1 INTRODUCTION

Owing to the structural stability, sesquioxides of Y and Gd are excellent matrices for creating phosphors with various emission spectra when doped with photoluminescence (PL) activators based on rare earth elements (REE). The easily accessible and relatively cheap Eu3+ and Tb3+ activators are widely used REE activators, which make it possible to obtain efficient, strong, and single-center emission in the red (mean wavelength 612 nm), green (542 nm), and blue (487 nm) regions of the visible spectrum [14]. These colors are close to the main ones, by mixing of which one can create different shades of the emitted color.

The bixbyite-type lattice, in which cations, including Eu3+ and Tb3+, are located in the C3i centrosymmetric and C2 noncentrosymmetric sites [58], is a characteristic feature of the matrices of the cubic modification of C-Ln2O3. For polycrystalline nanoscale powders, the state of activators at the boundaries (surfaces) of crystallites is of particular importance [9, 10]. There are no C3i sites in this state because of distortion of the short-range order of the crystal lattice. The monoclinic modification of B‑Ln2O3 sesquioxides is characterized by three different noncentrosymmetric positions of REE elements in the crystal lattice [1113]. Corresponding changes in the structure of the first coordination sphere of the PL activator cations, for example, the Eu3+ cation surrounded by (O2–)6 ions, lead to redistribution of the excitation energy over radiative electronic transitions and energy losses due to electron–phonon interactions [1416]. This causes shifts in the maxima of the emission bands and changes in their relative intensities and, finally, in the perception of the color of displays, monitors, and object images. The well-known optical effect of warm–cold lighting can serve as an example of the visual perception of such changes in radiation.

The coexistence of terbium ions with different oxidation states Tb3+ and Tb4+ is a specific feature of PL compounds with the Tb3+ activator. The Tb4+ state does not create emission centers. Thus, it is necessary to eliminate the oxidizing atmosphere and use either an inert atmosphere or a vacuum in order to synthesize the corresponding phosphors. Another option requires varying the chemical composition or the lattice structure in order to provide the thermodynamic stability of Tb3+ ions at the temperature of the synthesis and under a nonequilibrium oxygen pressure. In connection with the above peculiarities, it must be admitted that a fundamental comprehensive analysis of the chemical composition of the sublattices of Gd, Tb, and Eu cations in the phosphors based on Gd2(1 ‒ x)O3:TbxEux solid solutions has not yet been carried out from the standpoint of the thermodynamic state of the system, namely, the Tb3+,4+ oxidation state of the cations and the redistribution of the Gd3+, Tb3+ (Tb4+), and Eu3+ cations in the bixbyite-type lattice and at the boundaries (surfaces) of crystallites. Thus, the objectives of this study were as follows:

—sol–gel synthesis of Gd2(1 – x)TbxEux(OH)y(CO3)z · nH2O solid solutions with their subsequent annealing at temperatures of 700, 900, and 1200°C to the corresponding C-cubic and B-monoclinic modifications of Gd2(1 – x)O3:(Tb3+,4+)x(Eu3+)x compounds, and characterization of the chemical and structural compositions of the obtained solid solutions;

—analysis of the localization of Eu3+ and Tb3+ ions over the C3i and C2 cationic sites and the Cs surface states in the Gd2(1 – x)O3:(Tb3+,4+)x(Eu3+)x lattice and analysis of the distribution of the emission excitation energy over radiative electronic transitions according to the PL spectra.

2 EXPERIMENTAL

To obtain PL compounds based on Gd2(1 – x)O3: (Tb3+,4+)x(Eu3+)x solid solutions, we used the sol–gel method [14, 17, 18]. Nitrates Gd(NO3)3 · 6H2O, Tb(NO3)3 · 5H2O, and Eu(NO3)3 · 5H2O synthesized from the corresponding oxides with at least 99.9% of the main component were used as reagents. The synthesis was carried out starting from aqueous solutions of mixtures of the above salts, the compositions of which corresponded to the ones necessary for the formation of sol–gel products of solid solutions of hydroxocarbonates Gd2(1 – x)TbxEux(OH)y(CO3)z · n(H2O). For this, 0.2 M solutions of nitrates that c-orresponded, in percentage, to the compositions of the required oxides were used. A twofold excess of a solution of special-purity grade NaOH in bidistilled water was used a precipitant. The synthesis was carried out on a setup with sputtering of the starting reagents to achieve homogeneity in the volume of the working   solution. The resulting precipitate of Gd2(1 ‒ x)TbxEux(OH)y(CO3)z · nH2O was washed until neutral pH of the rinsing water was achieved and dried in air at 50°C. The product was divided into parts to form polycrystalline samples weighing 1 g each, which then were annealed at 700, 900, and 1200°C for 1–2 h until formation of the final products, i.e., solid solutions Gd2(1 – x)O3:(Tb3+,4+)x(Eu3+)x, where x = 1 and 2.5 mol %. Further, we will use notation Gd2O3: Tb3+,Eu3+ (x mol %) for brevity. The concentrations of Tb3+ and Eu3+ were chosen equal to 1 and 2.5 mol %, since the Tb3+ and Eu3+ ions acquire PL excitation energy independently at Tb3+ concentrations of <1.5 mol %, while the Tb3+ ions serve as PL coactivators at concentrations of >1.5 mol % and partially transfer the excitation energy from the 5D4 level to the 5D1 of Eu3+ cations, for example, in the CaWO4 matrix [19]. A concentration of 2.5 mol % was found to be the most effective concentration for Eu3+ (1–2 mol % in [20, 21]). In [1, 22], it was recommended to use Eu3+ concentrations of 1.25 mol % in the Y2O3 and Gd2O3 matrices for the effective emission of Eu3+ cations, but with low concentrations of the Tb3+ coactivator at a level of 0.0025–0.0050 mol %. Our goal was to study the distributions of the Eu3+ and Tb3+ cations over the lattice sites and to analyze the paths of the transfer of the PL excitation energy over the sublattices comprised of the C3i and C2 sites; therefore, equal concentrations of these ions were chosen.

The X-ray diffraction phase analysis of the samples was performed on a Shimadzu XRD-7000 diffractometer (CuKα radiation, Ni filter, 2θ range 5°–70°, step 0.03°, and 2s accumulation). IR spectra were recorded on a SCIMITAR FTS 2000 IR spectrometer in the frequency range of 4000–375 cm–1.

The luminescence spectra were studied using a Cary Eclipse fluorescence spectrophotometer (VARIAN). The excitation wavelength was chosen according to the diffuse reflection spectra (DRS) and the excitation luminescence spectra of single vibronic levels (SVL) of the main emission bands of the Tb3+ (542 nm) and Eu3+ (612 nm) ions. According to preliminary DRS experiments, all polycrystalline samples have a high optical absorbance of about 70–80% in the ultraviolet region. Wavelength λex = 280 nm was chosen on the basis of the transfer of excitation energy over the maximum possible number of known channels for these compounds, namely, charge transfer channel Gd3+ → Tb3+; and electron transfer channels 4f → 5d Tb3+, 5D4 (Tb3+) → 5D1, 2(Eu3+) [2], 4f → 4f (Eu3+) [1, 19, 2225], and O2– → Eu3+ [10, 26] for efficient emission of Tb3+ and Eu3+ ions via the corresponding radiative transitions. In comparison with the cubic modification, the transmittance of the monoclinic modification at a wavelength of λ = 280 nm was lower by a factor of almost 2, and the excitation efficiency was the same for the band with a maximum at 612 nm and eight times higher for the band with a maximum at 550 nm.

3 RESULTS AND DISCUSSION

Figure 1 shows the X-ray diffraction patterns of the studied samples.

Fig. 1.
figure 1

X-ray diffraction patterns of samples 1–8.

The unit cell parameters of the main phase of C‑Gd2O3:Tb3+,Eu3+ (x mol %), the sizes of the co-herent scattering regions (CSRs), and accompanying phases are given in Table 1 for the samples obtained by annealing the sol–gel products of Gd2(1 ‒ x)TbxEux(OH)y(CO3)z · nH2O under various conditions.

Table 1. Results of processing the XPD data of the annealing products for the samples synthesized by the sol–gel method

All samples, except for samples 1, 6, and 11, are mainly cubic modifications of C-Gd2O3:Tb3+,Eu3+ (x mol %) with the crystal lattice belonging to space group \(Ia\bar {3}\). The lattice parameter of sample 5 is closest to that of the Gd2O3 matrix. Taking into account the known values of the radii of Eu3+ (0.109 nm), Gd3+ (0.115 nm), and Tb3+ (0.093 nm) ions for a coordination number of 6 in oxides [27, 28], it can be assumed that the Tb3+ and Eu3+ ions substitute preferably the Gd3+ ions of the matrix in the C3i sites, since the volume of the corresponding coordination sphere for the C3i sites is smaller than that for the C2 sites.

In the X-ray diffraction pattern of sample 1, which is a sol-gel product annealed in air at 700°C for 1 h, twelve reflections belong to the phase of cubic oxide C-Gd2O3 listed in the Inorganic Crystal Structure Database (ICSD) under entry # 183130 (ICSD-96207), eighteen reflections can be attributed to the Gd(OH)3 phase (ICSD-152449) in an amount of less than 5 wt %, and eight reflections correspond to the presence of the Gd(CO3)OH phase (ICSD-200093) at a level of about 1 wt %. The same pattern has been observed earlier [16, 17]. Reflections are shifted within 2θ = ±0.03° relative to their position in pure C-Gd2O3. This is explained by the fact that all phases are solid solutions of oxides based on neighboring REE ions, i.e., Eu3+, Gd3+, and Tb3+. In addition, the terbium ion in the oxide has an effective oxidation state of Tb3.64+ in air [29], which corresponds to the mean oxidation state of the cations in compound Tb11O20. All reflections are broadened and there is a wide halo in the region of small diffraction angles, which indicates the small size of crystallites and the presence of an amorphous phase.

Sample 2 is obtained from sample 1 by successive annealing in an H2 atmosphere for 2 h. The sample is comprised of the phase of solid solution C-Gd2O3: Tb3+,Eu3+ (1 mol %) and the impurity phases of tetragonal Tb7O12 (two strong reflections at 2θ = 61.719° and 64.350°, space group \(R\bar {3}\), ICSD-73822) and Gd(CO3)OH (one reflection at 2θ = 61.908°) with the diffraction peak intensities noticeably higher than the background level. Other reflections cannot be probably resolved within the technique used, and some of the reflections overlap with the reference reflections of C-Gd2O3. The reflections have deviations with respect to the 2θ angles, since the solid solutions of three oxides of Eu3+, Gd3+, and Tb3+ have a bixbyite structure, which contains two types of positions of REE cations in the unit cell, namely, eight ions in the centrosymmetric C3i positions and twenty four ions in the noncentrosymmetric C2 positions. Moreover, the distribution of elements over the positions depends on the ionic radius [3034], on the temperature of synthesis or homogenizing annealing [14], and on the conditions of compound synthesis, which can be the sol-gel process [1, 22, 33], the hydrothermal method [3], the solid-phase sintering [34], etc. The CSR size determined by the Debye–Scherrer model is 36 ± 3 nm for the isometric shape of crystallites.

Sample 3, which is a sol–gel product annealed in air at 900°C for 1 h, contains the phase of cubic solid solution C-Gd2O3:Tb3+,Eu3+ (1 mol %). In consequence of the fact that the sample was annealed in an air atmosphere with subsequent free cooling, the structure contains Tb3.4+ ions, since seven reflections correspond to the trigonal Tb7O12 phase (see Fig. 1), three reflections at 2θ = 29.239°, 48.683°, and 57.804° correspond to the main reflections of the cubic TbO1.81 phase that belongs to space group \(Fm\bar {3}m\) (ICSD-28916). The presence of these phases does not contradict the thermography data [29], which indicate a change in the stoichiometry of terbium oxide and, consequently, a change in the effective oxidation state of terbium ions after heating in air. The established reflections of the oxide phases of Tb3.4+ are located quite close to the reflections of the C-Gd2O3 phase and are partially overlapped with them, changing their reference intensity. This gives grounds to assume that semicoherent interfaces are formed between the phases, by analogy with compounds that have incommensurate sublattices [35, 36]. Moreover, impurity phases can be represented by polysynthetic twin layers of nanometer sizes.

Sample 4 synthesized from sample 3 by annealing at 900°C for 1 h in H2 is comprised of one phase of solid solution C-Gd2O3:Tb3+,Eu3+ (1 mol %). The Tb7O12 phase is absent, since the Tb4+ ions in the oxide are reduced to Tb3+ in the H2 atmosphere. The coherent scattering region is 36 ± 3 nm for sample 3, and increases to 86 ± 9 nm in the [222] and [004] directions, to 71 ± 8 nm in the [044] direction, and to 54 ± 5 nm in the [226] direction for sample 4.

Sample 5 was obtained by annealing a sol-gel product at 1200°C for 1 h and mainly contains the cubic phase of solid solution C-Gd2O3:Tb3+,Eu3+ (1 mol %) with a coherent scattering region at the level of that for the reference micrometer-size indicates an increased degree of crystallinity. For pure C-Gd2O3, the cubic bixbyite lattice is a basic one lattice up to 1200°C [30]. However, this modification is a thermodynamically metastable modification according to the conclusions drawn in [37]. An impurity phase is found, four reflections of which are close to those of the Tb7O12 phase.

Sample 6 obtained from sample 5 by annealing at 1200°C for 1 h in H2 is comprised of the monoclinic B‑Gd2O3:Tb3+,Eu3+ (1 mol %) phase belonging to space group C2/m and an impurity phase of trigonal Tb7O12 (four reflections) that contains Tb3.4+ cations. The appearance of the B-Gd2O3:Tb3+,Eu3+ (1 mol %) phase after annealing in a reducing H2 atmosphere at 1200°C is explained by the specific behavior of the stability of the cubic and monoclinic phases of REE oxides [29, 37, 38]. To check this feature, sample 6 was additionally annealed at 1200°C for 1 h in air (sample 9) and sample 5 was annealed at 1200°C in high-purity grade Ar for 1 h (sample 10). In both cases, phases of monoclinic B-Gd2O3:Tb3+,Eu3+ (1 mol %) containing an impurity of the trigonal Tb7O12 phase were obtained. Thus, polymorphic transition to the monoclinic phase occurs after annealing for 3 h in total. This confirms the well-known conclusion [37] that the cubic structure of sesquioxides of cerium group, including C‑Gd2O3, is a metastable phase at low temperatures and can be converted into the monoclinic phase at elevated temperatures with a sufficient annealing time. The C-Gd2O3 lattice is a basic one lattice at a relatively low concentration of dopants. However, we have previously found that solid solutions of rare-earth oxides based on C-Gd2O3 can be stabilized by Eu3+ impurities in high concentrations [15].

To test the effect of an increased concentration of PL activators, samples 7 and 8 with Tb3+ and Eu3+ concentrations of 2.5 mol % each were synthesized. Both samples contain the phase of the cubic modification of C-Gd2O3:Tb3+,Eu3+ (2.5 mol %) with CSR of 120 and 150 nm, respectively. The formation of the cubic modification of sample 7, which is a sol–gel product annealed in air at 1200°C for 1 h, and its retention after additional annealing in H2 at 1200°C for 1 h are explained by the stabilizing effect of Eu3+ and Tb3+ ions. The resulting C-Gd2O3 matrix is a metastable phase in terms of thermodynamics [15, 37] and is probably stabilized by terbium sesquioxide (Tb3+), which is thermodynamically stable in a cubic modification at these temperatures. The absence of the Tb7O12 impurity phase and, consequently, Tb4+ cations is probably associated with the formation of stable Tb3+–O2––Eu3+ clusters due to exchange interactions with concentrations Eu3+ and Tb3+ ions above 1.5 mol %, as was established for the CaWO4 matrix [19].

The IR spectra of the samples are shown in Fig. 2. With an increase in the annealing temperature in the range of 700–1200°C, dehydroxylation of the samples and a decrease in the content of carbonate ions occur, which is reflected in a decrease in the intensity of absorption bands of molecular water in the range of wavenumbers 1400–3600 cm–1, and absorption bands of O–H groups and asymmetric stretching vibrations of C–O groups at 1510 and 1400 cm–1, respectively [33]. The bands in the region of 390–600 cm–1 are related to stretching vibrations of Ln–O bonds in the C‑Gd2O3 [39, 40], C-Eu2O3 [41], and C-Tb2O3 [42] sublattices and Ln–O bonds of the monoclinic lattice and impurity phases (sample 6). It is noteworthy that the absorption bands of O–H groups appear after annealing at high temperatures of 1200°C in an atmosphere of H2 (sample 6), which is probably explained by the fact that the reduction of Tb4+ ions (or the reduction of an effective oxidation state of 3.64+ in the Tb11O22 matrix or 3.40+ in the Tb7O12 matrix to Tb3+). In this case, oxygen is released and forms adsorbed hydroxyl groups by reacting with H2. A control experiment with additional annealing of sample 5 in an Ar atmosphere showed the absence of O–H groups (sample 10).

Fig. 2.
figure 2

Infrared spectra of samples 1–10.

The PL spectra normalized with respect to intensity are shown in Figs. 3a and 3b.

Fig. 3.
figure 3

(a) Photoluminescence spectra of samples 1–6 in the form of 3D histograms of the maxima of the main emission bands of Tb3+ ions at 487, 513, 542, 550, and 584 nm and Eu3+ ions at 595, 612, and 622 nm, and (b) general views of the spectra of samples 7–10; the spectra of samples 8, 9, and 10 are raised by 10, 20, and 40 intensity units, respectively.

The spectra of the samples are given by histograms of the main PL band maxima (Fig. 3a), which give a clear idea of the change in the PL intensities from sample to sample. The real spectra of samples 1–5 are similar to the spectra of samples 7 and 8 (Fig. 3b) with different ratios of band intensities, which are given in Tables 2 and 3. Shifts of the band maxima along the radiation wavelength are noted, which is associated with a change in the crystal field in the coordination spheres of the PL activators. This may be a result of changing the real lattice structure (defectiveness) and redistributing the cations over the C2 and C3i states and over the surface Cs states.

Table 2. Intensities of the PL bands of the samples in relative units after the decomposition of the spectra into components
Table 3. Intensity ratios of the main emission bands of Eu3+ and Tb3+ ions for samples obtained under different annealing conditions with different concentrations of PL activators

The obtained results indicate that the intensity of emitted radiation in the reference bands of Tb3+ and Eu3+ ions increases with an increase in the annealing temperature. This is a result of dehydration of compounds and an increase in the degree of crystallinity of the samples. Moreover, the effect of dehydration in a multiphase system on the redistribution of the excitation energy over the bands can hardly be explained at the existing level of theoretical knowledge.

The analysis of the I612/I595 ratio of the PL band intensities of Eu3+ ions, which arise from radiative transition 5D07F2 at 612 nm and radiative transition 5D0 → 7F1 at 595 nm, respectively, is a well grounded analysis of the redistribution of cations over the C2 and C3i sites. This ratio is proportional to the ratio of concentrations of Eu3+ ions in the indicated lattice positions and is called the “asymmetric ratio” (AR) [43, 44].

Sample 2 after annealing of sample 1 in H2 is characterized by sharp increases in the intensity of the band of Tb3+ at 490 nm by a factor of 16, in the intensity of the superposition of bands (doublet [45]) at 542 and 550 nm by a factor of 18, and in the intensity of the band at 585 nm by a factor of 40. Moreover, the emission band of Eu3+ at 595 nm increases by an order of magnitude with an increase in the intensity of the band at 612 nm merely by a factor of 5. It is understandable that the Tb4+ ions reduced to Tb3+ are located preferably at the C3i lattice sites. The concentration of Eu3+ ions increases predominantly in the C3i sites because of transferring these ions into the bulk of the material with a decrease in the specific surface area of crystallites, in which they occupied the noncentrosymmetric Cs sites. Moreover, the probability of incorporating these ions into the C3i sites with a smaller volume of the surrounding sphere is rather high in accordance with the following descending order of radii of the cations: Gd3+ > Eu3+ > Tb3+ [27, 46]. The noted violation of the so-called “lanthanide contraction” of neighboring Eu3+ and Gd3+ ions has been recorded in the literature, and we have earlier observed it for solid solutions of oxides in the (GdxY1 – x)2O3:Eu3+ system [15]. This feature of Eu3+ ions was also found in the anomalous shift of the Raman band of Eu3+ ions in the region of 340–380 cm–1 for the series of lanthanides in the spectra of the corresponding cubic sesquioxides [47], which can be explained by an enhanced electron–phonon interaction. In particular, the PL excitation of Eu3+ ions occurs through electronic transitions 5D4 (Tb3+) → 5D1, 2 (Eu3+). Energy is predominantly transferred along the C3i sublattice within the energy of the 5D47F6 transition (487 nm), since the lowest increase in the intensity of this band was found after annealing of the precursor, i.e., sample 1. The AR value naturally decreases substantially and practically reaches the equilibrium value for the bixbyite lattice (AReq = 3).

Annealing of the sol-gel product (sample 3) in air at 900°C leads to an increase in the emission of all radiative transitions in Tb3+ by a factor of 4–5 compared to sample 1, while the emission of the 5D07F1 (595 nm) transition in the Eu3+ ion increases proportionally. The proportional distribution of energy over radiative transitions indicates that the distribution of PL activator Tb3+ over local cation sites is preserved and an increase in the radiation efficiency is associated with an increase in the degree of crystallinity of the sample. It should be noted that the annealing process is accompanied by a decrease in the concentration of Tb4+ ions due to reduction of them to Tb3+ with a decrease in the partial pressure of oxygen relative to its equilibrium value at a given temperature [29]. Excitation of the emission of the 5D07F2 transition increases only by a factor of 2.5 and is probably associated with an increase in the crystallinity of the samples in comparison with sample 1 (see CSR, Table 1). In this case, the AR value decreases by a factor of 2, which indicates an increase in the concentration of Eu3+ ions in the C3i lattice sites. This is a result of the transfer of some of these ions from the surface to the bulk of the material with a decrease in the specific surface area of crystallites. Moreover, the transfer of excitation energy of the emission band of Eu3+ at 612 nm and, to a lesser extent, of the band at 595 nm occurs within the energy of the 5D47F6 transition of Tb3+ (487 nm), since the total emission energy of the doublet of bands at 542 and 550 nm increased by a factor of 8 with an increase in the emission energy of the band at 487 nm merely by a factor of 5. In this case, the ratio between the AR and the I612/I487 intensities decreases proportionally by a factor of 2 in comparison with sample 1 at a constant I595/I487. The transfer of PL excitation of Tb3+ ions to Eu3+ ions has been noted earlier [1, 19].

Sample 4 obtained from sample 3 after annealing in H2 is characterized by efficient reduction of Tb4+ to Tb3+ and a sharp increase in the degree of crystallinity (CSR increases, see Table 1). Accordingly, the intensity of all PL bands increases moderately (by a factor of 2–3). At the same time, relative increases in the emission intensities of the 5D47F6 transitions (487 nm) by a factor of 4.5 and of the 5D47F5 doublet by a factor of 5.7 in total are observed. A slightly larger increase in the emission intensity of the 5D07F1 transition is associated with the transfer of Eu3+ ions from the Cs sites on the crystallite surface into the bulk of the material to the C3i sites.

Annealing of a sol–gel product at 1200°C in air (sample 5) gives rise to a further increase in the concentration of Tb3+ ions [29] and a decrease in the specific surface area of crystallites relative to the results of annealing at 900°C (sample 3). According to this, the intensities of all emission bands increase by a factor of 5–7 for the most part. The emission intensity of the 5D47F6 transition at 487 nm and the total intensity of the 5D47F5 doublet transitions at 542 and 550 nm in Tb3+ ions show the largest increases by factors of 8 and 12, respectively. At the same time, the AR value decreases insignificantly and the I612/I487 ratio slightly increases. The latter corresponds to the transfer of enhanced PL excitation from the Tb3+ activator to Eu3+ ions through the C2 sublattice in addition to the excitation of emission due to 4f → 4f transitions of Eu3+ ions. In actual fact, the I595/I487 ratio slightly decreases (Table 2).

Sample 6 obtained from sample 5 after annealing in H2 comprises a monoclinic phase based on the B‑Gd2O3 matrix. A significant increase in the total intensity of the emission bands of Tb3+ at 487 and 550 nm is noted, which is naturally associated with the reduction of Tb4+ ions to Tb3+ ions that are effective PL site. In this case, the intensity of the emission band at 622 nm (taking into account shifts to the region of 622–624 nm) increases very strongly, which can be explained by the superposition of the 5D07F2 radiative transitions of Eu3+ ions [48] (the low-energy component of the doublet) and the 5D47F3 transitions of Tb3+ ions. The conversion of gadolinium sesquioxide into the monoclinic modification is a normal phenomenon at 1200°C [37]. In this case, the emission band of Eu3+ ions in the C3i lattice sites at 595 nm is due to the presence of centrosymmetric sites in the lattice of trigonal Tb7O12. Consequently, the trigonal phase is doped with Eu3+ ions. It is noticeable that an unreferenced PL band appears at 513 nm in the spectrum of the obtained monoclinic phases, which is probably formed by doublet emission of the 5D47F6 transition of the Tb3+ ion.

Sample 7 with an increased content of PL activators at a level of 2.5 mol %, which is the cubic modification of solid solution C-Gd2O3:Tb3+,Eu3+ (2.5 mol %) of the bixbyite type, is formed as a result of annealing of a sol–gel product in air at 1200°C for 1 h. In this case, the intensities of the bands of Tb3+ at 487, 542, and 550 nm decrease by a factor of 2.5–3, the intensity of the PL band of Eu3+ ions at 596 nm increases by a factor of 5, and the intensity of the emission band at 612 nm remains the same relative to sample 5, which is also synthesized by annealing at 1200°C. These effects are associated with the onset of the well known effect of PL quenching of Tb3+ ions at concentrations of >1.5 mol % [19]. A relatively larger increase in the intensity of the emission band of the 5D07F1 transition of Eu3+ ions at 596 nm in comparison with sample 5 is explained by a two-fold higher concentration of PL activators with a noticeable decrease in the degree of crystallinity of the sample due to an increase in the specific surface area of crystallites. This is associated with an increase in the concentration of Tb3+ ions in the surface states, which leads to the transfer of Eu3+ ions from the Cs surface states into the bulk of the material to the C3i sites. These conclusions are based on the well known belief that the Tb3+ ion is an inhibitor of the growth of crystallites in a similar chemical system [1, 9]. Indeed, a decrease in the CSR is observed in sample 7 compared to sample 5 (see Table 1). Apparently, all additionally provided Eu3+ ions occupy the C3i positions, since the intensity of the band at 612 nm remains at the level that was in sample 5. Subsequent annealing of sample 7 in H2 for 1 h (sample 8) leads to increases in the emission band intensities of the 5D47F5 transitions of Tb3+ at 542 and 550 nm and of Eu3+ at 612 nm by a factor of 2.0–2.5 relative to sample 5 with a lower activator concentration, reaching the maximum PL efficiency of both activators for the cubic modification of a solid solution with the C‑Gd2O3:Tb3+,Eu3+ (2.5 mol %) composition.

The maximum intensities of the PL bands of Tb3+ at 487 and 542 nm are achieved in samples 6, 9, and 10, which consist of the monoclinic phase. Sample 10 is obtained by additional annealing of sample 5 in Ar for 1 h. The efficient photoluminescence is probably related to the fact that the cations coordinate seven O2– ions, in the monoclinic modification [46], which contributes to the more efficient transfer of excitation energy to Tb3+ ions. For Eu3+ ions, the transition from the octahedral environment can lead to the transfer of excitation from the 5D2 level of Eu3+ to the closely lying 5D4 level of Tb3+ when the excitation energy of photons is high (280 nm). As a result, the intensity of the band of the monoclinic phase at 612 nm remains at the maximum level for the cubic phase. However, the band of the 5D07F1 transition in Eu3+ at 595 nm has a slightly higher intensity. The presence of this band is associated with the impurity of the trigonal phase of solid solution Tb7O12:Eu3+(x) (space group \(R\bar {3}\)), in which half of the Tb3+ cations occupy the centrosymmetric C3i positions. The spectra of the monoclinic phase (phases 6, 9, and 10) are characterized by a sharp decrease in band intensity ratio (610–614)/(622–624) < 1 relative to this parameter for cubic phases in the range from 3.5 to 5.7 (Table 3). This is explained by a decrease in the oxygen pressure relative to the equilibrium pressure at 1200°C for Tb7O12. However, this effect has been observed earlier in the synthesis of compound B-Gd2O3:Eu3+ (1.6 mol %) [13] when the integrated intensity of the PL spectrum increased as in our case. Sample 6 annealed in an H2 atmosphere is characterized by significant increases in band intensity ratios (610–614)/(541–543), (622–624)/(541–543), (487–491)/(541–543), and (548–552)/(541–543) compared to those for the samples of cubic modification C. This indicates the excitation energy in the emission doublet of the 5D07F2 transition of the Eu3+ ion at 610 and 622 nm is mainly transferred through the high-energy component (541–543 nm) of the 5D47F5 transition doublet of the Tb3+ ion. This is consistent with an increase in the (548–552)/(541–543) emission intensity ratio, which generally indicates increased phonon–electron interactions in sample 6. Moreover, this occurs preferentially in the sublattice of noncentrosymmetric C2 sites of the Tb3+ and Eu3+ cations, since the spatial symmetry of these sites is lowered. In actual fact, the (610–614)/(541–543) band intensity ratio is 2.2 for the noncentrosymmetric C2 sites, while the (594–596)/(541–543) band intensity ratio equals 1.5 for the centrosymmetric C3i sites.

4 CONCLUSIONS

Solid solutions Gd2(1 – x)TbxEuxO3 are characterized by photoluminescent properties that are determined by the modification of the lattice of the Gd2O3 matrix (which can be either cubic or monoclinic), the concentrations of the Tb3+ and Eu3+ activators, the composition of impurity phases, and the distribution of activators over the centrosymmetric (C3i) and noncentrosymmetric (C2) cationic sites of the bixbyite lattice and over the Cs surface states. Phase transformations accompanied with dehydration, dehydroxylation, and decomposition of carbonates occur during the annealing of precursors, i.e., products Gd2(1 ‒ x)TbxEux(OH)y(CO3)z · n(H2O) (x = 1 mol %) of the sol–gel process in the temperature range of 700–1200°C. Solid solutions Gd2O3:(Tb3+)x(Eu3+)x that are mainly formed contain microinclusions of lamellas of the Tb7O12:Eu3+ phase. These microinclusions are eliminated after additional annealing in a hydrogen atmosphere as a result of the reduction of Tb4+ cations into Tb3+ cations. The photoluminescence of such products at λex = 280 nm is characterized by the predominance of the emission efficiency of the bands of Tb3+ ions in the blue and green regions of the spectrum. The Tb3+ and Eu3+ photoluminescence activators preferably substitute Gd3+ ions in the C3i positions, and the excitation energy of the emission of the 5D07F1, 2 transitions of Eu3+ ions is partially transferred from Tb3+ ions mainly along the C3i sites of the sublattice, despite the fact that the concentration of Tb3+ and Eu3+ cations is less than 1.5 mol %. Thus, exchange interactions of activators also exist when their concentrations are 1 mol %. A change in the intensity of PL bands of activators and an increase in the number of Eu3+ cations in the C3i positions correlate with a decrease in the specific surface area of crystallites (an increase in the degree of crystallinity of the samples). Regardless of the air/hydrogen/argon atmosphere, the annealing of the samples at 1200°C for 2 h or longer leads to a phase transition of the matrix into the monoclinic modification with a Tb7O12:Eu3+ impurity, while the emission intensity of Tb3+ ions reaches very high relative values in the bands at 487 and 550 nm. At the same time, the excitation energy of emission from the band of Eu3+ cations at 596 nm is transferred owing to a decrease in the emission intensity of the high-energy band of the doublet of the 5D47F5 transition of the Tb3+ ion at 542 nm.

Samples with an activator concentration of 2.5 mol % have a stable cubic modification of solid solution C‑Gd2O3:Tb3+,Eu3+ (2.5 mol %). The degree of crystallinity of the samples slightly decreases owing to the appearance of crystal growth inhibitors on the surface of crystallites, which are Tb3+ ions in the Cs surface states. A corresponding increase in the concentration of Eu3+ ions in the bulk of crystallites, predominantly in the C3i positions, leads to the practical equality of the intensities of the bands of Eu3+ ions at 612 nm and the bands of Tb3+ ions at 542 and 550 nm.