Structure and crystallization behavior of Al containing glasses in the CaO–B₂O₃–SiO₂ system

Abstract

The aim of the present work is to investigate the effects of Al₂O₃ content on both the structure and crystallization behavior of CaO–B₂O₃–SiO₂ (CBS) glass composition using NMR, FT-IR and DSC. The density experimental results showed that the density (ρ) of glasses decreased linearly with increasing Al₂O₃ content, while the molar volume (V_m) of the glasses increased linearly. The NMR revealed that the Al occurred in the forms of Al^IV, Al^V and Al^VI units. With the increasing Al₂O₃ content, the amount of Al species show little change, the relative amount of B^IIIa and B^IIIs units increases while the B^IV units decrease, and the Q⁴ structural units increases clearly but the Q³ decreases. Further, crystallization peaks of CBS glasses shift to higher temperatures, indicating that Al₂O₃ addition can suppress crystallization of CBS glasses.

---

1. Introduction

During the second half of the 19th century, borosilicate optical glass was first studied to offer systematic insight into the relationships between the composition of glass and its physical and chemical properties. After this, CaO–B₂O₃–SiO₂ (CBS) glasses have been a popular study point over the years due to their excellent properties in a number of fields, for instance, high emissivity coating materials,¹ optical glasses for smart window applications,² or as bioactive materials to prevent microfouling,³ or also as storage materials for dealing with nuclear waste⁴ and electronic glasses in low temperature co-fired ceramics (LTCC) technology.⁵

In order to satisfy the need of the technical applications in different fields, the optimization in glasses is carried out in composition and structure of the glasses to achieve the required physical, mechanical and dielectric properties. Introduction of intermediate oxides, such as ZnO,^3,6 ZrO₂ (ref. 7) and Al₂O₃,⁸ in borosilicate glass can be an effective way to change structure of glass. They can act either as glass modifier or glass former, depending upon their amount in glass composition. The oxide addition in glass system will lead to the change of NBOs, resulting in the formation of different structural units within the glass network. This will make great contribution optimize the chemical, thermal stability and various properties (physical properties and dielectric properties) of glasses. Colak et al.⁶ reported the dramatic effect (modifier or network former) of ZnO in borate glass structure due to the non-bridging oxygen atoms. The ZnO played as a network former when having more than 5% (weight%) ZnO in glasses, on the contrary, the role was modifier. Khan et al.⁷ did some research about the modification in yttrium calcium borosilicate glasses units by ZrO₂ instead of CaO. Interestingly higher content of ZrO₂ ≥7.5 mol% may be acts as glass former and below 7.5 mol%, it may act as glass modifier. Also, the variation in structure leads to the change of optical and dielectric properties. Neuville et al.⁹ found that Al/Si tetrahedral distribution in the glass network in different Qⁿ units depending on the join R=CaO/Al₂O₃. The results revealed the remarkable function of aluminum in CaO–Al₂O₃–SiO₂ (CAS) ternary glass system. Aluminoborosilicate glasses consist of structural units, such as BO₃, BO₄,¹⁰ SiO₄,¹¹ AlO₄ and AlO₆ (ref. 12) species. The transformation of network formers in aluminosilicate glasses can make the structure of CaO–B₂O₃–SiO₂ glass meet the application need.

However, so far, there is no report about utilizing FT-IR spectra, NMR, DSC and XRD to research the structure and crystallization of calcium aluminoborosilicate glasses with varying Al content. Therefore, in this paper, we present a detailed report on the structure and crystallization of CaO–B₂O₃–SiO₂ glass with Al₂O₃ content, especially the relationship between the Al content and the impacts on the units of glass network, which will provide guiding significance for further optimization of the glass structure.

2. Experimental

2.1 Preparation of glass samples

The (1 − x/100) × (35CaO–37SiO₂–28B₂O₃) + xAl₂O₃, where x = 0, 1, 2.5, 5, 7.5 and 10 mol%, were prepared using analytically pure CaCO₃, H₃BO₃, SiO₂, and Al₂O₃ as the starting materials. The stoichiometric amounts of chemicals were mixed homogeneously and transferred to Pt crucibles to melt at a temperature of 1400 °C for 2 h in air. After melting, the melt was quickly quenched into deionized water and ground, followed by screened through a 500-mesh stainless steel wire screen to obtain glass powders with particle size less than 25 μm.

2.2 Characterization of the samples

The XRD patterns of crushed powder samples were recorded within angular range 10–80° using PANalytical X'Pert PRO with Cu Kα radiation (λ = 1.54 Å). The 2θ angle scans were made at step width of 0.02° and a time per step of 0.5 s.

The density of all glasses was examined using AccuPyc®II 1340 (Micromeritics USA). The data was tested ten times to obtain an accurate density value. Fourier transform infrared spectra (FT-IR, Thermo Nicolet Smart-380) were carried out in the range 400–2000 cm⁻¹ at room temperature with a resolution of 2 cm⁻¹ on glass powders. Samples were mixed with spectroscopic grade dry KBr powder and then compressed the mixtures to obtain pellets for FT-IR measurements.

A Differential Scanning Calorimetry (DSC, STA409 PG/PC, Netzsch, Germany) was carried out to determine the glass transition temperature (T_g) and the crystallization behavior. The test was conducted in a flowing atmosphere of dry air from ambient temperature to 1000 °C with a heating rate of β = 5 °C min⁻¹. The DSC experiments were carried out using alumina crucibles with a-Al₂O₃ powder as a reference material.

¹¹B, ²⁷Al and ²⁹Si magic angle spinning nuclear magnetic resonance spectroscopy (MAS-NMR, Varian Infinity-plus 400) were performed on all glasses to determine the roles of B, Si and Al in glasses system. The definite test condition refers to ref. 13.

3. Results and discussion

3.1 The XRD analysis

Fig. 1 presented the X-ray diffraction conducted on the glasses, which showed the diffusion peaks and no crystalline phases, indicating the amorphous state of all glasses.

Fig. 1 X-ray diffraction patterns of glasses.

3.2 The FT-IR analysis

The FT-IR spectra of all glasses for x = 0, 1, 2.5, 5, 7.5 and 10 mol% between 400 and 2000 cm⁻¹ are shown in Fig. 2(a), respectively. The broad bands showed in FT-IR absorbance spectra confirm the amorphous nature (as showed in Fig. 1) and wide distribution of Qⁿ (Qⁿ notation indicates n bridging oxygen per tetrahedron) units of glasses. All experimental glass compositions showed similar FT-IR spectra and the absorption bands are broad resulting in different component peaks to form overlap easily. Therefore, it is necessary to utilize the deconvolution based on Gaussian function to identify each component peak of all bands in the FT-IR spectra to realize their assignments. Fig. 2(b)–(d) represent the deconvoluted spectra of glasses for x = 0, x = 5 and x = 10 mol% as examples of the glasses investigated, respectively. The high frequency band around 1620 cm⁻¹ is attributed to the molecular water.¹⁴ The band near 1450 cm⁻¹ and 1380 cm⁻¹ are characteristic of antisymmetric stretching vibration of [BO₃].¹⁵ The band at 1234 cm⁻¹ is due to the corresponding mode of borate triangles with non-bridging oxygen (NBO) atoms, which are mainly BO₂O⁻ triangles with two bridging and one non-bridging oxygen atom.¹⁶ It is notable that the intensity of the band at 1234 cm⁻¹ increases gradually with the Al₂O₃ content increasing, revealing the enhancement of borate triangles with NBOs proportion in glasses. The band located around 1030 cm⁻¹ assigned to the merging bands of the SiO₄ tetrahedral,⁴ and the band shifted to lower wavenumber, the wavenumber of this band location are separately 1052, 1048, 1043, 1041, 1036, 1024 cm⁻¹ from x = 0 to x = 10, respectively, indicating the addition of Al₂O₃ may change the combination mode in glasses. The band near 910 cm⁻¹ is due to B–O stretching in BO₄ units¹⁷ and its intensity becomes weaker as the increasing Al₂O₃ content. Additionally, the relative areas of the peak corresponding to SiO₄ and BO₄ units were calculated according to the assignment of the IR spectra, showed in Fig. 3(a). It reveals that the amount of SiO₄ increase while the BO₄ units reduce in glass system with the increasing x. The appeared band of 722 cm⁻¹ is attributed to stretching vibrations of Al–O bond in [AlO₄] tetrahedral¹⁸ and the bending vibration of the B–O–B bonds in the borate network;^18,19 the band located around 500 cm⁻¹ and 467 cm⁻¹ are attributed to the Al–O stretching vibrations in [AlO₆] octahedral and the bending vibration of Si–O–Si.^19,20 Their relative areas are depicted in Fig. 3(b). The results show that the bending vibration of Si–O–Si weakens and the Al–O stretching vibrations increase.

Fig. 2 (a) The FT-IR spectra of the glasses; deconvoluted FT-IR spectra of the glasses based on Gaussian function, (b) x = 0, (c) x = 5, (d) x = 10.

Fig. 3 Dependence of the relative area of SiO₄, BO₄ group (a) and Al–O, Si–O–Si (b) on Al₂O₃ content. The lines are guide for the eyes.

3.3 MAS-NMR spectroscopy

The ¹¹B MAS-NMR spectra of the glasses with different Al₂O₃ content are shown in Fig. 4(a). The spectra show relatively broad peaks located at 11 ppm and 4 ppm are corresponding to asymmetric (B^IIIa, boron with 1 or 2 bridging oxygens) and symmetric trigonal (B^IIIs, boron with 0 or 3 bridging oxygens) boron units,²¹ respectively. The units are showed in the Fig. 5(d), the ball-and-stick models for species in glasses. And another sharper peak centered 0 ppm associated with tetrahedral (BO₄, B^IV) boron species,²¹ which is similar to borosilicate mineral superstructures danburite units [B(OB)(OSi)₃].²² It is noteworthy that the intensity of [BO₃] units increases distinctly with the increasing of x, indicating the rising amount of [BO₃] units in glasses. In order to quantitatively figure out the distribution of B^III and B^IV units of boron in glasses, the ¹¹B MAS-NMR spectra were deconvoluted and the example of x = 5 was showed in Fig. 5(a), and their NMR parameters, the isotropic chemical shift (δ_iso) and the relative amount of structure units, were showed in Table 1. The δ_iso of all boron units shifts to a higher value obviously with the increasing x, i.e. the value shifts from 11.97 to 12.47 for B^IIIa, from 3.17 to 3.64 for B^IIIs and from −0.28 to −0.04 for B^IV. In addition, the relative amount of B^IIIa units increases from 26.44% to 32.7% and B^IIIs is from 44.68% to 54.98%, while the B^IV units decrease from 28.88% to 12.32% with the increasing of Al₂O₃ content (Table 1). The results are in accordance with previous outcome revealed by the band at 1234 cm⁻¹ and 928 cm⁻¹ in FT-IR.

Fig. 4 NMR spectra of (a) ¹¹B, (b) ²⁷Al and (c) ²⁹Si of experimental glasses.

Fig. 5 Deconvolution of (a) ¹¹B, (b) ²⁷Al and (c) ²⁹Si NMR spectra for glass with x = 5. (d) The ball-and-stick models for species in glasses.

Table 1 NMR parameters for ¹¹B deconvolution

	Boron site	δiso (ppm)	Amount (%)
x = 0	B^IIIa	11.97	26.44
	B^IIIs	3.17	44.68
	B^IV(1B,3Si)	−0.28	28.88
x = 5	B^IIIa	12.33	28.14
	B^IIIs	3.47	53.38
	B^IV(1B,3Si)	−0.17	18.48
x = 10	B^IIIa	12.47	32.70
	B^IIIs	3.64	54.98
	B^IV(1B,3Si)	−0.04	12.32

---

Fig. 4(b) revealed the ²⁷Al MAS-NMR spectra of the glasses with x = 1, 5 and 10. The spectra showed a narrow peak centered at 55 ppm are associated with Al^IV unit²³ and the peak is sharper with the increasing of x, which shows the enhancement of Al units in glasses. In addition, the peaks are asymmetrical and broader demonstrating the presence of higher Al coordination,²³ such as Al^V and Al^VI units. In order to quantitatively clarify the fractions and distribution of different Al coordination in glasses, the spectral deconvolution was performed using Gaussian by fitting one line shape and an example of x = 5 was presented in Fig. 5(b), and the isotropic chemical shift (δ_iso) and the relative amount of different Al coordination units were showed in Table 2. The peaks having chemical shifts around 55, 33 and 6.7 ppm can be attributed to Al^IV, Al^V and Al^VI species (as showed in Fig. 5(d)),¹³ respectively. With the increasing of x, the relative amount of Al^IV units decreases from 61.82% to 57.04% and the Al^VI is from 10.85% to 10.08%, while the Al^V increases from 27.31% to 32.87%.

Table 2 NMR parameters for ²⁷Al deconvolution

	Aluminum site	δiso (ppm)	Amount (%)
x = 1	Al^IV	57.14	61.82
	Al^V	33.27	27.31
	Al^VI	6.74	10.85
x = 5	Al^IV	57.16	61.92
	Al^V	33.27	27.44
	Al^VI	6.74	10.63
x = 10	Al^IV	55.31	57.04
	Al^V	32.67	32.87
	Al^VI	6.74	10.08

---

The ²⁹Si MAS-NMR spectra for x = 0, 5 and 10 were showed in Fig. 4(c). Deconvolution of ²⁹Si spectra was performed using Gaussian to quantitatively analyze the existence form and fractions of Si units. An example of ²⁹Si deconvolution of x = 5 is presented in Fig. 5(c) and their NMR parameters are presented in Table 3. The results showed that two peaks around −84 and −90 ppm for all glasses, which are characteristic of the Q³ and Q⁴ structural units of silicon (as showed in Fig. 5(d)), respectively. The chemical shifts of Q⁴ structural units shift from −97.01 to −87.79 ppm and Q³ has no obvious change. Also, the relative amount of Q⁴ structural units increases clearly from 5.78% to 37.4%, while the Q³ decreases from 94.21% to 62.59% for x = 0 to x = 10.

Table 3 NMR parameters for ²⁹Si deconvolution

	Silicon units	δiso (ppm)	Amount (%)
x = 0	Q^3	−84.47	94.21
	Q^4	−97.01	5.78
x = 5	Q^3	−83.67	72.36
	Q^4	−94.03	27.63
x = 10	Q^3	−83.06	62.59
	Q^4	−87.79	37.4

---

The previous structural studies (FT-IR, NMR) have shown that B occurs as B^IIIa, B^IIIs and B^IV(1B,3Si) species, Al occurs in 4, 5, and 6-fold coordination (Al^IV, Al^V and Al^VI) and Si presents in Q³ and Q⁴ units in the glasses. This shows the intermediate oxides Al₂O₃ may partly act as glass network former in glasses researched in this paper. The structural changes (Table 1) that the B^IV species resolve into symmetric BO₃ species and non-bridging oxygen can be shown as:¹³

BO₄ ↔ B^IIIs + NBO (1)

The variation occurred in B^IIIa and B^IIIs species can be written as the following reaction:¹³

B^IIIs + NBO ↔ B^IIIa (2)

Moreover, the SiO₄ units with n bridging oxygen (Qⁿ) may combine with the non-bridging oxygen to form the Qⁿ⁻¹ unit. The change can be presented as:¹³

Qⁿ + NBO ↔ Qⁿ⁻¹ (3)

The shift of Al^IV, Al^V and Al^VI species with NBO can be represented as:

Al^V ↔ Al^IV + NBO (4)

As increasing of Al₂O₃ content, the relative amount of Al^IV units decreases while the Al^V increases, and the Al^VI remains stable (Table 2). These results shows that the Al^IV units can capture more NBO to form the Al^V units, however, the Al^IV units have not enough NBO to shape the Al^VI units. The NBOs captured by Al^IV units may be originated in the dissociation of BO₄ species (eqn (1)), which was consistent with the result revealed in Table 1. At the same time, due to the competition of Al units to gain the more NBOs, the relative amount of Q⁴ structural unit increases while the Q³ decreases (Table 3). This can account for the eqn (2). Therefore, the increasing concentration of Al substitution in glasses will lead to the decrease of NBOs, which will make the amount of BO₃, Q⁴ units increase while BO₄ and Q³ units decrease.

3.4 Thermal, phase and physical properties of glasses

Fig. 6 presents the DSC plots of all the glasses measured at β = 5 °C min⁻¹. The curves show 2 crystallization peaks (T_P1 and T_P2), related to the formation of the crystalline phase of wollastonite²⁴ and borocalcite,²⁵ respectively. With the increasing of Al₂O₃ content, the T_P1 and T_P2 shift in the direction of the higher temperature and gradually disappear. Table 4 presents the values of T_g (glass transition temperature), T_P1 and T_P2. The T_g is obtained from the onset of slope change of the DSC curves. The values of T_g for all glasses don't appear to be much different considering the error factor. The first crystallization peak T_P1 successively increases with Al₂O₃ content from 805 to 945 °C, however, the second crystallization peak T_P2 increases obviously with increasing x up to x = 5 mol%, and then disappears with the further increasing x in the whole measuring temperatures.

Fig. 6 DSC curve of the studied glasses.

Table 4 Characteristic temperatures and physical properties for all glasses

	Tg (°C)	First peak TP1 (°C)	Second peak TP2 (°C)	ρ (g cm^−3)	Vm (cm^−3)	ρB units (g cm^−3)
x = 0	745	805	833	2.845 ± 0.004	26.8	3.18
x = 1	744	816	867	2.832 ± 0.004	27.0	—
x = 2.5	746	818	902	2.824 ± 0.003	27.3	—
x = 5	744	845	955	2.810 ± 0.003	27.6	2.87
x = 7.5	747	884	—	2.801 ± 0.004	27.9	—
x = 10	750	945	—	2.785 ± 0.003	28.2	2.83

---

Fig. 7(a) shows the XRD patterns of all glasses isothermally treated once for 1 h at temperature of 850 °C with x = 0, 2.5, 7.5 and 10. It could be observed that the peak of crystallization become weaker gradually with the increasing of x, and the glass with x = 10 remains the glassy state. The main diffraction peaks for x = 0, 2.5 and 7.5 can be indexed to the CaSiO₃ (JCPDS No. 10-0487), and a few diffraction peaks corresponding to CaB₂O₄ (JCPDS No. 01-0833) phase. This result conforms to the change of crystallization peaks (T_P1 and T_P2) in DSC analysis as shown in Fig. 6. The Fig. 7(b)–(d) show the XRD patterns of glasses with x = 0, 5 and 10, respectively. All samples were isothermally treated four times for 1 h at temperature of 850 °C. Fig. 7(b) revealed that the patterns of sample with x = 0 implemented repeated heat treatment were about the same. This result illustrates the sample was crystallized completely when heated at the first time. For sample with x = 5, the repeated heat treatment makes the crystallization more complete. However, for sample with x = 10, the crystallization was emerged until the third time heat treatment. This observation confirms previous findings, that chemically weak inclusions do not provide active foreign nucleation substrates.²⁶ One of the most remarkable things about these patterns is that the crystalline phases, not just the CaSiO₃ and CaB₂O₄ phase, the most is the Ca₂SiO₄ (JCPDS No. 49-1672). These results revealed that the addition of Al₂O₃ in glasses may hinder the crystallization of CBS glasses. This can be ascribed to the addition of Al₂O₃ dopants, dissolved in glass system, modifies the thermodynamic equilibrium of the system and this change is mainly entropy driven and also slowdown the kinetics of crystallization.²⁷ The addition of Al₂O₃ leads to the reconnection of fractured silicon oxygen tetrahedron and a large extent of atomic rearrangements. This will make the glass need higher thermal energies to crystallize, resulting the increasing of crystallization temperature (Fig. 6 and Table 4).

Fig. 7 X-ray diffraction patterns of crystallized glasses at temperature of 850 °C. (a) The patterns glasses isothermally treated one time for x = 0, 2.5, 5 and 10. The patterns of glasses isothermally treated four times (b) x = 0; (c) x = 5; (d) x = 10.

The density and the molar volume (V_m) of all glasses for x = 0, 1, 2.5, 5, 7.5 and 10 also showed in Table 4. The molar volume (V_m) could be calculated using the following formula:²⁸

where M and ρ are molecular weight and density of the glass, respectively. As shown in Table 4, the density (ρ) of glasses decreased linearly form 2.845 to 2.786 g cm⁻³ with increasing Al₂O₃ content, while the molar volume (V_m) of the glasses increases linearly. The physical properties of borosilicate glasses depend on the structural units in glasses, which are closely bound up with BO₃, BO₄, Qⁿ units and Al species. Based on the deconvolution results in FT-IR, the relative areas of the peak corresponding to 722 cm⁻¹ in sample with x = 0 were 3.53%, while the relative areas of the peaks corresponding to 722 cm⁻¹ and 500 cm⁻¹ in sample with x = 10 were 4.94% and 3.27%, respectively. Therefore, the relative areas of Al–O in sample with x = 10 were approximately 4.5%, and the relative areas of B–O and Si–O were 48% and 46%, respectively. So, the Al species make up only a small percentage in glass system. In addition, the Al species changed a little (showed in Table 2), both of which cause a small contribution to the change of density. The change (Q⁴ units increase Q³ units decrease) occurred in Qⁿ units will make the density increase. The impacts of B units on the density of glasses can be calculated according to eqn (6):where w and ρ are the percentage and density of B^IIIa, B^IIIs and B^IV units, respectively. The value of w₁, w₂ and w₃ was showed in Table 1. The density of the symmetric BO₃ unit is ∼1.87 g cm⁻³, for the BO₄ tetrahedron ∼6.77 g cm⁻³ and for the asymmetric BO₃ unit ∼8.10 g cm⁻³.²⁹ The densities of B units for samples with x = 0, 5 and 10 were calculated and was listed in Table 4 based on the eqn (6). It can be seen that the ρ_{B units} decreases with the increasing of x value, and its influence outweighs the Qⁿ units. These factors make the density of glasses decrease with the increasing Al₂O₃ content.

4. Conclusions

The current study using NMR, FT-IR, DSC and XRD investigated the role of Al on structure and crystallization of CBS glass when added at a small concentration. The Al partly plays the network former role in CBS glasses and the Al goes into the glass network in 4-fold, 5-fold and 6-fold coordination whereas B goes in as both 4- and 3-fold coordination. The Si occurred in the form Q³ and Q⁴ units. With the increasing of Al₂O₃ content, the amount of BO₃, Q⁴ units increase while BO₄, Q³ units decrease. This is due to the NBOs competition occurred in Al, B and Si units. The density (ρ) of glasses decreased linearly with increasing Al₂O₃ content, while the molar volume (V_m) of the glasses increases linearly. The glass transition temperatures for all glasses don't appear to be much different, while the crystallization peaks of glasses containing Al shift to higher temperatures and disappear. At the same time, the peaks of crystallization become weaker gradually with the increasing of x. These results indicate that Al can hinder the crystallization of CBS glasses.

Acknowledgements

The authors thank the National Center for Magnetic Resonance in Wuhan acquiring the MAS-NMR measurement. This work was supported by the fund of the Basic Applied Research Foundation of Yunnan Province, China (Grant No. 2016FD125, 2016FB083), the 551 project of Kunming and Science &Technology Program of Yunnan Province (No. 2014DC019).

References

1. G. Shao, X. Wu, Y. Kong, X. Shen, S. Cui, X. Guan, C. Jiao and J. Jiao, J. Alloys Compd., 2016, 663, 360–370.
2. M. Y. Hassaan, H. M. Osman, H. H. Hassan, A. S. El-Deeb and M. A. Helal, Ceram. Int., 2017, 43, 1795–1801.
3. M. Klinger-Strobel, O. Makarewicz, M. W. Pletz, A. Stallmach and C. Lautenschlager, J. Mater. Sci.: Mater. Med., 2016, 27, 175.
4. M. Sitarz, J. Non-Cryst. Solids, 2011, 357, 1603–1608.
5. M. Ma, Z. Liu, F. Zhang, F. Liu, Y. Li and R. Bordia, J. Am. Ceram. Soc., 2016, 99, 2402–2407.
6. S. Cetinkaya Colak, I. Akyuz and F. Atay, J. Non-Cryst. Solids, 2016, 432, 406–412.
7. S. Khan, G. Kaur and K. Singh, Ceram. Int., 2017, 43, 722–727.
8. J.-Z. Liu, X.-F. Wu, N.-X. Xu, Q.-L. Zhang and H. Yang, J. Mater. Sci.: Mater. Electron., 2015, 26, 8899–8903.
9. D. R. Neuville, L. Cormier and D. Massiot, Chem. Geol., 2006, 229, 173–185.
10. S. Stefanovsky, K. Fox and J. Marra, MRS Proceedings, 2013, 1518, 53–58.
11. A. K. Yadav and P. Singh, RSC Adv., 2015, 5, 67583–67609.
12. A. Saini, A. Khanna, V. K. Michaelis, S. Kroeker, F. González and D. Hernández, J. Non-Cryst. Solids, 2009, 355, 2323–2332.
13. S. Sen, Z. Xu and J. Stebbins, J. Non-Cryst. Solids, 1998, 226, 29–40.
14. K. Singh, I. Bala and V. Kumar, Ceram. Int., 2009, 35, 3401–3406.
15. J. Wan, J. Cheng and P. Lu, J. Wuhan Univ. Technol., Mater. Sci. Ed., 2008, 23, 419–421.
16. A. Winterstein-Beckmann, D. Möncke, D. Palles, E. I. Kamitsos and L. Wondraczek, J. Non-Cryst. Solids, 2014, 405, 196–206.
17. K. Rao, Structural chemistry of glasses, Elsevier, 2002.
18. N. Santha, T. Nideep and S. Rejisha, J. Mater. Sci.: Mater. Electron., 2012, 23, 1435–1441.
19. G. J. Mohini, N. Krishnamacharyulu, G. Sahaya Baskaran, P. V. Rao and N. Veeraiah, Appl. Surf. Sci., 2013, 287, 46–53.
20. H. Shao, H. Q. Zhou and X. D. Shen, Adv. Mater. Res., 2011, 189–193, 4466–4471.
21. A. Gaddam, H. R. Fernandes and J. M. F. Ferreira, RSC Adv., 2015, 5, 41066–41078.
22. B. G. Parkinson, D. Holland, M. E. Smith, A. P. Howes and C. R. Scales, J. Phys.: Condens. Matter, 2007, 19, 415114.
23. S. H. Risbud, R. J. Kirkpatrick, A. P. Taglialavore and B. Montez, J. Am. Ceram. Soc., 1987, 70, C-10–C-12.
24. J. H. Jean, C. R. Chang and C. D. Lei, J. Am. Ceram. Soc., 2004, 87, 1244–1249.
25. C. R. Chang and J. H. Jean, J. Am. Ceram. Soc., 1999, 82, 1725–1732.
26. R. Müller, R. Meszaros, B. Peplinski, S. Reinsch, M. Eberstein, W. A. Schiller and J. Deubener, J. Am. Ceram. Soc., 2009, 92, 1703–1708.
27. A. Gaddam, H. R. Fernandes, M. J. Pascual, J. M. F. Ferreira and L. Pinckney, J. Am. Ceram. Soc., 2016, 99, 833–840.
28. A. Gaddam, H. R. Fernandes and J. M. Ferreira, RSC Adv., 2015, 5, 41066–41078.
29. H. Doweidar, K. El-Igili and S. A. El-Maksoud, J. Phys. D: Appl. Phys., 2000, 33, 2532.

---