Influence of Y₂O₃ Concentration on the Optical Properties of Multicomponent Glasses and Glass–Ceramics

Abstract

The optical properties and structural characterization of multicomponent silicate glasses of low Al₂O₃ and different Y₂O₃ concentrations have been studied. These glasses have also been crystallized to obtain glass–ceramic materials, and their properties have been characterized. The obtained glasses were transparent and their refractive indexes increased with Y₂O₃ concentration. After a heat treatment at 930 °C for 10 min, these glasses maintained their transparency, but a brown color appeared, and after 30 min, those glasses with high Y₂O₃ concentrations turned opaque or white in color. These processes of crystallization for obtaining the new glass–ceramics have been studied by means of FTIR and Raman spectroscopies, and the crystallized materials were characterized with XRD and FE-SEM techniques. These glasses and glass–ceramics have also been characterized by means of UV–vis spectroscopy, and the corresponding optical properties (reflectance, color, band-gap) have been determined as a function of the Y₂O₃ concentrations and the structural properties.

1. Introduction

Glasses, glass–ceramics (GCs) and even ceramic materials containing Y₂O₃ in their compositions have gained high importance during the last few years. This is due to the excellent properties shown by the Y³⁺ ions in the material structure. In the case of glasses, it is well known that the Y³⁺ ions can act as network modifiers or glass formers depending on if they break the Si-O-Si bonds to form non-bridging oxygens (NBOs) or if they take part in the glass structure by bonding with O atoms and forming bridging oxygens (BOs), respectively. Therefore, the function of Y₂O₃ is described as intermediate oxide. For silicate glasses, if the glass structure contains NBO due to the presence of modifier cations (mainly alkaline or alkaline-earth ions), the high-field strength Y³⁺ cations can attract O²⁻ ions and might be bonded to one, two or three O²⁻ to form Si-O-Y bonds strengthening the glass structure [1,2]. However, if the Y₂O₃ concentration exceeds a certain amount, the high field strength of Y³⁺ cations is so strong that it leads to agglomerates, provoking phase separation in the glass and/or phase crystallization [3].

For GC materials, the Y₂O₃ acts as a nucleating agent that leads to forming different crystalline phases that can be in low or high concentrations in the material structure [4]. The sizes of the crystalline phases formed inside the GC structures can be from a few nanometers to micrometers, and therefore, the final material could be transparent, translucent or even opaque [5]. The incorporation of Y₂O₃ into glasses and GC not only modifies their optical properties (transmittance, refractive index, etc.) but can also provide high mechanical properties such as hardness, toughness, elastic modulus and binding strength [6,7], as well as excellent chemical resistance [7], important modifications in the dielectric properties [8] and thermal properties [9,10], etc. These interesting properties have led us to study and analyze the possibility of using Y₂O₃-containing glasses and GCs in a wide range of applications, such as photoelectronic devices [11], mobile display devices [1], sealants for solid oxide fuel cell technology [12], electrical insulation in the electronics industry [8], laser materials [13], radiation shielding [14,15], etc.

The intermediate role of Y₂O₃ has been employed for preparing transparent GC materials [3]. For the obtention of transparent GCs, it is necessary that the glass matrix and the formed crystal phases pose similar refractive indexes, and if this is not achieved, then the sizes of the precipitated crystals must be below the visible wave length and the crystalline volume fraction must be low. Most of the GCs contain crystalline phases whose crystal sizes are in the visible region, i.e., 300–800 nm, which leads to scattering visible light and therefore opacity. Several oxides such as SnO₂, ZrO₂, TiO₂ and P₂O₅ can act as nucleating agents, leading to glass crystallization after specific heat treatments, and therefore, such oxides have been used for improving the crystallization of B₂O₃–Al₂O₃–SiO₂, ZnO–Al₂O₃–SiO₂, MgO–Al₂O₃–SiO₂, Li₂O–2SiO₂ and Li₂O–Al₂O₃–SiO₂ glass–ceramics [7,16]. The refractive index of these GCs is below 1.6 and can be improved by increase in the concentration of such nucleating agents, including other oxides with high refractive indexes, such as La₂O₅, Nb₂O₅, Ga₂O₃, BaO, etc., reaching values close to 1.8 [17]. However, glasses and GCs containing such rare earth oxides are very expensive, and it is necessary to obtain transparent GCs with multicomponent oxides of more common raw materials.

The transparency of GC materials is lost if the size of the precipitated crystalline phases increases with temperature. In general, it is difficult to obtain GCs with high transparency at temperatures higher than 800–1000 °C. Zhao et al. prepared transparent Na₂O–Al₂O₃–B₂O₃–SiO₂ glass–ceramics with Y₂O₃ and ZrO₂ additions (1%mol) at 800 °C [1], Zheng et al. obtained Y₂O₃–Al₂O₃–SiO₂ (15–35%wt.) with P₂O₅ additions at 700 °C [3], Liu et al. prepared Sm³⁺-doped Na₂O-Y₂O₃–SiO₂–P₂O₅ transparent glass–ceramics (containing 7% (mol)) Y₂O₃ at 750 °C [11] and Huang et al. obtained transparent MgO–Al₂O₃–SiO₂ glass–ceramics by adding ZrO₂ and SnO₂ (4% and 2% mol, respectively) with heating at 1080 °C [5].

In a previous work [18], we prepared multicomponent GCs with 14% Al₂O₃ in their compositions and with different Y₂O₃ concentrations, and it was observed that while the pristine glasses were transparent, the final GCs were white and opaque. This opacity was due to two kinds of simultaneous crystallizations: one occurred on the surface and the other in the bulk, and both increased with Y₂O₃ concentration. In this new work, we have prepared new glasses with only 3% Al₂O₃ to improve the mobility of the nucleating ions in the glass structure and thus understand which of the two types of crystallization predominates over the other. The effect of Y₂O₃ on the structural and optical properties of a multicomponent oxide glass crystallized at 930 °C has been studied. Crystallization occurs at low Y₂O₃ concentrations; however, the transparency is maintained due to the low crystallite size developed, while for high Y₂O₃ concentrations, white opaque GCs are obtained.

2. Materials and Methods

2.1. Samples Preparation

The raw materials employed in this work were silica quartz (99.5%), carbonates of Li, Na, K, Ba, Ca and Mg (all of 99.9%), and oxides of Zn, Zr, Al and Y (all of 99.9%). All the glasses were prepared with the same compositions, except that SiO₂ was replaced by Y₂O₃ in concentrations of 3, 6 and 12% in weight in order to analyze its influence on crystallization behavior and the corresponding optical properties. The corresponding mixtures of the raw materials were mixed for 2 h in an agata mortar and then put into an electric furnace to treat at 900 °C for 1 h to carry out the calcination of carbonates. Then, the temperature was raised to 1500 °C and maintained for 2 h. The melted glasses were then poured into metallic molds that were previously preheated at 350 °C for 30 min. Then, all the glass samples were transferred to a muffle at 600 °C for 4 h, and then they were cooled to 25 °C at a heating rate of 2 °C/min. As will be described below, the crystallization process of these glasses was carried out in an electrical furnace at temperatures and times that were previously selected from thermal analysis. Finally, both glasses and glass–ceramics were cut into pieces of 2 × 2 × 0.5 cm or were ground into powders for their characterization.

2.2. Characterization Techniques

The analyzed chemical composition of the prepared glasses is shown in

Table 1. Chemical compositions (wt%) of studied glasses.

Oxide	AK3A-0Y	AK3A-1Y	AK3A-3Y	AK3A-6Y	AK3A-12Y
SiO2	42.70	40.42	38.13	32.72	28.80
Y2O3	0	0.83	2.72	5.53	11.12
Al2O3	2.15	2.78	2.52	2.97	2.71
ZrO2	3.10	3.31	3.01	3.57	3.40
Li2O	0.32	0.73	0.24	0.76	0.24
Na2O	1.28	1.53	0.90	1.54	2.01
K2O	1.14	1.33	1.20	1.56	1.21
MgO	7.91	8.43	8.31	8.82	9.21
CaO	15.20	14.91	16.60	16.50	17.02
BaO	19.61	19.7	19.40	19.6	19.40
ZnO	3.29	3.35	2.82	3.42	2.21
SnO2	3.20	3.17	3.40	2.62	2.30

. This analysis was carried out by X-ray fluorescence (Philips, MagiX, Eindhoven, The Netherlands) except that of Li, which was analyzed by inductively coupled plasma–optical emission spectroscopy (ICP-OES, Agilent 5800 ICP-OES). The structure of the glasses was characterized by Fourier Transform Infrared (FTIR) and Raman spectroscopies and with x-ray diffraction (XRD). For FTIR analysis (Perkin-Elmer, XB spectrum, Waltham, MA, USA), the KBr pellet technique was employed, where 1 mg of powdered sample was mixed with 300 mg of KBr. For each FTIR spectrum, 10 scans were measured and the background was subtracted. For all times, the spectral resolution was 2 cm⁻¹. Raman spectra (Renishaw InVia, Gloucestershire, UK) were measured on fractured surface samples with a spectrometer that used a 514 nm laser with a 20 mW Ar source. A total of 20 scans were measured for each sample at a resolution of 2 cm⁻¹ and the background was always subtracted at the end of the measurements. XRD diffractograms were obtained using a Bruker D8 Advance with a Cu K_α1 (1.540598 Å) anode working at 40 kV, with the measurements being carried out in the 2θ range of 5–70°. Thermal characterizations were carried out by means of Differential Thermal Analysis (SDT Q600 TA Instruments, New Castle, DE, USA) from 25 to 1200 °C at a heating rate of 10 °C.min⁻¹ in flowing air. Transmittance and reflectance spectra of the glass and GC samples were measured with a UV–vis spectrophotometer (PerkinElmer, Lambda 950, Waltham, MA, USA) in the wavelength range of 280–2500 nm. At least 4 measurements were carried out in different areas of the samples that were previously surface polished with SiC paper (4200 grid) and cleaned with alcohol and finally dried in an oven at 50 °C. Microstructures of fractured surfaces were observed by Field-Emission Scanning Electron Microscope (FE-SEM, Hitachi S-4700, Chiyoda, Tokyo, Japan). They were previously treated with HF (5%) for 5 s, washed with ethanol and water several times and finally dried in an oven at 50 °C for 2 h.

3. Results

3.1. Y₂O₃-Containing Glasses

Table 1 gives the chemical composition of the different glasses prepared in this work. The concentration of Y₂O₃ increases from 0% to 11%, while that of SiO₂ decreases in the same concentrations because the other oxides present similar concentrations in all of them. These glasses were transparent with a little brown color, which can be due to the presence of SnO₂ and ZrO₂ in their compositions.

Figure 1a shows the FTIR and Raman spectra of the prepared glasses. These spectra present wide bands like other yttrium silicate glasses [8,19,20]. In the FTIR spectra shown in Figure 1, it can be observed that for the glass without Y₂O₃, the main peak that appears at about 1000 cm⁻¹ changes to 930 cm⁻¹ for the glass with a higher Y₂O₃ concentration. This peak is assigned to the Si-O-X bonds where X = Al, Zr, Y and the other elements also contribute to this observed decrease [21]. The other two peaks are due to Si-O bonds that appear at 798 cm⁻¹ and 466 cm⁻¹ in pure silica glass [21]; however, for the glasses prepared in this work, they appear at lower and higher frequencies, respectively. Thus, the first one is observed at 715 cm⁻¹ and decreases to 670 cm⁻¹, while the second one moves from 490 cm⁻¹ to 507 cm⁻¹. These changes indicate important structural modifications after the incorporation of Y₂O₃ in the glass network.

The Raman spectra of Figure 1b present some similarities with respect to the FTIR spectra, as the main peak or wide band at about 975 cm⁻¹ moves to 860 cm⁻¹ and becomes narrower. This wide band is assigned to Qⁿ units, where Q represents SiO(Si) bonds and n (between 0 and 4) is the number of bridging oxygens (BOs), being n = 4 for Si-O-Si tetrahedral bonds and n = 0 for Si⁴⁺ bonds, i.e., for non-bridging oxygens (NBOs). The effect of Y₂O₃ in the Raman spectra is more important than for the FTIR ones, showing that Y₂O₃ gives important changes in the NBO/BO ratio of these glasses. It is well known that in pure silica glass, the Q⁴, Q³, Q², Q¹ and Q⁰ units give bands at about 1200, 1080, 1025, 960 and 870 cm⁻¹, respectively [6], and therefore, the structure of the prepared glass without Y₂O₃ is mainly formed by Q²-Q³ units. However, the red-shift of the main Raman band with the addition of Y₂O₃ leads to the formation of Q², Q¹ and Q⁰ units, indicating the increase in the formation of NBO. It can be observed that the Raman spectra can be divided in two groups; the first one corresponds to those compositions with 0 to 3% of Y₂O₃ where the main band appears at 975 cm⁻¹, and the second one corresponds to the Y₂O₃ concentrations of 6% and 12% where the main band shifts to 860 cm⁻¹. These behaviors can be assigned to the effect of Y₂O₃ in the glass structure, acting as network formers for 1% and 3% Y₂O₃ concentrations and as network modifiers for 6% and 12% concentrations. A similar conclusion has been obtained by Singh et al. [22], who found that Y₂O₃ acts as network former for concentrations lower than 5%, and for higher concentrations, it acts as a network modifier or intermediate oxide.

The wide bands of the FTIR and Raman spectra and the absence of any sharp peak in them suggest the vitreous structure of the prepared glasses. This result is consistent with the XRD patterns in Figure 2. In this case, we only observed the wide hump at about 2θ = 28°–30° due to the vitreous state with the absence of any peak due to crystalline phases. This result, which is in accordance with FTIR and Raman results, indicates that the incorporation of Y₂O₃ in the glass structure only produces changes in the glass network at short or medium bond lengths. The location of the XRD hump is at about 28.5° for the glasses containing 0, 1% and 3% of Y₂O₃ and about 29.7° for glasses containing 6% and 12% of Y₂O₃, and this result, like those of FTIR and Raman, shows an important effect of Y₂O₃ as its concentration increases.

These glasses have also been characterized by UV–vis–NIR spectroscopy in the transmittance mode (Figure 3). In these spectra, the two above-mentioned behaviors can also be observed: one for the glasses containing 0–3% Y₂O₃, where the transmittance presents a hump at about 550 nm reaching 82–88%, then decreases in the 800–1000 nm region and increases again, reaching 80% of transmittance. The second behavior is found for those glasses containing 6% and 12% of Y₂O₃ that reach 90% of transmittance between 550 and 2500 nm. These changes are due to Y₂O₃ that modifies the glass network structure, as has been observed by Raman and FTIR spectroscopies.

To determine the glass stability of the prepared glasses with respect to the presence of Y₂O₃ and their tendency to crystallization, the glass samples have also been studied by DTA (Figure 4). The DTA curves show the presence of exothermic peaks (T_p) at high temperatures that are associated with a crystallization process. In Figure 4, the mentioned two behaviors can also be observed: one for Y₂O₃ concentrations between 0% and 3% and the other one for Y₂O₃ concentrations of 6% and 12%. In the first case, one main peak appears between 800 and 950 °C, while for the 6% Y₂O₃, two close peaks appear between 700 °C and 900 °C. For the 12% Y₂O₃, two such peaks appeared overlapped, leading to a main peak at 800 °C with a shoulder at about 820 °C. In the DTA curves, we can also observe a change in the curvature between 580 and 620 °C that is associated with the glass transition temperature (T_g) of each glass, and at temperatures higher than 1000 °C, the small endothermic peaks are associated with their corresponding liquidus temperatures (T_l). There are more than 30 equations that have been proposed for determining the glass resistance to crystallization, and they can be grouped into those that calculate the glass stability (GS) or those that calculate the glass forming ability (GFA). In general, both properties are well correlated. Although GS refers to the stability against crystallization, GFA is related to the formation of a glass [23,24]. The most-used equation for calculating GS is that proposed by Hrübý (Equation (1)):

K_H = (T_x − T_g)/(T_l − T_x)

(1)

And for GFA, one of the most used equations is the following:

β = T_xT_g/(T_l − T_x)²

(2)

Other authors have proposed the use of the temperature differences between onset and glass transition:

ΔT = T_x − T_g

(3)

In these equations, T_x corresponds to the onset temperature of the crystallization peak.

Table 2. Glass transition (T_g), peak crystallization (T_p) and melting (T_l) temperatures. Glass stability parameters (GFA, GS).

	Tg (°C)	Tx (°C)	Tp1 (°C)	Tp2 (°C)	Tl (°C)	ΔT1 (°C)	β	KH	ΔT
AK3A-0Y	620	858	821	907	1092	201	9.72	1.02	0.50
AK3A-1Y	580	811	834	875	1066	254	7.23	0.91	0.48
AK3A-3Y	581	821	851	852	1028	270	11.13	1.13	0.54
AK3A-6Y	602	744	787	846	1039	185	5.15	0.48	0.32
AK3A-12Y	609	700	757	820	1032	148	3.87	0.27	0.22

collects the results obtained for the prepared glasses and the calculated glass stability parameters. In general, glasses containing 0–3% Y₂O₃ present higher stability with respect to crystallization and better ability for glass formation. The increase in Y₂O₃ to 6% and 12% results in a decrease in the glass stability of these materials. In accordance with Juisti et al. [23], for K_H values higher between 0.7–1.1, the glasses can be classified as good materials against crystallization, while for K_H values between 0.2 and 0.7, they are reluctant to poor materials. The glasses containing 6% and 12% of Y₂O₃ can be classified in this group.

3.2. Y₂O₃ Glass–Ceramics

Firstly, the kinetics of the crystallization of the above discussed glasses has been determined by DTA analysis. The DTA curves have been measured at four heating rates (vc = 5, 10, 15 and 20 °C·min⁻¹) and the peak temperatures have been determined for the glass samples of AK3A-0Y and AK3A-12Y, i.e., with 0% and 12% of Y₂O₃, respectively. Although there are several methods for calculating the activation energy (E_a) of the crystallization process, the most-used Kissinger’s method [25] follows the following equation:

ln(vc/T_p²) = −E_a/RT_p + constant

(4)

Figure 5 shows the plots of the Kissinger’s method for the AK3A-0Y and AK3A-12Y glasses, and the calculated activation energies were 318 kJ.mol⁻¹ and 227 kJ.mol⁻¹, respectively. This result indicates that the addition of Y₂O₃ promotes the glass crystallization, and this is in accordance with the GS and GFA data of Table 2.

All the prepared glasses have been heat-treated to obtain glass–ceramic materials at temperatures higher than those at which the exothermic peaks in the DTA curves appear. They were treated at 930 °C for 10 and 30 min for developing all possible crystals, but similar results were obtained at both treating times, and only the characterizations carried out for glasses heated for 30 min will be presented. The obtained glass–ceramic materials containing 0%, 1% and 3% of Y₂O₃ were transparent, while those containing 6% and 12% of Y₂O₃ were opaque with a white color. This result indicates the formation of crystalline phases in the 6% and 12% Y₂O₃ glass–ceramics; however, all samples would be expected to be crystallized because they had been treated at higher temperatures than those at which crystals appear, according to their crystallization peaks determined by DTA (Figure 4). These glass–ceramics have been characterized as commented on below.

Figure 6 shows the FTIR and Raman spectra of the corresponding glass–ceramics. Here, as can be observed from those glass–ceramics obtained from the glasses containing 0%, 1% and 3% of Y₂O₃, both FTIR and Raman spectra are close to the as-prepared glasses, where the main band appears at 1000 cm⁻¹ and is maintained for all the samples, although a new band appears at about 900 cm⁻¹ in the glass–ceramics containing 6% and 12% of Y₂O₃. The main changes appear for the glass–ceramics containing 6% and 12% of Y₂O₃ where both FTIR and Raman spectra present narrow bands, mainly due to the formation of crystalline phases. In the FTIR spectra, the narrow bands appear at 692 cm⁻¹, 570 cm⁻¹, 520 cm⁻¹ and 433 cm⁻¹, while in the Raman spectra, the narrow bands appear at 889 cm⁻¹, 706 cm⁻¹, 520 cm⁻¹ and 406 cm⁻¹, and three shoulders appear at 340–360 cm⁻¹, 850 cm⁻¹ and 940 cm⁻¹.

Both FTIR and Raman spectra can be divided into three spectral regions where the different Si-O- vibrations of the glass network appear: (a) the high-frequency region (800–1200 cm⁻¹) due to symmetric stretching motions of BO of Si-O-T bonds, where T = Si, Al, Y, Zr, etc. (in this region, the vibrations of NBO denoted as Qⁿ⁻¹ (n = 4, 3, 2, 1) also appear and are discussed in Figure 1); (b) the intermediate-frequency region (500–800 cm⁻¹) due to asymmetric stretching motions of BO between two tetrahedra, i.e., T-O-T bonds; and (c) the low-frequency region (200–500 cm⁻¹), where the lattice vibrations of those cations that do not form part of the silica network appear, such as Ca²⁺, Mg²⁺ or even Y³⁺ and Zr²⁺ and Sn²⁺, which act as network modifiers. The bending vibrations of O-Si-O bonds and the Al-O bending vibration in [AlO₆] also appears here [26].

In the high-frequency region, the GC containing 0, 1 and 3% of Y₂O₃ present a wide band where all Qⁿ units are present. The maxima of this wide band is around 960 cm⁻¹ and moves to 930 cm⁻¹ when assigned to Q² units with the increase in Y₂O₃, and this result indicates the formation of Si-O-Y bonds [19]. For the GC containing 6 and 12% of Y₂O₃, the Raman spectra change radically, and now the main peak that appears at 889 cm⁻¹ presents two shoulders at 850 cm⁻¹ and 920–950 cm⁻¹ which can be assigned to Q¹, Q⁰ and Q² units, respectively. The intensity of the 920–950 cm⁻¹ shoulder decreases with the Y₂O₃ concentration, and this indicates that both Q² unit Si-O-Y bonds also decrease, indicating the intermediate role of this oxide in these glasses and glass–ceramics [19].

In the intermediate-frequency region, the GC containing 0, 1 and 3% Y₂O₃ present wide- and low-intensity bands, while those of 6% and 12% Y₂O₃ present two low-intensity bands at around 520 and 570 cm⁻¹ that originate from the stretching vibrations of Al-O-Al and Si-O-Al bonds in calcium aluminate glasses. The peak at 570 cm⁻¹ is also a contribution of Mg-O, Al-O and Si-O bonds that appear in magnesium-barium-aluminosilicate glass–ceramics containing ZnO and BaO [27,28]. Finally, in the low-frequency region, the peak at 406 cm⁻¹ and the wide shoulder around 340–360 cm⁻¹ is assigned to Ca-O bonds that appear in Zinc-Calcium silicates [29]

The formation of these crystalline phases has been characterized by XRD (Figure 7), and the presence of several peaks due to the crystallization of the material can be observed. This crystallization appears in the sample containing 3% of Y₂O₃ (i.e., GC-3Y) and increases with the Y₂O₃ concentration (i.e., samples GC-6Y and GC-12Y). It was observed that the main crystalline phases formed in the GC-3Y also appeared in the glass–ceramics with higher concentrations of Y₂O₃, but other crystalline phases also developed in them. The main crystalline phases that appear in the GC-3Y sample are anorthite (CaAl₂Si₂O₈: PDF 41-1486), Aluminum Yttrium oxide (Al₂Y₄O₉: PDF 80-1696), Potassium silicate (KAlSiO₄: PDF 12-0134) and Calcium Magnesium silicate (Ca₂Mg(Si₂O₇): PDF 83-1815), and this result is due to the chemical composition of the former glasses (Table 1). The XRD patterns of the GC-6Y and GC-12Y samples present other crystalline phases, the main ones being those of Barium Calcium Magnesium silicate (BaCa₂Mg(SiO₄)₂: PDF 31-0128) and Zirconium Yttrium oxide (Zr_0.82Y_0.18O_1.91: PDF 37-0107) and small amounts of Barium Zinc silicate (BaZn₂Si₂O₇: PDF 23-0843) and Calcium Aluminate (Ca₃Al₂O₆: PDF 38-1429). A small peak also appears, probably due to coesite (SiO₂: PDF 83-1413). These results indicate that the increase in Y₂O₃ in the glass matrix promotes the incorporation of BaO and MgO into crystalline phases, while for low concentrations of Y₂O₃, such oxides remain in the glassy matrix as intermediate former oxides. In addition, the increase in the Y₂O₃ concentration also leads to the increase in the Zirconium Yttrium oxide and Barium Calcium Magnesium silicate crystalline phases, as can be observed in Figure 7. A semiquantitative analysis of these XRD patterns has shown that the GC-3Y sample is 62% glass and 38% crystalline, the GC-6Y sample is 15% glass and 85% crystalline, and finally, the GC-12Y sample presents 12% glass and 88% crystalline. Therefore, there is an important change in crystallinity when the Y₂O₃ concentration increases from 3% to 6% in the glass composition.

The obtained glass–ceramics have also been characterized by UV–vis–NIR spectroscopy in the transmittance and/or reflectance modes (Figure 8). Because the glass–ceramics containing 0%, 1% and 3% of Y₂O₃ were transparent, their spectra presented low reflectivity and they were measured in the transmittance mode, while the other glass–ceramics presented high opacity and white color, and the spectra were measured in the reflectance mode. These glass–ceramics are not highly reflective materials because the maximum reflectivity is only higher than 50–70% in the visible spectral region and tends to decrease in the near-IR region.

The formation of the crystalline phases has been observed by FE-SEM analysis. Figure 9 and Figure 10 show different magnifications of the glass–ceramic materials obtained at 930 °C for 30 min. These images show a microstructure formed by the presence of crystalline phases even in the GC-0Y sample, and such, crystals are well-developed in the glass–ceramic containing only 1% Y₂O₃ (GC-1Y). It is interesting to note that the above characterization techniques (FTIR, Raman, XRD and UV–vis) have identified the presence of crystalline phases for the samples containing at least 3% or 6% of Y₂O₃; however, from FE-SEM, a wide crystallized microstructure was observed even in the 1% Y₂O₃ sample. This result indicates that such crystallites are of very small size and lower than visible light wave length. The formed crystals present globular-spherical shapes with irregular and smooth surfaces, and their size increases with the Y₂O₃ concentration. For the GC-1Y and GC-3Y samples, it is observed that the globular crystals are irregular agglomerates formed by small crystals, while for the GC-6Y and GC-12Y samples, the globular forms are very smooth, well defined and separated. Both the size and concentration of the crystals in the glass–ceramic samples lead to the differences obtained in the UV–vis–NIR spectra and therefore, the GC-6Y and GC-12Y glass–ceramics are opaque due to the high size of the globules. A careful analysis of these globular forms is carried out in the next section.

4. Discussion

As has been discussed above, the concentration of Y₂O₃ influences the structural, microstructural and optical properties of both glasses and glass–ceramic materials. With respect to the structural characteristics, both FTIR and Raman spectra showed a shift of the main peaks to lower wavelength frequencies with the increase in the Y₂O₃ content. A similar result was found by Kaur et al., and it has been assigned to the structural rearrangement of the Si-O-Si bonds leading to a decrease in the local symmetry due to the variation of the Si-O-Si bond angles [29]. When the Y₂O₃ concentration increasing the glass network is broken because Y³⁺ ions act as network modifiers, this leads to the formation of NBO, as Singh et al. have shown for Y₂O₃ concentrations higher than 5% [22]. FTIR spectra of Figure 1 show a small shift of the main band position from 988 cm⁻¹ to 970 cm⁻¹ when the Y₂O₃ concentration increases from 0% to 3%, and it shifts to 937 cm⁻¹ and 930 cm⁻¹ for concentrations of 6% and 12%, respectively. These results indicate that in the pristine glasses, Y₂O₃ probably acts as a network modifier for as little as 1% in the glass composition.

The Raman spectra also corroborate this conclusion. The Raman spectra (Figure 1b) show two features: the first one is the shift of the main band from 966 cm⁻¹ to 863 cm⁻¹, appearing at the highest frequency for Y₂O₃ concentrations from 0 to 3% and at the lowest frequency for those of 6% and 12%, and the second feature is the small increase in the intensity of the shoulder that appears at about 880 cm⁻¹. This shoulder has been assigned to the Q² units in Yttrium glasses and glass–ceramics [29] and confirms the formation of two NBOs when Y₂O₃ is incorporated into the glass or glass–ceramic structure. Therefore, this oxide can act as a network modifier from its low concentration in the pristine glasses or in their corresponding heat-treated glass–ceramics. In summary, the structure of the pristine glasses changes when they are heat-treated at 900 °C for 10 or 30 min to form glass–ceramic materials. As the Y₂O₃ concentration increases to 3%, diffraction peaks assigned to the crystalline phases anorthite and Zirconium Yttrium Oxide appear. However, when the Y₂O₃ concentrations increase to 6%, the above crystalline phases do not increase, but other new ones now appear (Barium Calcium Magnesium Silicate, calcium silicate and Lithium Yttrium Oxide) which are maintained in intensity in the glass–ceramic containing 12% Y₂O₃.

The microstructure of these crystalline phases has been observed by FE-SEM (Figure 9 and Figure 10) and presents globular or spherical shapes. The size and concentration of these globules increase with the Y₂O₃ percentage in the glass composition. A deep analysis of the globular microstructure has been carried out at the highest magnifications, as is shown in Figure 11. Here we can observe very small crystals grouped to form high globules with irregular shapes (Figure 11b), and in the glass–ceramics with 6% and 12% Y₂O₃ concentrations, the above-mentioned smooth globular forms that were observed at lower magnifications also appear (Figure 10). These smooth globules are well differentiated from the irregular ones that are formed by the aggrupation of small crystals, and therefore, it can be concluded that these smooth globules may be due to glass phases separate from the glass matrix due to a phase separation process [18]. Similar results have been obtained in different borosilicate glasses where phase separation appears for Y₂O₃ concentrations between 0 and 15% and such phase separation increases with Y₂O₃ [22,29].

In Figure 11c, we can observe the globular phases, the continuous glass matrix and the small crystals forming independent phases in the whole glass–ceramic material. In particular, we can observe a broken smooth globular phase that does not contain any crystal inside (Figure 11d) and confirms the assignation to separated phases in these glass–ceramics. The presence of these smooth globular shapes leads to the opacity of the GC-6Y and GC-12Y samples, as is shown in the UV–vis spectra of Figure 8.

The UV–vis–NIR spectra have been used for determining the differences in the optical properties between the Y₂O₃ glasses and their respective glass–ceramics. From the UV–vis spectral region of Figure 3 and Figure 8, the band-gap (E_g) values have been calculated. From the transmittance spectra (Figure 3 and Figure 8a), the optical absorption coefficients α(ν) can be determined at different wavelengths by using the following equation [30]:

α(ν) = [2 − log(I_t/I_o)]/d

(5)

where I_o and I_t are the incident and transmitted light, respectively, while d is the thickness and ν is the photon’s frequency. In accordance with Mott and Davies from α(ν), it is possible to calculate the optical band-gap energies, E_g [31]:

α(ν)h.ν = C(h.ν − E_g)ⁿ

(6)

where h is the Planck constant, C is also a constant and n is an index that depends on the mechanism of interband transitions. It is known that for allowed direct or indirect transitions, n = 1/2 or 2, respectively, while for forbidden direct or indirect transitions, n = 3/2 or 3, respectively. On the other hand, for the reflectance spectra (Figure 8b), it is necessary to combine the Kubelka–Munk function (F(R_∞)) and Tauc’s relation [32], and so Equation (6) is now:

F(R_∞) = C(hν − E_g)ⁿ/hν

(7)

F(R_∞) = (1 − R_∞)²/2R_∞

(8)

where R_∞ is the ratio between the reflectance of the sample and that of an infinitely thick specimen (R_∞ = R_sample/R_standard).

It is known that the band gap energy is related to the energy needed to excite an electron from the valence band to the conduction band, and in the case of studding the structure of glasses and glass–ceramics, the determination of E_g is important in predicting the effect of the incorporation of metal oxides into the network. When the metal oxide acts as a network modifier, it produces the weakening of the glass structure due to the formation of NBOs. This leads to a red-shift of E_g, and a similar result occurs when an ion can adopt two kinds of coordination and the lower coordination leads to lower E_g values [33].

For the UV–vis transmission spectra of Figure 3 and Figure 8a, the plot of α(ν)h.ν vs. hν is used for the calculation of E_g, as is shown in Figure 12. At the same time, from the reflectance spectra of Figure 8b applying Equation (7), the corresponding E_g values for the opaque glass–ceramic materials have been determined, as can be observed in Figure 13. In both figures, we observe a linear increase in α(ν) or F(R_∞) with increasing energy, and the intersection point of the linear fit in the x-axis has been used to estimate the E_g values that are collected in

Table 3. Optical parameters of the prepared glasses and glass–ceramics.

Glasses	Eg (eV)	Eu (eV)	Cut-Off (nm)	Glass–Ceramics	Eg (eV)	Eu (eV)	Cut-Off (nm)
AK3A-0Y	3.36	0.20	340	GC-0Y	4.10	0.65	280
AK3A-1Y	3.43	0.20	340	GC-1Y	3.39	0.50	324
AK3A-3Y	3.23	0.25	348	GC-3Y	3.73	0.21	251
AK3A-6Y	3.32	0.24	334	GC-6Y	2.44	0.26	294
AK3A-12Y	3.12	0.21	358	GC-12Y	2.78	0.34	303

. For those glasses containing Y₂O₃, a decrease in E_g with the Y₂O₃ concentration is observed, and this corresponds to the incorporation of Y³⁺ as a network modifier. In the case of the glass–ceramics, for those that present transparency, i.e., GC-0Y, GC-1Y and GC-3Y, and for that GC-6Y with opacity, the E_g values also tend to decrease with the increase in Y₂O₃ in their compositions; however, for that GC-12Y which is also opaque, its E_g value presents a little increase. It is also observed that for the opaque glass–ceramics, their E_g values are lower than those of the respective glasses with similar Y₂O₃ concentrations. On the contrary, the transparent glass–ceramics present higher E_g values than their counterpart glasses with similar Y₂O₃ concentrations. This result indicates that during the heat treatment carried out to crystallize the prepared glasses, some NBOs have turned to BOs while crystals are formed, and for Y₂O₃ concentrations of 6% and 12%, the above-mentioned phase separation has continued forming NBOs.

As has been discussed, the addition of Y₂O₃ leads to an increase in the number of NBOs due to its role as a network modifier and to the formation of different crystalline phases after heat treatment at 930 °C for 30 min. The incorporation of Y₂O₃ into the glass structure can also affect the structural disorder of glasses or glass–ceramic materials. This effect can be analyzed by the Urbach edge behavior that can be measured by the UV–vis spectra [34]. It is well known that the optical absorption coefficient α(ν) is used to calculate the Urbach energy (E_u) in accordance with the following:

α(ν) = C exp[(hν − E_g)/E_u)]

(9)

For the UV–vis transmission spectra of Figure 3 and Figure 8a, the plot of ln(α(ν)) vs. hν is used, while for the reflectance spectra of Figure 8b, the plot of ln(F(R_∞)) vs. hν is used for the calculation of E_u. The calculated E_u values for both glasses and glass–ceramics are collected in Table 3. In general, the increase in E_u is associated with structural disorder or amorphousity, and this occurs when Y₂O₃ concentration increases in as-prepared glasses. On the other hand, for the glass–ceramics, the E_u values are higher than those found for the respective glasses with the same Y₂O₃ concentration, and this suggests an increase in the disorder after the heat treatment carried out for crystallization. As has been discussed above, the heat treatment not only leads to the formation of different crystalline phases but also to a wide phase separation observed by FE-SEM images (Figure 10 and Figure 11); however, for the crystallized glasses, the E_u values decrease with the Y₂O₃ concentration, and this indicates that the formation of crystals reduces the degree of order of the glass–ceramic structure.

Finally, the obtained E_g values were used to semiempirically estimate the linear (n₀) and non-linear refractive indexes (n₂), as well as the metallization criterion (M) according to [35,36,37]:

(n₀² − 1)/(n₀² + 2) = 1 − (E_g/20)^1/2

(10)

n₂ = B/E_g⁴ (B = 1.26 × 10⁻⁹ esu eV⁴)

(11)

M = (E_g/20)^1/2

(12)

The calculated values are given in

Table 4. Optical parameters of the prepared glasses and glass–ceramics.

Glasses	n0	n2 (×10−11 esu)	M	Glass–Ceramics	n0	n2 (×10−11 esu)	M
AK3A-0Y	2.32	0.99	0.41	GC-0Y	2.19	0.45	0.45
AK3A-1Y	2.31	0.91	0.41	GC-1Y	2.31	0.95	0.41
AK3A-3Y	2.35	1.16	0.40	GC-3Y	2.25	0.65	0.43
AK3A-6Y	2.33	1.04	0.41	GC-6Y	2.55	3.55	0.35
AK3A-12Y	2.37	1.33	0.39	GC-12Y	2.45	2.11	0.37

. The metallization criterion is usually employed to differentiate a metal and a non-metal, which is related to the insulating behavior of the studied material. M values are between 0 and 1, being M = 0 for a metal and M = 1 for an insulator. In general, silicate and borate glasses containing large amount of glass-forming oxide have large M values between 0.50 and 0.70 [36], and the results of Table 4 fall below this situation for both glasses and glass–ceramics with different Y₂O₃ concentrations. A little decrease in M is observed for the 6% and 12% Y₂O₃ concentrations and especially in the cases of glass–ceramics, as corresponds to the lower E_g values of Table 3, and this indicates a given tendency for metallization in these glass–ceramics. These results show that the wide range of different oxides forming the studied glasses induces increases in the width of both valence and conduction bands and therefore increases in the tendency to metallization.

Regarding the values of n₀ and n₂, it should be taken into account that they are calculated from Equation (10), which uses the values of E_g (Table 3) obtained from the UV–vis spectra, and therefore, they are between 300 and 500 nm, i.e., in the visible spectral region. The tendency of n₀ and n₂ is to increase with the Y₂O₃ concentration in both glasses and glass–ceramics, which may be due to an increase in NBOs in the glass matrix [38], as was stated from FT-IR and Raman analysis. It is especially significant in the case of the GC-6Y and GC-12Y samples, in which separate phases and crystals were observed to form (Figure 10 and Figure 11). In these two samples, the Y₂O₃ produces an increase in the formation of structural defects such as NBOs, phase separation and crystallization which result in a loss of transparency and a greater tendency to metallization.

5. Conclusions

Multicomponent glasses with increasing Y₂O₃ concentrations were prepared and characterized, and after a heat treatment, they crystalized to develop glass–ceramic materials. The effect of Y₂O₃ on both the glass structure and properties depends on its concentration. Two types of Y₂O₃-dependent behaviors were observed, one for concentrations from 0 to 3% and the other for concentrations of 6 and 12%. Up to a concentration of 3%, the T_g decreased from 620 °C to 580 °C, and for concentrations of 6% and 12%, the T_g increased to 602 °C and 609 °C, respectively. The crystallization peaks followed the opposite trend: they increased from 821 °C to 851 °C and then decreased to 787 °C and 757 °C for such Y₂O₃ concentrations. This coincided with both the ability to form glass and the stability of the glass network, as well as with the intermediate behavior of the Y₂O₃, which acted as a network former for 1% to 3% concentrations and as a network modifier for 6% and 12% concentrations. The heat treatment at 930 °C led to the crystallization and phase separation in such glasses. The crystallization process is favored with Y₂O₃, as is observed by the lowest activation energy E_a of the 12% Y₂O₃ glass with respect to the one without Y₂O₃. While the crystallization process occurs even for the lowest concentration of Y₂O₃ and increases with it, phase separation is only observed for Y₂O₃ concentrations of 6% and 12%. For the 3% of Y₂O₃ concentration, the main crystalline phases are Potassium silicate, Calcium Magnesium silicate and Aluminium Yttrium oxide, and the vitreous phase is as high as 62% with respect to the crystalline one. For the 6% and 12% Y₂O₃ concentrations, the main crystalline phases are Barium Calcium Magnesium silicate and Zirconium Yttrium oxide, and the percentages of vitreous phases are only 15% and 12%, respectively. Therefore, Y₂O₃ promotes the incorporation of BaO into the Calcium Magnesium silicate lattice while forming a new phase with Al₂O₃. The high percentage of crystallinity and the presence of phase separation in the glass–ceramics with 6% and 12% of Y₂O₃ led to their opacity, while for lower Y₂O₃ concentrations, transparency was maintained similarly to the former glasses. The optical properties, such as the linear and non-linear refractive indexes (n₀ and n₂) as well as the band-gap energies (Eg) and Urbach edges (E_u), follow the same trends due to the intermediate behavior of Y₂O₃ and the formation of crystalline and phase-separated phases as the Y₂O₃ increases above a concentration greater than 3%. The metallization criterion indicates that all developed glass and glass–ceramics behave in a similar way to silicate and borosilicate glasses, although a little increase in metallization was observed for the 6% and 12% Y₂O₃ concentrations.