Energy Transfer Study on Tb³⁺/Eu³⁺ Co-Activated Sol-Gel Glass-Ceramic Materials Containing MF₃ (M = Y, La) Nanocrystals for NUV Optoelectronic Devices

Abstract

In the present work, the Tb³⁺/Eu³⁺ co-activated sol-gel glass-ceramic materials (GCs) containing MF₃ (M = Y, La) nanocrystals were fabricated during controlled heat-treatment of silicate xerogels at 350 °C. The studies of Tb³⁺ → Eu³⁺ energy transfer process (ET) were performed by excitation and emission spectra along with luminescence decay analysis. The co-activated xerogels and GCs exhibit multicolor emission originated from 4fⁿ–4fⁿ optical transitions of Tb³⁺ (⁵D₄ → ⁷F_J, J = 6–3) as well as Eu³⁺ ions (⁵D₀ → ⁷F_J, J = 0–4). Based on recorded decay curves, it was found that there is a significant prolongation in luminescence lifetimes of the ⁵D₄ (Tb³⁺) and the ⁵D₀ (Eu³⁺) levels after the controlled heat-treatment of xerogels. Moreover, for both types of prepared GCs, an increase in ET efficiency was also observed (from η_ET ≈ 16% for xerogels up to η_ET = 37.3% for SiO₂-YF₃ GCs and η_ET = 60.8% for SiO₂-LaF₃ GCs). The changes in photoluminescence behavior of rare-earth (RE³⁺) dopants clearly evidenced their partial segregation inside low-phonon energy fluoride environment. The obtained results suggest that prepared SiO₂-MF₃:Tb³⁺, Eu³⁺ GC materials could be considered for use as optical elements in RGB-lighting optoelectronic devices operating under near-ultraviolet (NUV) excitation.

1. Introduction

The development of materials dedicated to photonic applications is a frontier area of current materials engineering research [1,2]. Thus, considerable efforts are still being made to improve the optical parameters of such materials, e.g., the tuning of emissions in the desired spectrum range. In this case, the willingly studied pathway for generation a white light is related to the mixing of three primary colors—red, green and blue (RGB)—in optical materials. Such red-green-blue multicolor visible light can be achieved via up-conversion of near-infrared radiation (NIR) [3,4,5,6,7] or via the conversion of near-ultraviolet (NUV) photons [8,9,10]. Since rare-earth ions (RE³⁺) exhibit a broad range of emission in the visible (VIS) spectral scope via interactions with NIR and NUV irradiation, they are considered as essential parts in the development of white-light-emitting RGB materials [11,12]. The first of abovementioned ways of generating RGB emission can be realized via NIR up-conversion excitation in doubly (e.g., Yb³⁺/Er³⁺ [3,4], Yb³⁺/Tm³⁺ [4]) and triply doped (e.g., Yb³⁺/Er³⁺/Tm³⁺ [5,6], Tb³⁺/Tm³⁺/Yb³⁺ [7]) optical systems. For example, for Yb³⁺/Er³⁺/Tm³⁺ triply doped β-NaYF₄ microrods, Er³⁺ ions are responsible for the generation of red (the ⁴F_9/2 → ⁴I_15/2 transition) as well as green (the ²H_11/2 → ⁴I_15/2 and the ⁴S_3/2 → ⁴I_15/2 transitions) emissions through a two-photon absorption process involved in Yb³⁺ → Er³⁺ energy transfer. Simultaneously, due to the three-photon assisted process of Yb³⁺ → Tm³⁺ energy transfer, Tm³⁺ ions are able to emit blue light (the ¹D₂ → ³F₄ and the ¹G₄ → ³H₆ transitions) [5]. The second of the abovementioned routes can be realized via the successful conversion of NUV irradiation into VIS light. In this matter, Eu³⁺ ions are treated as a red or reddish-orange light source (the ⁵D₀ → ⁷F_J transitions, J = 0–4) and Tb³⁺ ions are considered as one of the most important sources of green emission (the ⁵D₄ → ⁷F₅ transition). Therefore, co-doping with Tb³⁺/Eu³⁺ ions seems to be a promising strategy for the generation of multicolor luminescence, which plays a key role in RGB optical materials [13,14,15,16].

Another particularly important point in the field of substantial enhancement, the luminescence of RE³⁺ ions is related to the selection of a suitable host lattice with low phonon energy. Since the phonon energies of fluorides—usually in a range from 400 up to 500 cm⁻¹ [17]—are significantly lower compared to phosphates (~1250 cm⁻¹) [18] or borates (~1350 cm⁻¹) [19], they are considered as great candidates for generating an efficient and long-lived luminescence of RE³⁺. Among fluorides, special attention should be paid to YF₃ and LaF₃ crystal phases characterized by wide band gap (>10 eV) and exceptionally low-phonon energies equal to ~358 as well as ~350 cm⁻¹, respectively, in which M³⁺ (M = Y, La) cations from crystal lattices can be easily substituted by RE³⁺ ions without any charge compensation [20,21,22,23]. Due to the above reasons, the oxyfluoride glass-ceramic materials (GCs) containing fluoride nanocrystals are considered as an interesting class of advanced optical materials, which are frequently reported in the literature [24,25,26,27,28]. Indeed, they successfully combine the advantages of the individual fluoride crystal phase with good mechanical strength and thermal durability of oxide hosts [29,30]. It should also be noted that due to the compliance with the principles of green chemistry, the systematic elimination of PbF₂ (characterized by exceptionally low-phonon energy equals to ~250 cm⁻¹ [31]) during preparation is currently a very important aspect. Thus, taking the above considerations into account, Tb³⁺/Eu³⁺ co-doped oxyfluoride GCs containing selected fluoride phases (e.g., YF₃ and LaF₃) seems to be a good choice for the generation of efficient visible emissions.

The conventional melt-quenching method followed by controlled heat-treatment at specified time and temperature conditions is currently the most widely used technique for fabricating the class of oxyfluoride GCs [32,33,34,35,36]. On the other hand, the high melting temperatures of glass-forming components (e.g., 1450 °C, 1500 °C [37,38,39]) increase the risk of volatilization of the fluoride compounds, which may adversely affect the crystallization process of the fluoride fraction. Therefore, an alternative route to obtain oxyfluoride GCs is the sol-gel technique, characterized by low-temperature processing [40,41,42]. This method is based on hydrolysis, condensation and polycondensation reactions of organometallic precursors, usually alkoxysilanes Si(OR)₄ (R = –CH₃, –C₂H₅, etc.) in a liquid phase at room temperature [43]. The in-situ crystallization of fluoride phases is possible due to introduction of a fluorinating reagent into reaction system at initial stages, whose role is commonly played by trifluoroacetic acid (TFA) [44]. Since the sintering of sol-gel materials is carried out at significantly lower temperatures (usually <500 °C) than the conventional melt-quenching of glasses, a risk of volatilization of fluorides is adequately lower, meaning that sol-gel processing is preferred. Lower energy consumption also makes the sol-gel technique particularly advantageous and attractive from an environmental friendliness point of view. Simultaneously, to the best of our knowledge, the investigation of Tb³⁺/Eu³⁺ energy transfer process in oxyfluoride GCs is extremely rarely described in the available literature. In the literature, there is only one excellent description concentrated on YF₃:Tb³⁺, Eu³⁺ nanocrystalline-based GCs fabricated from 44SiO₂-28Al₂O₃-17NaF-(10 − x)YF₃-TbF_3−xEuF₃ (x = 0, 0.1, 0.25, 1) glasses during their heat-treatment at 670 °C per 2 h, as far as we know [45]. Due to the above reasons, it seems to be justified to study the Tb³⁺ → Eu³⁺ energy transfer in sol-gel GCs containing MF₃ (M = La or Y) nanocrystals.

In the present work, the sol-gel oxyfluoride GCs materials containing MF₃:Tb³⁺, Eu³⁺ (M = Y, La) nanocrystals were successfully fabricated during controlled heat-treatment at low temperature (350 °C per 10 h) and characterized by detailed luminescence measurements. The studies were performed by means of excitation and emission spectra along with lifetime measurements. Based on photoluminescence results, the interactions between Tb³⁺ and Eu³⁺ dopant ions were systematically investigated and the incorporation of dopant ions into the fluoride environment was also analyzed.

2. Materials and Methods

The xerogels co-doped with Tb³⁺ and Eu³⁺ ions were synthesized using the low-temperature sol-gel method. All reagents used during the described procedure were taken from Aldrich Chemical Co. The sol-gel synthesis was started with the introduction of tetraethoxysilane (TEOS, 98%), ethyl alcohol (98%), deionized water (from Elix 3 system, Millipore, Molsheim, France) and acetic acid (99.5–99.9%) into round-bottom flasks. The molar ratio of components was equal to TEOS:C₂H₅OH:H₂O:CH₃COOH = 1:4:10:0.5 (90 wt.%). To perform hydrolysis and to initialize a condensation reaction, the components were stirred for 30 min. During the next step, the appropriate amounts of acetates, i.e., M(CH₃COO)₃ (M = Y, La; 99.9%) as well as Tb(CH₃COO)₃ (99.999%) and Eu(CH₃COO)₃ (99.999%) were weighed and dissolved in trifluoroacetic acid (TFA, 99%) and obtained mixtures were added into TEOS-based solutions. The molar ratio was equal to CF₃COOH:M(CH₃COO)₃:Tb(CH₃COO)₃:Eu(CH₃COO)₃ = 5:1:0.05:0.05 (10 wt.%) (M = Y, La). The resultant solutions were mixed for another 60 min. After sol-gel synthesis, the obtained liquid sols were dried at 35 °C for 7 weeks to form colorless and transparent solid xerogels (denoted in the text as XGs). Their further transformation into glass-ceramic materials containing YF₃ and LaF₃ nanocrystals was realized by controlled heat-treatment in a muffle furnace (FCF 5 5SHP produced by Czylok, Jastrzębie-Zdrój, Poland) at 350 °C per 10 h (the temperature was raised by 10 °C/min from room temperature). After this procedure, the samples were slowly cooled down to room temperature (denoted in the text as SiO₂-MF₃, M = Y, La). The successful formation of fluoride nanocrystals (LaF₃ NCs: P6₃/mmc, ICDD PDF-2 No. 08-0461; YF₃ NCs: Pnma, ICDD PDF-2 No. 32-1431) was verified using X-ray diffraction (XRD, X’Pert Pro diffractometer, Panalytical, Almelo, The Netherlands) and the nanocrystals imaging was done via high-resolution transmission electron microscope (HR-TEM, JEOL JEM 3010, Tokyo, Japan). The results are shown in Figure 1. The average diameters of fabricated nanocrystals were estimated to 8.1 nm for LaF₃ and 15.4 nm for YF₃. The in-situ formation of LaF₃ and YF₃ nanocrystals during the thermal decomposition of La(CF₃COO)₃ as well as Y(CF₃COO)₃ in applied heat-treatment conditions (350 °C, 10 h) was confirmed and reported in details in our previous works [46,47].

The excitation and emission spectra as well as decay curves were recorded on Horiba Jobin Yvon FluoroMax-4 spectrofluorimeter (Horiba Jobin Yvon, Longjumeau, France) supplied with 150 W Xe lamp. 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. Excitation Spectra of Fabricated Tb³⁺,Eu³⁺ Co-Doped Sol-Gel Materials

Figure 2 and Figure 3 present the photoluminescence excitation (PLE) spectra of fabricated Tb³⁺/Eu³⁺ co-doped xerogels. The PLE spectra were recorded in the spectral range from 340 to 520 nm and monitored at λ_em = 543 nm and λ_em = 612 nm emissions (the ⁵D₄ → ⁷F₅ green line of Tb³⁺ ions and the ⁵D₀ → ⁷F₂ red line of Eu³⁺, respectively). The recorded bands were attributed to the 4f⁸–4f⁸ and 4f⁶–4f⁶ intra-configurational transitions from both of optically active ions. The bands originating from Eu³⁺ were assigned to the following transitions: ⁷F₀ → ⁵D₄ (363 nm), ⁷F₀ → ⁵G_J, ⁵L₇ (from 372 nm to 389 nm), ⁷F₀ → ⁵L₆ (394 nm) and ⁷F₀ → ⁵D₂ (464 nm). Among the group of excitation bands originated from Tb³⁺ ions, the transitions were ascribed to the electronic transitions from the ⁷F₆ ground level into the subsequent upper-lying states: ⁵L₉ (352 nm), ⁵L₁₀ (370 nm), ⁵D₃ (378 nm) and ⁵D₄ (488 nm).

The selection of excitation parameters for further emission measurements and to study the Tb³⁺ → Eu³⁺ energy transfer process was done from the near-UV (NUV) irradiation area (<400 nm) due to the greater intensity of recorded excitation lines than in visible light (VIS) scope (>400 nm). Indeed, the optical elements considered to be used in RGB-lighting devices should operate under NUV excitation. Since the ⁷F₀ → ⁵L₆ excitation band of Eu³⁺ is the most intense, we decided to perform the luminescence measurements for Eu³⁺ ions using λ_exc = 394 nm wavelength.

As the band associated to the ⁷F₆ → ⁵L₉ transition of Tb³⁺ ions does not coincide with any excitation peak of Eu³⁺, the choice of λ_exc = 352 nm wavelength as the excitation source for the generation of the Tb³⁺ → Eu³⁺ energy transfer seems to be reasonable. Additionally, in order to compare the luminescence behavior of sol-gel samples co-doped with Tb³⁺/Eu³⁺ ions with samples singly doped with Tb³⁺ ions, all photoluminescence measurements for Tb³⁺ ions were performed using the λ_exc = 352 nm excitation line.

It should be also noted that for prepared GC samples, the ⁷F₀ → ⁵L₆ excitation line was split into two separated components, the maxima of which were located at 394 and 397 nm (shown in inset of Figure 2 and Figure 3). Similar results were reported by A.C. Yanes et al. [48] for sol-gel glass-ceramics with 89.9SiO₂-10LaF₃:0.1EuF₃ (mol %) composition. Based on numerous photoluminescence measurements (excitation and emission spectra recorded at different temperatures: room temperature (RT), 100 and 13 K; analysis of emission spectra under excitation at 393 and 396 nm), it was clearly proven that the two PLE components are strictly related to the distribution of Eu³⁺ between the SiO₂ sol-gel host (393 nm component) and LaF₃ nanocrystals (396 nm component). Thus, in the case of our studied sol-gel samples, we also assumed that the origin of such a split after the controlled heat-treatment of xerogels (XG-Y/Tb³⁺,Eu³⁺ and XG-La/Tb³⁺,Eu³⁺) is related to the partial migration of rare-earths from silicate sol-gel host into MF₃ (M = Y, La) nanocrystals. It is quite interesting that deeper splitting of the ⁷F₀ → ⁵L₆ band was observed for SiO₂-LaF₃:Tb³⁺,Eu³⁺ GC and, in consequence, two strong components with maxima at 394 and 397 nm are visible. For the SiO₂-YF₃:Tb³⁺,Eu³⁺ GC sample, the 397 nm component is visible as a weak shoulder. This effect may suggest more efficient segregation of dopant ions in LaF₃ than YF₃ nanocrystals. Analogous results were reported earlier by us for singly doped SiO₂-YF₃:Eu³⁺ [46] and SiO₂-LaF₃:Eu³⁺ [47] GC sol-gel systems.

3.2. Influence of Controlled Heat-Treatment at 350 °C on Tb³⁺ → Eu³⁺ Energy Transfer

The photoluminescence (PL) spectra recorded for prepared silicate xerogels are illustrated in Figure 4 (XG-Y/Tb³⁺ and XG-Y/Tb³⁺,Eu³⁺) as well as in Figure 5 (XG-La/Tb³⁺ and XG-La/Tb³⁺,Eu³⁺). The excitation of Eu³⁺ ions using the λ_exc = 394 nm wavelength resulted in the appearance of luminescence bands located within the reddish-orange light area: ⁵D₀ → ⁷F₀ (578 nm), ⁵D₀ → ⁷F₁ (592 nm), ⁵D₀ → ⁷F₂ (611 nm), ⁵D₀ → ⁷F₃ (645 nm), and ⁵D₀ → ⁷F₄ (698 nm). Since the local framework around Eu³⁺ ions in sol-gel host is non-symmetric, the most intense emission line corresponds to the ⁵D₀ → ⁷F₂ electric-dipole transition and red-to-orange ratio (R/O) values for fabricated xerogels are relatively high (R/O = 3.01 for XG-Y/Tb³⁺,Eu³⁺; R/O = 2.78 for XG-La/Tb³⁺,Eu³⁺).

The PL spectra recorded upon excitation at λ_exc = 352 nm for xerogels singly doped with Tb³⁺ ions revealed two emission bands in the bluish-green spectral scope, i.e., ⁵D₄ → ⁷F₆ (488 nm) and the most prominent ⁵D₄ → ⁷F₅ (543 nm) line. Two another emission bands of Tb³⁺ ions were detected in the yellowish-red range: ⁵D₄ → ⁷F₄ (584 nm) and ⁵D₄ → ⁷F₃ (619 nm). The excitation of Tb³⁺/Eu³⁺ co-doped xerogels using λ_exc = 352 nm wavelength led to generating the characteristic emission lines of Tb³⁺ ions; however, it should be noted that some spectral broadening of the ⁵D₄ → ⁷F₄ as well as the ⁵D₄ → ⁷F₃ bands was observed. Such broadening is a consequence of the energy transfer process from Tb³⁺ to Eu³⁺, which resulted in the appearance of additional luminescence coming from Eu³⁺ dopant (⁵D₀ → ⁷F₁ and ⁵D₀ → ⁷F₂ transitions) [49,50,51,52]. In general, the spectral matching of donor’s emission (Tb³⁺) and acceptor’s excitation (Eu³⁺) regions is a fundamental condition for energy transfer occurrence. In this way, upon irradiation using λ_exc = 352 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 takes place. Since there is spectral overlapping between the ⁵D₄ → ⁷F₅,⁷F₄ emissions of Tb³⁺and the ⁷F₁ → ⁵D₁ and the ⁷F₀, ⁷F₁ → ⁵D₀ excitation bands of Eu³⁺, the energy could be successfully transferred from Tb³⁺ into Eu³⁺ ions as follows [53,54]:

⁵D₄ (Tb³⁺) + ⁷F₁ (Eu³⁺) → ⁷F₅ (Tb³⁺) + ⁵D₁ (Eu³⁺) (1),

⁵D₄ (Tb³⁺) + (⁷F₀, ⁷F₁) (Eu³⁺) → ⁷F₄ (Tb³⁺) + ⁵D₀ (Eu³⁺) (2).

Hence, among the characteristic emission lines from Tb³⁺ ions, additional bands originated from Eu³⁺ can also be recorded. The matching of Tb³⁺ emission and Eu³⁺ excitation as well as illustration of energy levels involved in Tb³⁺ → Eu³⁺ energy transfer process are depicted in Figure 6.

The PL spectra recorded for glass-ceramic materials are shown in Figure 7 (SiO₂-YF₃:Tb³⁺ and SiO₂-YF₃:Tb³⁺,Eu³⁺) and Figure 8 (SiO₂-LaF₃:Tb³⁺ and SiO₂-LaF₃:Tb³⁺,Eu³⁺). For both types of prepared co-doped GC, a well-resolved Stark structure of recorded luminescence bands of Eu³⁺ ions was observed, which points to crystalline-like environment around them. For SiO₂-YF₃:Tb³⁺, Eu³⁺ GCs the following maxima are located at: 586/592/594 nm (⁵D₀ → ⁷F₁), 614 nm/619 nm (⁵D₀ → ⁷F₂), 650 nm (⁵D₀ → ⁷F₃) and 690/692/698 nm (⁵D₀ → ⁷F₄), while for SiO₂-YF₃:Tb³⁺, Eu³⁺ GCs, the maxima of individual emission bands were detected at the following wavelengths: 590 nm (⁵D₀ → ⁷F₁), 612 nm/618 nm (⁵D₀ → ⁷F₂), 649 nm (⁵D₀ → ⁷F₃) and 680/688/692 nm (⁵D₀ → ⁷F₄). Such clear splitting is a consequence of the partial segregation of Eu³⁺ ions inside fluoride crystal lattices. Indeed, when Eu³⁺ ions are inserted into the crystal lattice, the subsequent energy levels get split by the crystal-field effect and the number of sub-levels depends on the local site symmetry. It is reported in the literature that in YF₃ and LaF₃ crystal lattices, Eu³⁺ ions occupy C_s and C_2v point symmetry sites, respectively [55,56]. If Eu³⁺ ions occupy C_s and C_2v site symmetry, the J term of the ⁷F_J levels should split into three (J = 1), five (J = 2), seven (J = 3) and nine (J = 4) sub-levels [57]. However, observation of such strong splitting for glass-ceramic systems is quite difficult due to the partial distribution of Eu³⁺ ions within the amorphous sol-gel host. Moreover, compared to xerogels, a significant increase in intensity of the ⁵D₀ → ⁷F₁ magnetic-dipole transition band was observed, while the intensity of the ⁵D₀ → ⁷F₂ electric-dipole transition is strongly inhibited. Generally, the ⁵D₀ → ⁷F₁ magnetic-dipole transition is orbitally allowed and practically insensitive to symmetry in the local environment around Eu³⁺ ions. Conversely, the ⁵D₀ → ⁷F_0,2–4 electric-dipole transitions are forbidden in centrosymmetric sites due to the same parity of energy levels. However, if Eu³⁺ ions are located in non-centrosymmetric sites, the transitions became allowed as forced electric-dipole transitions due to the mixing of wavefunctions of 4f⁶ sublevels with different J values. Therefore, the ratio of integrated luminescence intensity of the ⁵D₀ → ⁷F₂ electric-dipole transition to the ⁵D₀ → ⁷F₁ magnetic-dipole transition can inform us about local symmetry around Eu³⁺ ions and is called the R/O ratio. A decrease in R/O-ratio value clearly suggests that local symmetry around Eu³⁺ ions is closer to an inversion center. The calculated R/O ratio values after controlled heat-treatment are relatively low (R/O = 0.42 for SiO₂-YF₃:Tb³⁺,Eu³⁺; R/O = 0.89 for SiO₂-LaF₃:Tb³⁺,Eu³⁺), which clearly points to the partial migration of Eu³⁺ ions from sol-gel host into more symmetric fluoride crystal phases.

For the prepared SiO₂-MF₃:Tb³⁺ GC samples, the characteristic emission bands corresponding to the transitions from the ⁵D₄ excited level into the ⁷F₆, ⁷F₅, ⁷F₄ and ⁷F₃ lower-lying states were detected. In the case of SiO₂-MF₃:Tb³⁺,Eu³⁺ co-doped GCs, the mutual coincidence of the luminescence lines originating from both of the rare-earths was clearly observed after excitation at the λ_exc = 352 nm line. Therefore, an appearance of characteristic emission bands coming from Eu³⁺ ions upon excitation of Tb³⁺ confirms the occurrence of Tb³⁺ → Eu³⁺ energy transfer.

It should be particularly pointed out that the intensity of Tb³⁺ emission bands decreased, which was accompanied by the simultaneous enhancement of Eu³⁺ luminescence. Such an effect is much more observable for GC samples than for precursor xerogels. Hence, it could suggest that both of the optically active dopants were incorporated into the crystal lattices of the YF₃ as well as LaF₃ phases, causing there to be a shorter distance between neighboring Tb³⁺-Eu³⁺ pairs in comparison with the sol-gel host. Thus, the Tb³⁺ → Eu³⁺ seems to be more efficient in glass-ceramic materials, which results in a much more intense emission of Eu³⁺ ions.

3.3. Luminescence Decay Analysis of the ⁵D₄ (Tb³⁺) and the ⁵D₀ (Eu³⁺) Levels in Sol-Gel Materials

To further optical examination of prepared xerogels and GC materials, the luminescence decay analysis of the ⁵D₄ (Tb³⁺) and the ⁵D₀ (Eu³⁺) excited states was performed. The analysis of decay profiles allows for the deeper characterization of the Tb³⁺ → Eu³⁺ energy transfer process and for establishing the relation between luminescence lifetimes and distribution of rare-earths within sol-gel materials.

Firstly, we compared the luminescence lifetimes of the ⁵D₄ level of Tb³⁺ ions in singly doped sol-gel materials (Figure 9) and materials co-doped with Tb³⁺/Eu³⁺ ions (Figure 10). The decay curves were recorded upon NUV excitation (λ_exc = 352 nm) and monitoring a green emission related to the ⁵D₄ → ⁷F₅ transition of Tb³⁺ (λ_em = 543 nm). For xerogels, the curves are well-fitted to mono-exponential functions given by:

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

(1)

where τ corresponds to the luminescence decay time. It was observed that the τ (⁵D₄) estimated for XG-Y/Tb³⁺ and XG-La/Tb³⁺ is quite comparable and equals 0.97 ± 0.01 and 0.89 ± 0.02 ms, respectively. Such negligible differences in luminescence lifetime values are caused by considerable similarity of chemical environment around Tb³⁺ in fabricated silicate xerogels. Indeed, the non-radiative relaxation between subsequent J states might occur due to the interaction of electronic levels of RE³⁺ ions with suitable vibrational modes in their nearest surrounding. As was proven earlier by infrared measurements, some liquids (residual solvents used during sol-gel process and products of homo- and heterocondensation reactions) are retained in xerogel pores [46]. In this way, Tb³⁺ and Eu³⁺ ions are coordinated by CF₃COO⁻ anions (C=O groups: ~1650 cm⁻¹, C–F bond: ~1200 cm⁻¹) and OH groups (>3000 cm⁻¹). According to energy gap law, an increase in the non-radiative decay rate is assisted by the decreasing number of phonons needed to cover the energy gap, ∆E. Since OH groups are characterized by higher vibrational energy than CF₃COO⁻ anions, we assumed that OH groups with higher phonon energies play a major role in non-radiative relaxation from the ⁵D₄ state. In this way, a maximum of five OH phonons are needed to cover the ⁵D₄ → ⁷F₀ energy gap of Tb³⁺ ions (∆E = 15,000 cm⁻¹); therefore, the probability of non-radiative depopulation from the ⁵D₄ state in xerogels is relatively high.

A comparison of luminescence lifetimes of the ⁵D₄ state for xerogels singly doped with Tb³⁺ ions and co-doped with Tb³⁺/Eu³⁺ ions reveals a slight shortening when Tb³⁺ ions coexist with Eu³⁺ ions in the same host. Indeed, it was reported that there was a decrease from 0.97 ± 0.01 to 0.82 ± 0.03 ms for XG-Y/Tb³⁺,Eu³⁺ and from 0.89 ± 0.02 to 0.75 ± 0.03 ms for XG-La/Tb³⁺,Eu³. Therefore, such an effect is a clear evidence of Tb³⁺ → Eu³⁺ energy transfer occurrence. The energy transfer efficiency (η_ET) between Tb³⁺ and Eu³⁺ ions could be expressed by the following Equation [54,58]:

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

(2)

where τ and τ₀ correspond to intrinsic the decay lifetime of the ⁵D₄ excited level of Tb³⁺ in the presence and absence of acceptor ions (Eu³⁺) in the host lattice, respectively. Indeed, the shortening of the donor’s lifetime when the acceptor coexists is a measure of its ability for energy transfer. The energy transfer efficiencies for prepared xerogels are depicted in

Table 1. Luminescence lifetimes of the ⁵D₄ state of Tb³⁺ (τ) and ET efficiencies (η_ET) for fabricated silicate xerogels.

Sample	τ(5D4) (ms)	ηET (%)
XG-Y/Tb3+,Eu3+	0.82 ± 0.03	15.5
XG-Y/Tb3+	0.97 ± 0.01	-
XG-La/Tb3+,Eu3+	0.75 ± 0.03	15.7
XG-La/Tb3+	0.89 ± 0.02	-

. The calculated efficiencies of Tb³⁺ → Eu³⁺ energy transfer for both of co-doped xerogels are comparable and they are close to η_ET ≈ 16%. The comparability in η_ET values is a consequence of the chemical similarity of the nearest framework around rare-earths (Tb³⁺, Eu³⁺) in xerogels. Such relatively low efficiencies are correlated with the long interionic distances between Tb³⁺ and Eu³⁺ ions dispersed within silicate sol-gel host.

For glass-ceramic samples the recorded decay curves are well-fitted to double-exponential functions, which can be expressed by following Equation:

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

(3)

where τ₁ is the decay time of a short lifetime component, τ₂ is the decay time of a long component, while the A₁ and A₂ parameters are amplitudes at t = 0. The double-exponential character of the decay curves clearly means that Tb³⁺ ions are distributed between two different surroundings in which decay processes take place with variable rates. The first of them is a silicate sol-gel network in which the luminescence lifetimes of the ⁵D₄ state of Tb³⁺ ions are shorter (τ₁ decay component) due to presence of high-vibrational energy modes in their local framework (numerous Q³ units of SiO₄ tetrahedrons (~1050 cm⁻¹) and residual Si–OH groups (>3000 cm⁻¹)). The second surrounding is according to fluorides, i.e., YF₃ and LaF₃ nanocrystals. Due to their low-phonon energies close to ~358 (YF₃ phase) and ~350 cm⁻¹ (LaF₃ phase), the radiative relaxation from the ⁵D₄ excited state is strongly promoted, and thus, the luminescence lifetimes are considerably prolonged (τ₂ decay component). In other words, the change in decay profiles from mono-exponential functions (for xerogels) into double-exponential functions (for GCs) is evident proof, which allowed us to conclude that dopant ions are located either in a silicate sol-gel host or fluoride nanocrystals.

It is quite interesting that the τ₂(⁵D₄) component for SiO₂-LaF₃:Tb³⁺ GC is longer (6.96 ± 0.27 ms) compared with the τ₂(⁵D₄) component for SiO₂-YF₃:Tb³⁺ GC (4.85 ± 0.40 ms). Since the phonon energies of YF₃ and LaF₃ fluoride phases are almost the same, we assumed that such difference in τ₂ lifetime values may suggest that the migration of Tb³⁺ ions from silicate sol-gel host into LaF₃ phase could be more efficient than into YF₃ nanocrystals. It should be also pointed out that co-doping with Eu³⁺ ions lead to the considerable shortening of τ₂(⁵D₄) components due to occurrence of Tb³⁺ → Eu³⁺ ET process (3.10 ± 0.24 ms for SiO₂-YF₃:Tb³⁺,Eu³⁺ GC and 2.97 ± 0.26 ms for SiO₂-LaF₃:Tb³⁺,Eu³⁺ GC). To estimate the energy transfer efficiencies, we used the average luminescence lifetimes of the ⁵D₄ state for glass-ceramic materials singly doped with Tb³⁺ and co-doped with Tb³⁺/Eu³⁺ ions calculated from the following Equation [59]:

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

(4)

where A₁ and A₂ are fitting constants, and τ₁ and τ₂ are short and long decay components, respectively.

To determine the percentage contribution of short- (τ₁) and long-lived (τ₂) components involved in the total decay process, the following equations were used:

$$\%,{\mathsf{\tau }}_1=\frac{{\mathrm{A}}_1}{{\mathrm{A}}_1+{\mathrm{A}}_2}\times 100\%,$$

(5)

$$\%,{\mathsf{\tau }}_2=\frac{{\mathrm{A}}_2}{{\mathrm{A}}_1+{\mathrm{A}}_2}\times 100\%.$$

(6)

For both SiO₂-YF₃:Tb³⁺ as well as SiO₂-YF₃:Tb³⁺,Eu³⁺ GCs, the percentage contributions of τ₁ and τ₂ are quite comparable and equal nearly 50%. A slightly different observation was attained for GCs containing LaF₃ nanocrystals. The difference in contribution of τ₁ and τ₂ decay components is slightly more visible for GCs containing LaF₃ nanocrystals. For the sample singly doped with Tb³⁺ ions, the contribution of the τ₂ component is slightly larger (56.42%) than that of the τ₁ component (43.58%); meanwhile, co-doping with Eu³⁺ ions results in an increase in the τ₁ component’s contribution (57.16%) and slight decrease in τ₂ contribution (42.84%). The luminescence lifetimes of the ⁵D₄ state of Tb³⁺ ions, their percentage contribution (%, τ_n(⁵D₄)), average lifetimes (τ_avg) and the Tb³⁺ → Eu³⁺ ET efficiencies (η_ET) for fabricated GCs are depicted in

Table 2. Luminescence lifetimes of the ⁵D₄ state of Tb³⁺, their contribution and energy transfer (ET) efficiencies (η_ET) for fabricated glass-ceramics.

Sample	τ(5D4) (ms)	%,τn(5D4)	τavg(5D4) (ms)	ηET (%)
SiO2-YF3:Tb3+,Eu3+	0.86 ± 0.10 (τ1) 3.10 ± 0.24 (τ2)	49.79% (τ1) 50.21% (τ2)	2.62	37.3
SiO2-YF3:Tb3+	0.94 ± 0.12 (τ1) 4.85 ± 0.40 (τ2)	51.50% (τ1) 48.50% (τ2)	4.18	-
SiO2-LaF3:Tb3+,Eu3+	0.67 ± 0.07 (τ1) 2.97 ± 0.26 (τ2)	57.16% (τ1) 42.84% (τ2)	2.44	60.8
SiO2-LaF3:Tb3+	1.34 ± 0.12 (τ1) 6.96 ± 0.27 (τ2)	43.58% (τ1) 56.42% (τ2)	6.23	-

.

It was observed that the average luminescence lifetime for SiO₂-LaF₃:Tb³⁺,Eu³⁺ GC (τ_avg(⁵D₄) = 2.44 ms) is 2.6-fold shorter than for SiO₂-LaF₃:Tb³⁺ GC (τ_avg(⁵D₄) = 6.23 ms), while for SiO₂-YF₃:Tb³⁺,Eu³⁺ GC (τ_avg(⁵D₄) = 2.62 ms), it is 1.6-fold shorter than for SiO₂-YF₃:Tb³⁺ GC (τ_avg(⁵D₄) = 4.18 ms). The greater shortening of τ_avg(⁵D₄) luminescence lifetime for SiO₂-LaF₃:Tb³⁺,Eu³⁺ GC than for SiO₂-YF₃:Tb³⁺,Eu³⁺ GC indicates a more efficient Tb³⁺ → Eu³⁺ energy transfer. Indeed, the calculated ET efficiency value for SiO₂-LaF₃:Tb³⁺,Eu³⁺ (η_ET = 60.8%) is about 1.6-fold higher than for SiO₂-YF₃:Tb³⁺,Eu³⁺ (η_ET = 37.3%). It should also be noted that the transformation from xerogels into glass-ceramic materials during controlled heat-treatment strongly determines the Tb³⁺ → Eu³⁺ energy transfer efficiency, η_ET, which increased from ~16% (xerogels) up to 37.3% (SiO₂-YF₃:Tb³⁺,Eu³⁺ GCs) and 60.8% (SiO₂-LaF₃:Tb³⁺,Eu³⁺ GCs).

The luminescence decay curves of the ⁵D₀ state of Eu³⁺ ions in xerogels and SiO₂-YF₃:Tb³⁺,Eu³⁺ as well as SiO₂-LaF₃: Tb³⁺,Eu³⁺ glass-ceramics were recorded for λ_em = 611 nm (Figure 11). Similarly, as in the case of decay curves recorded for the ⁵D₄ state of Tb³⁺ ions, the curves of the ⁵D₀ level are well-fitted to mono-exponential functions for xerogels and to double-exponential functions for GCs. Relatively short luminescence lifetimes for xerogels are caused by the coordination of Eu³⁺ ions by high-vibrational OH groups, which are involved in the non-radiative depopulation of the ⁵D₀ level (only four phonons are needed to cover energy gap, ∆E = 12,500 cm⁻¹). Since controlled heat-treatment of xerogels resulted in the partial migration of Eu³⁺ ions inside fluoride nanocrystals, we distinguished two significantly different luminescence lifetimes (τ₁, τ₂). From the decays, it could be concluded that there is a clear correlation between the designated luminescence lifetimes and energy transfer efficiencies (η_ET) related to the relative distribution of rare-earths between silicate sol-gel host and fluoride environment. The higher energy transfer efficiency for SiO₂-LaF₃:Tb³⁺,Eu³⁺ (η_ET = 60.8%) results in longer decay components (τ₁(⁵D₀) = 0.74 ± 0.04 ms, τ₂(⁵D₀) = 6.84 ± 0.29 ms) than for SiO₂-YF₃:Tb³⁺,Eu³⁺ (τ₁(⁵D₀) = 0.83 ± 0.08 ms, τ₂(⁵D₀) = 5.67 ± 0.39 ms) with lower Tb³⁺ → Eu³⁺ energy transfer efficiency (η_ET = 37.3%). Indeed, it was observed that the average luminescence lifetime for SiO₂-LaF₃:Tb³⁺,Eu³⁺ (τ_avg(⁵D₀) = 5.87 ms) is longer than the average lifetime for SiO₂-YF₃:Tb³⁺,Eu³⁺ (τ_avg(⁵D₀) = 4.59 ms). Such an effect could also indicate that the interionic distance between neighboring Tb³⁺ and Eu³⁺ ions in LaF₃ nanocrystals seems to be shorter than in the YF₃ crystal lattice. In other words, this could suggest that the segregation of rare-earths inside the LaF₃ crystal phase is greater, which promotes the Tb³⁺ → Eu³⁺ ET. The luminescence decay times, their percentage distribution as well as average lifetimes for both types of fabricated GC sample are depicted in

Table 3. Luminescence lifetimes of the ⁵D₀ state of Eu³⁺ and their contributions for fabricated glass-ceramics.

Sample	τ(5D0) (ms)	%, τn(5D0)	τavg(5D0) (ms)
SiO2-YF3:Tb3+,Eu3+	0.83 ± 0.08 (τ1) 5.67 ± 0.39 (τ2)	66.30% (τ1) 33.70% (τ2)	4.59
SiO2-LaF3:Tb3+,Eu3+	0.74 ± 0.04 (τ1) 6.84 ± 0.29 (τ2)	63.58% (τ1) 36.42% (τ2)	5.87

.

The optical characterization of fabricated co-doped sol-gel samples was supplemented with calculations of quantum yields, Ф_Eu. The quantum efficiencies were calculated from Ф = k_R/k formula in which k is the total decay rate constant (k = 1/τ) and k_R is the radiative rate constant estimated from the following Equation [60]:

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

(7)

In above equation, A_MD,0 is according to the Einstein spontaneous emission coefficient for the ⁵D₀ → ⁷F₁ transition (14.65 s⁻¹ [61]); I_tot is the sum of integrated intensities of the ⁵D₀ → ⁷F_J (J = 0–4) luminescence bands of Eu³⁺; I_MD is the integrated intensity of the ⁵D₀ → ⁷F₁ magnetic-dipole transition and n is the refractive index. The refractive index of YF₃ and LaF₃ nanocrystals is almost the same and equals n = 1.55–1.56. Such values are comparable with previously published results [62,63]. It is quite interesting that the refractive index of LaF₃ nanocrystals is slightly lower than that of the corresponding LaF₃ single crystal, as was proven by Z. Wang et al. [64]. Finally, the calculated quantum yields are depicted in

Table 4. Quantum yields (Ф_Eu) calculated for fabricated xerogels and glass-ceramic materials.

Sample	ФEu (%)
XG-Y/Tb3+,Eu3+	9.2
XG-La/Tb3+,Eu3+	8.1
SiO2-YF3:Tb3+,Eu3+	73.0
SiO2-LaF3:Tb3+,Eu3+	49.5

.

The quantum efficiencies for xerogels are comparable and equal to 9.2% for XG-Y/Tb³⁺,Eu³⁺ and 8.1% for XG-La/Tb³⁺,Eu³⁺. Much greater differences in quantum efficiencies were denoted for fabricated glass-ceramics. In fact, the Ф_Eu calculated for SiO₂-YF₃:Tb³⁺,Eu³⁺ GC equal to 49.5%, meanwhile for SiO₂-LaF₃:Tb³⁺,Eu³⁺ GC sample, the Ф_Eu value achieve even 73.0%. The observed significant difference is clearly related to Tb³⁺ → Eu³⁺ energy transfer, the efficiency of which is much greater for SiO₂-LaF₃:Tb³⁺,Eu³⁺ (η_ET = 60.8%) compared with SiO₂-YF₃:Tb³⁺,Eu³⁺ glass-ceramics (η_ET = 37.3%). Based on the current literature, the luminescence quantum yields for Tb³⁺,Eu³⁺ co-doped nanocrystals are usually in a range from 32% to 61% [65]. It is very interesting that the Ф_Eu value for SiO₂-YF₃:Tb³⁺,Eu³⁺ is in good accordance with the presented data, but the SiO₂-LaF₃:Tb³⁺,Eu³⁺ GC sample is characterized by a greater quantum efficiency value. Furthermore, as was proven by N. Shrivastava et al. [66], the quantum yields for LaF₃:Eu³⁺ nanocrystals could vary from 67% up to 85% depending on the Eu³⁺ concentration. Therefore, the obtained results suggest that fabricated sol-gel glass-ceramics could be considered as quite good candidates for visible light-emitting devices.

4. Conclusions

In this work, SiO₂-MF₃:Tb³⁺,Eu³⁺ (M = Y, La) glass-ceramic materials were prepared via sol-gel method during the controlled heat-treatment of silicate xerogels at 350 °C. The performed systematic photoluminescence measurements confirmed that fabricated sol-gel materials exhibit multicolor emission due to the coexistence of the characteristic emission bands originating from both dopant ions, i.e., Tb³⁺ (⁵D₄ → ⁷F_J, J = 6–3) and Eu³⁺ (⁵D₀ → ⁷F_J, J = 0–4) due to Tb³⁺ → Eu³⁺ energy transfer occurrence. Based on performed luminescence decay analysis from the ⁵D₄ (Tb³⁺) and the ⁵D₀ (Eu³⁺) excited levels, the correlation between luminescence lifetimes and the distribution of rare-earth ions between the silicate sol-gel host and fluoride nanocrystals was clearly proven. Hence, it was suggested that more preferable segregation of rare-earth ions inside LaF₃ nanocrystals occurred. Moreover, the transformation from xerogels into glass-ceramic materials during controlled heat-treatment determines the strong influence on Tb³⁺ → Eu³⁺ energy transfer efficiency, η_ET, which increased from ~16% (xerogels) up to 37.3% and 60.8% (SiO₂-YF₃:Tb³⁺,Eu³⁺ and SiO₂-LaF₃:Tb³⁺,Eu³⁺ glass-ceramics, respectively). The obtained luminescent results clearly suggest that the prepared sol-gel glass-ceramic materials could be considered as promising candidates for use as optical elements in RGB-lighting optoelectronic devices operating under NUV excitation.