Investigation of Isothermal Treatment on the Structural, Microstructure and Physical Properties of Li2O-Al2O3-SiO2 Glass-Ceramic

The present work aims to investigate the effects of isothermal treatment on the structural, microstructure and physical properties of Li 2 O-Al 2 O 3 -SiO 2 glass-ceramic. Sintering temperature plays a major role in producing the desired lithium aluminosilicate (LAS) glass-ceramic crystalline phases. This work also aims to achieve a low thermal expansion coecient β-spodumene (LiAlSi 2 O 6 ) crystalline phase with improved density and lower porosity, which can be useful for the applications with thermal shock properties. The LAS glass-ceramic was fabricated by the melt-quenching technique at 1550 °C for 5 h before being isothermally sintered at an elevated temperature of 900 to 1200 °C for 30 min. The evolution of LAS glass-ceramic crystalline phases was identied using differential thermal analysis and the β-spodumene exothermic peak appeared at 999 °C. Based on the X-ray diffraction results, the complete transformation of β-spodumene from high-quartz solid solution (β-quartz) occurred at 1000 °C. However, the sintering temperature did not change the crystalline phase when sintered above 1000 °C, but the lattice parameter of the crystal structure was slightly altered. Moreover, it was observed that the LAS glass-ceramic grain size increased with temperature, whereby the smallest average grain size recorded (0.61 µm) for LAS glass-ceramic sintered at 1100 °C. Meanwhile, the fully densied LAS glass-ceramic at 1100 ° C was measured at 2.47 g/cm 3 with 0.52% porosity. The isothermal treatment at elevated temperature indicated that sintering at 1100 °C provided a denser, less porous, and small average grain size which is preferred for thermal shock resistance applications.


Introduction
Glass-ceramics are inorganic and non-metallic material that consists of residual glassy phases, containing polycrystalline solid material phases prepared by controlled crystallisation of glasses via different processing method [1]. A glass-ceramic contains at least 5-95% of crystallinity [2] because it comprises of the inherently amorphous structure of parent glass and crystalline phases obtained by the second heat treatment. This heat treatment controls the nucleation and crystallisation growth to improve the properties of glass-ceramic such as toughness and strength [3,4]. Lithium aluminosilicate (LAS) glass-ceramic is a glass-ceramic system that possesses excellent properties of low, zero or even negative coe cient of thermal expansion (CTE) as well as excellent thermal shock resistance, chemical durability and high mechanical strength [5][6][7].
LAS glass-ceramic is often used in applications that require minimal thermal expansion at a higher temperature such as cooktop panels and high-temperature furnace windows [8][9][10]. The zero or negative CTE of LAS glass-ceramic is possessed by crystalline phases with highly anisotropic CTE, named βquartz [11,12], β-spodumene [13][14][15] and β-eucryptite [16,17]. β-spodumene is a solid solution derived from keatite with CTE ranging between 0.5-1.0 × 10 − 6 °C − 1 [18,19]. Having said that, the relationship between CTE and thermal shock resistance is reversed. The lower the CTE, the higher the thermal shock resistance of the LAS glass-ceramic [20]. β-spodumene can be classi ed as the most stable crystalline phase at high temperature, which is especially important for cooktop panel and furnace window applications requiring uctuating high-temperature changes during services. This stable β-spodumene crystalline phase can withstand thermal expansion while preventing cracks that could cause thermal shock.
In LAS glass-ceramic system, additional oxides are usually added to facilitate the fabrication of LAS. Of which alkali oxide and alkali earth oxide are often used as glass network modi er, to lower the viscosity of molten glass such as NaO [21], BaO [22], MgO [23], ZnO [24] and CaO [25]. Additionally, nucleating agents like ZrO 2 [26], TiO 2 [27] or a combination both are added [10] to promote nucleation before crystallisation to produce ne-grained LAS glass-ceramic. Without nucleating agents, coarser grains will be produced leading to potential microcracks that would deteriorate the mechanical properties when used in high thermal applications [28].
A comprehensive study on the chemical composition of parent glass and the thermal treatment should be assessed as these factors can in uence the nal properties of the crystalline phase. Whereby, the nal crystalline phase determines the performance of LAS glass-ceramic at high thermal applications. For the production of the nal crystalline phase, β-spodumene can be obtained via conventional techniques namely melt quenching, sintering and crystallisation of the mixed powder of parent glass [29]. A good quality product can be achieved using high purity starting materials of glass and melted to a higher temperature before quenching to obtain glass frit or cullet. Sintering process begins with the noncrystalline (amorphous) phase before crystallisation occurs. As the temperature rises, crystallisation and sintering processes occur simultaneously to achieve crystalline phases and densi cation [30].
To date, many LAS parent glass composition has been modi ed to produce LAS glass-ceramic, for instance, the addition of oxides in LAS glass-ceramic such as B 2 O 3 [31], BaO [22], ZnO [32] and MgO [33].
These additives were intended to improve the sinterability [25], induce crystallisation process [34], increase exural strength properties [24], enhance CTE properties to the smallest possible value [21,25,29] and improve thermal shock resistance [35] of the LAS glass-ceramic. Apart from the addition of oxides, the sintering temperature also plays a major role in the formation of the desired crystal phases.
Generally, β-quartz which is formed below 900 °C is changed to β-spodumene crystalline phase above 900 °C [1,36]. According to a LAS glass-ceramic study [9] conducted using varying sintering temperatures for 24 hours, β-spodumene which appeared at 710 °C is considered to be of low temperature compared to other studies [5,30,37]. Lower sintering temperature requires longer sintering time to form β-spodumene, which can be time-consuming. On the other hand, the formation of β-spodumene crystal phase was achieved via non-isothermal sintering temperature at above 1000 °C, led to poor sinterability and increased porosity of the LAS glass-ceramic [6]. This observation demonstrated that the formation of the intended crystalline phase does not depend strictly on the alteration in the composition of the parent glass, but also the sintering temperature. Therefore, the present work focuses on the effects of sintering temperatures on the structural, microstructure and physical properties of LAS glass-ceramic to determine the best sintering temperature resulting in increased density and less porosity.

Glass-ceramic preparation
The LAS glass-ceramic was fabricated using reagent grade SiO 2 , Al 2 O 3 (Sigma, USA), P 2 O 5 , ZnO (Sigma-Aldrich, USA), Li 2 CO 3 , CaCO 3 , MgO (Merck, Germany), ZrO 2 , TiO 2 , BaO (Aldrich, USA), and Na 2 CO 3 (Bendosen, Malaysia). The chemical composition of the glass-ceramic is summarised in Table 1. Approximately 300 g of all raw materials were mixed and ball-milled on a horizontal roller for 6 h with a ball to powder weight ratio (BPR) of 2:1 to yield homogeneity. The raw materials in the powder form were then melted in an alumina crucible at 1550 °C for 5 h with a heating rate of 10 °C/min in an electrically heated furnace (Lenton Tube EHF 1800, Japan). The molten glass was then quenched in a distilled water bath at room temperature to produce glass cullet for milling. The glass cullet was milled in a tungsten vial in a planetary ball mill containing tungsten balls at 400 rpm for 3 h. The BPR was 10:1. The powder that was obtained was then sieved through mesh size of 200 µm, whereby the mean particle size of the glass powder was determined using laser diffraction particle size analyser (Malvern Instruments, United Kingdom). Next, the glass powder was pressed into cylindrical pellets (Ф 13 mm × 5 mm) using uniaxially hydraulic pressing (Specac Atlas ®, USA) at a pressure of 75 MPa for 20 s using 2 wt.% oleic acid. Subsequently, the glass pellets were heated at 5 °C/min up to 500 °C for 2 h to remove the binder followed by further heating at elevated temperature (  The glass transition temperature (T g ) and crystallisation temperature (T c ) were determined using differential thermal analysis (DTA) (Rigaku, Thermo plus Evo TG 8120, Japan). Approximately 15 mg of LAS glass powder was heated non-isothermally under air at 5 °C/min at room temperature up to 1400 °C, in an alumina crucible. The crystalline phase transformation of glass-ceramics was identi ed using X-ray diffraction (XRD) (D8 Bruker Advance Diffractometer, England) with Cu-Kα radiation (λ = 1.5406 Ǻ at 20 kV and 30 mA) in an angular (2θ) range from 10 to 70°. The functional group of glass-ceramic powders were determined using Fourier transform infrared spectroscopy (FTIR) (Perkin Elmer Spectrum One, USA).
The spectrum was recorded between 400 -4000 cm -1 under a transmittance mode. The surface morphology of the glass-ceramics was observed under a scanning electron microscopy (SEM) (FEI Quanta 650 FEG, Japan) and a eld emission scanning electron microscopy (FESEM) (Zeiss Supra 35 VP, Germany). The atomic percentage of LAS glass-ceramic elements was determined using a FESEM with the combination of energy-dispersive X-Ray spectroscopy (EDX). The EDX result was unable to detect the Lithium (Li) element because the machine is unable to detect elements below the beryllium (Be). The glass-ceramics were thermally etched at 50 °C below its sintering temperature in the furnace for 3 h and coated with a thin layer of gold prior to microstructural analysis. The grain size of the glass-ceramics was determined from FESEM micrographs through the linear intercept method in accordance with the ASTM E112 [38] in ImageJ (open source) software.

Physical evaluations
The density of the glass-ceramic was determined using the Archimedes method following ASTM C830 [39]. Five measurements for every sample were recorded, while the bulk density and apparent porosity were calculated based on equation 1 and 2: where W d is weight of dry cylindrical pellet, W s is weight of suspended pellet, W w is weight of soaked pellet.

Thermal analysis of LAS parent glass
Non-isothermal DTA of LAS parent glass is illustrated in Fig. 1. The T g of parent glass was estimated at 529 °C. Moreover, 3 exothermic T c peaks (T C1 , T C2 and T C3 ) were observed at 828, 999, and 1238 °C, respectively, suggesting the formation of different phases when the heat was increased. The rst exothermic peak (T C1 ) corresponded to the hexagonal high quartz solid solution (s.s.) (LiAlSi 2 O 6 ) phase.
The second and third crystallisation peaks at 999 and 1238 °C were due to the formation of tetragonal βspodumene (LiAlSi 2 O 6 ) and β-eucryptite (LiAlSiO 4 ) phases, respectively. During the early stages of crystalline phase formation, the binary oxide phases of Li 2 Si 2 O 5 and LiAlO 2 were initiated from Li 2 CO 3 reaction with SiO 2 and Al 2 O 3 . The binary oxides then expanded the reaction by forming LiAlSi 2 O 6 and LiAlSiO 4 ternary oxide phases [40]. Based on the DTA ndings, the transformation of crystalline phases in temperature-in uenced LAS glass was revealed. Therefore, different sintering temperatures were selected starting at 900 °C to provide su cient thermal treatment for crystallisation.
ZrTiO 4 was observed at 30.64° around 1000 and 1100 °C, while ZrO 2 appeared around 30.27° when heated at 1150 and 1200 °C. The diffraction peaks of these secondary phases were often very weak and broad, due to the low amount of very small-sized grains. Having said that, the ZrTiO 4 and ZrO 2 act as nucleation agents for the subsequent LAS phases. The ZrTiO 4 or ZrO 2 that were formed as tiny crystals were embedded in a shell enriched with Al 2 O 3 following the completion of nucleation. After which, crystallisation of the LAS crystal phase (e.g. β-spodumene, high quartz s.s) took place [43,44]. The negrain microstructure was observed due to the presence of nucleation agents [42] which was proven by the grain size obtained using FESEM (Fig. 7). Furthermore, the absence of a nucleating agent peak at 900 °C in Fig. 2a might be due to the decreased concentration of Li + . The smaller radius and high eld energy of Li + in a network make it accessible to be surrounded by regular oxygen ions which restrain nucleation and crystallisation growth. In this gure, only A2, A3, A4 and A5 samples were considered due to the similarity in the major crystalline phase. The main β-spodumene peak slightly shifted to lower angles as temperature increased, possibly due to the increased Si 4+ and Al 3+ . This observation could also be supported by the lattice parameter of β-spodumene which was extracted from Rietveld re nement analysis. The tetragonal β-spodumene structure possessed lattice parameters of a = b ≠ c space group of P 41 21 2 and angles α = γ = 90°, where the lattice volume (V) = a 2 ×c. Fig. 3a represents the lattice parameter cell (a, c) and the V. The gure demonstrated that a and c were inversely proportional to decreasing V as the temperature increased.
The reduction in V could presumably be due to increasing Si 4+ and Al 3+ . This observation can also be veri ed via EDX element analysis through atomic percentage obtained by Si (silicon), Al (aluminium) and O (oxygen) as illustrated in Fig. 3b. However, the decrease in a of LAS glass-ceramic sintered from 1000 to 1200 °C was probably due to the augmentation of Li  Next, Fig. 4 represents the atomic percentage of nucleating agents (Ti and Zr) and glass modi ers (Na, Ba, Mg, Zn and Ca) as a function of sintering temperature. The trend of atomic percentage of Ti is opposite to that of Zr as illustrated in Fig. 4a. The enriched TiO 2 was produced at early sintering stages due to the smaller ionic radius of Ti 4+ that grows at a lower temperature until the remaining TiO 2 and ZrO 2 which were established in the glass matrix were later precipitated as ZrTiO 4 crystal. In general, nucleation precipitates in glass matrix (residual glass) and occurs within the LAS periphery [47]. As illustrated in Fig. 2a, the nucleating agent appeared beyond 1000 °C. For instance, ZrTiO 4 appeared at 1000 and 1100 °C, while ZrO 2 appeared at 1150 and 1200 °C. Moreover, there was a moderate and steady growth of nucleating agents, as indicated by a low-intensity peak almost at 30°. The low-intensity peak appeared due to the formation of a diffusion barrier among nucleant droplets and LAS crystals following the crystallisation of ZrTiO 4 [12]. On the contrary, the only nucleating agent found to be present in A4 and A5 samples was ZrO 2 , which was further veri ed by the increased atomic percentage of Zr in both A4 (1150) and A5 (1200) with the depletion of Ti (Fig. 4a).
Meanwhile, Fig. 4b displays the atomic percentage of Na and Ba with increasing sintering temperatures.
Modi er oxides of Na 2 O and BaO are used in glass-ceramic purposely to weaken the ionic glass network.
The disruption of Si-O-Si network was due to greater ionic size of Na and Ba which reduced the glass viscosity to permit crystallisation and densi cation. Generally, the same trend was also apparent in Fig.  4b, where the atomic percentage of Na and Ba increased at 1100 ° C (A3) before declining. Therefore, this observation explained the occurrence of crystallisation and full densi cation at 1100 ° C, where Na and Ba could have possibly disrupted the Si-O-Si network and any associated bonds. This phenomenon diminishes the degree of glass polymerisation to improve the crystallisation process. Meanwhile, the effects of sintering temperatures on the atomic percentage of Mg, Zn and Ca are illustrated in Fig. 4c. The atomic percentage of each element displayed a dissimilar trend. An increase in Ca is observed at 1100 °C (A3) before decreasing. While the atomic percentage of Zn increased as temperature the increased before decreasing at 1200 °C, and the atomic percentage of Mg uctuated as temperature increased. However, the atomic percentage of Mg increased signi cantly from 1000 ° C to 1100 ° C. This increase is probably due to the formation of β-spodumene facilitated by the presence of MgO, which also weakened the silica network [30,48]. Apart from NaO and BaO, MgO also weakens the silica network.
On the other hand, the increase in Zn in A4 and A5 did not affect the crystallisation process much but disrupted other compositions that gradually altered the entire crystalline process by forming a crystalline phase variation. This observation is in agreement with another study [29], where ZnO made densi cation achievable and less porous. Besides MgO and ZnO, CaO is also useful in promoting the formation of crystalline phase [49,50] by Ca function alone or in a combination with MgO to strengthen the glassy phase [50].

FTIR analysis
The structural changes in LAS glass-ceramic at elevated sintering temperatures can be achieved via FTIR.
The intended studies were performed within the wavenumber region range of 4000-400 cm −1 , however, vibrations was observed at 462 cm -1 regions in this study [53].
The FTIR spectra for LAS glass-ceramic heat-treated at elevated sintering temperatures illustrated four distinct absorption bands appearing at wavenumber (cm -1 ) 1300 -900, 850 -680, 640 -520 and 500 -420. A strong bond absorption band around 1300 and 900 cm -1 appeared in all glass ceramics corresponding to the asymmetric stretching vibration of Si-O-Si bonds in the SiO 4 tetrahedra network [54,55] and symmetrical stretching vibration of Si-O-Al [55]. The asymmetric stretching vibration represents the typical framework silicate [3], while the symmetrical stretching vibration indicates the replacement of SiO 4 tetrahedron with AlO 4 tetrahedron [55]. Based on this observation, the absorption band shifted to higher wavelength region (towards 1000 cm -1 ) with an increase in sintering temperature from 900 to 1200 °C. This change was caused by the substitution of AlO 4 tetrahedra with SiO 4 tetrahedra,  [57]. This band formed in this region also represents the vibration of Al-O covalent bonds in AlO 4 tetrahedra β-spodumene [56] and symmetric Si-O-Si bending vibrations [22] at 718 cm -1 . Next, the absorption band at around 640 -520 cm -1 was due to the stretching vibration of AlO 6 units in Al-O-Al and the vibration of TiO 4 tetrahedron [57]. Furthermore, a weak peak (637 cm -1 ) which appeared at the beginning of this band for samples A3 to A5 indicated the bending vibrations of the bridging oxygen (Si-O-Si, Si-O-Al) [3]. Meanwhile, a pronounced peak at 567 cm -1 for glass-ceramics indicated the presence of β-spodumene associated with AlO 4 tetrahedra vibration [56]. The characteristic band that appeared at 445 cm -1 indicated the bending vibrations of Si-O-Si in the SiO 4 tetrahedral network [54][55][56] and stretching vibration of Ti-O-Ti [57].

Microstructure analysis
The microstructure of LAS glass ceramics characterised through SEM and FESEM were illustrated in Fig.  6 and Fig. 7, respectively. In general, samples A1 and A2 indicated uniform dispersion of spherical particles sized between 1 -2 µm ( Fig. 6a and b). At 900 °C, the LAS glass-ceramic which was at the initial sintering stage demonstrated a slight growth at the neck due to the lattice and grain boundary diffusion. Referring to Fig. 3b, the progressive neck growth was observed when the sintering temperature was increased to 1000 °C. This observation could be due to a higher sintering diffusion rate. Ultimately, the necking growth will form grain boundaries and diminish pores as the temperature is increased. Moreover, sample A2 denoted by Fig. 6b displayed the adherence of distinct small white particles to the larger particles. The small white particles are assumed to correspond to nucleating agents (TiO 2 and ZrO 2 ) which adhere to the hosting large dark particles of LiAlSi 2 O 6 . Compared to sample A1 in Fig. 6a, there were no small white particles associated with the XRD result (Fig. 2a) and no peaks for nucleating agents. This microstructure behaviour of A1 and A2 samples are supported by the density results in Table  4. Although the primary sintering process was dominant at 900 to 1000 °C, full densi cation was not achieved.
Based on Fig. 7a, b and c, the microstructure evolution of the LAS glass ceramics heat-treated at 1100, 1150, and 1200 °C are presented. According to the FESEM images, the grain size of LAS glass ceramics gradually increased as the sintering temperatures were increased, with grain boundaries consisting of polygonal shape and lamellar structure. The lamellar structure is believed to consist of β-spodumene layers with positive and negative thermal expansions indicating a low anisotropic CTE of LAS glassceramic [58]. Meanwhile, the larger grain size of LAS glass ceramics as increased temperatures could be attributed to the nucleation and development of the new phase, which leads to an increase in granular grain formations [49]. Previous literature stated that the microstructure of β-spodumene was approximately 0.3-1.0 µm [59] and 1-2 µm [60]. In this study, the mean grain size of sintered LAS glass ceramics as a function of temperature is tabulated in Table 3. The mean grain size of the samples A3, A4 and A5, in particular, increased to 0.61, 0.71 and 1.65 µm, respectively.
The ne-grained microstructure is reported to fall within the range of 0.1 -1.0 µm [59]. Hence, based on this reference, samples A3 and A4 were categorised as ne-grained microstructures, while sample A5 has a coarser-grain microstructure. The ne-grain microstructure may lead to enhanced mechanical properties related to the grain boundary surface area [61,62], indicating better resistance to thermal shock.
Meanwhile, the coarser grain boundary deteriorates the mechanical properties of LAS glass-ceramic [14].
The grain size of sample A3 and A4 did not differ much but produced a signi cant effect on the density of the sample A3 as indicated in Table 4. The high sintering temperature of 1200 °C holds an enormous effect on the grain size of the sample A5 (Fig. 7c), displaying a notably coarser microstructure than sample A3 and A4. , causing a rapid surface diffusion rate to form irrepressible grain boundaries.  Table 4 indicates the physical properties such as density, linear shrinkage and porosity for the sintered LAS glass-ceramic at different temperatures. The densi cation of LAS glass-ceramic depends on the sintering temperature. Hence, increasing the sintering temperature from 900 °C to 1100 °C (A1 to A3) demonstrated a progressive glass-ceramic sintering behaviour which dominated the crystallisation process. As the temperature increased, the viscosity of the glass structure decreased, further promoting a viscous ow sintering mechanism [63,64]. Therefore, a higher diffusion rate initiates the formation of neck growth between individual particles of the compacted sample. Grain boundary diffusion permits the growth of the centre connection between the individual particles by reducing the volume of the glassceramic sample and converting the surface area to grain boundaries (Fig. 6). Meanwhile, this study also demonstrated a densi cation progression from lower to a higher temperature. The fully densi ed LAS glass-ceramic in this work measuring 2.47 g/cm 3 (sample A3) was sintered at 1100 °C.
The sintering temperature is considered low for complete densi cation of LAS glass-ceramic compared to full densi cation reported in a previous study [65], which was achieved at 1350 °C. The density of sample A3 was also in the range of LAS glass-ceramic in a previous study [47]. The improvement in density is likely to be due to the presence of high-density spodumene as a major phase obtained in XRD [14]. On the other hand, apart from being a nucleating agent, TiO 2 also facilitates the densi cation of LAS glassceramic by increasing volume crystallisation, to yield crack-free denser LAS glass-ceramic [59]. Having said that, a higher atomic percentage of Ti in Fig. 4a sintered at 1100 °C contributed to the full densi cation of sample A3. Meanwhile, the density of samples A4 and A5 decreased as the atomic percentage of Ti declined. The viscous ow sintering inhibited at higher temperatures allows devitri cation to occur [63] and increases the degree of polymerisation. Therefore, the higher the degree of polymerisation, the higher the Si/O ratio. This pattern implies an increase in the viscosity of the glass network, thus, hindering densi cation. Based on Fig. 3b, an increase in the Si atomic percentage and a decrease in O atomic percentage indicated an increase in Si/O ratio as the sintering temperature increases. The density regression in samples A4 and A5 was due to the coarsening effect. Surface diffusion occurs when two particles change shape without centre approach which does not occur during the previous stages of sintering. In short, coarsening impedes centre contacts between the particles, limiting the growth of the neck structure and progressing the growth of grain without densi cation as the temperature rises. Hence, during coarsening, as the grain size increases the grain number is reduced [66]. Fig. 7c corroborates the above observation.
On the other hand, the linear shrinkage of the LAS glass-ceramic was signi cantly increased as the temperature increased, however, there was a slight reduction in linear shrinkage towards higher temperature. This phenomenon was due to the kinetic energy that allowed the particles of the materials to vibrate, diffuse and realign themselves into a periodic arrangement, leading to the occurrence of shrinkage [49]. According to a previous study [26], the maximum linear shrinkage for LAS glass-ceramic was estimated at approximately 20 % which was almost similar to the value achieved in this study. The reduced linear shrinkage of A4 and A5 samples suggested a restricted densi cation process at 1150 and 1200 °C, respectively. Moreover, porosity decreased as the sintering temperature increased (Table 4). Despite the lower density, the porosity of sample A5 was greatly reduced. The reduced porosity could be due to the lower dihedral angle of pores at higher temperatures, which stabilises the pores and constrains density. Meanwhile, the number of grains intersecting the pore will decrease as the sintering temperature increases, leading to the shrinkage of unstable pores [66]. Less porous LAS glass-ceramic at increased temperatures might be due to the presence of ZnO. As illustrated in Fig 4c, an increase in the atomic percentage of Zn observed at the increased temperature indicated lower porosity for all LAS glassceramic samples.

Conclusion
Li 2 O-Al 2 O 3 -SiO 2 glass-ceramic was isothermally treated at elevated temperatures from 900 to 1200 °C for 30 min to determine the appropriate sintering temperature to produce β-spodumene crystalline structure that creates a denser and less porous LAS glass-ceramic. The sintering temperature above 1000 °C did not alter the major crystalline phase of β-spodumene, however, the lattice parameter of the crystal structure was altered. The mean grain size of LAS glass-ceramic sintered at 1100 °C, 0.61 μm, could prevent possible microcracking to yield better thermal shock resistance. Meanwhile, sintering temperature also has a signi cant effect on densi cation and porosity followed by linear shrinkage. A high-density LAS glass-ceramic (2.47 g/cm 3 ) with optimal linear shrinkage (15.69%) at 1100 °C indicated grain growth and densi cation of the particles. This observation relates to the tetragonal β-spodumene crystal structure achieved by XRD characterisation and appeared in FESEM images as lamellar structure. All the above observations suggested that LAS glass-ceramic sintered at 1100 °C can be potentially used in high thermal shock resistance applications.