Structural and Photoluminescence Investigations of Tb³⁺/Eu³⁺ Co-Doped Silicate Sol-Gel Glass-Ceramics Containing CaF₂ Nanocrystals

Abstract

In this work, the series of Tb³⁺/Eu³⁺ co-doped xerogels and derivative glass-ceramics containing CaF₂ nanocrystals were prepared and characterized. The in situ formation of fluoride crystals was verified by an X-ray diffraction technique (XRD) and transmission electron microscopy (TEM). The studies of the Tb³⁺/Eu³⁺ energy transfer (ET) process were performed based on excitation and emission spectra along with luminescence decay analysis. According to emission spectra recorded under near-ultraviolet (NUV) excitation (351 nm, ⁷F₆ → ⁵L₉ transition of Tb³⁺), the mutual coexistence of the ⁵D₄ → ⁷F_J (J = 6–3) (Tb³⁺) and the ⁵D₀ → ⁷F_J (J = 0–4) (Eu³⁺) luminescence bands was clearly observed. The co-doping also resulted in gradual shortening of a lifetime from the ⁵D₄ state of Tb³⁺ ions, and the ET efficiencies were varied from η_ET = 11.9% (Tb³⁺:Eu³⁺ = 1:0.5) to η_ET = 22.9% (Tb³⁺:Eu³⁺ = 1:2) for xerogels, and from η_ET = 25.7% (Tb³⁺:Eu³⁺ = 1:0.5) up to η_ET = 67.4% (Tb³⁺:Eu³⁺ = 1:2) for glass-ceramics. Performed decay analysis from the ⁵D₀ (Eu³⁺) and the ⁵D₄ (Tb³⁺) state revealed a correlation with the change in Tb³⁺–Eu³⁺ and Eu³⁺–Eu³⁺ interionic distances resulting from both the variable Tb³⁺:Eu³⁺ molar ratio and their partial segregation in CaF₂ nanophase.

1. Introduction

In recent years, the materials doped with rare earths (RE³⁺) are considered to be indispensable in the development of optoelectronics, offering promising applications in LEDs [1], displays [2], lasers [3], or optical thermometry [4]. The proper adjustment of emission (i.e., generation of appropriate color purity and maintaining suitable luminescence lifetimes) usually requires the involvement of several RE³⁺ ions incorporated into the same host material [5,6,7,8,9,10,11,12]. Indeed, the mutual interactions between them—depending on the concentration of donor and acceptor as well as the type of host—allow for successful tailoring of the above-mentioned optical parameters.

Among numerous combinations of RE³⁺ in doubly- or triply-doped materials, the optical system consisting of Tb³⁺ and Eu³⁺ is a very promising strategy for the generation of multicolor luminescence, which plays a key role in the development of red-green-blue (RGB) optical materials. The Tb³⁺/Eu³⁺ energy transfer has been extensively explored and described in various types of phosphors, e.g., LaBWO₆ [13], Tb₂MoO₃O₁₂ [14], ScPO₄ [15], or KAlP₂O₇ [16]; meanwhile, such studies have not been as common for glass-ceramic materials so far, where Tb³⁺ and Eu³⁺ ions could be distributed between the amorphous glassy host and crystal phase characterized by different decay rates. The evidence of Tb³⁺/Eu³⁺ ET is stated by the shortening of a lifetime from the ⁵D₄ state of Tb³⁺ in the presence of acceptor ions (Eu³⁺). It is reported in the literature that the fluorescence decay becomes quicker with the increment of Eu³⁺ content, which accelerates the Tb³⁺/Eu³⁺ ET and makes it more efficient [17,18,19,20,21,22]. Furthermore, a comparative analysis of energy transfer efficiency, η_ET, in precursor glasses and glass-ceramics is also carried out to demonstrate the impact of controlled crystallization on Tb³⁺/Eu³⁺ ET. Such studies were performed in excellent work by Chen et al. [19] for 44SiO₂-28Al₂O₃-17NaF-(10-x)YF₃-1TbF₃-xEuF₃ (x = 0, 0.1, 0.25, 0.5, 1) (mol%) glasses and derivative glass-ceramics fabricated at 670 °C. The η_ET for a glass containing 0.1 mol% Eu³⁺ achieved as low value as 1.39%, which finally grew to 30.28% for a glass containing 1 mol% Eu³⁺. Further, the authors have clearly proven that a crucial role in Tb³⁺/Eu³⁺ ET plays is glass crystallization, which results in significant growth in η_ET values from 16.63% (for glass-ceramic containing 0.1 mol% Eu³⁺) up to 47.70% (for glass-ceramic containing 1 mol% Eu³⁺). Similarly, an impact of controlled ceramization on Tb³⁺/Eu³⁺ ET behavior was also studied by Hu et al. [21], who found that η_ET increased from 8.7% for glasses with composition 45SiO₂-20Al₂O₃-10CaO-24.04CaF₂-0.05TbF₃-0.01EuF₃ (mol%) up to 14.0% for nano-glass-ceramic produced at 700 °C.

It should be pointed out that the majority of glass-ceramics containing Tb³⁺ and Eu³⁺ ions characterized and described in literature was prepared by the conventional melt-quenching method followed by controlled heat-treatment at the specified conditions of time and temperature [17,18,19,21,22,23,24,25,26]. An alternative route to the fabrication of glass-ceramics is the sol-gel technique, which offers quite easy fabrication of materials with complex compositions, which could be difficult to obtain via the melt-quenching technique [27,28,29,30,31]. Moreover, particular research attention has been focused on oxyfluoride glass-ceramics, which possess higher chemical and mechanical stability than fluoride glasses and lower phonon energies than oxide glasses. Among fluorides, the calcium fluoride, CaF₂, is an optically isotropic crystal characterized by a broad region of transparency from 0.13 up to 9.5 μm, wide bandgap (12 eV), and relatively low phonon energy (~466 cm⁻¹) [32,33,34]. Those features of CaF₂ crystals are urgent to be a suitable medium for optically active rare earths, widely dedicated to fulfilling many sophisticated, active functions for optoelectronics. Indeed, the optical materials based on the CaF₂ phase are frequently applied to generate an efficient up- [35] and down-conversion luminescence [36] or white light emission [37]. Therefore, such materials could be successfully predisposed for use in laser technologies [38], bio-imaging [39], or to increase the efficiency of solar cells [40]. Moreover, according to our previous results for sol-gel nano-glass-ceramics containing divalent metals fluorides, MF₂ (M = Ca, Sr, Ba), singly-doped with Eu³⁺ ions, the most efficient segregation of Eu³⁺ inside fluoride crystal lattice, was reported for the SiO₂-CaF₂ system [41,42,43]. Indeed, a clear correlation was observed between the average decay time of the ⁵D₀ state and growing difference in ionic radius of Eu³⁺ and each of individual M²⁺ cation in the following order: Ca²⁺ → Sr²⁺ → Ba²⁺ (SiO₂-CaF₂:Eu³⁺: τ_avg = 11.92; SiO₂-SrF₂:Eu³⁺: τ_avg = 7.77; SiO₂-BaF₂:Eu³⁺: τ_avg = 4.08 ms). Therefore, due to the efficient and long-lived luminescence in fabricated SiO₂-CaF₂:Eu³⁺ nano-glass-ceramics, it seems that this material could be considered as a very promising host to study the Tb³⁺/Eu³⁺ ET. Indeed, we reported interesting results for Tb³⁺/Eu³⁺ ET for sol-gel systems containing selected trivalent metals fluorides, MF₃ (M = Y, La) [44]; hence, we performed such measurements for nano-glass-ceramics with divalent metal fluoride, CaF₂. Moreover, to the best of our knowledge, the investigation of Tb³⁺/Eu³⁺ ET in oxyfluoride sol-gel glass-ceramics is rarely described in the available literature. Indeed, it was examined only in SiO₂-SrF₂ [20] and SiO₂-BaGdF₅ [45] sol-gel nano-glass-ceramics; however, there are no reports on the SiO₂-CaF₂ system so far. Due to the above reasons, it seems justified to study the Tb³⁺/Eu³⁺ ET in silicate sol-gel glass-ceramics containing CaF₂ nanocrystals.

In this work, we fabricated and examined the series of sol-gel SiO₂-CaF₂ nano-glass-ceramics co-doped with Tb³⁺ and Eu³⁺ ions with the variable Tb³⁺:Eu³⁺ molar ratio (0.05:x, where x = 0.025, 0.05, 0.075, 0.1). The CaF₂ phase was obtained via in situ crystallization from Ca(CF₃COO)₂ at as low a temperature as 350 °C, and its presence was verified using XRD measurements and TEM microscopy. The changes in photoluminescence behavior of fabricated sol-gel materials have been described in association with the variable Tb³⁺:Eu³⁺ molar ratio, as well as the structural transformation from amorphous xerogels into nano-glass-ceramics. Based on photoluminescence results, the interactions between Tb³⁺ and Eu³⁺ ions were systematically investigated. Indeed, a clear correlation was observed between the R/G ratio and energy transfer efficiency (η_ET), as well as decay times of the ⁵D₄ state (Tb³⁺) as the Tb³⁺:Eu³⁺ molar ratio gradually decreased. Additionally, the lifetimes of the ⁵D₀ excited level of Eu³⁺ ions were also determined. The obtained sol-gel materials exhibited bright multicolor luminescence tuned when the Tb³⁺:Eu³⁺ molar ratio was changed.

2. Materials and Methods

The series of xerogels co-doped with Tb³⁺ and Eu³⁺ ions were synthesized using the sol-gel method described elsewhere [41,46]. All reagents were taken from Sigma Aldrich Chemical Company (St. Louis, MO, USA). After pre-hydrolysis of the mixture containing TEOS, ethyl alcohol, deionized water, and acetic acid in molar ratio equals 1:4:10:0.5 (90 wt.%), the solutions of Ca(CH₃COO)₂, Tb(CH₃COO)₃, and Eu(CH₃COO)₃ in water and trifluoroacetic acid (TFA) were added dropwise. For each sol-gel sample, a mixture of TFA and acetates was 10 wt.%, and the molar ratio equaled TFA:Ca(CH₃COO)₂:Tb(CH₃COO)₃: Eu(CH₃COO)₃ = 5:1:0.05:x (where x = 0, 0.025, 0.05, 0.075, and 0.1). The obtained sols were dried at 35 °C for seven weeks to form solid xerogels. The following xerogel samples were denoted as follows: XG_1Tb0.5Eu (x = 0.025), XG_1Tb1Eu (x = 0.05), XG_1Tb1.5Eu (x = 0.075), and XG_1Tb2Eu (x = 0.1). The glass-ceramic materials were obtained by controlled heat-treatment at 350 °C for 10 h. Such glass-ceramics were denoted as appropriate GC samples: GC_1Tb0.5Eu, GC_1Tb1Eu, GC_1Tb1.5Eu, GC_1Tb2Eu. The sol-gel samples singly-doped with Tb³⁺ ions were also prepared (XG_1Tb, GC_1Tb) to compare with the luminescence properties of Tb³⁺/Eu³⁺ co-doped materials.

The sol-gel network’s vibrational modes were identified using the Nicolet iS50 ATR spectrometer (Thermo Fisher Scientific, Waltham, MA, USA) in a frequency region 500–4000 cm⁻¹. The X-ray diffraction analysis of fabricated xerogels and glass-ceramics was performed using an X’Pert Pro diffractometer supplied by PANalytical with CuK_α radiation (Almelo, The Netherlands). The microstructure of fabricated glass-ceramics was observed using the JEOL JEM 3010 electron transmission microscope operated at 300 kV (JEOL JEM 3010, Tokyo, Japan). The excitation and emission spectra, as well as luminescence decay curves were recorded on Horiba Jobin Yvon FluoroMax-4 spectrofluorimeter supplied with 150 W Xe lamp (Horiba Jobin Yvon, Longjumeau, France). The spectra were recorded with 0.1 nm resolution, and the decay curves were recorded with 2 μs accuracy. All structural and luminescence measurements were carried out at room temperature.

3. Results and Discussion

3.1. Structural Characterization: XRD, TEM, and IR Spectroscopy

In general, fabricated sol-gel materials’ structural properties strongly determine the local environment around Tb³⁺ and Eu³⁺, which are crucial in understanding any changes in their photoluminescence behavior (i.e., emission spectra and decay profiles). The detailed studies in this aspect (the structural evolution from sols, through gels, and xerogels, up to nano-glass-ceramics) were systematically investigated and described for the similar SiO₂-LaF₃:Eu³⁺ system in our previous work [47]. Therefore, to explain the luminescence features of fabricated SiO₂-CaF₂:Tb³⁺, Eu³⁺ samples, a brief comment on their structural properties was also presented below.

The performed heat-treatment of xerogels is responsible for both in situ crystallization of CaF₂ nanophase (verified by XRD and TEM measurements, Figure 1) and evolution of the silicate sol-gel network (indicated by IR-ATR measurements, Figure 2). Such measurements were performed for representative XG_1Tb1Eu and GC_1Tb1Eu co-doped samples. As was demonstrated in Figure 1, a broad halo pattern was recorded for xerogel, which indicates an amorphous nature without long-range structural order. The diffraction reflexes characteristic for the CaF₂ phase crystallized in Fm3m space group (ICDD, PDF-2 No. 65-0.0535) were observed after controlled ceramization. The broadening of recorded diffraction lines indicates the crystallization of the CaF₂ phase in nanoscale, the average crystal size of which was estimated to be 8.1 ± 0.4 nm from the Scherrer formula:

$$\mathrm{D}=\frac{\mathrm{K}\mathsf{\lambda }}{\mathrm{B}\cos \mathsf{\theta }}$$

(1)

where D is the crystal size, K is a constant value (in our calculations, we took K = 1), λ is the X-ray wavelength (1.54056 Å, CuK_α), B is a half of the diffraction line, and θ is the diffraction angle [48]. Another method that allowed us to determine the crystallites’ size, which is an extension of the Scherrer equation, was the Williamson–Hall method, based on plotting the following dependence for several reflexes derived from the same crystalline phase:

$$\mathsf{\beta }\cos \mathsf{\theta }=\frac{\mathrm{K}\mathsf{\lambda }}{\mathrm{D}}+4\left(\frac{\Delta \mathrm{a}}{\mathrm{a}}\right)\sin \mathsf{\theta }$$

(2)

in which β is a half of the diffraction line and (Δa/a) refers to the lattice deformation [49]. From the Williamson–Hall plot’s linear fit, the crystallite size was estimated to be 16.6 ± 1.5 nm (the “chi-square” regression coefficient was equaled to 0.97). Since the Williamson–Hall approach considers crystal imperfections and lattice distortion, as well as apparatus factors, this method allowed us to determine the mean crystallite size more reliably than the Scherrer method (indeed, the latter does not consider such distortions of the crystal lattice). Indeed, the size of CaF₂ nanocrystals from the TEM image (inset of Figure 1) was in more prominent agreement with the data obtained by the Williamson–Hall method (16.6 ± 1.5 nm) than by the Scherrer equation (8.1 ± 0.4 nm). It should be noted that the average crystal size estimated by the Williamson–Hall method was two-fold larger than from the Scherrer equation. Taking into account the factors related to the prepared sample itself, we suppose that one of the reasons for the discrepancy in the values estimated by various methods, apart from the differences in the ionic radii of Ca²⁺ (1.00 Å) [50], Tb³⁺ (1.04 Å), and Eu³⁺ (1.07 Å) [51], could be related to the charge compensation when trivalent cations were substituting divalent Ca²⁺ in the CaF₂ crystal lattice. Indeed, to balance an excess of the positive charge introduced by the RE³⁺ ion, the F^- anions were distributed in interstitial positions [52]. Such interstitial F^- anions induced a local distortion in the crystal lattice due to the repulsion between them. Therefore, we assumed that the charge compensation effect might contribute to the identified difference in crystal size.

Besides the crystallization of the CaF₂ nanophase, the track of structural changes inside the sol-gel host is also important to explain Tb³⁺ and Eu³⁺ dopant ions’ luminescence properties. The sol-gel systems were dynamic during controlled heat-treatment at 350 °C performed for the next 10 h because it induced evaporation of residual solvents (water and organic liquids) from a microporous structure and polycondensation reaction with the participation of Q² ([Si(O_1/2)₂O₂]⁻²), Q³ ([Si(O_1/2)₃O]⁻), and Q⁴ units ([Si(O_1/2)₄]) (in the description “O_1/2” is corresponding to each oxygen atom, which is involved in the formation of Si–O–Si bridge; however, “O” is according to a non-bridging oxygen atom; therefore higher value of “n” index is, the less Si–OH unreacted groups are). To examine such structural evolution, the IR-ATR spectra were recorded in the frequency region from 500 to 4000 cm⁻¹ (Figure 2), and the bands were assigned to appropriate vibrational modes based on literature data [53,54,55,56,57,58]. Generally, we distinguished three primary regions classified by the functional groups causing the characteristic vibrations: 3750–2500 cm⁻¹ (OH groups and C–H bond), 1820–1510 cm⁻¹ (C=O bond, Si–OH groups, and adsorbed H₂O), and 1275–500 cm⁻¹ (silicate host). For a more in-depth interpretation of oscillations that occurred in the sol-gel network, deconvolution was performed. The peak fitting during the deconvolution was done using a Gauss function with a squared regression coefficient of R² ≥0.998.

Firstly, we analyzed the origin of deconvoluted bands (denoted as A–E) identified within the 1275–620 cm⁻¹ spectral region. It was established that band A (~1200 cm⁻¹) was corresponding to the TO₄ mode of created Si–O–Si siloxane bridges, band B (~1140 cm⁻¹) was correlated with oscillations within Q⁴ structural units ([Si(O_1/2)₄]), and band C (~1070 cm⁻¹) could be associated with the TO₃ mode of Si–O–Si siloxane bridges. Bands D (~1030 cm⁻¹) and E (~960 cm⁻¹) were related to Q³ and Q² units’ oscillations, respectively [53,54,56]. According to the literature, bands A and B could also be interpreted as vibrations originated from the C–F bond inside –CF₃ groups in trifluoroacetates [55]. It was clearly visible that the intensities of bands A, B, and E decreased during the transformation of xerogels into glass-ceramics. Indeed, since Ca(CF₃COO)₂ underwent thermal decomposition during controlled ceramization [41], the vibrations from the C–F bond disappeared for glass-ceramics, and bands A and B should have originated only from vibrations inside the Si–O–Si siloxane bridges. Further, a decrease in the intensity of band E clearly pointed to the conversion of Q² structural units into Qⁿ (n = 3, 4) units as a consequence of polycondensation reaction. The additional weak bands located at the 616 and 561 cm⁻¹ frequency region indicated some cyclic structures inside the sol-gel host [53,56].

Finally, we analyzed the broad band recorded in a frequency region from 3750 up to 2500 cm⁻¹. The deconvolution revealed the presence of three bands originated from different types of OH moieties: Geminal or vicinal Si–OH groups (~3630 cm⁻¹, band F), hydrogen-bonded Si–OH groups (~3450 cm⁻¹, band G), and hydrogen-bonded OH groups originated from residual water and organic compounds (~3200 cm⁻¹, band H). It should be noted that deconvolution also revealed the fourth component band (~2990 cm⁻¹, band I), which corresponded to the vibrations of C–H bonds [53,54,56]. It was observed that an indicated broad band was much more intense for xerogels than for glass-ceramics (indeed, to show the deconvolution better, the band’s intensity was fivefold enlarged). In fact, such a relatively strong band for xerogels was a consequence of “trapping” of water and organic liquids inside pores via hydrogen-bonding created with unreacted Si–OH groups. During ceramization, the band was significantly reduced, which clearly evidenced successful evaporation of water and organic liquids as well as a continuation of polycondensation reaction. The conclusions from the above observations could also be confirmed by the behavior of the infrared signal located at ~1660 cm⁻¹, which was attributed to the vibrations of the C=O bond, Si–OH surface groups, and adsorbed water [55,57,58]. The indicated band was well-observable for xerogels, and it almost completely disappeared for glass-ceramics. According to IR and XRD results, the graphical visualization of sol-gel evolution during performed ceramization at 350 °C, as presented in Figure 3.

3.2. Luminescence Properties of Fabricated Sol-Gel Materials

3.2.1. Determination of Local Symmetry Using Spectroscopy of Eu³⁺ Ions as Spectral Probes

The emission spectra recorded for fabricated XG_1TbxEu xerogels using λ_exc = 395 nm excitation line were shown in Figure 4. The characteristic ⁵D₀ → ⁷F_J luminescence bands of Eu³⁺ ions were detected in the reddish-orange light area and their maxima were located at following wavelengths: 578 (J = 0), 592 (J = 1), 615 (J = 2), 647 (J = 3), and 698 nm (J = 4). A gradual increase in their intensity was observed, when the Tb³⁺:Eu³⁺ molar ratio systematically changed from 1:0.5 to 1:2. It is clearly visible that among series of recorded bands, the ⁵D₀ → ⁷F₂ one was the most prominent line for all xerogels before their controlled ceramization, despite the Tb³⁺:Eu³⁺ molar ratio. Indeed, Eu³⁺ ions were frequently used as spectral probes due to the characteristic nature of their transitions. The ⁵D₀ → ⁷F₁ is a magnetic-dipole transition (MD) in nature, the intensity of which is rather independent of the host. In contrast, the ⁵D₀ → ⁷F_0.2–4 are electric-dipole transitions (ED) known to be forbidden by the Laporte selection rule and may occur due to mixing the 4f orbitals with the opposite parity at the low-symmetry sites. Among ED transitions, the ⁵D₀ → ⁷F₂ one has a hypersensitive nature, and its intensity is easily affected by the local vicinity: It is promoted in low-symmetric frameworks; meanwhile, it is inhibited in more symmetric environments. Hence, we could infer about the symmetry based on the ratio of integrated intensities of the bands mentioned above, which is well-known in literature as the R/O ratio (I(⁵D₀ → ⁷F₂)/I(⁵D₀ → ⁷F₁)) [59,60]. XRD results confirmed the amorphous nature of xerogels without long-range order, so we expected relatively high asymmetry in the immediate vicinity of Eu³⁺. Indeed, the calculated R/O ratio values hesitated from 3.50 (XG_1Tb2Eu) to 3.91 (XG_1Tb0.5Eu). On this occasion, it should also be noted that comparability in calculated R/O ratio values clearly pointed to a chemically similar environment of Eu³⁺ ions in all synthesized xerogels.

The emission spectra recorded for GC_1TbxEu glass-ceramics upon excitation at λ_exc = 394 nm were presented in Figure 5. Similar as for xerogels, a series of characteristic bands corresponding to the intraconfigurational transitions within 4f⁶ manifold were registered: ⁵D₀ → ⁷F₀ (577), ⁵D₀ → ⁷F₁ (592), ⁵D₀ → ⁷F₂ (612), ⁵D₀ → ⁷F₃ (648) and ⁵D₀ → ⁷F₄ (683/689/698 nm). It was clearly observed that the intensity of the emission bands successfully increased with the growing content of Eu³⁺ ions (as the Tb³⁺:Eu³⁺ molar ratio changed from 1:0.5 to 1:2). For each glass-ceramic, the orange emission band corresponding to the ⁵D₀ → ⁷F₁ MD transition maintained the greatest intensity and dominated the ⁵D₀ → ⁷F₂ ED red band. Generally, an almost six-fold decline of R/O ratio value was observed, which equaled close to 0.64 after controlled ceramization (compared with xerogels for which the R/O ratio value was approximately equal to 3.70). Hence, such a decrease in the R/O ratio value pointed to significant changes in the symmetry in the immediate vicinity of Eu³⁺ ions, as well as a change in the nature of the bonding character from covalent to more ionic [61]. When the nearest framework of Eu³⁺ ions was more symmetric (which usually accompanied the migration of Eu³⁺ from an asymmetric amorphous structure without long-range order to crystalline lattice), the probability of the ⁵D₀ → ⁷F_0,2–4 electric-dipole transitions successfully decreased [60]. Indeed, the identified decrease in the R/O ratio value was strictly accompanied by partially entering of optically active ions into precipitated CaF₂ fluoride crystal fraction with long-range order. In other words, such a decline was direct evidence that Ca²⁺ cations from fluoride crystal lattice were successfully substituted by Eu³⁺ ions. It should also be noted that a Stark splitting characteristic in the crystal-field was not observed because part of the Eu³⁺ ions was still distributed in an amorphous sol-gel host [60].

3.2.2. Studies of Tb³⁺/Eu³⁺ Energy Transfer in Sol-Gel Materials with the Variable Tb³⁺:Eu³⁺ Molar Ratio

To select the excitation line appropriate for Tb³⁺/Eu³⁺ ET studies, the photoluminescence excitation spectra for XG_1Tb1Eu co-doped representative xerogel were presented in Figure 6. The spectra were recorded for λ_em = 543 and λ_em = 612 nm emissions of Tb³⁺ (the ⁵D₄ → ⁷F₅ green line), and Eu³⁺ (the ⁵D₀ → ⁷F₂ red line), respectively. The recorded excitation bands of Tb³⁺ ions were associated to the following f-f intraconfigurational transitions: ⁷F₆ → ⁵L₉ (352), ⁷F₆ → ⁵L₁₀ (370), ⁷F₆ → ⁵D₃ (379), as well as ⁷F₆ → ⁵D₄ (487 nm). Simultaneously, the recorded bands were ascribed to the transitions of Eu³⁺ ions from the ⁷F₀ ground state into the subsequent higher energy levels: ⁵D₄ (363), ⁵G_J, ⁵L₇ (from 371 to 390), ⁵L₆ (395), and ⁵D₂ (465 nm). Since the ⁷F₆ → ⁵L₉ transition of Tb³⁺ ions did not coincide with any excitation peak of Eu³⁺, we decided to select the λ_exc = 351 nm excitation line to study the Tb³⁺/Eu³⁺ ET process.

The photoluminescence spectra recorded for XG_1Tb xerogel (upon excitation at λ_exc = 351 nm wavelength) as well as for the XG_1Tb1Eu co-doped representative sample (recorded under excitation at λ_exc = 395 and λ_exc = 351 nm lines) are depicted in Figure 7. The spectra recorded for the XG_1Tb sample revealed two emission bands in bluish-green spectral scope, i.e., ⁵D₄ → ⁷F₆ (488) and ⁵D₄ → ⁷F₅ (542 nm) of the predominant intensity, as well as two other emission bands of Tb³⁺ ions in the yellowish-red area: ⁵D₄ → ⁷F₄ (582) and ⁵D₄ → ⁷F₃ (620 nm). When the XG_1Tb1Eu co-doped sample was excited by the λ_exc = 395 nm line, only the characteristic 4f⁶-4f⁶ emission bands originated from Eu³⁺ ions (⁵D₀ → ⁷F_J, J = 0–4) were recorded. A tune in the excitation wavelength to λ_exc = 351 nm also led to the generation of the characteristic emission lines of Tb³⁺ ions. Such coexistence of emission lines originated from both optically active dopants is due to the energy transfer process from Tb³⁺ to Eu³⁺ [17,18,19,20,21,22]. In the case of the band recorded in a red spectral scope, a slight shift was observed of a maximum from 620 (XG_1Tb) to 618 nm (XG_1Tb1Eu), which was caused by overlapping the weak ⁵D₀ → ⁷F₂ band of Eu³⁺ ions (with a maximum at 615 nm) with the ⁵D₄ → ⁷F₃ band of Tb³⁺ ions (with a maximum at 620 nm). In general, the spectral matching of the donor’s emission (Tb³⁺) and the acceptor’s excitation (Eu³⁺) regions was a fundamental condition for energy transfer occurrence [62]. In this way, upon irradiation using the λ_exc = 351 nm line from NUV spectral region, Tb³⁺ ions could be successfully pumped into the ⁵L₉ level, and then, the non-radiative de-activation to the ⁵D₄ state took place. The excitation energy from the ⁵D₄ state (Tb³⁺) could be successfully transferred into the ⁵D₁ or the ⁵D₀ level (Eu³⁺). Hence, among characteristic emission lines from Tb³⁺ ions, additional bands originated from Eu³⁺ can also be recorded.

The emission spectra of Tb³⁺/Eu³⁺ co-doped xerogels with the varying Tb³⁺:Eu³⁺ molar ratio recorded upon excitation at λ_exc = 351 nm are presented in Figure 8. A slight decrease was observed in the emission intensity of the ⁵D₄ → ⁷F₆ and the ⁵D₄ → ⁷F₅ bands of Tb³⁺ ions when the Tb³⁺:Eu³⁺ molar ratio gradually decreases. Hence, the R/G ratio values ((I(⁵D₀ → ⁷F₂)(Eu³⁺)+I(⁵D₄ → ⁷F₃)(Tb³⁺)/I(⁵D₄ → ⁷F₅)(Tb³⁺)) were estimated. For the XG_1Tb sample, the R/G ratio was defined as the ratio of integrated intensities of the ⁵D₄ → ⁷F₃ red band and the ⁵D₄ → ⁷F₅ green emission line [63]. In the case of the XG_1TbxEu co-doped samples, an additional contribution of luminescence originated from Eu³⁺ ions into total red emission should also be taken into account. Hence, the R/G ratio was calculated as (I(⁵D₄ → ⁷F₃)(Tb³⁺) +I(⁵D₀ → ⁷F₂)(Eu³⁺)/I(⁵D₄ → ⁷F₅)(Tb³⁺)) and its increase can be interpreted as a growing share of Eu³⁺ ions in total generated multicolor luminescence. Indeed, a slight increase in R/G ratio values was determined in the following order: From 0.09 (XG_1Tb), through 0.15 (XG_1Tb0.5Eu), 0.17 (XG_1Tb1Eu), 0.26 (XG_1Tb1.5Eu), to 0.30 (XG_1Tb2Eu). Such an increment of the R/G ratio suggests more efficient Tb³⁺/Eu³⁺ ET when the Eu³⁺ content gradually grew, since the Tb³⁺:Eu³⁺ molar ratio changed from 1:0.5 to 1:2. Nevertheless, such a small increase in the R/G ratio resulted from the relatively large interionic distances between Tb³⁺ and Eu³⁺ ions, characteristic for the amorphous xerogel host.

Additionally, an increasing background for prepared sol-gel samples was also observed, especially at wavelengths <540 nm. Such a background was associated with the wide band from the silicate sol-gel host, as was shown by other authors, e.g., Tomina et al. [64], for different types of Eu³⁺-loaded aminosilica spherical particles. The authors suggest that such a band could result, i.e., from the charge transfer on Si-O bonds or defect from the silicate network. They have proven that such a wide band’s intensity is strictly related to the type of complexes formed by Eu³⁺ ions with amine ligands connected with the silicate sol-gel host. A similar effect was reported by Kłonkowski et al. [20], who synthesized sol-gel glass-ceramics containing SrF₂ singly- and doubly-doped with Tb³⁺/Eu³⁺. The broad band’s origin was explained by defects, like dangling bonds, inside silicate sol-gel host.

The emission spectra recorded for GC_1Tb (upon excitation at λ_exc = 351 nm wavelength) and GC_1Tb1Eu co-doped the representative sample (recorded under excitation at λ_exc = 394 nm and λ_exc = 351 nm lines) are depicted in Figure 9. Similarly as for xerogel, for the GC_1Tb sample, the characteristic emission bands corresponding to the transitions from the ⁵D₄ excited level into the ⁷F₆ (488), ⁷F₅ (542), ⁷F₄ (581), and ⁷F₃ (621 nm) lower-lying states were detected. In the case of the GC_1Tb1Eu sample, the coexistence of the luminescence lines originating from both rare-earth dopants was clearly observed after excitation at λ_exc = 351 nm line. Therefore, an appearance of characteristic emission bands coming from Eu³⁺ ions upon excitation of Tb³⁺ confirmed the occurrence of Tb³⁺/Eu³⁺ ET. It should be particularly pointed out that the intensity of Tb³⁺ emission strongly decreased, accompanied by significant enhancement of Eu³⁺ luminescence. Additionally, the maxima of bands recorded the in 570–630 nm spectral scope were shifted from 581 (for GC_1Tb sample) to 592 nm (for GC_1Tb1Eu sample) for an orange band and from 621 (for GC_1Tb sample) up to 612 nm (for GC_1Tb1Eu sample) for a red band. Indeed, an enhancement of Eu³⁺ emission via Tb³⁺/Eu³⁺ ET was much more effective for glass-ceramics than for xerogels.

The emission spectra recorded under λ_exc = 351 nm for GC_1TbxEu co-doped glass-ceramics are depicted in Figure 10. Based on the collected data, it was established that intensities of the ⁵D₄ → ⁷F_J (J = 5,6) bands of Tb³⁺ in the bluish-green spectral scope are strongly dependent on the Tb³⁺:Eu³⁺ molar ratio. Indeed, the intensity of the Tb³⁺ emission was strongly reduced when the concentration of the acceptor gradually increased, and such an effect was simultaneously accompanied by a well-observable increase in the intensity of Eu³⁺ emission within the reddish-orange spectral scope. The observed correlations in mutual intensities of characteristic emission bands were accompanied by an adequate increase in R/G ratio values from 0.14 (GC_1Tb) and 0.80 (GC_1Tb0.5Eu), through 1.60 (GC_1Tb1Eu), 2.47 (GC_1Tb1.5Eu), and up to 3.76 (GC_1Tb2Eu). Therefore, the increment in the calculated R/G ratio values was more dynamic for glass-ceramics than for xerogels, for which only a slight increase was reported when the Tb³⁺:Eu³⁺ molar ratio decreased. Such a correlation was undoubtedly associated with the decreased interionic distance between Tb³⁺ and Eu³⁺ ions due to their successful entering into the CaF₂ crystal lattice.

For Tb³⁺ ions, the G/B ratio analysis defined as I(⁵D₄ → ⁷F₅)/I(⁵D₄ → ⁷F₆) could also be treated as a useful tool for characterization of the symmetry around Tb³⁺ dopant ions [65]. Since the ⁵D₄ → ⁷F₅ line is a magnetic-dipole in nature and the ⁵D₄ → ⁷F₆ transition is an electric-dipole one, the G/B ratio value should have changed when xerogels were transformed into glass-ceramic counterparts. Hence, the G/B ratio values should have been higher in more centrosymmetric sites [66]. Indeed, some changes in emission lines originated from Tb³⁺ ions could also be observed, similarly as for Eu³⁺. On the other hand, it should be pointed out that the G/B ratio was not as sensitive a spectroscopic probe as the R/O ratio calculated for Eu³⁺ optically active ions. The G/B ratio was calculated for samples singly-doped with Tb³⁺ ions, and the ratio changes from 2.95 (XG_1Tb) to 3.80 (GC_1Tb). The results were consistent with the data presented by us earlier in our previous work, concentrating on Tb³⁺-doped sol-gel materials’ photoluminescence behavior [67]. Based on structural changes undergone during controlled ceramization at 350 °C, the identified differences in G/B ratio values were clearly related to the migration of Tb³⁺ ions from the amorphous silicate sol-gel network into CaF₂ nanocrystals formed during controlled heat-treatment.

3.2.3. Effect of Changing in the Tb³⁺:Eu³⁺ Molar Ratio on Decay Times of the ⁵D₄ (Tb³⁺)

The further evaluation of Tb³⁺/Eu³⁺ ET in fabricated sol-gel materials was based on the decay analysis of the ⁵D₄ excited state of Tb³⁺ ions. Firstly, the interpretation of collected data allowed us to establish a clear correlation between the decay profile (mono- or double-exponential) and type of sol-gel material (i.e., xerogel or glass-ceramic). Indeed, the curves recorded for xerogels were well-fitted to a first exponential decay mode described by the following equation:

$$\mathrm{I}\left(\mathrm{t}\right)={\mathrm{I}}_0\exp \left(-\mathrm{t}/\mathsf{\tau }\right)$$

(3)

where I(t) and I₀ are the luminescence intensities at time t and t = 0, respectively, while τ is the luminescence decay time [68]. Factually, in our xerogels, the rare-earths were chemically bonded with OH moieties and CF₃COO⁻ anions in complex compounds [69]. It should be noted that high vibrational energies characterize such ligands, i.e., >3000 (OH groups) and ~1200, ~1140 cm⁻¹ (CF₃COO⁻ anions) as was demonstrated in the Structural characterization: XRD, TEM, and IR spectroscopy section (Figure 2). According to the energy gap law, the effective phonons with maximum energy located in a local surrounding of RE³⁺ ions (ħω_max) generate the strongest effect on decay times [70]. In this case, since OH moieties’ vibrational energy was the highest, they played a major role in the non-radiative depopulation of excited states. For glass-ceramics, the decay curves were well-fitted to a second exponential decay mode, which can be expressed by the equation:

$$\mathrm{I}\left(\mathrm{t}\right)/{\mathrm{I}}_0={\mathrm{A}}_1\exp \left(-\mathrm{t}/{\mathsf{\tau }}_1\right)+{\mathrm{A}}_2\exp \left(-\mathrm{t}/{\mathsf{\tau }}_2\right)$$

(4)

where A₁ and A₂ are amplitudes, while τ₁ and τ₂ are the decay times of short and long lifetime components, respectively [68]. The double-exponential decay profile, as well as considerable differences in τ₁ and τ₂ values, allowed us to conclude about the distribution of rare-earths between two chemically distinct surroundings characterized by different phonon energies. In fact, part of the RE³⁺ ions migrated during the controlled heat-treatment into the CaF₂ crystal lattice, and formed inside the amorphous sol-gel network as a new chemical environment with low phonon energy (~466 cm⁻¹). Due to such a low phonon energy of the CaF₂ lattice, the multiphonon non-radiative depopulation of excited states was strongly restricted. However, the remainder of rare-earths was located in an amorphous sol-gel host. According to IR-ATR spectra recorded for glass-ceramics (Figure 2), it was observed that an intensity of the broad infrared signal originated from OH moieties was significantly reduced; therefore, a major role in non-radiative relaxation was attributed to Q³ groups (~1030 cm⁻¹). Nevertheless, their phonon energy was greater than that of CaF₂ crystal lattice. Such differences in phonon energies in the nearest surrounding of rare-earths determined the variable rates of radiative depopulation of their excited states: In silicate sol-gel host, the lifetimes were shorter (τ₁ components), while in the CaF₂ crystal lattice, the lifetimes were prolonged (τ₂ components). Based on such distinguished lifetime components and their relative contributions to the total radiative decay profile, the average luminescence lifetime could be calculated using the following formula [71]:

$${\mathsf{\tau }}_{\mathrm{avg}}=\frac{{\mathrm{A}}_1{\mathsf{\tau }}_1^2+{\mathrm{A}}_2{\mathsf{\tau }}_2^2}{{\mathrm{A}}_1{\mathsf{\tau }}_1+{\mathrm{A}}_2{\mathsf{\tau }}_2}$$

(5)

The luminescence decay curves of the ⁵D₄ state (Tb³⁺) recorded for XG_1Tb, GC_1Tb, as well as for individual XG_1TbxEu, GC_1TbxEu co-doped sol-gel samples are measured and plotted in Figure 11 and Figure 12, respectively. The decay curves were recorded upon λ_exc = 351 nm excitation and monitoring λ_em = 541 nm green luminescence of Tb³⁺ ions. A slight shortening of the decay lifetime was observed for xerogels from 1.18 (XG_1Tb) to 1.04, 1.01, 0.96, and 0.91 ms for XG_1Tb0.5Eu, XG_1Tb1Eu, XG_1Tb1.5Eu, and XG_1Tb2Eu, respectively. For glass-ceramics, a change in the Tb³⁺:Eu³⁺ molar ratio from 1:0.5 to 1:2 resulted in significantly more efficient shortening of the average decay time of the ⁵D₄ state from 4.75 (GC_1Tb) to 3.75 (GC_1Tb0.5Eu), 2.59 (GC_1Tb1Eu), 1.92 (GC_1Tb1.5Eu), and 1.55 ms (GC_1Tb2Eu) (the individual values of τ₁ and τ₂ components are depicted in

Table 1. Measured lifetimes of the ⁵D₄ state (Tb³⁺), energy transfer efficiencies, and R/G ratio values for prepared xerogels.

Xerogel	τ(5D4) (ms)	ηET (%)	R/G
XG1Tb	1.18	-	0.09
XG1Tb0.5Eu	1.04	11.9	0.15
XG1Tb1Eu	1.01	14.4	0.17
XG1Tb1.5Eu	0.96	18.6	0.26
XG1Tb2Eu	0.91	22.9	0.30

and

Table 2. Measured lifetimes of the ⁵D₄ state (Tb³⁺), average decay times, energy transfer efficiencies, and R/G ratio values for prepared glass-ceramics.

Glass-Ceramic	τm(5D4) (ms)	τavg(5D4) (ms)	ηET (%)	R/G
GC1Tb	1.40 (τ1) 5.67 (τ2)	4.75	-	0.14
GC1Tb0.5Eu	1.32 (τ1) 4.74 (τ2)	3.75	25.7	0.80
GC1Tb1Eu	0.62 (τ1) 2.99 (τ2)	2.59	45.5	1.60
GC1Tb1.5Eu	0.37 (τ1) 2.08 (τ2)	1.92	59.6	2.47
GC1Tb2Eu	0.20 (τ1) 1.66 (τ2)	1.55	67.4	3.76

). Shorter luminescence lifetimes for xerogels than for glass-ceramics (when we compare the samples with the same Tb³⁺:Eu³⁺ molar ratio) were caused by the coordination of Tb³⁺ ions by high-vibrational OH groups, involved in the non-radiative depopulation of the ⁵D₄ level. Their remarkable removal during controlled heat-treatment and partial segregation of Tb³⁺ inside CaF₂ nanocrystals with low-phonon energy allowed the share of non-radiative processes in relaxation to be reduced significantly; hence, the lifetimes from the ⁵D₄ state for glass-ceramics were longer.

Since the shortening in luminescence lifetimes was caused by an energy transfer from Tb³⁺ towards Eu³⁺ ions, the ratio of luminescence lifetimes of the ⁵D₄ state of Tb³⁺ ions in the presence (τ) and the absence of Eu³⁺ ions (τ₀) could be used as a valuable tool to estimate the energy transfer efficiency [72]:

$${\mathsf{\eta }}_{\mathrm{ET}}=\left(1-\frac{\mathsf{\tau }}{{\mathsf{\tau }}_0}\right)\times 100\%$$

(6)

For xerogels, the η_ET values increased from 11.86% (XG_1Tb0.5Eu) through 14.41% (XG_1Tb1Eu), 18.64% (XG_1Tb1.5Eu) to 22.88% (XG_1Tb2Eu). Compared to xerogels, a prompt increase in η_ET values has been noted for glass-ceramic materials, which reached 25.69% (GC_1Tb0.5Eu), 45.47% (GC_1Tb1Eu), 59.58% (GC_1Tb1.5Eu), and 67.37% (GC_1Tb2Eu). Hence, it was easily observed that the energy transfer efficiencies estimated for glass-ceramic samples were noticeably higher than for analogous xerogels with the same Tb³⁺:Eu³⁺ molar ratio, which was mainly caused by migration of rare-earths into the CaF₂ crystal lattice during controlled heat-treatment, where the interionic Tb³⁺–Eu³⁺ distances were significantly shorter than in the amorphous sol-gel host. Additionally, those results clearly indicated a correlation between η_ET and a change in the Tb³⁺:Eu³⁺ molar ratio from 1:0.5 to 1:2 in fabricated sol-gel materials, which was undoubtedly associated with a higher probability that more Eu³⁺ ions could be located adjacent to Tb³⁺. Indeed, since the Tb³⁺/Eu³⁺ ET is characterized by dipole-dipole interactions [13,16], ET’s probability is proportional to 1/R⁶ (R is the average distance between Tb³⁺ and Eu³⁺). On this occasion, when Tb³⁺ and Eu³⁺ dopants were segregated inside CaF₂ nanocrystal lattice, the interionic distances of Tb³⁺–Eu³⁺ pairs were vastly shortened notably if the Tb³⁺:Eu³⁺ molar ratio changed (from 1:0.5 to 1:2).

For better readability, the correlation between the R/G ratio, η_ET, and τ(⁵D₄) lifetimes for prepared sol-gel samples are graphically presented in Figure 13 and depicted in Table 1 (for xerogels) and Table 2 (for glass-ceramics). The R/G ratio gradually increased when the Tb³⁺:Eu³⁺ molar ratio changed from 1:0.5 to 1:2, pointing to the increasing share of emissions originated from Eu³⁺ ions, along with gradual growth in η_ET values and shortening of the τ(⁵D₄) decay times of Tb³⁺ ions. For xerogels and glass-ceramic materials, such a relation was due to the increasing content of Eu³⁺ ions in accordance with Tb³⁺. The changes in the values of the parameters mentioned above for glass-ceramics materials were much more significant for each change of the Tb³⁺:Eu³⁺ molar ratio, which resulted from the preferential segregation of optically active ions into the CaF₂ nanophase.

3.2.4. The Luminescence Decay Analysis of the ⁵D₀ State of Eu³⁺ Ions

The characterization of luminescence properties of fabricated Tb³⁺/Eu³⁺ co-doped sol-gel materials was supplemented by decay analysis of the ⁵D₀ level (Eu³⁺) upon excitation at λ_exc = 394 wavelength and monitoring λ_em = 592 nm (Figure 14). The decay times were also depicted in

Table 3. Measured lifetimes of the ⁵D₀ state (Eu³⁺) in xerogels (λ_exc = 395 nm excitation).

Xerogel	τ(5D0) (ms)
XG1Tb0.5Eu	0.37
XG1Tb1Eu	0.43
XG1Tb1.5Eu	0.44
XG1Tb2Eu	0.45

(for xerogels) and

Table 4. Measured lifetimes and calculated average decay times of the ⁵D₀ state (Eu³⁺) in fabricated glass-ceramics (λ_exc = 394 nm excitation).

Glass-Ceramic	λexc = 394 nm
	τm(5D0) (ms)	τavg(5D0) (ms)
GC1Tb0.5Eu	0.98 (τ1) 9.04 (τ2)	8.40
GC1Tb1Eu	1.33 (τ1) 9.52 (τ2)	8.59
GC1Tb1.5Eu	1.34 (τ1) 8.79 (τ2)	7.94
GC1Tb2Eu	1.35 (τ1) 7.76 (τ2)	6.96

(for glass-ceramics).

For xerogels, the τ(⁵D₀) lifetime values hesitated from 0.37 (XG_1Tb0.5Eu), 0.43 (XG_1Tb1Eu), 0.44 (XG_1Tb1.5Eu), to 0.45 ms (XG_1Tb2Eu). The growing content of Eu³⁺ caused such a slight increase in decay times in prepared xerogels due to changing the Tb³⁺:Eu³⁺ molar ratio from 1:0.5 to 1:2. The relatively short luminescence lifetimes were caused by numerous OH groups in the immediate vicinity of Eu³⁺ ions in the silicate xerogel host. Interestingly, it was found that the τ_avg(⁵D₀) lifetimes in glass-ceramics exhibited no evident and straightforward correlation with the increasing content of Eu³⁺ ions as was found for xerogels. Indeed, the partial segregation of optically active dopants in CaF₂ nanocrystals was responsible for the effective shortening of average distances between them and may have caused competition between radiative and non-radiative processes. Comparing the individual τ_avg(⁵D₀) lifetime values when the λ_exc = 394 nm wavelength was used as an excitation source, it was easy to observe that changing the Tb³⁺:Eu³⁺ molar ratio from 1:0.5 to 1:1 promoted the slight lifetime prolongation (τ_avg(⁵D₀) = 8.40 for GC_1Tb0.5Eu and τ_avg(⁵D₀) = 8.59 ms for GC_1Tb1Eu). Meanwhile, a further change in the Tb³⁺:Eu³⁺ molar ratio (1:1.5 and 1:2) caused shortening of the calculated average decay time (τ_avg(⁵D₀) = 7.94 for GC_1Tb1.5Eu and τ_avg(⁵D₀) = 6.96 ms for GC_1Tb2Eu). Since the R/O ratio values were almost the same for all fabricated glass-ceramic samples (from 0.59 to 0.64), we could assume that the relative distribution of Eu³⁺ ions between CaF₂ nanocrystals and the amorphous sol-gel host was comparable in any case. Simultaneously, it also meant that the content of Eu³⁺ ions in precipitated CaF₂ nanocrystals should have been proportional to the total concentration of Eu³⁺ introduced during the performed synthesis. Such a relation of the decrease in τ_avg(⁵D₀) values, when the Tb³⁺:Eu³⁺ molar ratio equaled 1:1.5 and 1:2, could be explained by the cross-relaxation process. In this case, an excited Eu³⁺ ion made a downward transition (⁵D₂ → ⁵D₁ and/or ⁵D₁ → ⁵D₀), whereas a coupled unexcited neighboring Eu³⁺ ion made an upward transition (⁷F₀ → ⁷F₄ and/or ⁷F₀ → ⁷F₃) [73]. Such non-radiative relaxation depended on the separation between Eu³⁺ interacting ions; hence, the shortening in the interionic distance inside CaF₂ nanocrystals (promoted when the Tb³⁺:Eu³⁺ molar ratio exceeds 1:1) would be predominantly responsible for such a decrease in τ_avg(⁵D₀). To compare, in the case of Tb³⁺ ions, we excluded an involvement of the cross-relaxation process on luminescence lifetimes of the ⁵D₄ state based on our previous results for SiO₂-PbF₂:Tb³⁺ sol-gel glass-ceramics, for which we reported the non-radiative relaxation mechanism when the molar ratio of Tb³⁺ (in accordance to Pb²⁺ cations) exceeded 0.6:1 [67]. On the other hand, since fabricated sol-gel samples were Eu³⁺ low-concentrated, such shortening of the τ_avg(⁵D₀) decay times when the Tb³⁺:Eu³⁺ molar ratio equaled 1:1.5 and 1:2 could be caused by lattice defects, which are well-known as quenching channels [74]. Indeed, a charge compensation induced the formation of vacancies inside the crystal lattice, the number of which would be greater if greater amounts of trivalent dopant ions entered into CaF₂ nanocrystals [75]. Hence, the defects could be responsible for effective faster depopulation of the ⁵D₀ state when the content of Eu³⁺ grows, resulting in shortening of the τ_avg(⁵D₀) decay times. Moreover, it is interesting to note that the τ_avg(⁵D₀) lifetime was prolonged when the Tb³⁺:Eu³⁺ molar ratio was achieved 1:1, and then reduced when the molar ratio equaled to 1:1.5 and 1:2, whereas the luminescence intensity was still increased. A similar effect was also observed for Eu³⁺-doped silicate hybrid materials [76]. The luminescence intensity successfully grew from 0.1 mol% up to 1 mol%; however, it was reported that the lifetimes of the ⁵D₀ state gradually reduced from 617 (for 0.1 mol% Eu³⁺-doped sample) to 275 μs (for 1 mol% Eu³⁺-doped sample). In the case of our fabricated sol-gel samples, the experimental results from luminescence decay analysis have clearly proven that the variable molar ratio of Tb³⁺:Eu³⁺ and controlled crystallization of amorphous xerogels could be responsible for modulating the character of interionic processes.

Based on recorded emission spectra and performed decay analysis of the ⁵D₀ state of Eu³⁺ ions, the quantum yields, Φ_Eu, were calculated using Φ = k_R/k equation. In this equation, k is the total decay rate constant (k = 1/τ(⁵D₀)), whereas k_R is the radiative rate constant. The value of k_R was estimated from the following relation [77]:

$${\mathrm{k}}_{\mathrm{R}}={\mathrm{A}}_{\mathrm{MD},0}{\mathrm{n}}^3\left(\frac{{\mathrm{I}}_{\mathrm{tot}}}{{\mathrm{I}}_{\mathrm{MD}}}\right)$$

(7)

where I_tot is the sum of integrated intensities of the ⁵D₀ → ⁷F_J (J = 0–4) emission bands of Eu³⁺, I_MD is the integrated intensity of the ⁵D₀ → ⁷F₁ magnetic-dipole transition, n is the refractive index of the host and A_MD,0 denotes the Einstein spontaneous emission coefficient for the ⁵D₀ → ⁷F₁ transition and its value for sol-gel systems is equal to 14.65 s⁻¹ [78]. Similar to previous reports for CaF₂ thin films [79] and CaF₂ nanoparticles produced by the fluorolytic sol-gel process [80], the refractive index of CaF₂ nanocrystals was close to n = 1.44. The quantum efficiencies for xerogels were very similar and hesitated from 9.3% to 10.5%. These values were changed drastically during controlled heat-treatment when xerogels were transformed into glass-ceramic systems. The quantum efficiencies achieved the following values: 75.6% (GC_1Tb0.5Eu), 76.1% (GC_1Tb1Eu), 69.0% (GC_1Tb1.5Eu), and 60.0% (GC_1Tb2Eu). Our calculations are in good agreement with results reported by Sun et al. [81] for Eu³⁺-doped CaF₂ thin films, for which the highest quantum efficiency was estimated to 64.24%.

4. Conclusions

In summary, the Tb³⁺/Eu³⁺ ET was systematically investigated in a series of xerogels and glass-ceramics containing CaF₂ nanocrystals and the variable Tb³⁺:Eu³⁺ molar ratio. The transformation of amorphous xerogels into glass-ceramics was successfully carried out at as low a temperature as 350 °C. A particular emphasis was placed on determining the correlation between the photoluminescence properties of prepared sol-gel materials and controlled crystallization, as well as the change in the Tb³⁺:Eu³⁺ molar ratio. The following points have been established:

• Using spectroscopy of Eu³⁺ ions as spectral probes, it was found that optically active dopants were preferably segregated inside the lattice of CaF₂ nanocrystals during controlled heat-treatment of initial xerogels. Indeed, the ⁵D₀ → ⁷F₁ MD transition occupied the predominant advantage for glass-ceramics, which resulted in an almost six-fold decline in R/O ratio values from approximately 3.70 (for amorphous xerogels) to 0.64 (reported after controlled ceramization);
• The growing R/G ratio (from 0.09 to 0.30 for xerogels, and from 0.14 to 3.76 for glass-ceramics) was observed when the Tb³⁺:Eu³⁺ molar ratio changed from 1:0.5 to 1:2. Notably, in glass-ceramics, the emission of Tb³⁺ ions visibly gradually weakened, while luminescence of Eu³⁺ ions occupied the predominant advantage, significantly enhancing the reddish-orange emission;
• Performed decay analysis revealed an interesting dependence of decay times on change in the Tb³⁺:Eu³⁺ molar ratio, as well as partial segregation of Tb³⁺ and Eu³⁺ ions inside CaF₂ nanocrystals formed during controlled heat-treatment at 350 °C. Indeed, a well-observable gradual shortening in τ(⁵D₄) lifetimes for Tb³⁺ ions when the Tb³⁺:Eu³⁺ molar ratio changed from 1:0.5 to 1:2 was reported for xerogels (from 1.18 to 0.91 ms) and glass-ceramics (from 4.75 to 1.55 ms), and it was accompanied by an adequate increase in η_ET (from 11.9% to 22.9% for xerogels and from 25.7% to 67.4% for glass-ceramics). Higher η_ET values for the glass-ceramics resulted from a significant reduction in interionic distances between Tb³⁺ and Eu³⁺ ions inside the CaF₂ crystal lattice;
• The decay analysis of the ⁵D₀ state (Eu³⁺) clearly revealed that the partial crystallization induced a remarkable prolongation of τ_avg(⁵D₀) lifetimes even to 8.59 ms when the Tb³⁺:Eu³⁺ molar ratio equals 1:1, however, the further change in Tb³⁺:Eu³⁺ caused a slight shortening of decay times (7.94 when Tb³⁺:Eu³⁺ = 1:1.5, and 6.96 ms when Tb³⁺:Eu³⁺ = 1:2), which indicated a competition between radiative and non-radiative processes.