λ_em = 618.8 nm (blue dots).

The above data disclosed two, spectroscopically very different, Eu³⁺ ions. Yet, the host offers, as we discussed, as many as four symmetry sites for the dopant. A closer look at the low-temperature emissions (Fig. 4d) makes it clear that the ⁵D₀ → ⁷F₁ transition upon 257 nm excitation produces more lines than might result for one Eu³⁺ site. Furthermore, the ⁵D₀ → ⁷F₀ emission feature around 574.2 nm (upon 325 nm excitation) is also split into at least 2 lines. Thus, despite the decay times of the various emissions giving only two different values (see Fig. 5), the low-temperature PL spectra actually indicate the presence of more emitting centers. Next, low-temperature and higher-resolution experiments will shed more light on that problem.

In the already-mentioned silicates, α-Sr₂SiO₄ (ref. 21) and Ba₂SiO₄,²⁴ La₂Si₂O₇,²⁵ similarly peculiar Eu³⁺ luminescent lines around 574 nm, with all the characteristics of the ⁵D₀ → ⁷F₀ emission, were recorded. Their abnormally high energy was postulated to result from the presence of an extra interstitial O²⁻ ligands not bound to silicon. However, other authors claimed that this ⁵D₀ → ⁷F₀ luminescence is merely related to Eu³⁺ located in a different crystallographic site.^22,23 However, they did not offer a clear explanation as to why the ⁵D₀ → ⁷F₀ emission should appear at such short wavelengths.

⁵D₀ → ⁷F₀ luminescence around 574–576 nm was also reported for La_0:95Eu_0:05Ca₄(PO₄)₃O, whose crystallographic structure may be derived from an apatite structure.²⁶ In an analogous composition with Bi replacing La, the high-energy ⁵D₀ → ⁷F₀ luminescence was not present. It was concluded that the different electronic structure of Bi³⁺ vs. La³⁺ in the former reduced the covalent character of the Eu³⁺–O²⁻ bond with oxygen not bound to P. Furthermore, in²⁷ it was shown that in La₃SbO₇:Eu such a short-wavelength ⁵D₀ → ⁷F₀ luminescence is also observed while in Gd₃SbO₇:Eu it was not. Crystallographic considerations led to the conclusion that the more polarizable environment in the La-composition was responsible for that effect. Also in ref. 28 a narrow line around 575 nm is reported, although it is not discussed in depth. This may be attained when an interstitial oxygen atom is bound only to the emitting ion and/or the Eu³⁺ enters the site with a rather loose arrangement of ligands. In fact, both conditions are fulfilled in the case discussed here of Sr₂GeO₄.

To proceed with our analysis of the origin of the different Eu³⁺ emitting ions in Sr₂GeO₄, we performed additional more precise, higher-resolution experiments at 20 K to get a more comprehensive picture of that problem. The high-resolution experiments were performed for the whole range of wavelengths in which the various emissions appeared (570–720 nm) and exploiting 350–550 nm excitation range of the f–f transitions. But, in the following part of the paper only the most relevant fractions, where the differences are most clearly seen, are presented.

Fig. 6a shows PLE spectra of various lines related to the ⁵D₀ → ⁷F₁ transitions and corresponding PL spectra (Fig. 6b). Table 2 lists the characteristic excitation and emission lines and should be helpful in further considerations. Note that only lines unambiguously assigned to different sites are listed. In both PLE and PL spectra the features overlap. Despite the complexity it is clear from Fig. 6 that as many as four different excitation and emission spectra can be distinguished in the higher-resolution experiments at 20 K. Due to very low local symmetries of the emitting centers, we did not succeed in assigning the PL and PLE spectra to specific Eu symmetry sites. We do not exclude the possibility that it might be achievable with the help of computational chemistry.

Fig. 6 Site-selective PLE (a) and PL (b) spectra of Sr₂GeO₄:0.1% Eu³⁺ recorded at 20 K. PLE spectra were measured for four emission lines: 586.0 nm (black), 588.5 nm (green), 591.9 nm (blue) and 594.1 nm (red) – marked by circles (a). For PL spectra four selective excitation lines were used: 392.2 nm (red), 393.0 nm (black), 393.3 nm (green) and 393.7 nm (blue) – marked by circles (b).

Table 2 Line positions on the excitation spectra for four different Eu³⁺ sites and corresponding characteristic emission lines

	Excitation (nm)	Characteristic emission lines (nm)
Regular emission	392.2	583.6; 584.1; 594.1
	393.0	586.0; 593.2
	393.3	588.6; 591.0
	393.7	587.5; 589.0; 591.7
Irregular – emission	460.2	574.25
	460.7	574.80
	461.0	575.45
	461.9	576.55

---

Let us stress that they are all positioned within the typical range of wavelengths characteristic for Eu³⁺ transitions. While the achievable resolution is not sufficient to record fully site-selective spectra, with either PLEs or PLs, the differences observed within the range of the ⁵D₀ → ⁷F₁ transitions are adequate to conclude that as many as four Eu³⁺ ions of different local symmetries contribute to the luminescence in this range.

The above findings are remarkable. Already four Eu³⁺ emitting centers had been found within the regular set of PL lines, those best seen upon the nonselective excitation around 257 nm, see Fig. 4. Next lines shall be seen due to the high-resolution PLE and PL spectra focused on the peculiar emission around 574 nm efficiently excited around 325 nm (see Fig. 4).

Fig. 7 depicts the results of site-selective PLE and PL spectra of the ⁵D₀ → ⁷F₀ transitions with emission around 574 nm. PLE spectra are presented in the range of 458–464 nm (⁵D₀ → ⁷F₂ transitions, Fig. 7a) and PL emissions in the range of 572–579 nm (Fig. 7b). The results are straightforward. For the chosen excitation wavelengths, clearly distinct ⁵D₀ → ⁷F₀ emissions are observed. Their respective ⁷F₀ → ⁵D₂ excitation transitions are also significantly different and well-resolved. The positions of the characteristic lines are given in Table 2. Thus, for the peculiar luminescence with the CT excitation around 325 nm and ⁵D₀ → ⁷F₀ transitions at unusually short wavelengths, as many as four different Eu³⁺ emitting centers were also found.

Fig. 7 Site-selective PLE (a) and PL (b) spectra of Sr₂GeO₄:0.1% Eu³⁺ recorded at 20 K. PLE spectra were measured for four emission lines: 574.25 nm (black), 547.8 nm (red), 575.45 nm (blue) and 576.55 nm (green) (a). For PL spectra four selective excitation lines were used: 460.2 nm (black), 460.7 nm (red), 461.0 nm (blue) and 461.9 nm (green) (b).

Hence, as many as eight Eu³⁺ emitting centers were clearly resolved. They form two groups, each consisting of four Eu³⁺ sites sharing similar characteristics. The first group might be termed regular centers, as their emissions locate within the typical range of wavelengths for Eu³⁺ luminescence. The other four centers, let us term them irregular, are characterized by the peculiarly short wavelength of the ⁵D₀ → ⁷F₀ luminescence, located around 574–577 nm. Taking into account the characteristic properties of the centers, the regular ones should be assigned to the Eu³⁺ ions occupying the Sr²⁺ sites, , and adopting their local symmetries in Sr₂GeO₄. This does not explain how the next four sites of Eu³⁺ are then formed. The answer to this question should also rationalize the peculiar spectroscopic characteristics of the four irregular sites.

The low-energy (long-wavelength, peak around 325 nm) CT transition of the Eu³⁺ ions emitting around 574–577 nm may be understood by assuming that the dopant excessive charge is compensated by an interstitial O²⁻ ligand, , located in the first coordination sphere of the emitting ion. Since it would not be bound to Ge, its bond with Eu³⁺ would show a more covalent character than in the case of the regular oxygen atoms.²⁹ This would, obviously, lower the energy needed to accomplish the → Eu³⁺ charge transfer transition, exactly as experimentally observed.

Furthermore, since may in general locate randomly within the host lattice (though some tendency to trace positively charged sites may occur by means of their coulombic interaction), only a fraction of all the Eu³⁺ ions encounter the extra ligand in their first coordination sphere. And only Eu³⁺ ions within the – associates would show the peculiar luminescent properties with the low-energy CT transitions and the short-wavelength ⁵D₀ → ⁷F₀ luminescence. Since both the regular and irregular sets of sites are composed of four Eu³⁺ centers it is apparent that Eu enters each of the four Sr sites in Sr₂GeO₄, and within each of them the interstices may be located.

The above explanation is consistent with the published data on similar specific luminescence of silicates – Sr₂SiO₄:Eu, Ba₂SiO₄:Eu and La₂Si₂O₇:Eu.²⁵ Obviously, the additional ligand, not bound to Ge, around Eu³⁺ will strongly affect the position of the dopant electronic levels. This would be the privileged – bond which would be responsible for the peculiar luminescence.

The peculiarly short wavelengths of the ⁵D₀ → ⁷F₀ transitions were often (though not always) connected with their abnormally high intensities. Sometimes they dominated the emission spectrum, although not always as strongly as in our case (see Fig. 4b and d). This is also curious as, according to the Judd–Ofelt theory, these transitions are strongly forbidden both as electric dipole and magnetic dipole transitions.^30–33 Such an anomalously potent ⁵D₀ → ⁷F₀ luminescence of Eu³⁺, basically breaking the Judd–Ofelt theory rules, was reported for a number of compositions,^34–36 among them Ba₂SiO₄,²⁴ La₂Si₂O₇,²⁵ BiCa₄(PO₄)₃,²⁶ CuLaO₂,³⁷ BaFCl,^38,39 Ba₄Y₂ZrWO₁₂,⁴⁰ Y₃SbO₇,²⁷ LaOCl,^41,42 LaOBr.⁴³ The effect was also demonstrated in some apatites and their derivatives, like Ca₁₀(PO₄)₆(OH)₂, Ca₁₀(PO₄)₆O,⁴⁴ LaCa₄(PO₄)₃O,²⁶ M₅(PO₄)₃X (M = Ca, Sr; X = F⁻, Cl⁻, Br).⁴⁵ An important case is BaFCl:Eu³⁺ for which calculations were performed and it was shown that it is an oxygen-substituting F⁻ in the first coordination sphere which is responsible for the irregular Eu³⁺ luminescence.³⁸

It was noted that the irregular Eu³⁺ luminescence with the unusually high ⁵D₀ → ⁷F₀ emission intensity and energy often occurs in so-called charge-unbalanced materials, where the Eu³⁺ activator replaces a doubly charged metal ion site and when oxygen-compensating sites are likely to appear.^39,46 Nevertheless, not all cases fall under this rule, which is clear from the above list of compositions. However, the strong ⁵D₀ → ⁷F₀ emission located at unusually high energy is regularly connected with a high energy of the ⁵D₀ → ⁷F₁ luminescence and large crystal field splitting of the ⁷F₂ level.³⁹ From Fig. 4d it is also clear that all these characteristics are presented by the irregular Eu³⁺ ions in Sr₂GeO₄. This observation clearly points to the importance of the effect that the local crystal field exerts on the irregular Eu³⁺ ions in their nearest neighborhood. Let us note that the O²⁻ → Eu³⁺ CT represents a temporary reduction of the dopant to the Eu²⁺ state with a hole left behind at the oxygen atom. Then, the activator electron cloud expands altogether by about 20%,¹⁷ which necessarily leads to local distortion of the lattice and its substantial relaxation.

According to the Judd–Ofelt theory, the (so-called hypersensitive) ⁵D₀ → ⁷F₂ electric dipole (ED) transitions are sensitive to the local symmetry site, while the ⁵D₀ → ⁷F₁ magnetic dipole (MD) transitions are not. Consequently, their intensity ratios serve as a probe of the local symmetry site of the emitting Eu³⁺ ions. The Judd–Ofelt theory predicts that the ⁵D₀ → ⁷F₀ transition rate is always low. As was clarified in ref. 39, so-called J-mixing cannot account for such a spectacular increase in the ⁵D₀ → ⁷F₀ transition rate as is observed in the irregular Eu³⁺ luminescence. However, it has already been shown about 50 years ago that in the case of some particular site symmetries, allowing the operation of linear crystal field parameters, the ⁵D₀ → ⁷F₀ transition may rise greatly in intensity.^47,48 This finally leads to an extension of the Judd–Ofelt theory proposed by Wybourne^49–51 and Downer.^52,53 As a result, it appears that even a small linear contribution to the crystal field Hamiltonian may boost the ⁵D₀ → ⁷F₀ transition probability if only the energy of charge transfer state is low enough.³⁹ This occurs rarely, and is known as a third-order ED mechanism, which makes the ⁵D₀ → ⁷F₀ transition rate much higher than in regular cases and allows its intensity to even beat those of the other ⁵D₀ → ⁷F_J components. This is the case we have in our Sr₂GeO₄:Eu and which is observed in other systems showing the irregular characteristics of the Eu³⁺ luminescence, which we cited above. On the other hand, the low-energy O²⁻ → Eu³⁺ CT transition, peaking >∼300 nm, is not common but rather extraordinary. Its regular position is below ∼300 nm, often around 230–260 nm. In orthophosphates it occurs at even shorter wavelengths (higher energies). To lower the CT transition energy, at least one of the O²⁻-ligands has to be easily polarizable and its electrons should not be engaged into bonds with ions other than the emitting Eu³⁺. Then, the energy needed to move an electron from the O²⁻-ligand to the Eu³⁺ ion orbitals may be lower than in other cases. Consequently, the conditions for the third-order mechanism leading to the unusually significant intensity of the ⁵D₀ → ⁷F₀ luminescence as described by Wybourne and Downer are fulfilled.

There is thus every reason to claim that in Sr₂GeO₄:Eu the four lines around 574–577 nm which dominate the spectra of the irregular Eu³⁺ ions reflect the presence of a charge-compensating O²⁻ ion in the first coordination sphere of Eu³⁺ dopants. This reduces the CT transition energy of that ion and in turn boosts the rate of the ⁵D₀ → ⁷F₀ transition well beyond those to the ⁷F_J levels with J > 0. What is quite unique in Sr₂GeO₄:Eu is that each of the four sites offered by the host experiences this effect. Thus the interstitial oxygen may locate in the first coordination sphere of each of the Eu³⁺ substituting either of the sites.

It is noteworthy that many of the host materials in which the ⁵D₀ → ⁷F₀ transitions of Eu³⁺ show these peculiar properties are rather loosely packed and/or contain chains formed by the M–O polyhedra, between which larger, unfilled spaces are present. These may be filled with oxygen interstices, facilitated when charge compensation is necessary. This is also true for Sr₂GeO₄:Eu³⁺, as indicated in Fig. 1. This further indicates that the occurrence of interstitial sites in hosts in which their interaction is restricted to the emitting ions may well be the main source of the peculiar behaviour of the ⁵D₀ → ⁷F₀ transitions. In such specific, highly polarizable environments the higher rate of the ⁵D₀ → ⁷F₀ compared to other ⁵D₀ → ⁷F_J transitions as well as its higher-energy locations were considered quite probable, as we have already mentioned above. Thus, the results presented in this paper and those already reported in the literature give the strong impression, quite well justified experimentally, that the abnormally high energy of the ⁵D₀ → ⁷F₀ transitions and/or their unusually high intensities are indeed connected with the presence of interstitial oxygen which is exclusively bonded/coordinated to the emitting center. Being easily polarizable, this oxygen lowers the energy of the CT transitions and also affects the rates of the f–f transitions.

Photoluminescence of Sr₂GeO₄:Eu²⁺

Samples prepared in a reducing atmosphere were expected to present a broad-band 5d → 4f Eu²⁺ luminescence. At room temperature, however, only a residual trace of Eu³⁺ luminescence could be recorded upon different excitations. This indicated that Eu was presumably reduced to Eu²⁺ (otherwise the Eu³⁺ luminescence would be strong), but the reduced form of the dopant does not emit light at RT. Indeed, after cooling the sample to 20 K, a broad-band intense emission peaking at 620 nm was recorded upon excitation at any wavelength in the range of 220–470 nm. See Fig. 8a. The emission band is assigned to the 4f⁶5d¹ → 4f⁷ (d → f) transition of Eu²⁺. The estimated Stokes shift reaches ∼7100 cm⁻¹. This is quite a considerable value, which more than doubles the Stokes shift in BaMgAl₁₀O₁₇:Eu²⁺, a commercial blue phosphor.^10,54 This is an important observation as a significant Stokes shift leads to significant thermal quenching.⁵⁵

Fig. 8 20 K PLE spectra of the luminescence monitored at 620 nm for Sr₂GeO₄:Eu²⁺ with two different concentration (0.1% and 2.0%) and PL spectra of the same samples under 360 nm excitation (a). Color coordinates of Sr₂GeO₄:Eu²⁺ (b).

Let us note that the long-wavelength excitation tail reaches zero-level intensity around 500 nm, while the short-wavelength tail of the emission onset appears around 530–535 nm, which gap already contributes ∼1200–1300 cm⁻¹ to the Stokes shift. Such a characteristic of the Eu²⁺ luminescence points to a very significant relaxation occurring around the Eu²⁺ ion(s) after the absorption/excitation but before emission. This accords with the significant temperature quenching of the Eu²⁺ emission, well below 300 K as we shall see shortly. A higher Stokes shift of Eu²⁺ luminescence may also occur in so-called anomalous Eu²⁺ emission when the 4f → 5d excitation elevates the electron to the 5d level already immersed within the host conduction band (CB). There, taking advantage of the 5d-CB coupling the excited electron escapes its mother ion to localize on a nearby defect, giving an excitonic state with the hole left at the Eu. Relaxation of this entity produces the luminescent photons. This, however, typically leads to a broader emission with a longer decay time, and does not lead to a gap between excitation and emission spectra, as presented and discussed in Fig. 8. The decay time of our Sr₂GeO₄:Eu²⁺ at 45 K reaches a value of τ ∼ 1.5 μs and the luminescence band full width at half maximum (FWHM) attains 2550 cm⁻¹ (100 nm). This value is barely larger than in the commercial BAM, in which the regular 5d → 4f Eu²⁺ luminescence has an FWHM of ∼2320 cm⁻¹.⁵⁴ Thus, we deduce that in Sr₂GeO₄:Eu²⁺ the emission is the regular 5d → 4f luminescence, and not the anomalous one.

As Fig. 9a shows, at 30 K the luminescence has already become slightly less intense and at 50 K its efficiency drops by more than 50% compared to the emission at 20 K. At 100 K the Eu²⁺ emission is almost totally gone. Furthermore, the decay trace at 100 K is strongly nonexponential (Fig. 9b) and its average value is as short as 〈τ〉 ∼ 80 ns (Table 3). With decreasing temperature, when the intensity of emission increases, the decay time gets longer and finally at 45 K its average value reaches 〈τ〉 ∼ 1.5 μs and the trace is almost monoexponential (Table 3).

Fig. 9 PL spectra under 360 nm excitation for Sr₂GeO₄:0.1% Eu²⁺ in the range of temperature 20–100 K. Inset shows the change of PL intensity in the range 20–300 K for this sample (a). DEC traces of the same sample excited at 350 nm, monitored emission 625 nm (b).

Table 3 DEC parameters for Sr₂GeO₄:0.1% Eu²⁺ λ_exc = 350 nm, λ_em = 625 nm recorded at different temperatures

Temperature (K)	τ1 (ns)	Rel1%	τ2 (ns)	Rel2%	〈τ〉 (ns)
100	13.7	31.5	108.0	68.5	78.2
75	159.6	27.7	614.1	72.3	488.2
45	350.3	6.8	1631.3	93.2	1543.8

---

Such a low-temperature quenching is characteristic of a large shift between the parabolas representing the ground and the emitting states. This leads to their crossing, see Fig. 10a, close to the minimum of the emitting level parabola, which opens a nonradiative pathway for relaxation of the excited electrons already at low temperatures. However, the thermal quenching may also result from a proximity of the emitting level and the conduction band, see Fig. 10b. Then, an electron excited to the 5d level would be prone to escape to the conduction band if a sufficient amount of thermal energy were to become available. At present, we cannot offer an unequivocal explanation of which of the two mechanisms really occurs in Sr₂GeO₄:Eu²⁺ leading to the strong thermal quenching of the Eu²⁺ luminescence. Nevertheless, while the position of the Eu²⁺ luminescence in Sr₂GeO₄:Eu²⁺ (Fig. 8b) would be suitable to improve the colour rendering index in white LEDs,⁵⁶ adding some red colour to the blue LEDs combined with the yellow YAG:Ce phosphor, and also the efficient f → d excitation around 440 nm would be very suitable, the thermal quenching documented and described above gives no hope for that.

Fig. 10 Simplified schemes of configuration coordinate diagrams explaining the postulated mechanisms of the Eu²⁺ luminescence thermal quenching due to a large Stokes shift of the ground and emitting state parabolas. At low temperatures radiative relaxation occurs. At higher temperatures the thermal population of higher vibronic states of the emitting level (blue line) allows for non-radiative flow of the electrons down to the ground electronic state (red line) over the parabolas' crossing point (a).⁵⁷ (b) shows an alternative thermal quenching mechanism with the emitting level located only slightly below the host conduction band (CB) to which the electron may escape at higher temperatures, precluding its radiative relaxation. See text.

Independent of the Eu content, for Eu²⁺ luminescence we do not see any indications of emission from different sites. This is in contrast to the materials doped with Eu³⁺ for which all four sites could be identified in Eu³⁺ emissions and they were even doubled due to the interstitial oxygen in the first coordination sphere of Eu³⁺, as discussed earlier. It is also different from Sr₂SiO₄:Eu, in which two well-separated Eu²⁺ emissions were reported.⁵⁸ However, since the silicate is not isostructural with the germinate,¹⁴ any deeper comparisons are not in fact substantiated. A speculative explanation of such behavior of Sr₂GeO₄:Eu²⁺ might be that excited 5d levels of all but one type of Eu²⁺ ion in the host are located within its conduction band. Consequently, only the dopants whose first excited 5d level locates below the conduction band would contribute to the d → f emission at sufficiently low temperatures. Those in other sites would be blind at any temperatures.

Conclusion

Orthorhombic powders of Sr₂GeO₄ doped with either Eu³⁺ or Eu²⁺ ions were prepared by means of the ceramic method in air or forming gas, respectively. XRD studies confirmed the formation of an orthorhombic phase, space group Pbn2₁, independent of the oxidation state of Eu and their concentration. The luminescence studies of the Sr₂GeO₄:Eu³⁺ revealed that all four available sites offered by the host are occupied by Eu³⁺ and contribute to the luminescence. All of those sites produce a typical red luminescence of Eu³⁺ with CT excitation peaking around 260 nm. However, some of the dopant ions give emission with the ⁵D₀ → ⁷F₀ luminescence located around 574 nm, which is an unusually short wavelength for that transition. Another peculiarity is that these Eu³⁺ ions show the CT excitation transition peaking around 325 nm, which is an exceptionally long wavelength for oxide hosts. Such behavior was explained by the presence of an interstitial oxygen, , bound to Eu³⁺ by a strongly covalent interaction, which was possible because, contrary to the regular O-atoms, the was not bound to the Ge. The effect was discussed with the help of the Wybourne and Downer extension to the Judd–Ofelt theory. Eu²⁺ ions show their allowed 5d → 4f emission only below 100 K. The reason for such a strong thermal quenching was postulated to come from either a nonradiative relaxation because of a strong shift between ground and emitting states parabolas or from low-energy separation of the emitting 5d level and the host conduction band. Unfortunately, the spectroscopic characteristics of both the Sr₂GeO₄:Eu³⁺ and Sr₂GeO₄:Eu²⁺ preclude their applicability in phosphor-converted white LEDs.

Acknowledgements

The authors are indebted to Dr Joanna Cybinska for helpful discussions. This work was supported by POIG.01.01.02-02-006/09 project co-funded by European Regional Development Fund within the Innovative Economy Program. Priority I, Activity 1.1. Subactivity 1.1.2.

References

1. S. H. M. Poort, W. Janssen and G. Blasse, J. Alloys Compd., 1997, 260, 93–97.
2. J. S. Lee and Y. J. Kim, J. Nanosci. Nanotechnol., 2012, 12, 8630–8634.
3. J. K. Park, M. A. Lim, C. H. Kim, H. D. Park, J. T. Park, S. Y. Choi, J. K. Park, A. Lim, C. H. Kim and H. D. Park, Appl. Phys. Lett., 2011, 82, 683–685.
4. L. Zhang, Z. Lu, P. Han, J. Lu, N. Xu, L. Wang and Q. Zhang, J. Am. Ceram. Soc., 2012, 95, 3871–3877.
5. J. Y. Woo, K. Kim, S. Jeong and C. S. Han, J. Phys. Chem. C, 2011, 115, 20945–20952.
6. Z.-J. Wang, Z.-P. Yang, Q.-L. Guo, P.-L. Li and G.-S. Fu, Chin. Phys. B, 2009, 18, 2068–2071.
7. X. Sun, J. Zhang, X. Zhang, Y. Luo and X. Wang, J. Rare Earths, 2008, 26, 421–424.
8. M. A. Tshabalala, F. B. Dejene, S. S. Pitale, H. C. Swart and O. M. Ntwaeaborwa, Phys. B, 2014, 439, 126–129.
9. J. K. Park, K. J. Choi, C. H. Kim, H. D. Park and S. Y. Choi, Electrochem. Solid-State Lett., 2004, 7, H15.
10. C. Ronda, Luminescence From Theory to Applications, WILEY-VCH, Weinheim, 2008.
11. T. Q. Khanh, P. Bodrogi, Q. T. Vinh and H. Winkler, LED Lighting, Technology and Perception, Weinheim, 2015.
12. R. Xie, N. Hirosaki, T. Suehiro and F. Xu, Chem. Mater., 2006, 18, 5578–5583.
13. Y. Q. Li, A. C. A. Delsing, G. De With and H. T. Hintzen, Chem. Mater., 2005, 17, 3242–3248.
14. F. Nishi and Y. Takeuchi, Z. Kristallogr., 1996, 211, 607–611.
15. M. Catti, G. Gazzoni, G. Ivaldi and G. Zanini, Acta Crystallogr., Sect. B: Struct. Sci., 1983, 39, 674–679.
16. M. Catti and G. Gazzoni, Acta Crystallogr., Sect. B: Struct. Sci., 1983, 39, 679–684.
17. R. D. Shannon, Acta Crystallogr., Sect. A: Cryst. Phys., Diffr., Theor. Gen. Crystallogr., 1976, 32, 751–767.
18. W. D. Kingery, H. K. Bowen and D. R. Uhlmann, Introduction to Ceramics, John Wiley & Sons, 2nd edn, 1976.
19. G. H. Dieke and H. Crosswhite, Spectra and energy levels of rare earth ions in crystals, Interscience Publishers Wiley, New York, USA, 1968.
20. J. Zeler, J. Cybińska and E. Zych, RSC Adv., 2016, 6, 57920–57928.
21. E. D. Bacce, A. M. Pires and M. R. Davolos, J. Alloys Compd., 2002, 344, 312–315.
22. S. K. Gupta, M. Mohapatra, S. Kaity, V. Natarajan and S. V Godbole, J. Lumin., 2012, 132, 1329–1338.
23. A. Nag and T. R. N. Kutty, J. Mater. Chem., 2004, 14, 1598–1604.
24. A. M. Pires, M. R. Davolos and O. L. Malta, J. Lumin., 1997, 74, 244–246.
25. B. Piriou, M. Richard-Plouet, J. Parmentier, F. Ferey and S. Vilminot, J. Alloys Compd., 1997, 262–263, 450–453.
26. N. Lakshminarasimhan and U. V. Varadaraju, J. Solid State Chem., 2004, 177, 3536–3544.
27. J. Wang, Y. Cheng, Y. Huang, P. Cai, S. Il Kim and H. J. Seo, J. Mater. Chem. C, 2014, 2, 5559–5569.
28. H. Xie, J. Lu, Y. Guan, Y. Huang, D. Wei and H. J. Seo, Inorg. Chem., 2014, 53, 827–834.
29. A. M. Srivastava, Opt. Mater., 2016, 58, 89–92.
30. G. S. Ofelt, J. Chem. Phys., 1962, 37, 511–520.
31. B. R. Judd, Phys. Rev., 1962, 127, 750–761.
32. B. G. Wybourne, J. Alloys Compd., 2004, 380, 96–100.
33. P. A. Tanner, Y. Y. Yeung and L. Ning, J. Phys. Chem. A, 2013, 117, 2771–2781.
34. L. Smentek and A. Kȩdziorski, J. Alloys Compd., 2009, 488, 586–590.
35. A. S. Souza, Y. A. R. Oliveira and M. A. Couto Dos Santos, Opt. Mater., 2013, 35, 1633–1635.
36. A. K. Parchur and R. S. Ningthoujam, RSC Adv., 2012, 2, 10859.
37. F. Fujishiro, M. Murakami, T. Hashimoto and M. Takahashi, J. Ceram. Soc. Jpn., 2010, 118, 1217–1220.
38. X. Y. Chen, W. Zhao, R. E. Cook and G. K. Liu, Phys. Rev. B: Condens. Matter Mater. Phys., 2004, 70, 1–9.
39. X. Y. Chen and G. K. Liu, J. Solid State Chem., 2005, 178, 419–428.
40. H. Xie, L. Xu, L. Qin, Y. Huang, D. Wei, S. Il Kim and H. J. Seo, Mater. Lett., 2014, 115, 18–20.
41. G. Blasse and A. Bril, J. Chem. Phys., 1967, 46, 2579.
42. J. Holsa and P. Porcher, J. Chem. Phys., 1981, 75, 2108.
43. J. Holsa and P. Porcher, J. Chem. Phys., 1982, 76, 2790.
44. B. Piriou, D. Fahmi, J. Dexpert-Ghys, A. Taitai and J. L. Lacout, J. Lumin., 1987, 39, 97–103.
45. R. Jagannathan and M. Kottaisamy, J. Phys.: Condens. Matter, 1995, 7, 8453–8466.
46. L. Gaigerova, O. F. Dudnik, V. F. Zolin and V. A. Kudryashova, Luminescence of Crystals, Molecules and Solutions, Plenum Press, New York, 1973.
47. W. C. Nieuwpoort and G. Blasse, Solid State Commun., 1966, 4, 227.
48. B. R. Judd, J. Chem. Phys., 1966, 44, 839.
49. B. G. Wybourne, Optical Properties of Ions in Crystals, Wiley Interscience, New York, 1967.
50. B. G. Wybourne, J. Chem. Phys., 1968, 48, 2596.
51. B. G. Wybourne, Eur. J. Solid State Inorg. Chem., 1991, 28, 19.
52. M. C. Downer, G. W. Burdick and D. K. Sardar, J. Chem. Phys., 1988, 89, 1787–1797.
53. G. W. Burdick, M. C. Downer and D. K. Sardar, J. Chem. Phys., 1989, 91, 1511.
54. E. Zych, W. Goetz, N. Harrit and H. Spanggaard, J. Alloys Compd., 2004, 380, 113–117.
55. G. Blasse and B. C. Grabmaier, Luminescent Materials, Springer-Verlag, 1994, vol. 1.
56. P. Mottier, LED for Lighting Applications, Wiley-ISTE, 2010.
57. E. Zych, Inorganic display materials, in Handbook of luminescence, display materials, and devices, 2003, vol. 2, pp. 251–300.
58. A. Madej and E. Zych, RSC Adv., 2015, 5, 104441–104450.

---