Effects of nitrogen on phase formation, microstructure and mechanical properties of Y–Ca–Si–Al–O–N oxynitride glass–ceramics

Journal of Non-Crystalline Solids 368 (2013) 79–85

Contents lists available at SciVerse ScienceDirect

Journal of Non-Crystalline Solids journal homepage: www.elsevier.com/ locate/ jnoncrysol

Effects of nitrogen on phase formation, microstructure and mechanical properties of Y–Ca–Si–Al–O–N oxynitride glass–ceramics Zhiwei Luo, Anxian Lu ⁎, Xiaolin Hu, Weizhen Liu School of Materials Science and Engineering, Central South University, Changsha 410083, PR China

a r t i c l e

i n f o

Article history: Received 9 January 2013 Received in revised form 25 February 2013 Available online 3 April 2013 Keywords: Oxynitride glass; Glass–ceramics; YAG; SEM; Mechanical properties

a b s t r a c t The effect of nitrogen substitution on the crystallisation of an oxynitride glass in the Y–Ca–Si–Al–O–N system has been studied. The appropriate heat treatment temperatures were selected according to the information provided by the differential scanning calorimeter (DSC) measurement. There is a significant increase in Tg and Tc with increasing nitrogen content. X-ray diffraction (XRD) and scanning electron microscope (SEM) analysis demonstrated that, for the oxide glass and oxynitride glasses containing 6 equiv.% and 12 equiv.% nitrogen, crystallisation results in the formation of irregular lath-shaped Ca4Y6O(SiO4)6 and stick-shaped anorthite. As the nitrogen content increases to 18 or 24 equiv.%, irregular plate-like yttrium-aluminium garnet (YAG) is identified as the main crystalline phase. As the nitrogen content increases to 30 equiv.%, microscopic needle-like crystals of Al6O3N4 become the only crystallised phase. The best composition, owing to the mechanical properties (e.g. flexural strength of 162 MPa and Vickers hardness of 8.5 GPa), was found to correspond to a nitrogen content of 24 equiv.%. © 2013 Elsevier B.V. All rights reserved.

1. Introduction Silicon oxynitride glasses were first discovered as grain boundary phases in silicon nitride based ceramics [1]. Silicon nitride ceramics are usually densified with the aid of various sintering additives (such as MgO, Y2O3, etc.). These additives in combination with silicon dioxide, which originates from the furnace atmosphere, or from the dissolution of silicon nitride particles in the melt, cause the formation of a liquid phase that transforms into a glass upon cooling [2]. Since it is very difficult to investigate experimentally the properties of grain boundary phases, research on them in bulk form was initiated [3]. Oxynitride silicate glasses are a branch of high performance glasses, obtained by incorporating nitrogen atoms into silicate or alumino-silicate glasses [4,5]. The unique properties of silicon oxynitride and SiAlON glasses have led to a search for areas of potential application [6–8]. The improvement in physical and mechanical properties when nitrogen is incorporated into silicate or alumino-silicate glasses is also realised in glass–ceramics. As with other silicate or alumino-silicate glasses, oxynitride glasses may be heat treated at the appropriate temperatures to crystallise as glass–ceramics and many studies of crystallisation of these types of glasses have been reported [9–14]. The crystallization improves the mechanical and thermal properties. The specific crystalline phases formed upon heat treatment, and the extent of their formation, determine the properties of the material [15–18]. In general, nitrogen increases the stability of oxynitride glasses. The ⁎ Corresponding author. Tel.: +86 731 88830351; fax: +86 731 88877057. E-mail addresses: [email protected], [email protected] (A. Lu). 0022-3093/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jnoncrysol.2013.03.005

conventional process to produce a glass–ceramic involves two steps: a lower temperature heat treatment of glasses, generally just above the glass transition temperature, to induce nucleation, followed by heating to a second higher temperature, the so-called crystallisation temperature, to allow growth of the pre-formed nuclei. The crystalline phases formed depend on both the composition of the parent glass and the heat-treatment process but in many cases, oxide phases form first leaving the residual glass more N-rich [19]. The present work investigates the influence of the nitrogen addition on phase formation, microstructure and mechanical properties of Y–Ca–Si–Al–O–N glass. XRD and SEM were used to study the crystallization behaviour of the glass–ceramics. An understanding of the roles played by nitrogen atom in this system is clearly vital to a coherent approach to improved material performance. 2. Experimental procedure 2.1. Preparation of materials A base Y–Ca–Al–Si–O oxide glass was prepared with a cation composition (in equiv.%) of 12Y:12Ca:61Si:15Al and 100 equiv.% oxygen. 6, 12, 18, 24 and 30 equiv.% N was substituted for oxygen in order to evaluate the effects of nitrogen on crystallization of oxide glass or oxynitride glasses. Y2O3 (99.99%), α-Si3N4 (99.8%), SiO2 (99.9%), CaCO3 (99.9%), and Al2O3 (99.9%) were used as raw material powders to prepare Y–Ca–Si–Al–O–N glasses. Dried powders were weighed and performed by mechanical agitation (using an attritor mill), mixed in isopropyl alcohol for 24 h, and then dried again. The mixture was

80

Z. Luo et al. / Journal of Non-Crystalline Solids 368 (2013) 79–85

melted at 1500–1650 °C for 2 h in a graphite crucible lined with BN powder to avoid any sealing between the glass and the crucible. Syntheses were performed in an atmosphere of an inert gas (N2 or N2 + Ar gas). Melts were finally poured into a preheated graphite mould and annealed in air at a temperature close to an estimated vitreous transition temperature. After the determination of the effective Tg by DSC analysis, glasses were annealed at Tg for 2 h and slowly cooled to room temperature. Heat treatments of the glass samples to prepare glass–ceramics were carried out in a silicon carbide horizontal furnace in a nitrogen atmosphere after DSC measurement.

2.2.6. Three-point bending strength Three-point bending strength was measured on a glass bar (4 × mm × 4 mm × 25 mm with accuracy ± 0.02 mm) by using an electronic multipurpose tester (CSS-44100 model, China). The crosshead speed was 0.5 mm/min. The glass–ceramic surfaces were polished to a 1 μm diamond finish. The measurement error for the hardness and the bending strength is within ±2%.

2.2. Characterization techniques

3.1. Parent glasses

2.2.1. Differential scanning calorimeter (DSC) For DSC analysis, about a 10 mg powder sample was placed in an alumina crucible and subjected to a heating rate of 10 °C/min from ambient temperature to 1400 °C in a flowing high purity argon environment. The onset point of an endothermic drift on the DSC curve corresponding to the beginning of the glass transition range is reported as Tg whilst the peak of the exotherm is taken as Tc. Errors in measurement are ±5 °C.

The appearances of the obtained parent glasses were primarily inspected by the naked eye. The prepared oxynitride glass samples are translucent and grey by the naked eye except for samples N0 and N30. Actually, the sample N0 is a high-quality clear colourless glass (with nitrogen-free), whilst the sample N30 is translucent dark. With the influence of nitrogen, the colour of the glasses is observed in the transparent colourless region and shifted to a darker grey region when nitrogen increases. Oxynitride glasses are as a rule less transparent than corresponding oxide glasses. Measured densities for the determined glass compositions are given in Table 1. The density of the Y–Ca–Si–Al–O–N glasses, as seen in Table 1, is found to have values ranging from 3.02 to 3.24 g/cm 3 and increases slightly with increasing nitrogen content. Addition of nitrogen leads to an increase in density for the silicon oxynitride glasses, which can be attributed to the formation of a significant fraction of three-coordinated N in the glass networks. To study the crystallization behaviour, the parent glasses were subjected to DSC measurements. The results of DSC curves are shown in Fig. 1. The DSC curve of the glass sample N0 shows welldefined glass transition ranges and two obvious exothermic peaks corresponding to crystallizations of two possible crystalline phases. However, the other glass samples only show one exothermic peak. And it is easy to see that the exothermic peaks of glass samples N6, N12 and N18 are clear and distinct but that of glass samples N24 and N30 become weak or absent. The glass transition temperatures range from 834 ± 5 °C to 941 ± 5 °C and the locations of main exothermic peaks are in the range 1035 ± 5 to 1166 ± 5 °C. It is clearly seen that the crystallization peaks (Tc) are shifted towards higher temperatures with increasing of nitrogen contents and the extracted Tc values are summarised in Table 1. Fig. 1 shows that there is a significant increase in Tg and Tc with increasing nitrogen content. Since an increase in nitrogen content by all accounts increases the crosslinking of the glass network, the Tg and Tc are expected to increase as the nitrogen content of the glass increases. The effect increases with increasing the amount of nitrogen and this may be related to the increase in viscosity of the glass which consequently leads to less mobility of the structural elements in the glass. Glasses were heat-treated in a resistance furnace under flowing nitrogen with a heating rate of 20 °C/min to near the nucleation temperature, with a hold for 5 h followed by heating at 10 °C/min to >Tc temperature with a hold for 10 h followed by removal from the

2.2.2. Measurement of density (ρ) The true density (ρ) of the oxynitride glass samples was evaluated by means of a gas pycnometer, operating with He gas on samples in powdered form. The results were averaged on 10 measurements. The experimental error in the measured value was ±0.01 g/cm 3. 2.2.3. X-ray powder diffraction (XRD) Crystalline phases in the prepared glass–ceramics were identified by an X-ray diffractometer (D/max 2500 model, Japan) with Cu-Kα radiation using an incident wavelength of 0.15406 nm (Cu-Kα). Voltage and current were selected as 40 kV and 50 mA, respectively. Date were collected from 2θ = 10° to 90° at a scanning rate of 8°/min. To identify the crystalline phases, XRD patterns were analysed by jade 6.0 software using ICDD-PDF2 database. 2.2.4. Scanning electron microscopy (SEM) SEM and environmental SEM imaging were carried out using an FEI Quanta 200 SEM, which was fitted with a Peltier-cooled stage during SEM operation. Energy dispersive X-ray analysis (EDX) was used to obtain a chemical analysis of particular crystals or regions. Before the SEM observations, these glass–ceramic samples were mounted in epoxy resin and polished to 1 μm with diamond slurries. The mirror surfaces of specimens were sputtered with a platinum coating. 2.2.5. Vickers micro-hardness Vickers hardness was measured using a micro-hardness tester (DHV-1000-CCD, China) with a pyramid shaped diamond indenter. The glass–ceramic surfaces were polished to a 1 μm diamond finish. A load of 2 kg was applied on polished glass–ceramic samples. At least 10 indentations were made on each sample, using the average indentation diagonal length to calculate hardness values in GPa.

3. Results and discussion

Table 1 Composition (in equiv.%), melting temperatures, colour, thermal properties, and density of the Y–Ca–Si–Al–O–N glasses. Samples no.

N0 N6 N12 N18 N24 N30

Composition (in equiv.%) Y

Ca

Si

Al

O

N

12

12

61

15

100 94 88 82 76 70

0 6 12 18 24 30

Melting temperatures

Colour

Tg (±5 °C)

Tc1 (±5 °C)

Tc2 (±5 °C)

Density (±0.01 g/cm3)

1500 1550 1580 1600 1600 1650

Transparent colourless Translucent and gray Translucent and gray Translucent and gray Translucent and gray Translucent and dark

834 867 874 890 910 941

1035 1052 1090 1117 1125 1166

1136

3.02 3.04 3.07 3.13 3.16 3.24

Z. Luo et al. / Journal of Non-Crystalline Solids 368 (2013) 79–85

Fig. 1. DSC recording for oxynitride glass samples N0, N6, N12, N18, N24 and N30.

furnace (seen in Table 2). Following heat-treatment of the glasses, the glass–ceramics produced were analysed by XRD and SEM. 3.2. Phase formation by XRD analysis Figs. 2, 3 and 4 show the X-ray diffraction patterns of the glasses after heat treatment at different temperatures. Phase developments in the samples with different nitrogen contents are different. The base oxide glass (sample N0) crystallised to form the crystal of calcium yttrium oxide silicate (Ca4Y6O(SiO4)6, PDF# 27-0093) and anorthite (CaAl2Si2O8, PDF# 89-1462) along with a small amount of residual glass. Addition of 6 or 12 equiv.% N to glass results in similar crystalline phases obtained compared with the oxide glass. As the nitrogen content is 0 or 6 equiv.%, Ca4Y6O(SiO4)6 is identified as the main crystalline phase; whilst the nitrogen content increases to 12 equiv.%, anorthite becomes the main crystalline phase, at the same time, the crystallinity of glass–ceramic sample N12 decreases. It has to be noted that, when the nitrogen content is ≤ 12 equiv.%, participation of nitrogen in the glass compositions does not change the phase types but it induces a change of the relative intensity ratio of these two phases. As is shown in Table 2, when the nitrogen content increases to 18 and 24 equiv.%, the phase formation of the oxynitride glass–ceramics

Table 2 Heat treatment condition (°C, h), crystalline phases, bending strength (±4 MPa) and Vickers hardness (±0.2 GPa) of Y–Ca–Si–Al–O–N glass–ceramics as a function of nitrogen content. Samples Heat treatment condition no.

Hardness Crystalline phases Bending strength (±0.2 GPa) (±4 MPa)

N0

900 °C, 4 h + 1050 °C, 8 h

N6

900 °C, 4 h + 1100 °C, 8 h

N12

925 °C, 4 h + 1150 °C, 8 h

Ca4Y6O(SiO4)6, Anorthite Ca4Y6O(SiO4)6, Anorthite Anorthite, Ca4Y6O(SiO4)6

N18

900 °C, 4 h + 1050 °C, 8 h 900 °C, 4 h + 1150 °C, 8 h 950 °C, 4 h + 1250 °C, 8 h

N24

N30

900 °C, 4 h + 1050 °C, 8 h 900 °C, 4 h + 1150 °C, 8 h 950 °C, 4 h + 1300 °C, 8 h

86

6.4

113

7.3

105

7.8

Y3Al5O12, Anorthite

148

8.1

Y3Al5O12, Anorthite

162

8.5

129

9.2

900 °C, 4 h + 1050 °C, 8 h 900 °C, 4 h + 1200 °C, 8 h 1000 °C, 4 h + 1350 °C, 8 h Al6O3N4

81

Fig. 2. XRD patterns of glass–ceramic samples N0, N6 and N12 obtained after respective heat-treatment conditions followed in Table 2.

took place at higher crystallization temperatures. In these two glass– ceramic samples, yttrium–aluminium garnet (Y3Al5O12-YAG, PDF# 71-0255) is identified as the main crystalline phase whilst anorthite (CaAl2Si2O8, PDF# 89-1473) becomes the secondary crystalline phase. As the nitrogen content increases to 30 equiv.%, a small quantity of aluminium oxide nitride (Al6O3N4, PDF# 48-1579) becomes the only crystal phase. These may be indicated that the crystallisation rate of the Y–Ca–Al–Si–O–N glass was suppressed by the involvement of nitrogen. It should be emphasised that some residual glass phase is also present in different proportions depending on the composition. In general, the crystallization of Y–Ca–Al–Si–O–N glass is strongly influenced by nitrogen content. With progressive nitrogen additions, significant restraint on the growth of the Ca4Y6O(SiO4)6 crystals occurs because of the increase of melt viscosity. The presence of Ca4Y6O(SiO4)6 as the major crystalline phase was also confirmed by other crystallisation studies of CaO–SiO2–Al2O3–Y2O3 glasses [2]. P. Lichvar et al. reported that the major phases identified in crystallised glasses containing 20 mol.% CaO and 6.66 mol.% Y2O3 were various crystallographic modifications of Ca4Y6O(SiO4)6, Y2Si2O7 and anorthite [2]. But in the present investigation, no Y2Si2O7 was detected. The major phases identified in crystallised glasses containing 18 or 24 equiv.% nitrogen were Y3Al5O12 (YAG) and anorthite, and no Ca4Y6O(SiO4)6 was detected. When the nitrogen content increases to 30 equiv.%, Al6O3N4 becomes the only crystallised phase. The XRD results discussed above show that crystallisation of Y–Ca– Si–Al–O–N glasses is a complex process, which is strongly influenced not only by the content of nitrogen, but also by the temperature and

Fig. 3. XRD patterns of glass–ceramic samples N18 and N24 obtained after respective heat-treatment conditions: (a) N18, 950 °C, 4 h + 1250 °C, 8 h; (b) N24, 950 °C, 4 h + 1300 °C, 8 h.

82

Z. Luo et al. / Journal of Non-Crystalline Solids 368 (2013) 79–85

time of heat treatment. In our study, the phases which crystallised, as the nitrogen content of the glasses was varied, were all known silicates or aluminosilicates suggesting that nitrogen may have substituted for oxygen in one or more of the crystalline phases or it may have remained in the residual glass. It can be concluded that, when the nitrogen content increases in the parent oxynitride glass, a nitrogen rich glassy phase also forms as well as YAG. 3.3. Microstructure by SEM and EDS

Fig. 4. XRD patterns of glass–ceramic sample N30 obtained after heat-treated: 1000 °C, 4 h + 1350 °C, 8 h.

The microstructures of the glasses after heat treatment at various temperatures were investigated. SEM photographs of the polished surfaces of the glass–ceramic samples are presented in Fig. 5. Agglomeration of crystals in the glass matrix is clearly observed. The microstructure of the oxide glass–ceramic of composition (in equiv.%) 24Y:24Ca:61Si: 15Al:100O (sample N0) is shown in Fig. 5(A) and the glass–ceramic sample crystallises to form small crystals of Ca4Y6O(SiO4)6 with length of b 1 μm (white regions) and a minor amount of anorthite (black

Fig. 5. SEM micrographs of all the glass–ceramic samples obtained after respective heat-treatment conditions followed in Table 2: (A) N0, 900 °C, 4 h + 1050 °C, 8 h; (B) N6, 900 °C, 4 h + 1100 °C, 8 h; (C) N12, 925 °C, 4 h + 1150 °C, 8 h; (D) N18, 950 °C, 4 h + 1250 °C, 8 h; (E) N24, 950 °C, 4 h + 1300 °C, 8 h; (F) N30, 1000 °C, 4 h + 1350 °C, 8 h.

Z. Luo et al. / Journal of Non-Crystalline Solids 368 (2013) 79–85

regions). It is difficult to distinguish the anorthite from the Ca4Y6O(SiO4)6 in this micrograph because of the too small size of the two crystals. Addition of 6 equiv.% N does not change the crystallisation products but the lath-shaped Ca4Y6O(SiO4)6 crystals have grown as a result of nitrogen addition to 5–10 μm in length, whilst the cross-linking columnar Anorthite crystals have grown to ≥20 μm in length and with diameters of b2–3 μm, and the columnar structure can be observed as shown in Fig. 5(B). From the energy dispersive X-ray spectroscopy seen in Fig. 6(A), it was found that the white irregular crystals correspond to Ca4Y6O(SiO4)6, whilst the darker stick-shaped crystals are due to anorthite. At 12 equiv.% N, the amount of Ca4Y6O(SiO4)6 decreases slightly but it still retains its lath-shaped appearance with crystals 5–10 μm in

83

length as can be seen (white regions) in Fig. 5(C). The cross-linking columnar anorthite crystals have still retained its columnar appearance of length ≥ 20 μm and diameters of b 2–3 μm. Thus, a similar microstructure is evident in the glass–ceramic sample containing 6 equiv.% N compared with 12 equiv.% N. The addition of 6 or 12 equiv.% nitrogen does not change significantly the phase assemblage or amounts of crystalline phases formed but the morphology of the thin lath-shaped Ca4Y6O(SiO4)6 crystals and the cross-linking columnar anorthite crystals grows to form bigger crystalline grains. This may also be due to the higher heat treatment temperature. Further nitrogen incorporation to 18 equiv.% leads to the disappearance of Ca4Y6O(SiO4)6 and reductions of the amounts of anorthite,

Fig. 6. Quantitative EDS scans of the selected area in glass–ceramic samples marked by circles and arrows: (A) N6; (B) N24.

84

Z. Luo et al. / Journal of Non-Crystalline Solids 368 (2013) 79–85

which provides sufficient Y and Al for crystallisation of YAG as shown by the white crystals in Fig. 5(D) and this is the dominant phase at high nitrogen concentrations. From the energy dispersive X-ray spectroscopy seen in Fig. 6(B), it was found that the white irregular plate-like crystals correspond to Y3Al5O12-YAG, whilst the darker stick-shaped crystals are due to anorthite. A similar microstructure is evident in the glass–ceramic containing 18 equiv.% N compared to 24 equiv.% N (seen in Fig. 5(E)). The light crystals are YAG and the darker black are anorthite and residual glass is evident as the grey phase in which some of the N is concentrated. At 30 equiv.% N, the homogeneous distribution of needle-like Al6O3N4 crystals becomes the only crystallised phase. The microstructure is shown in Fig. 6(F). These results suggested that the presence of nitrogen has influence on the crystallization of the Y–Ca–Si–Al–O–N glasses. It may also be consistent with that observed by XRD results. Effect of nitrogen on the crystallization behaviour of the crystal phase which leads to a change in its morphology has been described by the view of composition adjustment in the separated matrix phases of the glassy samples. Based on the microstructures of the oxynitride glass– ceramics, it can be concluded that progressive additions of nitrogen lead to significantly different morphologies and crystal sizes for Ca4Y6O(SiO4)6, anorthite and YAG phases and lead to different amounts of residual glass phase. When a higher N was incorporated, it is very difficult to induce oxynitride glass to crystallization. The SEM micrographs of all the glass–ceramic samples support the XRD results. When the nitrogen content is lower than 12 equiv.%, crystallisation results in the formation of irregular lath-shaped Ca4Y6O(SiO4)6 and stick-shaped anorthite. As it increases to 18 and 24 equiv.%, irregular plate-like yttrium–aluminium garnet (YAG) is identified as the main crystalline phase. As the nitrogen content increases to 30 equiv.%, microscopic needle-like crystals of Al6O3N4 become the only crystal phase. This difference could be related to the high viscosity of the oxynitride glass, due to its higher nitrogen content. 3.4. Bending strength and Vickers hardness Fig. 7 and Table 2 report the bending strength and Vickers hardness of all the glass–ceramic samples obtained after respective heat treatment. The obtained glass–ceramics from Y–Ca–Si–Al–O–N glasses exhibited good mechanical properties similar to that of previous oxynitride glass–ceramics [9,19,20], which in turn possess largely better properties than oxide glass–ceramics, as shown in Table 2. The highest value of the bending strength at 162 ± 4 MPa is observed in the

glass–ceramic containing 24 equiv.% N, whilst the highest value of the micro-hardness at ~9.2 ± 0.2 GPa is observed in the glass–ceramic containing 30 equiv.% N. Generally, Vickers micro-hardness of the glass–ceramics increases with N content. However, similar trends are not observed for the bending strength. Indentation micro-hardness measured for the obtained glass– ceramic materials was found to lie in the range 6.4 to 9.2 ± 0.2GPa which was carried out on glass–ceramic samples N0–N30 treated at different temperature. The Vickers hardness values increased as the content of nitrogen content increased, as Vickers hardness is one of the important properties of the oxynitride glass–ceramic, which is related to the amounts of the nitrogen of the residual glass phase. The hardness values of the Y–Ca–Al–Si–O–N glass–ceramics are similar to the results reported by other authors [19]. From Table 2, we can see that although the crystal content of the glass–ceramic sample N30 is lower along with increasing N content, the glass–ceramics obtained better hardness. According to Ref. [21,22], it is reported that finegrained glass–ceramics possess better hardness. However, in the present work, the results show that there is no similar tendency in the change of micro-hardness. Perhaps the main reason is that, however, oxynitride glass–ceramics process the relative lower crystallinity than corresponding silicate or alumino-silicate glass–ceramics. The bending strength of a composite may be a function of the bending strength of its different crystal components, their proportions and their orientations. The factors regulating mechanical properties are crystal type, crystallization rate and homogeneity of crystal size. As discussed above, the main crystalline phases of Y–Ca–Al–Si–O–N glass–ceramics with different N contents are different. The finegrained glass–ceramics possess better bending strength values together with finer crystal size. The possible reason for the higher bending strength values in glass–ceramics N18 and N24 is the homogeneity of crystal size and relatively higher amount of YAG crystalline phase. For glass–ceramic sample N30, though it possesses relative higher nitrogen content, the bending strength is still relatively lower. From XRD patterns (seen in Fig. 4), it can be seen that there is an observable glass phase remaining in the parent glass in sample N30. The remaining glass phase reduces continuity of the crystal phase and decreases the bending strength of the glass–ceramic sample. Various parameters such as the morphology, the size of crystalline particles, and their amounts control the mechanical strength of a particular glass–ceramic. The results show that there is no obvious tendency in the change of mechanical properties with chemical composition. Thus, the bending strength values of these glass–ceramics are shown to be more sensitive to changes in microstructure, phase morphology and crystal size, whilst Vickers hardness is related to the amounts of nitrogen in residual glass phase. 4. Conclusion

Fig. 7. The bending strength and Vickers hardness of all the glass–ceramic samples obtained after respective heat-treatment conditions followed in Table 2. Dotted lines are drawn as the guides to the eyes.

The effects of nitrogen on crystallisation of Y–Ca–Si–Al–O–N glasses have been studied. Gradual addition of nitrogen enhances the Tg and Tc of the oxynitride glasses and influences the crystallisation products and microstructures, including crystal size and morphology. Y–Ca–Si–Al–O– N oxynitride glass–ceramics containing different crystals were successfully prepared through the heat treatment process. For the oxide glass, crystallisation results in formation of calcium yttrium oxide silicate (Ca4Y6O(SiO4)6) and anorthite (CaAl2Si2O8) along with a small amount of residual glass. Addition of 6 or 12 equiv.% N results in similar crystalline phases obtained compared to the oxide glass but the morphology of the thin lath-shaped Ca4Y6O(SiO4)6 crystals and the cross-linking columnar anorthite crystals grows to form bigger crystalline grains. In oxynitride glasses containing 24 equiv.% N, irregular plate-like Y3Al5O12 (YAG) is the dominant crystalline phase, which possesses the highest bending strength and a relatively higher micro-hardness value. As the nitrogen content increase to 30 equiv.%, a small quantity of Al6O3N4 becomes the only crystal phase. The bending strength values of these

Z. Luo et al. / Journal of Non-Crystalline Solids 368 (2013) 79–85

glass–ceramics are shown to be more sensitive to changes in microstructure, phase morphology and crystal size, however, Vickers hardness is related to the amounts of nitrogen in residual glass. The best composition, owing to the mechanical properties (e.g. flexural strength of 162 MPa and Vickers hardness of 8.5 GPa), was found to correspond to a nitrogen content of 24 equiv.%. Acknowledgment The research was financially supported by the National Natural Science Foundation of China (No. 51272288) and the Graduate Scientific Research Innovation Project of Hunan Province, in China (No. CX2011B108). References [1] K.H. Jack, Review: SiAlONs and related nitrogen ceramics, J. Mater. Sci. 11 (1976) 1135–1158. [2] P. Lichvár, P. Šajgalík, M. Liška, D. Galusek, CaO–SiO2–Al2O3–Y2O3 glasses as model grain boundary phases for Si3N4 ceramics, J. Eur. Ceram. Soc. 27 (2007) 429–436. [3] R.A.L. Drew, S. Hampshire, K.H. Jack, Nitrogen glasses, Proc. Br. Ceram. Soc. 31 (1981) 119–132. [4] S. Hampshire, M.J. Pomeroy, Oxynitride glasses, Int. J. Appl. Ceram. Tecnol. 5 (2008) 155–163. [5] S. Hampshire, Oxynitride glasses, J. Eur. Ceram. Soc. 28 (2008) 1475–1483. [6] D. Criado, M.I. Alayo, M.C.A. Fantini, I. Pereyra, Study of the mechanical and structural properties of silicon oxynitride films for optical applications, J. Non-Cryst. Solids 352 (2006) 2319–2323. [7] S. Sakka, Structure, properties and application of oxynitride glasses, J. Non-Cryst. Solids 181 (1995) 215–224. [8] S. Hampshire, Oxynitride glasses: their properties and crystallization—a review, J. Non-Cryst. Solids 316 (2003) 64–73. [9] Z. Luo, A. Lu, G. Qu, Y. Lei, Synthesis, crystallization behavior, microstructure and mechanical properties of oxynitride glass–ceramics with fluorine addition, J. Non-Cryst. Solids 362 (2013) 207–215.

85

[10] J.L. Besson, D. Billieres, T. Rouxel, P. Goursat, R. Flynn, S. Hampshire, Crystallization and properties of a Si–Y–Al–O–N glass ceramic, J. Am. Ceram. Soc. 76 (1993) 2103–2105. [11] R. Ramesh, E. Nestor, M.J. Pomeroy, S. Hampshire, K. Liddell, D.P. Thompson, Potential of NdSiAlON glasses for crystallisation to glass–ceramics, J. Non-Cryst. Solids 196 (1996) 320–325. [12] X. Li, A. Lu, Crystallization and microstructures of Y–Si–Al–O–N glass–ceramics containing main crystal phase Y3Al5O12, Bull. Mater. Sci. 34 (2011) 767–774. [13] Y. Menke, L.K.L. Falk, S. Hampshire, The crystallisation of Er–Si–Al–O–N B-phase glass–ceramics, J. Mater. Sci. 40 (2005) 6499–6512. [14] P. Vomacka, R. Ramesh, S. Hampshire, Influence of zirconia addition on the crystallization kinetics of a Y–S–Al–O–N glass, J. Eur. Ceram. Soc. 16 (1996) 1253–1262. [15] S. Mollazadeh, B. Eftekhari Yekta, J. Javadpour, A. Yusefi, T.S. Jafarzadeh, The role of TiO2, ZrO2, BaO and SiO2 on the mechanical properties and crystallization behavior of fluorapatite–mullite glass–ceramics, J. Non-Cryst. Solids 361 (2013) 70–77. [16] S.A.M. Abdel-Hameed, F.H. Margha, Preparation, crystallization behavior and magnetic properties of nanoparticles magnetic glass–ceramics in the systems Fe2O3–CoO–MnO2, Fe2O3–NiO–MoO3 and Fe2O3–CoO–V2O5, J. Non-Cryst. Solids 358 (2012) 832–838. [17] H.A. Elbatal, Z.S. Mandouh, H.A. Zayed, S.Y. Marzouk, G.M. Elkomy, A. Hosny, Thermal, structure and morphological properties of lithium disilicate glasses doped with copper oxide and their glass–ceramic derivatives, J. Non-Cryst. Solids 358 (2012) 1806–1813. [18] D.S. Brauer, M.N. Anjum, M. Mneimne, R.M. Wilson, H. Doweidar, R.G. Hill, Fluoride-containing bioactive glass–ceramics, J. Non-Cryst. Solids 358 (2012) 1438–1442. [19] A.R. Hanifi, A. Genson, W. Redington, M.J. Pomeroy, S. Hampshire, Effects of nitrogen and fluorine on crystallization of Ca–Si–Al–O–N–F glasses, J. Eur. Ceram. Soc. 32 (2012) 849–857. [20] M.A. Sainz, P. Miranzo, M.I. Osendi, Sintering behaviour and properties of YAlSiO and YAlSiON glass–ceramics, Ceram. Int. 37 (2011) 1485–1492. [21] F. Peng, K. Liang, A. Hu, H. Shao, Nano-crystal glass–ceramics obtained by crystallization of vitrified coal fly ash, Fuel. 83 (2004) 1973–1977. [22] P.K. Maiti, A. Mallik, A. Basumajumdar, P. Guha, Influence of barium oxide on the crystallization, microstructure and mechanical properties of potassium fluorophlogopite glass–ceramics, Ceram. Int. 38 (2012) 251–258.