Thermal Evolutions to Glass-Ceramics Bearing Calcium Tungstate Crystals in Borate Glasses Doped with Photoluminescent Eu³⁺ Ions

Abstract

Thermal evolutions of calcium-tungstate-borate glasses were investigated for the development of luminescent glass-ceramics by using Eu³⁺ dopant in a borate glass matrix with calcium tungstate, which was expected to have a combined character of glass and ceramics. This study revealed that single-phase precipitation of CaWO₄ crystals in borate glass matrix was possible by heat-treatment at a temperature higher than glass transition temperature T_g for (100−x) (33CaO-67B₂O₃)−xCa₃WO₆ (x = 8−15 mol%). Additionally, the crystallization of CaWO₄ was found by Raman spectroscopy due to the formation of W=O double bondings of WO₄ tetrahedra in the pristine glass despite starting with the higher calcium content of Ca₃WO₆. Eu³⁺ ions were excluded from the CaWO₄ crystals and positioned in the borate glass phase as a stable site for them, which provided local environments in higher symmetry around Eu³⁺ ions.

1. Introduction

Monolithic materials of glass-ceramics have recently attracted a lot of interest because of their potential applications as toughened ceramics for biomedical uses [1,2,3], ionic conductors for energy conversions [4,5,6], and luminescence phosphors for efficient illuminations and display devices [7,8,9]. To design the synthesis of glass-ceramics with a targeted crystal, a choice of starting glass composition is to be carefully considered. In this study, the material evolution of glass-ceramic bearing calcium tungstate crystals in a glass matrix doped with red-luminescent Eu³⁺ ions was investigated. We aimed for precipitation of calcium tungstate crystals as a single phase in glasses. Additionally, control in local structures around Eu³⁺ ions simultaneously doped with the initial host glasses was attempted. There are two types of calcium tungstate crystals of interest as rare-earth phosphors [10,11,12], tetragonal CaWO₄ and monoclinic Ca₃WO₆. The former is composed of WO₄ tetrahedra [13], while the latter has a double perovskite structure with CaO₆ and WO₆ octahedra [14]. The question is whether Eu³⁺ ions are finally positioned after the precipitation of crystals in glass-ceramics based on calcium tungstates.

The choice of the host matrix is very important for this material elaboration. Among preliminary screening tests with silicate, borate, phosphate, etc., calcium borate was selected as a host matrix for its compatibility with the calcium tungstate crystals and low melting point. Ca₃WO₆ was used as a starting crystal, which is known to exhibit quite high red luminescent purity for Eu³⁺ ions [14]. In this study, glass samples were first prepared from 33CaO-67B₂O₃(in molar) glass and Ca₃WO₆ crystal with various compositional ratios, and their thermal and structural properties were investigated. Additionally, they were then examined to know which crystal phases were obtained by thermal treatment at a temperature higher than the glass transition temperature. This material elaboration could be utilized to develop luminescent glass-ceramics. Hence, thermal properties of glass transition temperature (T_g) and on-set crystallization temperature (T_x), and melting point (T_m) were here reported, as well as the results of elementary analysis. Structures of the pristine glasses were examined by Raman spectroscopy to elucidate the role of the unit structures of borate and tungstate in glass on the precipitation of calcium tungstate crystals. X-ray diffractometry showed that CaWO₄ crystals were precipitated in the borate glassy matrix rather than Ca₃WO₆, despite higher calcium content in the pristine glasses. Photoluminescence properties of Eu³⁺-doped (100 − x)(33CaO-67B₂O₃) − xCa₃WO₆ glass and glass-ceramics were also surveyed and the spectral change due to the precipitation of calcium tungstate was discussed in light of the asymmetry ratio of Eu³⁺ ions [15] derived from the luminescence intensity of electric dipole ⁵D₀-⁷F₂ transition against that of magnetic dipole ⁵D₀-⁷F₁ transition, which revealed the thermal evolution of local structures around Eu³⁺ ions to more stable and symmetric ones that were different from CaWO₄ crystal.

2. Experimentals

2.1. Glass Synthesis & Crystallization

H₃BO₃, CaCO₃, Eu₂O₃, and WO₃ were used as received for the sample preparation. In this study, (100 − x)(33CaO-67B₂O₃) − xCa₃WO₆ (x = 0, 1, 2, 4, 8, 12, 15 and 16 mol%) glasses and 85 (33CaO-67B₂O₃) − 15Ca_2.98Eu_0.02WO₆ glasses were synthesized by a melt-quenching method from homogeneous mixtures of Ca₃WO₆ or Ca_2.98Eu_0.02WO₆ powder in crystal and 33CaO-67B₂O₃ glass powder (denoted as 33CaB). The mixed powder was melted in a platinum crucible by heating at 1100 °C in a furnace for x = 1–8 mol% Ca₃WO₆ and quenched on a metallic plate. To gain glass samples over x = 12 mol%, a higher melting temperature of 1400 °C and water quenching were needed to apply a faster cooling rate.

The details of the respective powders used to prepare the glass samples studied (the pristine) are shown below: 33 CaB glass was synthesized by a melt-quenching method. Firstly, H₃BO₃ was heated at 100 °C for 24 h for the evaporation of adsorbed H₂O molecules. The mixture of baches powder of H₃BO₃ and CaCO₃ was melted in a platinum crucible at 1080 °C for 2 h in a furnace and quenched on a metallic plate. Ca₃WO₆ and Ca_2.98Eu_0.02WO₆ powders were synthesized by the traditional solid-state reaction method. A batch powder with a 3.2/1 molar ratio of CaCO₃/WO₃ was mixed for 12 h with ethanol and a ZrO₂ milling ball, and calcined at 1200 °C for 12 h in an alumina crucible, after drying at 80°C [16]. CaWO₆ powder also gained from the same powder calcination procedure. Ca_0.98Eu_0.02WO₄ powder synthesized by CaCO₃, WO_3, and Eu₂O₃ using a solid-state reaction method. The stoichiometric powders were mixed for 30 min with ethanol in a mortar and calcinated at 1100 °C for 6 h in an alumina crucible.

Glass-ceramics with 2–15 mol% Ca₃WO₆ contents were gained by a heat-treatment at a temperature higher than their respective glass transition point, 660 °C, 665 °C, and 635 °C applied for 2 and 12, 4 and 8, and 15 mol% samples for 90 h in a furnace, as shown in

Table 1. Summary of sample preparation: Quenching condition (N.Q.: Normal Quenching, W.Q.: Water Quenching), Ratio of Glass (G) and Crystal (C) inside the species quenched, Location of crystals (SurF.: Surface, IntF.:Interface between glass and crucible, V.C.: Volume Crystallization, n.d.: not detected), Results of XRD and Glass Transition temperature Tg for the obtained glasses (pristine), Heating condition (Temperature and Time) for fabricating glass-ceramics (GC), Results of XRD and Raman experiments, and Detected phases of crystals after the heat-treatment, for (100 − x)(33CaO-67B₂O₃) − xCa₃WO₆(or Ca_2.98Eu_0.02WO₆) samples. (-: no data; n.a.: not applied).

x	Quenching Condition	Ratio (G:C)	Location of Crystals	XRD of the Pristine	Tg	Heating Condition (Temp./Time)	XRD after Heating	Raman after Heating (Figure S1)	Phases of Crystals
0	N.Q.	10:0	n.d.	G	638.3 °C	800 °C /40 min	-	GC	CaB4O7, CaB2O4
1	N.Q.	10:0	n.d.	G	637.5 °C	800 °C /40 min	-	GC	CaB4O7, CaB2O4, CaWO4
2	N.Q.	10:0	n.d.	G	637.5 °C	660 °C /90 h	GC	GC	CaB4O7, CaB2O4, CaWO4
4	N.Q.	10:0	n.d.	G	637.7 °C	665 °C /90 h	G	GC+G	CaWO4
8	N.Q.	~10:0	SurF.	G	631.3 °C	665 °C /90 h	-	GC	CaWO4
12	W.Q.	9:1	IntF.	G	628.7 °C	660 °C /90 h	GC	GC	CaWO4
15	W.Q.	9:1	IntF.	G	601.4 °C	635 °C /90 h	GC	GC	CaWO4
16	W.Q.	6:4	V.C.	GC	-	n.a.	-	-	-
15, Eu3+	W.Q.	9:1	IntF.	G	-	635 °C /90 h	GC	-	CaWO4

.

2.2. Characterizations

X-Ray Diffraction (XRD) patterns were obtained with Bruker D8 Advance eco by using Cu Kα radiation to confirm glass formation and for the assignment of crystalline phases that were precipitated. Scanning Electron Microscopy (SEM) and Energy Dispersive X-ray Spectroscopy (EDS) (Quanta 200, FEI Co., Hillsboro, OR, USA) were adopted for sample observations and quantitative elemental analysis. Thermal properties were estimated from Differential Scanning Calorimeter (DSC) (NETZSCH DSC404F1). Raman spectroscopy using a green laser (532 nm) as a light source (Thermo Scientific NicoletTM Almega) was adopted for understanding the chemical unit structures in each of the samples. To investigate the luminescence properties, photoluminescence (PL) spectra were examined by a fluorescence spectrophotometer equipped with double monochromators (Czerny-Turner) in excitation and emission (Fluorolog3, Horiba Jobin Yvon), using a 450 W Xe-lamp as an excitation source.

3. Results and Discussion

3.1. Quenched Glasses of (100 − x)33CaB-xCa₃WO₆ (x = 0~16 mol%)

Figure 1 shows a photo-image of the glass samples and glass-ceramics at x = 4 and 8 mol% Ca₃WO₆. The clear and transparent glasses were gained in the range between x = 0 to 4 mol% Ca₃WO₆. The 8 mol% sample appeared partially crystallized, but was at most clear and transparent, as seen in Figure 1. While a normal quenching rate was applied for x =0−8 mol% to obtain the transparent glass samples, quit rapid quenching into water (called “water quenching” here) was needed over x = 12 mol%. However, even after the water quenching, the obtained glasses at x = 12–16 mol% still had a part of crystals. It was because the quenching was conducted so that the crucible with glass melt was put into water from the bottom of the crucible and roughly 80% of the crucible was sunk in the water bath. The top of the melt was not covered with water to avoid unexpected influences of water molecules on the structure of the obtained borate glasses. If the crucible was not completely cooled down to room temperature and then the crucible was taken from the water bath, the supercooled liquid was resultantly kept warm to allow it to be crystallized. Nevertheless, the crystallized portion was only just seen between the interface of glass melt and crucible for x = 12 and 15 mol% and it was found to be small. The volume ratio of glass and crystals was about 9:1 for x = 12 and 15 mol%, and about 6:4 for x = 16 mol%, as shown in Table 1. (Other possibilities will be discussed in Section 3.3. EDS elementary analysis for 33CaB-Ca₃WO₆ glasses.) Figure 2 shows an SEM image of the crystal part of the 15 mol% sample of 85(33CaO-67B₂O₃) – 15Ca₃WO₆. The observed morphology of the crystal is so unique, and the SEM image clarifies the line structure of the surface of the crystal, indicating uniaxial nuclear growth of the tetragonal CaWO₄ phase, which will be discussed later. Toward further experiments, the crystallized parts for 12 and 15 mol% were carefully removed, and then clear and transparent glassy parts were extracted. However, at x = 16 mol%, serious crystallization occurred and thus it was not used for the heat-treatment.

3.2. XRD Patterns of 33CaB-Ca₃WO₆ Glasses & Glass-Ceramics

Figure 3a shows XRD patterns of the samples (x = 0–15 mol%) mentioned above. All XRD patterns shown here have a broad halo characteristic that exhibits their amorphous natures. At x = 0 to 15 mol%, the samples are confirmed to be homogeneous without any crystalline components, which were thus used as pristine glass samples for the further heat-treatment. The above glass formation region is in good agreement with one reported in the previous research [17].

After the heat-treatment, the 2 mol% samples were completely devitrified from the surface to the center of the glass, while the 4 and 8 mol% samples heated became opaque just on the surface and white dots appeared on a part of the surface. The 12 and 15 mol% samples became white and appeared to be completely crystallized after the heat-treatment. The impacts of the heat-treatment were firstly examined by XRD methodology.

Figure 3b shows XRD patterns of the samples after the heat-treatment at a temperature higher than T_g for 90 h (See Table 1). In this figure, a theoretical XRD pattern of the crystal structure of CaWO₄ (ICSD ID No.15586) calculated by using RIETAN-FP [18] is also shown for comparison. From the XRD patterns, a crystalline phase of CaWO₄ was found at x = 12 and 15 mol% of Ca₃WO₆ (the photo-image of them are depicted in Figure S2 as Supplemental data), as well as at x = 15 mol% of Ca_2.98Eu_0.02WO₆. It is noted at x = 15 mol% for non- and Eu³⁺ doped samples, the peaks observed at 2θ ~ 31.5° and 65.6°, assigned to 004 and 008 reflections of CaWO₄ were enhanced in intensity when the patterns were compared with the powder pattern calculated. This result indicates a crystalline orientation to the c-axis for the precipitated CaWO₄ phase. It can also be mentioned that no crystalline secondary or impurity phases were observed at x =12 and 15 mol%. However, many X-ray reflections were detected for the 2 mol% sample after the heat-treatment, which were identified as CaB₄O₇ and CaB₂O₄ phases, as well as the CaWO₄ phase. x = 4 mol%, XRD data showed an amorphous state of the sample, but a small amount of CaWO₄ phase was found by Raman spectroscopy, which was shown in Figure S2 as Supplemental data. Moreover, it should be borne in mind that the single-phase precipitation of CaWO₄ crystal in borate glasses requires a Ca₃WO₆ content higher than 8 mol% against 33CaB glass.

3.3. EDS Elementary Analysis for 33CaB-Ca₃WO₆ Glasses

Figure 4 shows the glass compositions (B₂O₃, CaO, and WO₃) of the 33CaB-Ca₃WO₆ glasses (glass parts) for x = 8 to 16 mol%, estimated by EDS measurements. Qualitatively the decrease in B₂O₃ content and the increase in CaO and WO₃ with the addition of Ca₃WO₆ are matched with the variations of the theoretical contents. However, the experimental values of CaO and WO₃ contents became lower than the theoretical ones for the respective components. However, the experimental value of B₂O₃ content was higher than the theoretical one. It may be caused by the possible evaporation of CaO and WO₃ components during melting because the addition of Ca₃WO₆ to 33CaB glass required a higher melting temperature of 1400 °C. During the melt preparation, the 33CaB component was relatively faster changed to a liquid phase in the initial stage of melting, where Ca₃WO₆ crystals could stay in a solid-state. It can be noted that the deviation between the experimental and theoretical contents was reduced for the higher Ca₃WO₆ concentration. A mechanism causing this behavior is still not known, but it could be speculated that the CaO component from Ca₃WO₆ crystals was preferentially introduced into the 33CaB melt and then made it easy to further incorporate the remaining WO₃ component to the melt for the higher Ca₃WO₆ concentration, especially x = 16 mol%. In Section 3.1, we explained the partial crystallization of the quenched glasses for x = 12 to 16 mol% after the water quenching. The fact that the crystals were found, especially between the glass and crucible, it may result from the decomposition of Ca₃WO₆ to 2CaO and CaWO₄. If it is the case, the observation in Section 3.1 could be explained, such that the CaO component was incorporated into 33CaB melt, while the CaWO₄ was partially melted, but still, CaWO₄ crystals remained after the water quenching. As an alternative possibility, it is deduced that the CaO and CaWO₄ produced from the decomposition of Ca₃WO₆ crystals could be melt, but the high viscosity of the melt brought about the heterogeneity of the melt, and eventually, a CaWO₄ rich liquid was free to be precipitated at the lower part of the quenched glass.

3.4. Thermal Properties

Figure 5a shows the DSC curves of (100 − x) (33CaO-67B₂O₃) – x Ca₃WO₆ glasses. These curves exhibit the typical thermal behavior of glasses with a small endothermic peak, large exothermic peaks, and a large endothermic peak, representing glass transition temperature (T_g), on-set crystallization temperature (T_x), and melting temperature (T_m), respectively. The values of T_g and T_x as a function of Ca₃WO₆ concentration are shown in Figure 5b. The value of T_g slightly decreased from ~640 °C to ~630 °C, up to 12 mol% Ca₃WO₆ concentration and then dropped down to ~600 °C at 15 mol% Ca₃WO₆. The value of T_x increased from 760 °C to ~785 °C between 0 and 2 mol% Ca₃WO₆. For higher Ca₃WO₆ concentrations, the T_x value decreased drastically to ~740 °C until 4 mol% Ca₃WO₆ and then decreased more gently to ~700 °C for the further higher Ca₃WO₆ concentrations.

The parameter T (=T_x − T_g) and Hruby parameter, K_gl given by Reference [19]

$$K_{gl}=\frac{T_x-T_g}{T_m-T_x}$$

(1)

This is shown in Figure 5c to evaluate their glass stability. ΔT represents the temperature interval during nucleation. Thus, ΔT is also used to evaluate the stability of the glassy state against crystallization. The values of ΔT at 1–2 mol% Ca₃WO₆ concentration are higher than 100 °C, indicating that they are more stable against devitrification [20]. However, the ΔT values at Ca₃WO₆ concentrations over 12 mol% were lower than 80 °C, indicating that these Ca₃WO₆-rich glasses are easily crystallized in comparison to the cases of the lower Ca₃WO₆ concentrations. That is well-matched with the fact that CaWO₄ single-phase appeared by crystallization at 12 and 15 mol% Ca₃WO₆. K_gl also showed a similar tendency. The glass-forming ability is in general determined by measuring the critical cooling rates. Firstly, to be mentioned, it is proved that, for thermally stable glass-forming systems, the value of K_gl is more than 0.1 and hence thermal stabilization of glasses can be characterized by high values of K_gl and vice versa [21]. From previous experimental and theoretical researches on the various glass-forming compositions where the critical cooling rate was correlated with the glass stability (K_gl), it is also found that there exists an empirical relation between K_gl and critical cooling rate [22,23]. Therefore, it is plausible to use K_gl for judging the glass-forming ability [24]. The Ca₃WO₆ concentration dependence of K_gl is shown in Figure 5c. The values, K_gl of 12, 15 mol% Ca₃WO₆ are lower than K_gl of 1–8 mol% Ca₃WO₆, indicating that 12, 15 mol% Ca₃WO₆ have the low glass-forming ability. This is corresponding to the fact that the 12, 15 mol% samples needed a higher cooling rate, as using water to quench the glass melt, than the 1–8mol% samples. The higher and lower values of ΔT and K_gl at the 1, 2 mol% and 12, 15 mol% Ca₃WO₆ seem influenced by the same factor: The introduction of WO₃ as a network former to B₂O₃ glass network at the lower concentration would reinforce the resultant glass structures, however, the influence of CaO as an inhibitor to glass-forming must become dominant over WO₃ as a network former at the samples with 8–16 mol% Ca₃WO₆.

3.5. Raman Spectroscopy

Figure 6a,b show area normalized Raman spectra of 0–16 mol% Ca₃WO₆ glass samples and deconvoluted Raman spectra of 33CaO-67B₂O₃ glass. The peaks at 420–485 cm⁻¹ (Nos.2 and 3), 630 cm⁻¹ (No.4), 755–770 cm⁻¹ (No.6), and 950 cm⁻¹ (No.8) can be assigned to B–O stretching modes of borate symmetric and asymmetric vibrations, ring-type metaborate, four- and three-coordinated boron in diborate [25] and orthoborate group, respectively. Additionally, the peaks at 300–350 cm⁻¹ (No.1) and 850–920 cm⁻¹ (No.7) are assigned to the O-W-O bending mode in distorted γ–WO₆ and W–O stretching vibrations of W⁶⁺ = O double bonding [26], respectively. The deconvoluted peak intensity of Nos.1, 7, and 8, and Nos.2, 3, 4, and 6 as a function of Ca₃WO₆ concentration showed in Figure 7a,b, respectively. The peak intensities of Nos. 2, 3, 4, and 6 decrease with Ca₃WO₆ concentration, whereas the peak intensity of No.8 increases with the Ca₃WO₆ concentration. The decreasing peak intensities of Nos.2 and 3 indicate the decrease amount of boron in the glass host. This result corresponds to the decreasing number of B₂O₃ estimated by EDS. The borate ring groups, ring-type metaborate, and diborate have bridging oxygen and work as a glass network former. Orthoborate groups have non-bridging oxygens and are located at the end of glass networks. The above result that bridging oxygen and non-bridging oxygen decreased and increased with Ca₃WO₆ concentration, respectively, shows the host glass structures were strongly affected by Ca²⁺ ions working as a network modifier with the Ca₃WO₆ additions. Thus, the glass sample with lower Ca₃WO₆ concentration included much of the amount of borate ring structures in comparison to the samples with higher Ca₃WO₆ concentration, and it seems that an impurity phase CaB₄O₇, which includes a borate ring structure, appeared in the crystallized 2 mol% Ca₃WO₆ sample.

However, the peak intensity of Nos.1 and 7 exhibited an increasing tendency with Ca₃WO₆ concentration. Especially, peak No.7 showed a significant increase in comparison with peak No.1. In other words, a more significant change in the peak intensity about W = O double bonding than WO₆ bending mode was observed, implying an increase in the number of W⁶⁺ ions and a structural change to WO₄ tetrahedra was more dominant than to WO₆ octahedra. Thus, the CaWO₄ crystal phase composed of WO₄ tetrahedra would be precipitated after the heat-treatment, instead of the Ca₃WO₆ crystal phase with WO₆ octahedra.

To understand the structural changes of 33CaO-67B₂O₃ glass by Ca₃WO₆ addition, more discussion can be given in the following: The borate glass matrix was composed of ring-type metaborates and diborates with bridging oxygens at lower Ca₃WO₆ concentration. However, the addition of more Ca₃WO₆ contents promoted the structural conversion of them to orthborates (BO₃)³⁻ with three non-bridging oxygens, as shown in Figure 7a because the strong ionic field of Ca²⁺ ion is enough to break B-O-B networks of metaborate/diborate and produces the isolated structures of orthoborate. Figure 7a also elucidated a larger number of tungsten oxides of covalent WO₄ (No.7) working as a network formerly in comparison to WO₆ (No.1) working as a network modifier. This is interpreted so that the ionic nature of Ca²⁺ ions mainly affected the borate networks, while more covalent WO₄ entities can take part in the construction of a glass network with the borate structures with bridging oxygens to maintain the glassy structures after the quite rapid quenching. Even at x = 16 mol%, the B₂O₃ concentration was higher than WO₃ (B₂O₃/WO₃ = 3.51) and the increasing orthoborates would make the vitrification difficult from the viewpoint of structural chemistry. It should be borne in mind that the oxygen basicity was also increased with the introduction of Ca²⁺ ions as CaO from Ca₃WO₆ content, which promoted the production of W=O double bonds as well as O-W-O linkages. The W=O bond is not a part of the WO₆, but the WO₄ structure. From these considerations, it can therefore be deduced that the formation of WO₄ structures may support stabilizing their glassy state and also be important to model the structural changes of the calcium borate glasses obtained by the rapid water quenching with the addition of more Ca₃WO₆ contents.

3.6. Luminescence Properties

Figure 8 displays PL spectra and asymmetry ratio (∧) of 15 mol% Ca_2.98Eu_0.02WO₆ glass, glass-ceramic, and Ca_0.98Eu_0.02WO₄ normalized by PL intensity at 593nm. Three peaks of ~578 nm, ~590 nm, and ~615 nm originate from Eu³⁺ luminescence assignable to ⁵D₀–⁷F₀, ⁵D₀–⁷F_1, and ⁵D₀–⁷F₂ transitions, respectively [27]. Asymmetry ratio, which is defined as a PL intensity ratio of electric dipole ⁵D₀–⁷F₂ and magnetic dipole ⁵D₀–⁷F₁ intensities (∧ = I(⁵D₀–⁷F₂)/I(⁵D₀–⁷F₁)), estimated for discussion on the degree in asymmetry of local structure around Eu³⁺ ions [28,29,30]. It is well known that ⁵D₀–⁷F₂ transition intensity is much affected by local asymmetry around Eu³⁺ and ⁵D₀–⁷F₁ transition is independent of the local structure because of the respective natures of electric and magnetic dipole transitions [31]. Asymmetry ratios of 15mol% Ca_2.98Eu_0.02WO₆ glass, glass-ceramic and Ca_0.98Eu_0.02WO₄ are estimated as 3.0, 2.1, and 7.5, respectively. It is found that the asymmetry ratio of 15 mol% Ca_2.98Eu_0.02WO₆ sample is lower in the glassy state than that of the crystal and decreased through the heat-treatment, although despite CaWO₄ crystals, were precipitated in the glass-ceramics. The result suggests that the heat-treatment rearranged Eu³⁺ environments in the sample and eventually allowed the local symmetry of ligand structures around Eu³⁺ ion to be improved, and most of the Eu³⁺ ions were not positioned in the CaWO₄ crystal phase precipitated but in the glass matrix. This is interpreted such that Ca and W ions were used for the precipitation of CaWO₄ in a borate glass matrix and Eu³⁺ ions were forced to be located in a more ionic matrix of calcium borate glass with less W content. This is matched with the results of the Raman investigation. It is assumed that the glass structure becomes a more stable phase, because of ions transfer by heating.

To estimate how the doped Eu³⁺ ions were incorporated in the parent glass or the crystalline phase during the transformation of glass to the glass–ceramic system, Eu^{3+ 5}D₀ decay curves were examined and shown in Figure 9. It is seen that the luminescence decays were almost identical for the pristine glass and the glass-ceramic, while the Eu³⁺ ions in CaWO₄ crystal exhibited a faster ⁵D₀ decay curve. The evaluated lifetimes were 6.248 $\pm$ 0.007, 5.735 $\pm$ 0.008, and 1.479 $\pm$ 0.001 ms for the pristine glass, glass-ceramic, and Ca_0.98Eu_0.02WO₄ crystal, respectively. The PL lifetime is found to start to decrease a bit by the formation of the glass-ceramic, however, the values for the glass and glass-ceramic were almost the same, ~6 ms. However, the crystal Ca_0.98Eu_0.02WO₄ had a shorter lifetime of ~1.5 ms. The difference in the lifetime between the glass/glass-ceramic and the crystal corresponds to the behavior of the asymmetry ratios obtained from the PL data. As seen in a higher resolution PL detection (Figures S3 and S4 in Supplementary), sharp PL lines in ⁵D₀-⁷F₂ transition were more or less observed for the glass-ceramic, from which it is imagined that a part of Eu³⁺ ions could enter the CaWO₄ crystal but the majority of the ions was still in the glass matrix. In general, a higher asymmetry ratio of Eu^{3+ 5}D₀-⁷F_1,2 luminescence means an increase in the probability of ⁵D₀-⁷F₂ transition in electric dipole nature against that of ⁵D₀-⁷F₁ transition in magnetic dipole nature, which will induce the enhancement of total transition probability from ⁵D₀ level and give faster ⁵D₀ luminescence lifetime. This can well explain the difference in the behavior of the observed spectra and PL decays between the glass/glass-ceramic and the crystal. From the results, it can be mentioned that Eu³⁺ ions were still positioned in the glassy matrix even after the crystallization. More to be mentioned finally, the slight decrease in the PL lifetime and the decrease in the asymmetry ratio for the glass-ceramic in comparison to the pristine glass imply the presence of ion-ion interaction between Eu³⁺ ions and/or non-radiative transition in the glass-ceramic with higher symmetric sites for Eu³⁺ ions.

4. Conclusions

Synthesis of glass-ceramic, including the CaWO₄ phase, was succeeded by the heat-treatment at a temperature higher than T_g for (100 − x) (33CaB) − xCa₃WO₆ (x = 2–15 mol%). XRD data clarified the precipitation of CaWO₄ crystals in x = 2, 12, and 15 mol% samples by the heat-treatment. The 2mol% Ca₃WO₆ glass-ceramic sample included crystal phases of CaB₄O₇ and CaB₂O₄. DSC curves and Raman spectra were examined to understand the crystallization mechanism of CaWO₄, CaB₄O_7, and CaB₂O₄ in the glasses. The stability of glassy state ΔT increased in the range from x = 0 to 2 mol% and then decreased in the range of higher than 4mol% Ca₃WO₆ content. The above change corresponded with the numbers of glass network former of borates with bridging oxygens, which was elucidated by Raman spectra. The result suggested that the decreasing borate ring structures helped the crystalization of CaWO₄ without B₂O₃ components by the heat-treatment. Contrary to the structural evolution of borates, the W=O double bondings in WO₄ tetrahedra were significantly increased with the Ca₃WO₆ content, while the WO₆ octahedra bending vibrations tended to slowly increase, which was also found by Raman spectroscopy. This behavior was a cause of the induction of CaWO₄ composed of WO₄ tetrahedral units by the heat-treatment instead of Ca₃WO₆ with WO₆ octahedral units. PL investigation in the scope of asymmetry ratio of Eu³⁺ PL spectra demonstrated that Eu³⁺ ions doped in the 15mol% sample were surrounded by a ligand structure with higher local symmetry after the heat-treatment than Eu³⁺ ions in the pristine glass or CaWO₄ crystals, indicating that the heat-treatment forced Eu³⁺ ions to be located in the more ionic matrix of calcium borate glass with less W content.