Network coordination in low germanium alkaline-earth gallate systems: influence on glass formation

Abstract

Glass formation in a newly formulated low germanium alkaline earth gallate (LGMGa, M = Ca/Mg/Zn/Sr/Ba) system by a high temperature melt quenching route is reported here. The proportionate ratios of tetrahedral to octahedral coordination of both Ga³⁺ and Ge⁴⁺ network cations are found to be crucial for vitrification. The enhancement in octahedral coordination led to surface crystallization and phase separation in cast melts. This behavior has been attributed to the overall reduction in tetrahedrally coordinated network units due to the failure of Ca²⁺ ions in charge compensation of the [GaO₄]⁻ tetrahedron. Among different network modifying oxides, BaO in combination with CaO (mixed alkaline earth) containing compositions have proved to yield clear glass formation with improved tetrahedral networking. This has been explained though the basicity equalization concept and also due to efficient charge compensation of the [GaO₄]⁻ tetrahedron requiring lower field strength cations like Ba²⁺. The effective phonon energy of this low germanium calcium barium gallate (LGCBGa) glass is found to be around 650 cm⁻¹, with an extended infrared transparency up to around 7 μm, which is superior to most of the low phonon energy oxide glass systems, making it promising for photonic and mid-infrared luminescence applications. Furthermore, the observed dark yellow to brown coloration in the glass has been attributed to the precipitation of metallic nanoparticles, and this mechanism has been discussed in detail.

---

Introduction

Recent advancements in photonics have been concentrated towards the mid-infrared spectral region owing to their fascinating applications ranging from high precision molecular sensing, lasers for bio-medical applications, open space communications, surveillance, security and military applications. The potential of rare earth activated solid-state lasers for such applications demands high efficiency as well as high power output.¹ Glass plays an important role in this respect as a host for active ions. The basic requirements of glasses for such applications involve low phonon energy characteristics, low optical losses, and substantial thermal and mechanical strength. Although silicate glasses possess high thermal and mechanical durability, they become inferior in the mid-infrared region (beyond their natural minimum loss window of ∼1.5 μm) due to a high attenuation coefficient caused by high phonon characteristics, whereas fluorides and heavy metal oxide glasses exhibit other difficulties such as inferior physical properties, low damage threshold, and high scattering losses.^1,2 Therefore, an optimization has been sought to yield a more suitable low phonon energy glass, which may be possibly obtained based on the group IIIA metal oxides such as Al₂O₃ and Ga₂O₃.

Al₂O₃, being a glass network intermediate oxide, forms stable glasses with the aid of comparatively lower field strength cations such as alkali metal (group IA) or alkaline earth metal (group IIA) ions.² The structural chemistry of Al³⁺ ions prefers the [AlO₆] octahedron as the most stable polyhedral unit,³ but its moderate field strength gives the possibility for squeezing the oxygen coordination sphere down to a four-fold [AlO₄]⁻ tetrahedron,^3,4 as evidenced in the excess alkali containing silicate glasses and calcium-aluminate glasses.^2,5–7 The isovalent gallium (Ga³⁺) also follows the footsteps of Al³⁺ in a similar manner and forms a [GaO₄]⁻ tetrahedron to participate in the glass network structure.^8,9

The substitution of trivalent Al³⁺ over tetravalent Si⁴⁺ in the alkali silicate network gives rise to the polymerization of the anionic framework, where mono-valent alkali metal ions serve as charge compensators, leading to reduction of the overall density of non-bridging oxygens (NBOs) in the network. The inclusion of Al³⁺ in the alkali silicate network enhances both the viscosity and glass transition temperature (T_g) of glass along with added strength, chemical durability and thermal stability.² Furthermore, the lower bond strength of the Al^III–O bond in the [AlO₄]⁻ tetrahedron compared to the Si^IV–O bond reduces the lattice vibrational energy in low silica alkali/alkaline earth aluminate or binary alkali/alkaline earth aluminate glass.^10,11 The reduced phonon energy of aluminate glasses (∼800 cm⁻¹) extends the infrared window with the cut-off lying at around 6–6.5 μm, which is comparable to the heavy metal oxide glasses, besides exhibiting significantly low optical loss.¹² The substitution of Al³⁺ by Ga³⁺ in these networks can further extend the infrared transparency up to 7–8 μm, owing to the fairly low lattice vibrational energy of the Ga^III–O bond in the [GaO₄] tetrahedron (∼650 cm⁻¹).¹³

But, unlike Al³⁺, the glass forming ability of Ga³⁺ is comparatively poor. Shelby observed that gallosilicate glasses are more prone to phase separation than aluminosilicate glasses.⁸ The larger ionic radius of Ga³⁺ compared to Al³⁺ reduces the cationic field strength and makes it less favorable to form tetrahedral coordination with oxygen.³ In ideal conditions, one alkali ion per [M^IIIO₄]⁻ tetrahedron is sufficient for the charge compensation. But it is observed that a fraction of M³⁺ cations tends to occupy the octahedral coordination, which demands excess alkali ions to enhance the overall tetrahedral/octahedral ratio in the network.¹⁴ The basicity equalization concept of Duffy and Ingram gives a satisfactory explanation of this behavior.¹⁵ Accordingly, the structural units that have a group basicity close to the glass basicity are more abundant in the network. Miura et al.¹⁶ and Tanaka et al.¹⁷ could successfully predict the coordination chemistry of [BO₄] to [BO₃] proportionality in borosilicate glasses based on the basicity equalization concept. This concept provides clear guidelines for the structural chemistry of such glasses involving multi-coordinated network units.

The present work has been initiated to develop stable glass compositions based on Ga₂O₃ as the main network former, in association with alkaline earth cations and a small amount of GeO₂, in order to achieve low phonon characteristics for photonic applications. The glasses with low phonon energy characteristics are preferred hosts for luminescent ions in achieving better quantum efficiency, owing to their reduced non-radiative losses, apart from their extended infra-red transparency. Furthermore, some specific luminescence transitions, which are otherwise impossible in high phonon energy glass, could be achieved in such glasses. Thus, these gallate glass systems could be useful for the development of mid-infrared lasers and waveguides with low optical losses. Nishida et al.,^18,19 have also proposed the possibility of calcium-gallate glasses in optical memory devices. However, they have prepared such glasses by the rapid quenching of melt in ice-cold water and thus obtained samples with a transmission of about 40% only in the infrared region. Our initial experiments also revealed that the glass formation of the binary calcium-gallate system is quite poor. So, in this work, a series of compositional variations have been examined to achieve good glass formation in the gallate glass system. The low silica calcium aluminate (LSCA) system has been selected as a starting glass, owing to its superior properties among low phonon oxide glass systems.¹⁰ Since Ga³⁺ and Al³⁺ are iso-structural in the network, we have tried to maintain the similar structural network to get the advantages of LSCA glasses in the gallate glass system.

This work provides a detailed insight into the structural modifications due to the substitution of lower field strength Ga³⁺ and Ge⁴⁺ in the LSCA network, along with the effect of different alkaline earth metal ions on the glass formation ability. Infrared spectroscopy, differential thermal analysis and optical absorption spectroscopy have been adopted for thorough analysis of the structural and optical properties of prepared glasses. The results have been explained based on the glass basicity and the role of alkaline earth metal ions in the network, as well as their ability for charge compensation.

Experimental methods

Glass preparation

All the glasses, with chemical compositions as listed in Table 1, were prepared by a conventional melt quenching technique in a normal air atmosphere. High purity reagent grade chemicals (99.5% Al(OH)₃ (Sumitomo CHP-340S) from Solvadis Specialties GmbH, Germany; 99.8% SiO₂ from Brementhaler Quarzitwerk, GmbH, Germany; 99.99% Ga₂O₃ and SrCO₃, 99.999% GeO₂, 99.9% MgO from Sigma-Aldrich, Germany; 99.9% CaCO₃ and BaCO₃ from Loba-Chemie, India; and 99.9% ZnO from Fluka, Germany) were used for batch preparation. Each chemical batch was prepared to yield 15 g glass, mixed thoroughly using agate mortar and sintered at 800 °C for 1 h. The sintered batch was then taken into a pure Pt crucible and introduced to an electrical heating furnace at 1250 °C. The glass melting temperature has been optimized at 1500–1600 °C, depending on the respective chemical compositions, and heating was carried out for 2–3 h with intermittent stirring to ensure melt homogeneity. The melt was then poured onto a preheated (∼300 °C) iron mold and transferred to an annealing furnace. The annealing temperature was maintained at approximately 10 °C below the glass transition temperature for 1 h and then cooled slowly to room temperature. The obtained glasses were cut and polished to yield 2 mm thick clear glass plates for various characterizations.

Table 1 Chemical composition of glass samples investigated in the present study

Glass	Composition (mol%)
	Al2O3	Ga2O3	CaO	ZnO	SrO	BaO	MgO	SiO2	GeO2	La2O3
Series 1
G-0a	28	—	56	—	—	—	4	10	—	2
G-1a	21	7	56	—	—	—	4	10	—	2
G-2a	14	14	56	—	—	—	4	10	—	2
G-3a	7	21	56	—	—	—	4	10	—	2
G-4a	—	28	56	—	—	—	4	10	—	2
G-5a	—	28	58	—	—	—	4	8	—	2
G-6a	—	28	60	—	—	—	4	6	—	2
Series 2
G-4b	—	28	56	—	—	—	4	8	2	2
G-4c	—	28	56	—	—	—	4	6	4	2
G-4d	—	28	56	—	—	—	4	3	7	2
G-4e	—	28	56	—	—	—	4	—	10	2
Series 3
G-3e	7	21	56	—	—	—	4	—	10	2
G-4eMg	—	28	50	—	—	—	10	—	10	2
G-4eZn	—	28	28	28	—	—	4	—	10	2
G-4eSr	—	28	28	—	28	—	4	—	10	2
G-4eBa	—	28	28	—	—	28	4	—	10	2

---

Characterization

The densities of well-annealed glasses were measured by applying Archimedes' buoyancy principle, using water as the immersion liquid on a Mettler-Toledo digital mono-pan balance fitted with a density measurement kit. The refractive indices of the glasses were measured on a prism coupler (model: Metricon M2010, USA) attached with five diode lasers of different wavelengths (473 nm, 532 nm, 633 nm, 1064 nm and 1552 nm). The infrared reflection and transmission spectra of glasses were recorded in the 50–7800 cm⁻¹ wavenumber range on a FT-IR spectrophotometer (model: Frontier, FIR-FTIR, Perkin-Elmer, USA). The optical absorption spectra were recorded in the wavelength range 200–3000 nm on a UV-Vis-IR spectrophotometer (model: 3101, Shimadzu, Japan) with a spectral resolution of 0.5 nm. Differential thermal analysis (DTA) of the prepared glasses was carried out from room temperature to 1400 °C on a NETZSCH STA 449C, Germany, with a heating rate of 10 K min⁻¹. Approximately 100 mg of the sample was taken into a Pt-Ir crucible for DTA analysis. X-ray diffraction (XRD) analysis of prepared glasses was carried out on an X'pert Pro-MPD powder diffractometer, PANalytical, Almelo, Netherlands, to verify their amorphous nature. The glass microstructure was investigated by transmission electron microscopy (TEM) FEI Model Tecnai G2 30 ST, Hillsboro, OR, USA, equipped with an energy dispersive spectroscopy (EDS) system. Samples for TEM measurement were prepared by dispersing finely powdered glass in ethanol, followed by an ultrasonic agitation, and then its deposition onto a carbon-enhanced copper grid.

Results and discussion

Glass composition and its influence on structural properties

Table 1 enlists the chemical compositions of glasses sub-divided into three different series, which have been worked out in the present study. In Series 1, the Al₂O₃ contents in the glasses of composition (mol%) 56CaO–28Al₂O₃–4MgO–10SiO₂–2La₂O₃ were substituted by Ga₂O₃ in steps of 7 mol%. But with the increase in Ga₂O₃ contents in glass, a quick devitrification of the surface layer was observed during melt casting. Furthermore, the melt was observed to be less viscous for higher Ga₂O₃ containing glasses, which became more profound on subsequent reduction in SiO₂ content (in G-5a and G-6a samples). So, in order to maintain the network linkage in the glass, instead of reducing the SiO₂ contents, it was later replaced by GeO₂, yielding a set of four glasses comprised for Series 2. Even this replacement gave rise to another difficulty of phase separated melt, and the resulting glass was found to be partially crystallized. To avoid this problem of phase separation and improving glass formation, different combinations of glass modifying oxides were examined by introducing MgO, ZnO, SrO and BaO with partial substitution of CaO in glass composition, and these make up Series 3. It should be mentioned that good glass formation was achieved for the compositions containing an equal share of CaO and BaO as network modifying oxides (G-4eBa). The infrared reflectance spectroscopy and differential thermal analysis of glass samples were adopted to understand the structural modifications and glass stability factor brought about by compositional changes.

Fig. 1a shows the infrared reflectance spectra of the glasses in Series 1. It can be seen that the G-0a sample exhibits an intense band at 790 cm⁻¹ with two shoulders on both sides, one at 650 cm⁻¹ and the other peaking at 890 cm⁻¹, respectively. Table 2 presents the assignments of these bands, which are ascribed to the vibrational transitions of the Al–O bond in tetrahedral coordination, Al–O in octahedral units, and the Si–O vibrations in the non-bridging Q^(4−x) state as well as in the ‘isolated’ tetrahedron, respectively.^10,20 This 790 cm⁻¹ band was systematically reduced upon Ga₂O₃ substitution and two new bands appeared at around 636 cm⁻¹ and 503 cm⁻¹ due to the Ga–O bond vibration in the [GaO₄] tetrahedron and [GaO₆] octahedron, respectively.¹³ Furthermore, the vibrational band at ∼440 cm⁻¹ showed a red shift to 360 cm⁻¹, due to the fading effect in Al–O bending modes on Ga₂O₃ inclusion, which exhibits bending vibrations at around 360 cm⁻¹.¹³ The results suggest a systematic replacement of Al³⁺ in both the tetrahedral and octahedral coordination by Ga³⁺ cations. However, a close observation reveals that the relative intensities of tetrahedral to octahedral units in aluminate (G-0a) and gallate (G-4a) glasses exhibit a comparatively higher fraction of octahedral coordination in G-4a glass. Interestingly, this trend was reversed upon reducing the SiO₂ content, as can be seen from the spectra of the G-5a and G-6a glasses respectively, where the 636 cm⁻¹ band showed an enhancement and the 503 cm⁻¹ band went down compared to G-4a glass, implying improvement of the tetrahedral [GaO₄] coordination at the expense of the [GaO₆] units in the glass network.

Fig. 1 (a) Infrared reflectance spectra of Series 1, and (b) Series 2 glasses.

Table 2 Vibrational bands of different structural units in present glasses

IR band	Position (±10 cm^−1)	Description
1	890 cm^−1	Si–O stretching vibration in [SiO4]^x− tetrahedron
2	790 cm^−1	Al–O stretching vibration in [AlO4] tetrahedron
3	650 cm^−1	Al–O stretching vibration in [AlO6] octahedron
4	636 cm^−1	Ga–O stretching vibration in [GaO4] tetrahedron
5	503 cm^−1	Ga–O stretching vibration in [GaO6] octahedron
6	440 cm^−1	Al–O and Si–O bending vibrations
7	360 cm^−1	Ga–O bending vibrations
8	745 cm^−1	Ge–O stretching vibration in [GeO4]^x− tetrahedron
9	610 cm^−1	Ge–O stretching vibration in [GeO6] octahedron
10	407 cm^−1	Ge–O bending vibrations
11	>300 cm^−1	M^II–O vibrations in multi-polyhedral units

---

This strange cation coordination behavior has its explanation in the concept of glass basicity. Table 3 depicts the basicity of the glasses under study, along with some important physical properties. The basicity has been estimated by three different methods, based on (i) Pauling's electronegativity scale,^3,15 (ii) the oxide ‘optical’ basicity concept of Duffy and Ingram,^15,21 and (iii) the experimental method using measured refractive indices²² (Appendix I). The basicity values calculated from these methods show significant deviations from each other, which may be due to the different approaches used for their estimation.¹⁵ The theoretical basicity based on Duffy's method follows a similar trend to the experimental basicity values, increasing with the inclusion of Ga₂O₃ in the glass; but the basicity values estimated using Pauling's electronegativity scale show a decrease in glass basicity up to G-4a, and then a small increase for G-5a and G-6a glasses. Pauling defined electronegativity as the relative tendency towards electron attraction.³ Gallium, being more electronegative attracts a larger electron cloud in Ga–O bonding and thus reduces the total charge density of the anionic ligand, which in turn affects the overall electron donating power of the network and reduces the glass basicity. Furthermore, the reduction in acidic SiO₂ content in the G-5a and G-6a glasses again enhances the theoretical basicity of the glass. Since the metal ion favors higher coordination in an acidic environment, based on the above argument the proportionality of octahedral coordination increases from G-0a to G-4a glass due to the reduction in glass basicity, and then decreases on SiO₂ reduction, favoring the formation of tetrahedral coordination of Ga³⁺ cations. Further reduction in the theoretical basicity (Pauling) of the glass with the GeO₂ substitution for SiO₂ content in Series 2 compositions hints at the formation of more fractions of octahedrally coordinated units in the network, which might have resulted in phase separation of the glass melt. The infrared reflectance spectra shown in Fig. 1b give more insight into the structural modifications induced by GeO₂ inclusion. The Si–O stretching band at 890 cm⁻¹ diminished upon GeO₂ substitution, accompanied by a broadening of the 636 cm⁻¹ band with a slight blue shift in its peak position to 665 cm⁻¹.

Table 3 Physical nature of prepared glasses, density (d), average molecular weight (M_avg), refractive index (n_d), glass basicity (Λ) and thermal properties of present glasses

Glass	Nature of glass	Physical properties			Glass basicity (Λ)			Thermal properties
		d (g cm^−3)	M avg (g mol^−1)	n d	Pauling	Duffy	Exp^a	T g (±3 °C)	T p1 (±1 °C)	T p2 (±1 °C)	T m (±2 °C)	T x − Tg (±5 °C)
a Derived from experimental refractive index and oxide ion polarizability.(Glass + SC) refers to quick devitrification of upper surface layer of poured glass. (Glass + C) refers to partially crystallized glass – phase separated.
G-0a	Glass	3.068	74.1	1.676	0.697	0.747	0.799	806	941	970	1343	113
G-1a	Glass	3.361	80.1	1.690	0.689	0.760	0.795	786	918	943	1298	104
G-2a	Glass	3.558	86.1	1.708	0.679	0.772	0.824	761	880	915	1274	96
G-3a	Glass + SC	3.766	92.0	1.728	0.671	0.784	0.849	740	846	891	1250	91
G-4a	Glass + SC	3.973	98.0	1.751	0.662	0.797	0.874	738	829	865	1257	90
G-5a	Glass + SC	4.013	96.8	1.766	0.667	0.804	0.886	701	824	844	1228	93
G-6a	Glass + SC	4.067	95.6	1.774	0.672	0.812	0.889	670	778	821	1213	95
G-4b	Glass + C	4.038	98.9	1.772	0.661	0.800	0.881	736	839	—	1261	86
G-4c	Glass + C	4.083	99.8	1.774	0.660	0.803	0.881	738	834	—	1260	81
G-4d	Glass + C	4.093	101.1	1.780	0.659	0.807	0.894	739	827	—	1244	72
G-4e	Glass + C	4.106	102.5	1.783	0.658	0.812	0.905	730	816	960	1257	72
G-3e	Glass + C	—	96.5	—	0.666	0.800	—	736	860	—	1238	85
G-4eMg	Crystallized	—	101.5	—	0.647	0.804	—	690	722	801	1244	<50
G-4eZn	Crystallized	—	109.6	—	0.581	0.804	—	658	715	750	1238	<30
G-4eSr	Glass + C	—	115.8	—	0.670	0.828	—	726	814	—	1228	77
G-4eBa	Glass	4.546	129.7	1.789	0.686	0.837	0.964	710	839	861	1178	106

---

This is due to the superposition of the Ga–O bond vibration with the Ge–O stretching vibrational bands in the [GeO₄] tetrahedron peaking at around 745 cm⁻¹.²³ On critically observing the shift as well as the enhancement of the 636 cm⁻¹ band in Series 2 glasses, it is observed that both peak shifting and intensity enhancement is slowly retarded with the increase in GeO₂ content up to G-4d glass and then this trend is reversed for the G-4e glass sample. Usually, the structure of germinate glasses is complicated, since unlike SiO₂, GeO₂ exhibits two kinds of structural units, [GeO₄] and [GeO₆].^15,17 The observed changes in infrared spectra on GeO₂ inclusion can be correlated with the relative enhancement in octahedral [GeO₆] units in higher GeO₂ containing glasses (Series 2). This combined effect of increased proportions of both [GaO₆] and [GeO₆] units in Series 2 glasses reduces the network connectivity and results in partial devitrification of glasses, as observed in the present case.

This observation suggests that to obtain a good glass in the present Ga₂O₃ based composition, more Ga³⁺ should be in a tetrahedral coordination, which needs a more basic environment to facilitate an effective charge compensation of the [GaO₄]⁻ tetrahedron. The above assumption is strengthened by the observation of complete crystallization of the melts having MgO (G-4eMg) and ZnO (G-4eZn) as network modifiers in Series 3, which gives a more acidic nature to the network, and thus poor charge compensation. In this series, the addition of SrO to the glass composition eases this problem to some extent, but still the cast glass exhibits a small amount of partial crystallization. However, the BaO containing composition showed an excellent glass forming ability, resulting in clear glass without any crystallization, as can be noticed from the image of G-4eBa glass presented in the inset of Fig. 2a. Fig. 2a shows the infrared reflectance spectra of G-4eBa glass in comparison with the G-4e glass sample. It can be realized that the intensity of the 665 cm⁻¹ vibrational band shows significant enhancement on BaO inclusion in the network. Furthermore, the shoulder band at 745 cm⁻¹ is more pronounced in G-4eBa glass compared to G-4e. These changes indicate a relative enhancement in the tetrahedral coordination of both the Ga³⁺ and Ge⁴⁺ cations, which is confirmed by the deconvoluted components of vibrational bands depicted in Fig. 2b and 2c, respectively. The increased fraction of tetrahedral coordination units in G-4eBa glasses allows for better network linking and thus yields good glass formation.

Fig. 2 (a) Infrared reflectance spectra of G-4e and G-4eBa glasses along with their deconvoluted components (b) and (c), respectively (inset: image of the G-4eBa glass sample).

From all these observations, it can be understood that the theoretical basicity based on Pauling's electronegativity exhibits some definite correlation with the glass formation that has been achieved in the worked out compositions in the present investigation. The thermal analysis of the prepared glass samples also supports the same trend, as can be noticed from the glass stability factor (T_x − T_g) listed in Table 3 along with the glass transition temperature (T_g), crystallization peak temperatures (T_p1, T_p2) and melting temperature (T_m). The data clearly indicates that the glass stability factor is highest for the G-0a and G-4eBa samples, which practically could form clear glasses by an ordinary melt quenching technique. Another interesting observation of thermal analysis is that there is a decreasing trend in the glass transition and melting temperatures in the studied series of glasses, which can be attributed to the decreased bond strength with the modifications in chemical composition. The melting temperature is almost 165 °C lower in G-4eBa glass compared to the G-0a glass, which is one of the advantages of the gallate glass system, requiring a lower melting temperature despite maintaining similar thermal stability over crystallization. The differential thermal analysis (DTA) curves of some selected glasses are depicted in Fig. 3. It is evident that the crystallization peak profile undergoes striking changes with the composition modifications. The broad crystallization peak observed in G-0a glass becomes sharper in G-4a and G-4e glasses and then again becomes broad in G-4eBa glass. Such sharpening of the crystallization peak may arise due to the reduced viscosity of glass on Ga₂O₃ inclusion. But, on addition of Ba²⁺ into the network, the viscosity of the glass may again increase, due to the rise in network forming tetrahedral units in G-4eBa glass, accompanied by the presence of larger cations (Ba²⁺) in the network creating obstacles in the process of crystallization.²⁴ This effect has been experienced practically in the present glass preparation.

Fig. 3 DTA profiles of G-0a, G-4a, G-4e and G-4eBa glasses.

Optical transparency and presence of metal nanoparticles

The reduced effective phonon energy of the glasses on Ga₂O₃ substitution to Al₂O₃ followed by GeO₂ with SiO₂ has extended the infrared transparency in gallate glasses, as has been demonstrated in the infrared absorption spectra shown in Fig. 4. The infrared band edge of aluminate glass, G-0a, shifts significantly to the lower energy region in gallate glasses following the pattern G-0a > G-4a > G-4e > G-4eBa. The spectra exhibit a broad absorption band at around 3400 cm⁻¹, which is due to a small amount of OH⁻ impurity present in the glass. Another absorption band at around 1780 cm⁻¹ observed in G-0a and G-4a glasses is attributed to the second harmonic of the Si–O bond vibration in the non-bridging Q^(4−x) state or the ‘isolated’ tetrahedron, respectively. This absorption band could not be detected in G-4e and G-4eBa glasses, owing to the absence of SiO₂; instead, a small hump at around 1500 cm⁻¹ has been observed in the later glasses due to the second harmonic of the Ge–O bond vibrations. This overall shift in the vibrational band positions as a function of compositional modification from aluminate to gallate glasses has widened the infrared transparency and extended the IR cut-off from 1750 cm⁻¹ in G-0a glass to 1490 cm⁻¹ in G-4eBa glass.

Fig. 4 Infrared absorption spectra of G-0a, G-4a, G-4e and G-4eBa glasses (inset: infrared transmission of glasses).

Similarly to the infrared band edge, the UV band edge also exhibits a red shift on Ga₂O₃ inclusion in the glass, as can be seen from the UV-Vis-NIR absorption spectra of the glasses shown in Fig. 5. This shift in UV cut-off is expected, owing to the comparatively lower bond strength of Ga–O bonding compared to Al–O bonding. The optical transparency of these glasses is observed to be more than 80%. Apart from the UV band edge shift, a small absorption band could be detected in the absorption spectra of gallate glasses peaking at around 450 nm. The inset shows an enlarged view of absorption spectra, giving a clearer profile of the absorption band. It can be seen that the intensity of this absorption band increases from G-4a to G-4eBa glass, imparting a slightly brown colored tint to the glass. A similar observation has also been reported by Shaw and Shelby¹⁴ in barium gallosilicate glass, where they observed the appearance of a broad absorption spectrum in the visible region on increasing the BaO content in the glass.

Fig. 5 UV-Vis-NIR optical absorption spectra of G-0a, G-4a, G-4e and G-4eBa glasses (inset: enlarged view of the spectra in the UV-Vis region).

Since the raw materials used in the glass preparation were highly pure and the subsequent processing was carried out by taking sufficient precautions to avoid any contamination of transition or rare earth ion impurities, the only culprit behind this absorption band can be the surface plasmon resonance (SPR) of metal nanoparticles generated in the glass matrix.²⁵ So, to confirm the presence of metal nanoparticles in the glass matrix, X-ray diffraction patterns of glass samples were recorded, but it did not show any distinct diffraction pattern, suggesting the glasses are X-ray amorphous. However, TEM analysis of the glass sample displayed the presence of nanoparticles scattered throughout the glass matrix, as seen from the TEM image of G-4eBa glasses presented in Fig. 6a. From the TEM image, it is clear that the volume percentage of nanoparticles in the matrix is smaller, which might be a reason behind the absence of distinct diffraction peaks in the XRD pattern. However, the selected area electron diffraction (SAED) analysis under TEM revealed a lattice diffraction pattern, confirming the presence of nanoparticles, as shown in Fig. 6b. This pattern has been identified as being due to the metallic Ga⁰ nanoparticles (JCPDS file No. 02 0480) exhibiting different diffraction planes, as shown in Fig. 6b. The high resolution transmission electron microscopy (HR-TEM) image of the nanoparticle is shown in Fig. 6c along with the corresponding FFT pattern (Fig. 6d). The HR-TEM image shows a lattice plane with inter-planar spacing of 3.2 Å, further confirming the presence of Ga⁰ nanoparticles in glass. Germanium, Ge⁰, also exhibits the same inter-planar spacing of 3.2 Å in the cubic lattice. Also, the diffraction planes of metallic Ge⁰ match well with the SAED pattern shown in Fig. 6b (JCPDS file No. 03 478). However, Ga⁰ exhibits more diffraction planes compared to Ge⁰ and all of the diffraction planes of Ga⁰ could be matched, confirming its presence in the glass. However, the diffraction pattern and HR-TEM still also hint at the formation of Ge⁰ nanoparticles and their presence in the glass, accompanied by Ga⁰ nanoparticles.

Fig. 6 (a) Bright-field TEM image of G-4eBa glass, showing nanoparticles in the glass; (b) selected area electron diffraction (SAED) pattern; (c) HR-TEM image of nanoparticles; and (d) fast Fourier Transform (FFT) pattern.

In fact, the reducing environment of basic glasses is ideal for the formation of metal nanoparticles. The high basicity (experimental) of present glasses can spontaneously reduce the metal cations to metallic atoms, which then agglomerate to form nanoparticles. The tendency of such reduction of metal cations depends on the standard reduction potential (E⁰) of the reaction. Actually, the reduction potential is higher for germanium than for gallium ions, as illustrated in the following reactions:²⁶

Ge⁴⁺ + 4e⁻ → Ge⁰ (E⁰ = +0.124 V)

Ga³⁺ + 3e⁻ → Ga⁰ (E⁰ = −0.549 V)

So, this indicates that Ge⁴⁺ is more prone to form nanoparticles over Ga³⁺, and thus the coexistence of Ge⁰ nanoparticles along with Ga⁰ in the glass can be anticipated.

The deconvoluted components of the absorption spectrum of G-4eBa glass have been plotted in Fig. 7, representing the SPR bands of metal nanoparticles as well as the base glass spectrum. The spectrum indicates the presence of two distinct SPR peaks centered at 3.41 eV and 2.87 eV, respectively.^27,28 The combined SPR spectrum of the metal nanoparticles is presented in the inset, showing strong absorption in the UV-visible region of the spectrum. Although the presence of metal nanoparticles imparts brown colored tint in the glasses, their large surface area to volume ratio gives rise to an enhanced Coulombic field, advantageous for superior luminescent properties of dopant rare earth ions. It is also possible to minimize this SPR absorption band by introducing some suitable oxidizing agents like Sb₂O₃ into glasses, which oxidizes the metal nanoparticles and thus gives clear glass. This effect of oxidizing agents on metal nanoparticles in gallate glasses will be discussed in our next report.

Fig. 7 UV-Vis absorption spectra of G-4eBa glass along with its deconvoluted components (inset: SPR spectrum of nanoparticles).

Conclusion

In summary, stable glass formation has been achieved in a low germanium calcium barium gallate (LGCBGa) glass system with [GaO₄]⁻ serving as the main network unit. The problems of surface crystallization and phase separation encountered due to the substitution of Al³⁺ and Si⁴⁺ with the comparatively lower field strength of Ga³⁺ and Ge⁴⁺ in the glass network have systematically been addressed by thorough investigation of the structural chemistry of glasses. The results have been discussed through a correlation of network coordination, with the theoretical basicity of glass and glass formation. The failure of Ca²⁺ ions alone in the effective charge compensation of network cations indeed favored octahedral coordination and resulted in poor glass formation. The partial replacement of Ca²⁺ with Ba²⁺ enhanced the tetrahedral to octahedral coordination ratio in the glass network due to the comparatively more basic characteristics of Ba²⁺ ions, allowing easy electron donation in charge compensation.

Appendix I

The concept of basicity originated from the standard acid–base principle of matter, based on the Lewis acid–base theory.² An acid is a species (ion, atom or molecule) that will eagerly accept a pair of electrons to form a more stable product, and a base is the one that donates this pair of electrons to the acid. Thus a base will have an excess electron density, which in general represents its basicity (Λ). The O²⁻ ion in its free state exhibits the highest electron density, but when it comes in contact with metal cations, it shares a part of its electron cloud with the metal, which in turn reduces its ability to donate electrons. Since such sharing of the electron cloud depends on the electronegativity (χ) of the metal cation, as suggested by Pauling, the electronegativity exhibits a correlation with basicity (Pauling) given as,^10,21Here, X_i represents the fraction of oxygen atoms carried by individual metal cations (M_i) in oxide, over the total number of oxygen atoms present. The term in the denominator is referred to as the basicity moderating parameter, defining the electron attraction ability of an individual cation. Duffy and Ingram²¹ experimentally investigated the basicity of oxides of metal cations based on the shift in the absorption peak of p-block transition metal ions like Pb²⁺ doped in the respective oxide. Accordingly the ‘optical’ basicity can be estimated by using these values of independent oxide basicity (Λ_o) by adopting the relation^15,21

Another approach of direct experimental estimation of optical basicity is based on the determination of the refractive index (n), which in turn represents the electronic polarizability (α_O²⁻).²¹

Here V_m is the molar volume, α_i is the molar cation polarizability and N_O²⁻ is the number of oxygen ions in the composition.

Acknowledgements

Authors would like to thank Prof. I. Manna, Director, CSIR-CGCRI and Dr Ranjan Sen, Head, Glass Division for their kind encouragement and permission to publish this work that was carried out under TAPSUN-NWP0055 project. One of us (A. D. S.) is thankful to the CSIR, New Delhi for the award of Senior Research Fellowship to him.

References

1. S. D. Jackson, Nat. Photonics, 2012, 6, 423–431.
2. P. O. Mysen and P. Richet, Silicate glasses and melts, Elsevier, 2005.
3. L. Pauling, The Nature of the Chemical Bond, Third Edition, Cornell University Press, 1960.
4. M. Poulain, Nature, 1981, 293, 279–280.
5. P. McMillan, Am. Mineral., 1984, 69, 622–644.
6. R. A. Pires, I. Abrahams, T. G. Nunes and G. E. Hawkes, J. Mater. Chem., 2009, 19, 3652–3660.
7. D. R. Neuville, L. Cormier and D. Massiot, Chem. Geol., 2006, 229, 173–185.
8. J. E. Shelby, Phys. Chem. Glasses, 1994, 35, 147–152.
9. P. W. Angel, C. S. Ray and D. E. Day, J. Am. Ceram. Soc., 1990, 73, 2965–2969.
10. A. D. Sontakke and K. Annapurna, Spectrochim. Acta, Part A, 2012, 94, 180–185.
11. P. R. Foy, Fabrication and characterization of Calcium-aluminate glass fibers, PhD Thesis, The State University of New Jersey, 2008.
12. M. E. Lines, J. B. MacChesney, K. B. Lyons, A. J. Bruce, A. E. Miller and K. Nassau, J. Non-Cryst. Solids, 1989, 107, 251–260.
13. Y. Zhao and R. L. Frost, J. Raman Spectrosc., 2008, 39, 1494–1501.
14. C. M. Shaw and J. E. Shelby, Phys. Chem. Glasses, 1990, 31, 199.
15. J. E. Duffy and M. D. Ingram, J. Non-Cryst. Solids, 1976, 21, 373–410.
16. Y. Miura, H. Kusano, T. Nanba and S. Matsumoto, J. Non-Cryst. Solids, 2001, 290, 1–14.
17. Y. Tanaka, Y. Benino, T. Nanba, S. Sakida and Y. Miura, Phys. Chem. Glasses, 2009, 50, 289–293.
18. T. Nishida, T. Ichii and Y. Takashima, J. Mater. Chem., 1992, 2, 733–737.
19. T. Nishida, S. Kubuki and Y. Takashima, J. Non-Cryst. Solids, 1994, 177, 193–199.
20. P. Tarte, Spectrochim. Acta, 1967, 23A, 2127–2143.
21. J. A. Duffy and M. D. Ingram, J. Am. Ceram. Soc., 1971, 93, 6448–6454.
22. V. Dimitrov and T. Komatsu, J. Univ. Chem. Tech. Metal., 2010, 45, 219–250.
23. E. Culea, P. Pop, M. Bosca, T. Rusu, P. Pascuta and S. Rada, J. Phys.: Conf. Ser., 2009, 182, 012061.
24. H. Scholze, Glass: Nature, Structure and Properties, Springer-Verlag Publication, New York, Inc., 1991.
25. S. Eustis and M. A. El-Sayed, Chem. Soc. Rev., 2006, 35, 209–217.
26. W. M. Haynes and D. R. Lide, in CRC Handbook of Chemistry and Physics, ed. D. R. Lide, CRC Press, Boca Raton, FL, USA, 92nd edn, 2011–2012.
27. P. C. Wu, M. Losurdo, T.-H. Kim, M. Giangregorio, G. Bruno, H. O. Veritt and A. S. Brown, Langmuir, 2009, 25, 924–930.
28. P. Tognini, A. Stella, P. Cheyssac and R. Kofman, J. Non-Cryst. Solids, 1999, 249, 117–122.

---