Effect of Sodium Oxide on Structure of Lanthanum Aluminosilicate Glass

Abstract

Rare earth (RE) aluminosilicate glasses exhibit several favorable chemical, mechanical and thermal properties. As such, they are considered to be model systems for long-half-life actinides and are candidate containment materials for long-term immobilization of radioactive wastes. The aim of the present study was to investigate the effect of the substitution of sodium oxide on the glass transition temperature and structure of lanthanum aluminosilicate glasses. The primary objective was to elucidate the relationship between the substitution of Na₂O for La₂O₃ on the Tg reduction and structural characteristics of lanthanum aluminosilicate glass, including identifying changes in the main Qn species and local environments of Si and Al. The structure of SiO₂–Al₂O₃–La₂O₃–Na₂O glasses has not been studied previously, and, thus, this investigation is the first to assess the structural changes occurring when La₂O₃ is substituted by Na₂O. Three glasses were prepared with general composition (mol.%): 55SiO₂–25Al₂O₃–20M₂O_n (M = La or Na; n = 3 or 1). Glass G1 contains 20 mol.% La₂O₃; in G2, 15 mol.% of La₂O₃ was substituted by 15 mol.% Na₂O; and Glass G3 contains 20 mol.% Na₂O. The glasses were characterized by DSC to determine glass transition temperatures. As expected, as Na is substituted for La, T_g decreases substantially. Structural studies were carried out by FTIR spectroscopy, ²⁹Si, and ²⁷Al MAS NMR. As Na is substituted for La in these aluminosilicate glasses, the main goals that were achieved were the identification of Qⁿ species and also changes in the local environments of Si and Al: {Qⁿ_Si(mAl)} and {Qⁿ_Al(mSi)}.

1. Introduction

Aluminosilicate glasses are of paramount importance for contemporary technology, with current and potential applications in fibers for structural composites, glass-to-metal sealing materials for glass-to-metal systems and solid oxide fuel cells, optical and photochromic constituents, high-power laser gain media, flat-panel display substrates and bioactive materials [1,2]. Alkali and alkali earth aluminosilicate glasses are also useful simple model systems for naturally occurring complex magmas since silica and alumina are the main constituents of the Earth’s crust [3].

Many authors have studied the structure of these aluminosilicate glasses, and the general view is that both SiO₄ and AlO₄ tetrahedra are joined at their corners by bridging oxygens (BOs) to form frameworks with varying degrees of polymerization [1,2,3,4,5,6,7,8]. The Al⁽³⁺⁾O₄ tetrahedron has a charge deficit compared with a Si⁽⁴⁺⁾O₄ unit, and this can be charge compensated by, for example, a positively charged alkali ion such as Na⁺. If Na/Al = 1, then all (AlO₄)⁻ tetrahedra are compensated by the Na⁺ ions, and the network is fully connected with no non-bridging oxygens (NBOs). Network connectivity can be probed using magic-angle spinning (MAS)–nuclear magnetic resonance (NMR) spectroscopy. The chemical shift of ²⁹Si allows the degree of connectivity of the SiO₄ tetrahedra to be ascertained [1]. This is represented by the symbol Qⁿ, where n is the number of BO on a tetrahedron. As the number of BO decreases, the chemical shift is displaced to higher values. When Al substitutes for Si, an additional modification of the chemical shift occurs depending on the number m of Al atoms as second neighbors. The moieties for Si and Al are denoted by {Qⁿ_Si(mAl)} and {Qⁿ_Al(mSi)}, indicating m Si–O–Al or Al–O–Si bonds [2,9]. Several studies have shown that, for Al/(Al + Na) ratios ≤ 0.5, all Al³⁺ ions are found in AlO₄ tetrahedra, and Na ions are linked through NBOs to Q³ SiO₄ tetrahedra. When Al/(Al + Na) ratio > 0.5, then Al³⁺ ions are found to be in higher coordination states, i.e., ^[5]Al or ^[6]Al, and the environment of Na⁺ ions will change [3,4,5]. However, it has been reported that these higher coordination states (mainly ^[5]Al) have also been observed in tectosilicate compositions (Na/Al = 1) [6,7,8,9].

Rare earth (RE) aluminosilicate glasses [2,4,10,11] exhibit several favorable chemical, mechanical, and other properties, such as high hardness, refractive index, elastic modulus, and glass transition temperatures (T_g). In addition to the applications mentioned above, these glasses are being considered for long-term immobilization of radioactive wastes, and, effectively, they can be model systems for long-half-life actinides [1,2,10]. In particular, lanthanum and yttrium have been selected to simulate actinides because the ternary RE₂O₃-Al₂O₃-SiO₂ (RE = La or Y) phase diagrams are well known, and their physicochemical properties have been investigated [2,10]. These rare earth ions are more suited to structural studies using MAS NMR, as they avoid the problems of paramagnetism observed with other rare earth cations.

Investigations have shown that, for glasses with RE/Al < 3, there will be insufficient numbers of RE³⁺ cations to charge-compensate the (AlO₄)⁻ tetrahedra, and high levels of ^[5]Al or ^[6]Al are observed [8,10]. For glasses with RE/Al > 3, then effectively all AlO₄ tetrahedra should be charge-compensated, and RE cations act as modifiers, depolymerizing the network with Qⁿ_Si(mAl)→Qⁿ⁻¹_Si(mAl) conversions and formation of NBOs preferentially localized on SiO₄ tetrahedra [2,7,8,9]. Even so, there is evidence for small amounts of ^[5]Al and/or ^[6]Al in these compositions [12].

It should also be noted that the generalized structural models assume that Loewenstein’s rule [13], also known as the Al-avoidance principle, always pertains. It excludes ^[4]Al-O-^[4]Al constellations due to the excess negative charge of AlO₄ tetrahedra relative to SiO₄ and thereby maximizes the number of Si-O-Al bonds at each Qn unit. However, there are many exceptions to this rule; that is, the presence of ^[4]Al-O-^[4]Al linkages rather than ^[4]Si-O-^[4]Al or ^[4]Si-O-^[4]Si in aluminosilicate glasses have been reported [14,15]. MAS NMR of a series of La₂O₃–Al₂O₃–SiO₂ glasses [16] reveals Al speciations dominated by AlO₄ groups, with minor fractions of AlO₅ (5–10%) and AlO₆ (≲3%) polyhedra present; the amounts of ^[5]Al and ^[6]Al coordinations increase as the molar fraction of Si decreases. The glass structures exhibit a pronounced Al/Si disorder, including significant amounts of AlO₄–AlO₄ contacts [16,17].

While the structure and properties of SiO₂–Al₂O₃–La₂O₃ and SiO₂–Al₂O₃–Na₂O glasses have been studied extensively [1,2,3,4,5,6,7,8,9,10,11,12], it is clear from our in-depth review of the glass science literature that the structure of SiO₂–Al₂O₃–La₂O₃–Na₂O glasses has not been studied previously, and thus this investigation is the first to assess the structural changes occurring when La₂O₃ is substituted by Na₂O.

Structural characterization by MAS NMR of lanthanum sodium aluminoborosilicate glasses with compositions (mol.%): 55SiO₂–(25–x)Al₂O₃–xB₂O₃–15Na₂O–5La₂O₃ have been reported previously [18,19], in which lanthanum is intended to simulate other lanthanides, as well as minor actinides contained in nuclear waste.

These studies highlighted the narrow range of composition within which peraluminous vitreous matrices exhibit homogeneity after quenching, extending from 0% to 5% of B₂O₃. Infrared spectroscopy provided comprehensive characterization of the silicate and borate networks, emphasizing the presence of Q_n m(Al,B) units rather than Q_n m(Al) units alone in aluminoborosilicate glasses. NMR spectroscopy reveals that boron predominantly exists in triangular BO₃ units, with a small percentage in tetrahedral BO₄ units. Al predominantly occurs in 4-fold coordination, with some AlO₅²⁻ and AlO₆³⁻ species. The structural role of Na shifts towards primarily (AlO₄)⁻ charge compensation with increasing Al/B substitution.

The structure of these glasses is rather complex due to the mixing of the network-forming cations (Si, B, and Al), with the appearance of nanoscale phase separation for glasses with more than 5% B₂O₃ and the formation of a La-rich phase. Therefore, the present study was undertaken to investigate the structure of a lanthanum aluminosilicate glass with similar composition (but excluding boron), that is, in mol.%: 55SiO₂–25Al₂O₃–20La₂O₃. The effect of partially and wholly substituting Na₂O for La₂O₃ is investigated in order to understand the environment of these elements and bring more clarity to results from the previous aluminoborosilicate glasses. The primary objective is to elucidate how sodium oxide substitution for lanthanum oxide influences the glass transition temperature and structure of these specific glass compositions.

It is envisaged that lanthanum aluminosilicates without boron can also be model systems for actinides. As these glasses have high melting points, the use of sodium oxide as a modifying oxide to decrease the liquidus temperature and viscosity will be necessary and this will have an effect on the lanthanum aluminosilicate glass network structure. A ²⁷Al MAS NMR and Fourier-Transform Infrared (FTIR) study [20] on rare earth aluminosilicate glasses suggested that these glasses have a wide distribution of Qⁿ units. Thus, the compositions in the present study were chosen with the same contents of SiO₂ and Al₂O₃ and the same SiO₂:Al₂O₃ ratio as in the previous studies of aluminoborosilicate glasses [18,19], varying only the La₂O₃:Na₂O ratios. While the number of samples is limited, it will provide a baseline study for further work on the Na-La-aluminoborosilicate series of glasses in the future.

Therefore, the present study focuses on structural and coordination relationships combining information from FTIR and MAS NMR spectroscopy to provide direct information on Qⁿ and changes in the local environments of Si and Al: {Qⁿ_Si(mAl)} and {Qⁿ_Al(mSi)} as Na is substituted for La in these aluminosilicate glasses.

2. Materials and Methods

2.1. Glass Synthesis

Aluminosilicate glasses of 10 g were prepared with general composition (in mol. %): 55% SiO₂ (purity 0.997), 25% Al₂O₃ (purity 0.997), and 20% modifier oxides: G1—20% La₂O₃ (purity 0.997); G2—5% La₂O₃ + 15% Na₂O (from Na₂CO₃—purity 0.998); and G3—20% Na₂O (all reactants from Alfa Aesar, Illkirch, France). The starting powders undergo extensive grinding in an agate mortar to ensure a homogeneous mixture. These were melted in platinum crucibles placed inside an electric furnace.

In order to optimize the manufacturing process, XRD analysis was conducted on samples of G2 glass. These glasses were melted according to the following thermal program: raising the ambient temperature to 1200 °C over 2 h and 20 min, followed by maintaining the temperature at 1200 °C for 2 h. The same procedure was carried out on the other glasses until the disappearance of peaks at 1450 °C, aiming to determine a melting temperature that takes into account the presence of lanthanum and sodium in the compositions. Based on these results, it was decided that the melting temperatures that should be used were 1450 °C for G1, 1400 °C for G2, and 1300 °C for G3. During melting, a precise temperature control was maintained to ensure consistency, and significant weight loss was not observed.

2.2. Glass Characterization

2.2.1. Compositional Analysis

The oxide compositions were analyzed on polished samples, using Scanning Electron Microscopy coupled with Energy Dispersive Spectroscopy (SEM–EDS) (Noran System Six); the analyzed area was about 1 μm³.

2.2.2. X-ray Diffraction Measurements

In order to determine if the glasses were completely amorphous, samples were analyzed using X-ray diffraction (XRD) (Philips X’pert PanAlytical X-ray diffractometer in a Bragg–Brentano geometry, Almelo, The Netherlands) with monochromated Ni-filtered CuKα radiation (λ = 1.5406 Å) over a range of 2θ = 20°–70°, with a speed of 2.4°/min. Powder XRD patterns were obtained using acceleration voltage of 40 kV and a filament heating current of 30 mA.

2.2.3. Transmission Electron Microscopy Study

Nanoscale glass homogeneity was checked using a Philips (Eindhoven, The Netherlands) CM20 transmission electron microscope (TEM) operating at 200 kV and equipped with an Oxford Instruments (Oxford, UK) energy-dispersive X-ray spectroscopy (EDS) analyzer for all glasses. Samples were prepared from powder dispersed in absolute ethanol and deposited on a holey carbon film supported by a copper grid.

2.2.4. Differential Scanning Calorimetry Analysis

The glass transition temperature (T_g) of each sample was determined using Differential Scanning Calorimetry (DSC), employing a SETARAM (Caluire-et-Cuire, France) multi-HTC 1600 instrument. In this method, 0.3 g of powdered glass was placed in a Pt crucible and heated at a steady rate of 10 °C min⁻¹, under a continuous flow of argon, while an empty crucible was used as a reference. The temperature differential between the sample and the reference crucible was then recorded as a function of the energy supplied to both crucibles, allowing for the precise determination of the glass transition temperature.

2.2.5. Fourier Transform Infrared Spectroscopy

Infrared spectroscopy was carried out under vacuum with a Bruker (Billerica, Massachusetts, USA) Vertex 80v Fourier-Transform Infrared (FTIR) spectrometer with a working range from far infrared to UV (50–30,000 cm⁻¹). Spectrum acquisitions, in the range of interest (50–2000 cm⁻¹), were performed with an instrumental resolution of 4 cm⁻¹. The selected optical configurations include a Globar source, two beam splitters (multilayer [50–500 cm⁻¹], Ge/KBr [500–2000 cm⁻¹]), and the use of two detectors (Bolometer [50–500 cm⁻¹], DLaTGS/KBr [500–5000 cm⁻¹]). A reference measurement was conducted using an uncoated gold mirror. The glass samples, cut with parallel faces and optically polished, were subjected to reflection measurements to characterize their infrared response. The reflectivity spectra were analyzed in relation to the dielectric function, ε, employing the second Kirchhoff’s law and the Fresnel relation [18]. To extract optical functions, the experimental reflectivity spectra were fitted with a dielectric function expression tailored for disordered materials, consisting of absorption bands with Gaussian profiles to accommodate structural disorder effects. The work of reconstructing the spectra presented here was performed with FOCUS, version 1.0.17a, a curve-fitting software especially developed for the analysis of optical spectra [21]. This program is available free from the website of CEMHTI Laboratory, UPR 3079, Université d’Orléans, France. More details about the mathematical approach of the model used can be found in Ref. [21].

2.2.6. MAS (Magic-Angle Spinning) NMR (Nuclear Magnetic Resonance) Study

In the case of ²⁷Al (I = 5/2), MAS NMR spectra were obtained on a Bruker (Billerica, Mass., USA) Avance 750 MHz (17.6 T magnetic field) spectrometer equipped with high-speed MAS probe heads, using aluminum-free zirconia rotors of 2.5 mm diameter, spinning at the magic angle of 54.7°, at 30 kHz speed.

The ²⁹Si MAS NMR spectra were recorded using a Bruker Avance 300 MHz spectrometer (magnetic field = 7 T) operating at a resonance frequency of 59.63 MHz with Bruker MAS probe and a 4 mm ZrO₂ rotor spinning at the magic angle of 54.7° at 10 kHz speed. The spectra were obtained with 192 scans, with a pulse length of 4.5 μs (excitation pulse π/2) and a relaxation delay of 900 s. A ²⁹Si chemical shift in the spectra was referenced to TMS (tetramethylsilane) at 0 ppm. The NMR spectra were deconvoluted utilizing DM-fit [22] software, employing a Gaussian–Lorentzian line shape with a width of 11 ± 0.5 ppm for 29Si MAS NMR spectra. For ²⁷Al MAS NMR spectra, the CzSimple model lines [23], derived from the Gaussian Isotropic Model (GIM), were employed.

3. Results and Discussion

3.1. Compositional Effects on Glass Formation and Homogeneity

The quenched glasses were transparent and macroscopically homogenous. As shown in Figure 1, no crystalline phase was observed in the XRD patterns for the three glasses (G1 = 20La₂O₃; G2 = 5La₂O₃+15Na₂O; G3 = 20Na₂O), only the broad peak at low values of 2θ, showing that the samples are amorphous.

The analyzed compositions for the three glasses are compared with the starting compositions in

Table 1. Molar percentages of the prepared aluminosilicates glasses containing La₂O₃ and Na₂O (Exp.—expected composition; An.—analyzed composition). The uncertainty on measured values is ±0.1% molar.

		SiO2	Al2O3	Na2O	La2O3
G1	Exp. An. ±0.1	55 56.1	25 24.8	0 0.0	20 19.1
G2	Exp. An. ±0.1	55 56.3	25 24.7	15 14.7	5 4.3
G3	Exp. An. ±0.1	55 56.0	25 24.9	20 19.1	0 0.0

. Some slight loss of the modifying oxides occurred during melting, but no substantial differences were observed between the analyzed and starting compositions.

From the bright-field TEM observations, no crystallization or phase separation was observed for all three glasses (see Figure 2), and the samples were completely homogeneous.

DSC analyses of the glass series revealed a strong effect of the modifying oxide compositions on the glass transition temperature, T_g. Figure 3 shows a decrease in T_g from 874 °C for the G1 = 20La₂O₃ glass to 790 °C for the G3 = 20Na₂O glass.

The cationic field strength (CFS) is often used to provide some general insight into the relative strengths of the various M-O bonds (M = modifying cation) within a glass [24,25,26,27].

CFS = z_c/(r_c)²

(1)

where z_c is the charge on the modifying cation, and r_c is the ionic radius. For glasses containing mixed modifying oxides, an effective CFS [27] can be calculated based on the proportions of the modifying cations present. Figure 4 shows that T_g increases linearly as a function of effective cation field strength, as La₂O₃ is substituted for Na₂O within the glass composition.

The value of T_g for the 20 mol.% La₂O₃ glass (874 °C) is within the range of 840–900 °C, reported on similar glasses [2,28]. T_g decreases significantly when 15 mol.% of the La₂O₃ is substituted by Na₂O since Na has a much lower cation field strength than La. T_g is lowered again when all of the La₂O₃ is replaced by Na₂O. Higher T_g is expected from stronger bonding of structural units by La³⁺ in the glass, thus making segmental mobility more difficult.

Sodium oxide substitution for lanthanum oxide in aluminosilicate glasses significantly reduces the glass transition temperature (T_g), and this is illustrated very clearly when T_g is plotted against effective cation field strength in Figure 4. Increasing Na content has implications for the presence of non-bridging oxygens and the local environments of Si and Al atoms.

In waste-immobilization applications, it is crucial to understand these effects: a lower T_g facilitates processing at lower temperatures, while understanding and controlling Si and Al environments ensures optimal chemical stability and resistance to leaching, essential for safely encapsulating radioactive waste materials within the glass matrix and ensuring long-term environmental safety.

3.2. Compositional Effects on Structure of Glasses

The only apparent variable compositional parameter is La₂O₃ to Na₂O, with the SiO₂ and Al₂O₃ concentrations remaining constant in all three glasses. As the concentration of Al₂O₃ is 25 mol.%, then, assuming that all the Al adopts 4-fold coordination as (AlO₄)⁻ units, 50 positive charges are required to compensate for the deficient negative charges, i.e., 16.67 La³⁺ or 50 Na⁺. The values of 16.67 La³⁺ and 50 Na⁺ represent the number of cations required to balance the charge of the deficient negative charges resulting from the assumed 4-fold coordination of Al as (AlO₄)⁻ units. In the case of La³⁺, this value is calculated by dividing the total positive-charge requirement (50) by the valence of La³⁺. Therefore, it is necessary to round this value to a practical amount, resulting in 16.67 La³⁺ ions. Similarly, for Na⁺, the entire positive-charge requirement can be provided by 50 Na⁺ ions due to their monovalent nature.

Clearly, for G1 with 20 mol.% La₂O₃, there is an excess of La³⁺ cations needed to balance all the (AlO₄)⁻ units, and, thus, the formation of non-bridging oxygen (NBO) atoms is expected. For the G2 glass with 5 mol.% La₂O₃ and 15 mol.% Na₂O, there is still an excess of La³⁺ and Na⁺ cations needed to provide charge compensation for (AlO₄)⁻ units, meaning that some of the La and Na cations are acting as modifiers, and some NBOs will be formed.

However, as 20 mol.% La₂O₃ (G1) is fully substituted by 20 mol.% Na₂O (G3), the number of positive charges available for compensation of (AlO₄)⁻ units decreases considerably. Previous studies [3,4,5] have shown that, for Al/(Al + Na) ratios ≤ 0.5, all Al³⁺ ions are found in AlO₄ tetrahedra, and Na ions are linked through NBOs to Q³ SiO₄ tetrahedra. For G3 with 20 mol.% Na₂O, there are insufficient positive charges to compensate all (AlO₄)⁻ units, and, therefore, some Al³⁺ will be acting as modifier cations with a higher coordination than 4. Thus, changes will be expected in the FTIR and MAS NMR spectra with changes in cation composition.

3.3. Fourier-Transform Infrared Spectroscopy

Figure 5 shows the reflectivity spectra of the three glasses, with each spectrum exhibiting three broad bands across the range of 350–1200 cm⁻¹, and with the most intense bands in the range 800–1200 cm⁻¹, followed by moderately intense bands across the 350–500 cm⁻¹ range and weaker bands across the 500–800 cm⁻¹ region. These broad bands are clear indicators of the general disorder of the silicate network [29].

It is noted that the strong band for glass G1 (20 mol.% La₂O₃) peaks at ~900 cm⁻¹ [30], whereas the Na-containing glasses show maxima at ~1000 cm⁻¹. These intense broad bands may be attributed to stretching vibrations of (i) SiO₄ tetrahedra with specific numbers, n, of bridging oxygen (BO) atoms to m second-neighbor Al atoms (Qⁿ(mAl)) and (ii) AlO₄ tetrahedra. The bands in the region of 350–500 cm⁻¹ are caused by the bending vibrations of Si-O-Si and Si-O-Al linkages [31]. In the case of G1 and G2 glasses, the excess of positive charges from La³⁺ and Na⁺ ions means that Al is present mainly as (AlO₄)⁻ units. This is confirmed by the reflectivity bands in the region 500–800 cm⁻¹ (maximum at ~700 cm⁻¹) that are related to the Al-O bond-stretching vibrations with 4-fold coordination (^[4]Al) [20,32], suggesting that (AlO₄)⁻ units are dominant features in all three glasses.

Figure 6 shows the experimental reflectivity spectrum of the G2 glass (5La₂O₃ + 15Na₂O) and the data reproduced with the FOCUS fitting software [21], using the Fresnel formula and a dielectric function model based on causal Gaussian components [32].

Five Gaussian components are assigned to simulate the high-frequency stretching mode region between 800 and 1200 cm⁻¹. The silicate lattice has five modes. The first, ν₁, is the vibration mode, which reflects the general disorder. The other four contributions (Qⁿ(mAl)) (see

Table 2. Structural features and their wavenumbers obtained from reflectivity spectra of the three glasses reproduced with the FOCUS fitting software.

Glass (mol.%)	Wavenumber (cm−1)	Structural Features of SiO4	Approximate Contribution (%)
G1 = 20La2O3	820	Q3(3Al)	68
	880	Q4(4Al)	26
	910	Q4(3Al)	6
G2 = 5La2O3 + 15Na2O	909	Q3(3Al)	18
	958	Q4(4Al)	43
	1009	Q4(3Al)	27
	1067	Q4(0Al)	8
G3 = 20Na2O	930	Q3(3Al)	10
	970	Q4(4Al)	60
	1000	Q4(3Al)	30

for all three glasses) are as follows: ν₂, Q⁴(0A1); ν₃, Q⁴(3Al); ν₄, Q⁴(4Al); and ν₅, Q³(3Al).

All glasses show the contribution of Q⁴(3Al), Q⁴(4Al), and Q³(3Al) structural features.

The Si-O-Si bond symmetric and antisymmetric stretching modes are well known to be IR-active in the 800–1300 cm⁻¹ region [29,30,31,32], particularly related to Q⁴ and Q³ units.

The major species (68%) in G1 glass (20La₂O₃) is Q³(3Al), i.e., SiO₄ tetrahedra with three Si-O-Al linkages. As there is an excess of La³⁺ cations over those needed to charge-balance all the (AlO₄)⁻ units, the formation of non-bridging oxygen (NBO) atoms to form Q³ tetrahedra occurs linked to La³⁺ modifier cations. The remaining features, namely Q⁴(4Al) (26%) and Q⁴(3Al) (6%), are SiO₄ tetrahedra with no NBOs and four or three Si-O-Al linkages. G2 glass (5La₂O₃+15Na₂O) has also Q⁴(4Al) (43%) and Q⁴(3Al) (27%) features, as well as some Q³(3Al) (18%) but also some Q⁴(0Al) units, i.e., SiO₄ tetrahedra with no Si-O-Al linkages. In this case, 10 La³⁺ + 30 Na⁺ provide a total of 60 positive charges, which can compensate fully for the 50 (AlO₄)⁻ units, but the excess positively charged cations which act as modifiers are less than for G1. Some of the La and Na (total charge of 10⁺) cations act as modifiers, and this would explain the presence of Q³(3Al) features with one NBO.

In the case of G3, the predominant species are, once again, Q⁴(4Al) (60%) and Q⁴(3Al) (30%), with a smaller amount of Q³(3Al) (10%), and their associated NBOs. With no La, 40 Na⁺ can provide charge compensation for 40 (AlO₄)⁻ units, meaning that 10 Al³⁺ may be acting as modifier cations (i.e., up to 20% of the Al may be in ^[5]Al or ^[6]Al coordination). The levels of Q⁴(4Al) and Q⁴(3Al), i.e., SiO₄ tetrahedra with four or three Si-O-Al linkages, appear high but may be an indication of a certain ordering of Si and Al atoms. However, these are only qualitative observations, and, although we have tried to relate these results to specific structural features expected from the actual compositional variations, in order to clarify these findings, especially for the G3 glass, it is necessary to compare the data with results from MAS NMR.

3.4. MAS NMR Study

3.4.1. ²⁹Si MAS NMR

The ²⁹Si MAS NMR spectra for the three glasses—G1 = 20La₂O₃, G2 = 5La₂O₃ + 15Na₂O, and G3 = 20Na₂O—are shown in Figure 7. G1 shows a broad and symmetric signal extending between −75 and −105 ppm, with a peak maximum (δ_max) at −86 ppm. As Na substitutes for La, the peak maxima are displaced toward slightly more negative chemical shifts.

The chemical shift of the ²⁹Si spectra is sensitive to the degree of polymerization of the glass network [33], i.e., the nature of the SiO₄ Qⁿ units. Given the overlap of the chemical-shift domains between the different Qⁿ(mAl) species and the wide distribution of possible environments in a glass network, the attributions of the ²⁹Si MAS NMR signals are difficult. Usually, the spectra of ²⁹Si give indications only of the most representative species, as well as qualitative data concerning almost fully polymerized environments, taking into account the width of the peaks. However, indicative signal assignments have been proposed in the literature [17,34,35], based on spectrum calculations and component assignments, in accordance with published chemical-shift domains for Qⁿ(mAl) species.

The ²⁹Si MAS NMR spectra of the three glasses were qualitatively simulated using three components. The number of Gaussian components introduced to deconvolute the spectra is defined by the number of minimum components with similar FWHM (Full Width at Half Maximum) values, set at 11 ± 0.5 ppm for all components. The results showing Qⁿ(mAl) and (δ_max) are listed in

Table 3. Deconvolution of the ²⁹Si MAS NMR spectra in terms of the maxima in the chemical shifts (δ_max) and their associated Qⁿ(mAl) features.

Glass (mol.%)	δmax (ppm)	Intensity (%)	Structural Unit
G1 = 20La2O3	−83	59	Q3(2Al)
	−87	34	Q4(4Al)
	−98	7	Q4(2Al)
G2 = 5La2O3 + 15Na2O	−82	23	Q3(2Al)
	−88	41	Q4(4Al)
	−94	27	Q3(0Al)
	−110	9	Q4(0Al)
G3 = 20Na2O	−80	16	Q2(0Al)
	−90	64	Q4(3Al)
	−99	20	Q4(2Al)

, based on previous assignments from Diallo et al. [36].

In the La-aluminosilicate glass (G1), the main species present are Q³(2Al) and SiO₄ tetrahedra with one NBO and linked to two AlO₄ tetrahedra, consistent with the excess of La³⁺ cations over those needed to charge-balance all the (AlO₄)⁻ units. Thus, the formation of non-bridging oxygen (NBO) atoms to form Q³ tetrahedra will occur. The other main species are Q⁴(4Al) with no NBOs and four Si-O-Al linkages with a small fraction of Q⁴(2Al) features, i.e., SiO₄ tetrahedra with no NBOs linking to two (AlO₄)⁻ units.

In summary, the La-aluminosilicate glass has 34% SiO₄ tetrahedra with four Si-O-Al linkages. As Na₂O is substituted for 15% La ₂O₃, this increases to 41%, but there are also similar proportions of Q³ species with either two or zero Si-O-Al linkages. The combined La³⁺ and Na⁺ cations provide sufficient positive charges to compensate for all the (AlO₄)⁻ units, with some excess positive charges acting as modifiers, thus allowing for the formation of one NBO per tetrahedron (Q³).

The fully substituted Na-aluminosilicate glass (G3) has 64% SiO₄ tetrahedra with three Si-O-Al linkages: Q⁴(3Al) and smaller amounts of Q⁴(2Al) (no NBOs linking to two (AlO₄)⁻ units) and also a minor proportion of Q²(0Al) (SiO₄ tetrahedra with two NBOs and no Si-O-Al linkages). In this case, Na ions, acting as modifiers, will be in close proximity. With no La³⁺ ions present, the monovalent Na⁺ can provide charge compensation for 40 (AlO₄)⁻ units, leaving 10 Al³⁺ cations to act as modifiers, meaning that less of the AlO₄ is taking part in the glass network, and this is consistent with the types of features found, i.e., Q⁴(2Al) and Q²(0Al).

Overall, the results from ²⁹Si MAS NMR for G1 and G2 glasses are compatible with those from FTIR, but with some slight divergence in the case of G3, where it may be expected that Al may be in ^[5]Al or ^[6]Al coordination. Thus, it is important to compare results with ²⁷Al MAS-NMR.

3.4.2. ²⁷Al MAS NMR

The ²⁷Al MAS-NMR spectra for the three glasses are shown in Figure 8. The study was performed to identify the different coordination states of Al³⁺ ions: ^[4]Al at ~55–60 ppm, ^[5]Al in the range from 25 to 40 ppm, and ^[6]Al between −15 and 20 ppm) [37,38].

The intense ²⁷Al NMR resonance at ~55–60 ppm due to ^[4]Al is observed for all glasses. In all cases, this shows that Al primarily acts as a network-former, as implied by the results of FTIR and ²⁹Si MAS NMR, which have confirmed significant Si-O-Al linkages.

As shown in

Table 4. Deconvolution of the ²⁷Al MAS NMR spectra.

Glass (mol.%)	δiso 27Al (ppm)	FWHM (ppm)	Intensity (%)	Structural Unit
G1 = 20La2O3	60.8	11	90	AlO4
	37.8	9	9	AlO5
	5.6	11	1	AlO6
G2 = 5La2O3 + 15Na2O	64.9	9	98	AlO4
	41.6	9	2	AlO5
G3 = 20Na2O	60.1	11	87	AlO4
	40.2	9	13	AlO5

, for the La-aluminosilicate glass (G1), a significant amount of 4-fold coordinated Al was identified (^[4]Al), along with a minor amount of ^[5]Al and an insignificant amount of 6-fold coordinated ^[6]Al. It has been reported previously that higher coordinated Al is found in aluminosilicate glasses with high-field-strength cations, such as La [39], which, in this case, are providing excess positive charges which must be charge-balanced by (^[5]AlO5)³⁻ units. However, a higher coordination state for La³⁺ cations would negate this, leading to a higher level of ^[4]Al. It has been suggested [10] that preferential localization of La³⁺ occurs near to 4-fold coordinated Al atoms, with the possible formation of oxygen triclusters or an increase in La³⁺ coordination, leading to the stabilization of ^[4]Al species.

As Na is substituted for La, with a lower level of positive charges to be satisfied, Al is almost wholly in 4-fold coordination with a very small amount of 5-fold ^[5]Al (no AlO₆ units are observed). All of the (AlO₄)⁻ units should be compensated by the La³⁺ and Na⁺ ions, the excess of which creates some non-bridging oxygens leading to the formation of Q³ species, some with no Si-O-Al linkages.

For G3 with 20 mol.% Na₂O, there are insufficient positive charges to compensate all (AlO₄)⁻ units, and, therefore, a more significant amount of Al³⁺ (13%) will be acting as modifier cations with higher coordination (^[5]Al), as found by other investigators [5,10,30,35].

The structural features of each of these glasses suggest that they are quite stable for potential use for radioactive waste containment, consisting of highly bonded networks consisting of SiO₄ tetrahedra with four or three Si-O-Al linkages. Comparing these glasses with equivalent composition glasses in the SiO₂–Al₂O₃–B₂O₃–Na₂O–La₂O₃ system studied previously [18,19], where Al₂O₃ was substituted by B₂O₃, we observed that phase separation and crystallization occurred at substitution levels above 5% B₂O₃, which would negate their use for radioactive waste immobilization.

4. Conclusions

A La-aluminosilicate glass was prepared with the following composition (in mol. %): 20% La₂O₃, 55% SiO₂ and 25% Al₂O₃ (G1). The effects of the partial substitution of 15% La₂O₃ by 15% Na₂O (G2) and the full substitution of all the La₂O₃ by 20% Na₂O (G3) on the T_g and structure of the glass were investigated.

T_g decreases significantly from 874 °C for the La-aluminosilicate glass containing 20% La₂O₃ to 790 °C for the glass with complete substitution of Na₂O for the La₂O₃, suggesting weaker bonding within the glass network. Higher T_g is expected from stronger bonding of structural units by La³⁺ in the glass, thus making segmental mobility more difficult.

Structural studies using FTIR spectroscopy, ²⁹Si, and ²⁷Al MAS NMR have identified the main Qⁿ species and changes in the local environments of Si and Al: {Qⁿ_Si(mAl)} and {Qⁿ_Al(mSi)}. In all the glasses, (AlO₄)⁻ units are a dominant feature, showing that Al is mostly acting as a glass former within the network. A structural feature of all glasses is the presence of SiO₄ tetrahedra with four Si-O-Al linkages [Q⁴(4Al)]. In the La-aluminosilicate glass (G1), the major species are SiO₄ tetrahedra with one NBO and two Si-O-Al linkages [Q³(2Al)]. This is reduced substantially as Na₂O is substituted for La₂O₃ (G2), with the appearance of SiO₄ tetrahedra with one NBO and no Si-O-Al linkages [Q³(0Al)]. The combined La³⁺ and Na⁺ cations provide sufficient positive charges to compensate for all the (AlO₄)⁻ units, with some excess positively charged cations acting as modifiers, so allowing the formation of one NBO per tetrahedron, i.e., Q³ species with either two or zero Si-O-Al linkages. A small proportion of five-fold coordinated ^[5]Al is present in these glasses.

The fully substituted Na-aluminosilicate glass (G3) has substantially more SiO₄ tetrahedra with three Si-O-Al linkages [64% Q⁴(3Al)]. In this case, Na ions as modifiers will be in close proximity. With no La³⁺ ions present, the monovalent Na⁺ can provide charge compensation for 40 (AlO₄)⁻ units, leaving 10 Al³⁺ cations to act as modifier cations with 5-fold coordination (^[5]Al), thus meaning that less of the (AlO₄)⁻ is taking part in the glass network, and this is consistent with the types of features found, i.e., Q⁴(2Al) and Q²(0Al).

These findings underscore the importance of cation composition in controlling the thermal and structural properties of aluminosilicate glasses. Future research will explore the optimization of compositions, for example, glasses with SiO₂:Al₂O₃ = 2:1, with varying Na₂O/Al₂O₃ and La₂O₃/Al₂O₃ contents, to further elucidate the role of the two modifier cations in the structure and properties of these aluminosilicate glasses.