Mechanical Behavior of Ion-Exchanged Alkali Aluminosilicate Glass Ceramics

Abstract

Glass is one of the oldest and most versatile manmade materials and has been used for centuries. One of the areas of significant research and progress is that of high-mechanical-resistance glass. Several processes can be used to maximize the mechanical strength of glasses, namely thermal or chemical treatments. Glass ceramics, obtained through controlled crystallization, can enhance the mechanical properties of these materials and, as long as the crystals remain small enough, transparent glass ceramics can be obtained. On the other hand, ion exchange can strengthen the glass surface and reduce failure. As a result, ceramization followed by ion exchange can further enhance the mechanical characteristics of the parent glass. Aluminosilicate glasses and glass ceramics are known to present excellent transparency and chemical durability, plus good mechanical behavior; therefore, a lithium aluminosilicate glass composition was studied in order to obtain transparent glass ceramics, followed by an ion exchange process, and its mechanical properties were studied. Transparent glass ceramics were obtained, with an increase of ~15% in hardness over the parent glass. The glass ceramics were then subjected to an ion exchange treatment in a KNO₃ bath, in order to further enhance the mechanical properties, without hindering the optical transmittance. The combination of heat treatment and chemical treatment resulted in a cumulative hardness increase of ~25%, from 620 ± 10 HV, for the as-cast glass and from 773 ± 23 HV for the ion-exchanged glass ceramic. Regarding the indentation fracture toughness, values obtained for the glass ceramics were similar to those obtained for the cast glass, yielding no noticeable change. Indentation fracture toughness increased after the ion exchange treatment of the glass ceramic, since a much higher load was necessary to obtain a measurable indentation, indicating higher indentation fracture strength.

1. Introduction

In 1953, Stanley Donald Stookey, from Corning Glass Works^®, introduced the concept of glass ceramic (GC), where a mixed glass and ceramic material is formed through the controlled nucleation and crystallization of glass [1,2]. The production process of GC begins by producing glass, which then goes through a thermal process called ceramization that promotes internal crystallization and results in a pore-free, near-net-shape product [3]. Glass ceramics are now widely used materials due to the combination of the good mechanical properties of ceramics while obtaining non-porous complex shapes using common glass-forming processes. Depending on their chemical composition, microstructure, and processing technology, they offer a wide range of properties that can be tailored, yielding new and promising applications. Glasses and glass ceramics based on alkali aluminosilicates (AASs) usually present excellent transparency and chemical durability, plus good mechanical properties, and are currently used in industry in many different areas [4,5,6,7]. One of the areas with significant research and progress in AAS glass ceramics (AAS-GC) is in dental applications [8]; however, there is also a potential application in mobile phone displays or monitor screens, for example, as long as the glass ceramics remain transparent. In order for a GC to be transparent in the visible range, it needs to meet one of two criteria [9]: (a) the size of the crystals must be much smaller than the wavelength of visible light (less than one-tenth of the light wavelength); (b) the birefringence in the crystal and refractive index differences between the glass and the ceramic phases must be negligible. For the formation of a GC with a nanometric scale microstructure, it is necessary to promote nucleation through a high nucleation rate while avoiding grain growth. This control is possible through the introduction of nucleating agents such as TiO₂ and ZrO₂, and an adequate heat treatment.

Ion exchange or chemical tempering is a process where diffusion promotes an exchange of cations between the molten bath and the glass composition. If small-sized cations from the glass composition (such as Li⁺) are replaced by larger cations from the bath (like K⁺), then the glass structure becomes “stuffed” and compressive forces are inserted into the glass surface. In order to avoid relaxation, this process must occur below the glass transition [10]. In recent years, several reviews on this topic have been published [11,12,13,14,15,16].

The combination of both processes (ion exchange of a glass ceramic) can further increase the mechanical properties, as reported by Dejneka et al. [17], who presented ring-on-ring data on aluminosilicate glasses and ion-exchanged aluminosilicate glass ceramics, demonstrating that after ion exchange, the failure load for the glass ceramics increased by a factor of ~6. Beall et al. [18] reviewed different ion exchange mechanisms in aluminosilicate glass ceramics and the resulting mechanical properties, illustrating a wide range of possibilities. The results of the mechanical tests performed on ion-exchanged glass ceramics yielded a significant increase in the load to failure: from 380 to 460 N before ion exchange to values higher than 2300 N after chemical strengthening.

Starting from a previously reported [19] multicomponent alkali aluminosilicate glass (AAS) with nominal composition (wt%): 70% SiO₂—15%, Al₂O_3r—9%, Li₂O—3%, MgO—3%, and ZnO, to which nucleating agents (3% TiO₂ and 1% ZrO₂, in wt%) were added, a ceramization treatment of the glass and a subsequent chemical treatment of the glass ceramic, in a molten bath of KNO₃, were carried out in order to study the cumulative effect of the increase in hardness derived from ceramization and the increase in hardness derived from the ion exchange of the glass ceramic, without hindering the optical properties. The aim of this paper is to compare the optical and mechanical properties of the AAS glass plus the ion-exchanged AAS glass, whose results are presented in [19], with the results obtained here for the glass ceramics and ion-exchanged glass ceramics.

2. Experimental Methods

2.1. Glass and Glass Ceramic Preparation

Glass preparation follows the procedure reported in [19], where ~2 mm thick glass disks 25 mm in diameter were obtained. In this work, the glasses were heat-treated, in order to promote nanocrystallization, thus increasing the mechanical properties without loss of transmission, followed by an ion exchange process.

The raw materials were mixed and melted in a platinum crucible and then cast in pre-heated graphite molds, yielding ~2 mm thick glass disks 25 mm in diameter. The glass disks were then annealed and cooled in the furnace. For this composition, glass transition temperature occurs at about 525 °C [19]; therefore, annealing was performed at 485 °C (usually 40 to 50 °C below T_g). Further preparation details are reported elsewhere [19].

One important consideration in glass materials is the surface finishing. Manual polishing of these glasses with silicon carbide grinding papers down to 4000 grit (~5 µm) originated surface defects of the order of ~100 µm, as estimated from the fracture in the bending ring on the ring tests, and presented a fracture strength of 120 ± 26 MPa [20]. Therefore, the glass samples had to be further polished (lapping with 2000# carborundum disk and polishing down to 0.5 µm ceria), achieving a final thickness of 1.7 mm and average surface defects of the order of 10 µm, which increased the fracture strength of the glass to 310 ± 84 MPa [19]. The surface finish, in terms of average roughness (R_a) and mean roughness depth (R_z), obtained for both polishing methods was determined by AFM, as reported in

Table 1. Average roughness (R_a) and mean roughness depth (R_z) of AAS glasses.

Surface Finishing	Ra [nm]	Rz [nm]
Polished to 5 µm	3.05 ± 0.65	85.53 ± 8.61
Polished to 0.5 µm	0.59 ± 0.21	12.25 ± 9.57

.

After synthesis and polishing of the AAS glasses, ceramization was carried out through heat treatments performed between 570 and 650 °C for 1 h. These heat treatment conditions were chosen to promote nucleation without compromising the optical properties of the glass samples. A crystallization treatment was also performed at 700 °C, above the onset crystallization temperature (675 °C [19]).

2.2. Ion Exchange Process

The ion exchange conditions, for the AAS glass samples, were optimized in terms of hardness and UV-Vis transmission, by a treatment, in molten potassium nitrate, at 450 °C for 12 h, which yielded compressive layer depths of ~9 µm [19]. Therefore, similar tests were carried out for the glass ceramics, and combinations between temperature (420 and 450 °C) and ion exchange time (9, 12, and 30 h) were tested. The samples were kept in KNO₃ baths in an upright position to ensure that the ion exchange process took place on both surfaces, and for each test, new KNO₃ baths were used to avoid salt bath poisoning [21].

2.3. Sample Characterization

Density measurements were taken before and after the treatments carried out on the glass (thermal treatment and ion exchange) in order to monitor the possible effect on this property. The measurements were performed on a AM50 scale (Mettler Toledo, Greifensee, Switzerland) equipped with an assembly for the determination of density using Archimedes’ principle. Measurements were performed with distilled water, and the temperature was monitored during the measurements.

UV-Vis transmission measurements were taken before and after the treatments using a Helios Gamma instrument (Thermo Fisher Scientific, Waltham, MA, USA). Measurements were taken between 190 nm and 1100 nm, with a bandwidth of 2 nm. For comparison purposes, the average transmission value at 550 nm was considered.

X-ray diffraction (XRD) and Raman spectroscopy were carried out in order to characterize the different heat-treated samples. The XRD analyses were performed in powdered samples, on a D8 advance equipment (Bruker, Billerica, MA, USA) for XRD, with 2θ angles from 10 to 80°, a step size of 0.03°, and a step time of 1 s, with the samples rotating at 30 rpm. The Raman spectra of the bulk glasses were obtained on a LabRam HR 800 Evolution confocal Raman microspectrometer (Horiba Jobin-Yvon, Lille, France) using a 532 nm laser, focused with a 100× objective, with a power of 10 mW at the samples, and data were collected with 4 scans, 10 s each, with a 600 gr/mm grating, from 100 to 1800 cm⁻¹.

Hardness measurements were performed using Vickers microhardness in Duramin equipment (Struers France S.A.S, Champigny sur Marne, France) with the application of a 1.96 N load for 15 s before and after the treatments. The hardness values presented, in Kgf/mm², are the result of an average of at least 15 measurements.

Mechanical characterization, namely the measurement of elastic constants (Young’s modulus, E, shear modulus, G, Poisson’s coefficient, ν, indentation fracture toughness, K_C, fracture strength, σ_f, and Weibull modulus, m), was determined as described in [19]. Basically, the elastic constants were obtained by acoustic methods, while fracture toughness was estimated based on hardness measurements made using the Indentation Fracture Method and carried out in accordance with the JIS R 1607 standard [22], using Equation (1), where E is Young’s modulus, HV is the Vickers hardness, P is the load used, C corresponds to half the average length of the crack, and a is half the average indentation diagonal, respectively. According to this standard, the measurements are only valid when the cracks generated from the vertices of the indentation have no more than 10% difference between them and the length of the cracks are not less than 2.5 times the length of the diagonal. To obtain these Vickers hardness results, an AVK-C2 durometer (Mitutoyo, Kawasaki, Japan) was used with loads of 49, 98, or 196 N and loading times of 15, 20, or 25 s. On average, at least 15 measurements were considered.

$$K_C=0.018·{\left({\displaystyle \frac{E}{H·V}}\right)}^{{\displaystyle \frac{1}{2}}}·\left({\displaystyle \frac{P}{C^{\frac{3}{2}}}}\right)=0.026·{\displaystyle \frac{E^{\frac{1}{2}}·P^{\frac{1}{2}}·a}{C^{\frac{3}{2}}}}$$

(1)

The equibiaxial fracture stress (σ_f) was determined with a ring-on-ring assembly (support diameter equal to 20.2 mm and load diameter equal to 10.1 mm) on an INSTRON 5566 loading machine (INSTRON, Norwood, MA, USA) with a traverse speed of 5 mm/min. About 25 samples per experimental condition were tested, and the results were analyzed in terms of Weibull statistics [23], using the maximum likelihood method, which is considered to be less biased than the Weibull plot [24]. The results are presented in terms of survival probability with the Weibull modulus, m, which is a measure of the dispersion of the results (higher m value corresponds to lower dispersion of results). These tests were performed in accordance with the ASTM C1499-19 standard [25]. According to standard C1322-05b [26], it is possible to estimate the size of critical defects, “a”, responsible for inducing catastrophic failure, using Equation (2). By inserting the computed values of K_C and σ_f and considering Y = 1.59 (assuming that the rupture of the specimens occurs due to semi-elliptical surface defects of radius “a”), the size of the critical defects can be estimated:

$$K_C=Y·{\sigma }_f·\sqrt{a}$$

(2)

3. Results

3.1. AAS Glass Ceramic

The glass samples underwent thermal treatments of 1 h at temperatures between 570 and 650 °C. These treatments aimed to promote partial crystallization in AAS glass and determine the effect on hardness and other mechanical properties without hindering the transmission in the visible range. Examples of heat-treated glass disks are presented in Figure 1.

The results obtained, in terms of UV-Vis transmission and hardness, for the different heat treatments, are shown in Figure 2 and

Table 2. Vickers hardness results obtained for the heat-treated AAS glasses. The sample heat-treated at 700 °C for 1 h was cooled in the furnace in order to promote full crystallization while all the others were cooled at room temperature.

Heat-Treated Samples	Hardness (HV0.2)
AAS glass [19]	620 ± 10
570 °C; 1 h	673 ± 10
575 °C; 1 h	710 ± 14
580 °C; 1 h	754 ± 14
590 °C; 1 h	768 ± 17
700 °C; 1 h	806 ± 45

, respectively. Figure 2 indicates that a small decrease in transmission starts at 575 °C in the UV range, between 350 and 400 nm, but the transmission plateau in the visible range remains the same. For higher temperatures, the transmission plateau in the visible range decreases with increasing heat treatment temperature, causing a drastic decrease after 620 °C. The corresponding Vickers hardness values increase steadily for higher heat treatment temperatures.

The XRD results can be observed in Figure 3, where a small peak is observed at 2θ = 25.7°, for 570, 575 ion exchange, and 580 °C, which increases strongly for higher heat treatment temperatures. Data analysis was carried out using the Match! software (version 1.11), and the agreement between the phases found and the diffraction database is indicated by a Figure of Merit (FoM). The results obtained for 600 °C indicate the presence of LiAlSi₂O₆ (00-073-2336), with a FoM value of 0.836; however, at 700 °C, LiAlSi₃O₈ is the major crystalline phase, (00-035-0794), with a FoM value of 0.871, plus Li_xAl_xSi_(1−x)O₂, (00-040-0073), with an FoM of 0.724, although other minor phases, possibly different types of lithium aluminosilicates, can also be present.

The Raman spectra are presented in Figure 4 and are in line with the XRD results, since the Raman spectra of the 570 and 575 °C heat-treated glasses are fairly similar to the Raman spectrum of the AAS glass. Above 575 °C, the Raman spectra present bands indicating the presence of crystalline phases.

Since the 580 °C heat treatment revealed itself to be excessive, as it decreased the visible transmission plateau, and distinctive changes were observed in the Raman spectra and XRD diffractograms, the best heat treatment conditions yielding higher hardness with little transmission loss were considered to be 575 °C for 1 h. Mechanical characterization of the glass ceramic obtained with these conditions was then performed in 25 polished samples.

Table 3. Average results (n = 25 samples) obtained for the AAS glass ceramic heat-treated at 575 °C for 1 h.

ρ [g/cm3] (±0.01)	T [%] (500–550 nm)	HV0.2	E [GPa]	G [GPa]	ν (±0.01)	KC [MPa.m1/2]	σf [MPa]	Weibull Modulus
2.46	90	679 ± 14	85 ± 1	34 ± 1	0.24	1.6 ± 0.1	201 ± 92	2.89/2.32

resumes the findings for the AAS glass heat-treated at 575 °C for 1 h in terms of density, transmission plateau in the visible range, Vickers hardness, elastic constants, indentation fracture toughness, fracture strength, and Weibull modulus. In Table 3, two Weibull modules are presented. The two values presented correspond to different ways of determining its value: the first value corresponds to the traditional slope of the Weibull plot, while the second value corresponds to the maximum likelihood method (ASTM, C1239-06 [24]), which is considered to be less biased. The Weibull plots are presented in the Supplementary Materials.

3.2. Ion-Exchanged AAS Glass Ceramic

AAS glass samples heat-treated at 575 °C for 1 h were then subjected to ion exchange in KNO₃ in order to study the effect of the chemical treatment on the hardness and transmission in the visible range. The ion exchange conditions tested were 9 h, 12 h, and 30 h at 420 °C and 450 °C, respectively. Again, the objective is to maximize hardness while maintaining optical transmission in the visible range.

Regarding UV-Vis transmission, even after different chemical treatments, the glass ceramic samples present a high transmittance, around 90%, although for higher times at higher temperature, a decrease in transmission occurs, as observed in Figure 5 for 30 h at 450 °C. Figure 5 also presents the UV-Vis spectra obtained for sequential treatments performed on the same sample: as-cast (AAS glass); after ceramization, at 575 °C for 1 h (AAS-GC); and followed by ion exchange, (IE-AAS-GC), at 450 °C for 12 h.

The Vickers hardness results of the GC subjected to ion exchange are presented in

Table 4. Vickers hardness results for the AAS glass ceramic samples under different ion exchange conditions.

Ion Exchange Conditions		Hardness (HV0.2)
420 °C	9 h	720 ± 14
	12 h	763 ± 21
	30 h	745 ± 20
450 °C	9 h	755 ± 20
	12 h	773 ± 23
	30 h	763 ± 23

. Density and mass have also been monitored, but their variation is residual.

As observed in Table 4, ion exchange at 450 °C for 12 h presents the highest Vickers hardness value.

In order to estimate indentation fracture toughness, Vickers hardness measurements with higher loads were performed, but contrary to the fracture toughness measurements of ion-exchanged AAS glasses [19], for which measurements with 49 N for 15 s were able to produce indentations, whose c/a ratio was greater than 2.5, and met the requirement of JIS R 1607 [22], in the case of IE-AAS-GC, it was necessary to increase the loading conditions in order to obtain measurable indentations. In fact, the loading conditions for the IE-AAS-GC samples are higher than any other, which indicates a higher resistance to deformation. The loads and application times varied between 49, 98, and 196 N and 15, 20, and 25 s depending on the sample. According to JIR R 1607 [22], the results should have a c/a ratio above 2.5, but that could not be obtained, apart from the 450 °C for 12 h sample, as indicated in

Table 5. Indentation fracture toughness and c/a ratio observed in the Vickers hardness results for the IE-AAS-GC samples (since the c/a ratio is below 2.5, the obtained K_C values are just indicative).

Ion Exchange Conditions	Load [N]; Time [s]	KC [MPa·m1/2]	c/a
420 °C, 9 h	98; 20	3.18	1.86
420 °C, 12 h	98; 20	3.08	1.80
420 °C, 30 h	98; 20	2.60	2.12
450 °C, 9 h	98; 20	3.40	1.64
450 °C, 12 h	196; 15	1.97	3.08
450 °C, 30 h	Highly deformed indentations under any load and time

. Examples of indentations produced for the K_C measurement are presented in Figure 6, where images of the indentations produced elucidate the difficulty of the measurements. Figure 6e,f, for example, correspond to the same indentation, but due to its dimensions, two photographs had to be taken to observe the complete indentation. Since different loads and times had to be used to obtain measurable indentations, only a qualitative analysis can be carried out.

By analyzing the results present in Table 5, it is possible to conclude that, under similar conditions (98 N, 20 seg), the 450 °C for 9 h treatment presented a higher K_C measurement than the 420 °C treatments, and the load had to be increased for the longer 450 °C heat treatments, revealing higher fracture toughness for these samples. Among the samples treated at 450 °C, the 30 h exchange may be excluded as it produces a strongly distorted indentation compared to the 9 h and 12 h exchange times.

Based on the previous measurements, the ion exchange conditions selected for further mechanical characterization were 450 °C for 12 h, and

Table 6. Average results (n = 25) obtained for the ion-exchanged AAS-GC samples.

ρ [g/cm3] (±0.01)	T [%] (500–550 nm)	HV0.2	E [GPa]	G [GPa]	ν (±0.01)	KC [MPa.m1/2]	σf [MPa]	Weibull Modulus
2.46	88	773 ± 23	86 ± 1	35 ± 1	0.23	2.0 ± 0.2	362 ± 84	4.85/5.06 *

* The first value indicates the m_Wiebull obtained by linear regression, and the second value is obtained by the maximum likelihood method (ASTM, C1239-06 [24]). The Weibull plots are presented in the Supplementary Materials.

resumes the average results for 25 polished samples of AAS glass heat-treated at 575 °C for 1 h and submitted to ion exchange at 450 °C for 12 h.

4. Discussion

4.1. AAS Glass Ceramic Samples

The UV-Vis transmission spectra of the glass ceramic samples indicate a small decrease in the UV range for 575 °C, but there is little change in the visible transmission plateau for 570 and 575 °C heat treatments, although there is a significant variation in terms of hardness. At higher heat treatment temperatures, hardness continues to increase, but transmission decreases, particularly from 600 to 650 °C, as inferred from Figure 2 and Table 2.

The diffraction peak at ~25.7° of the heat-treated sample at 580 °C (Figure 3) was used to estimate the average crystal size. Using Scherrer’s equation [27] and taking into account the instrumental broadening of the equipment used, a value of ~18 nm was obtained. This value justifies the high transparency of the glass ceramic samples obtained after the heat treatment at lower temperatures (570 °C/575 °C, 1 h).

Regarding the XRD data, the results indicate that the first crystal phase to appear, LiAlSi₂O₆, gives rise to a new phase, LiAlSi₃O₈, between 600 and 700 °C, which becomes the major phase present at 700 °C, although other minor phases are also present. An amorphous halo is evident in the powder diffractograms, indicating the overall amorphous nature of the samples up to 600 °C.

In this study, special attention is given to the Vickers hardness results since this, in conjunction with fracture toughness, allows us to draw conclusions related to the effect of thermal treatments on the glass and thus helps in deciding which time–temperature pair enhances the mechanical properties the most.

Based on optical transmittance in the visible and hardness results, the ceramization conditions yielding the highest hardness values while maintaining the optical transmission in the visible region around ~90% are 575 °C—1 h. Higher ceramization temperatures originate higher hardness, but at the cost of transmission, as observed for the 580 and 590 °C results (Figure 2). The heat treatment at 575 °C for 1 h enhanced the hardness values by ~15% (~620 HV to ~710 HV).

When analyzing K_C, the average values obtained for AAS-GC (1.6 ± 0.1 MPa.m^1/2) are similar (within the error) to the ones previously obtained for AAS glass (1.7 ± 0.1 MPa.m^1/2) [19], indicating that ceramization had no effect on the indentation fracture toughness. This was not to be expected [17,18] and could suggest that although nucleating agents have been added to the composition, crystallization occurred mainly at the surface.

Regarding the fracture strength, there is an unexpected 35% drop from AAS glass to AAS-GC. The surface finish was the same for all samples, there was an increase in Vickers hardness, and the indentation fracture toughness is similar; this could be explained by the presence of tension stresses, probably due to local internal tensile stresses induced by the crystalline phase. It is known that when anisotropic crystallization occurs, triaxial tensile stresses can develop at the tip of the crystals, which could be the cause of the accentuated decrease in fracture strength.

4.2. Ion-Exchanged AAS Glass Ceramic Samples

The density measurements reveal a decrease in density for the glass ceramic samples from 2.49 to 2.46 ± 0.01, while the ion exchange treatment does not seem to affect the density values. This indicates that the density of the crystalline phase is lower than that of the glass, while the replacement of Li ions by K ions does not affect density. This is consentaneous with the XRD data, since heat treatments up to 600 °C reveal the presence of LiAlSi₂O₆, a phase that in the most common structure is called β-spodumene or spodumene-II present in the Li₂O–Al₂O₃–SiO₂ ternary equilibrium system and with a density of 2.374 g∙cm⁻³ [28].

Regarding the ion exchange process, it was already concluded in a previous work [19] that this chemical treatment does not affect the transmission of the samples, and the same was also observed in the present work (Figure 5). To study the effect of ion exchange in the AAS-GC samples, XRD measurements were also made in bulk glasses (Figure 7). Although an amorphous halo is still observed, the AAS-GC prepared at 575 °C for 1 h shows crystalline peaks with a higher intensity when compared with the powdered samples (Figure 3), with a correspondent crystallite size (calculated for the 25.7 ° peak of Figure 7) of ~59 nm (~27 nm for the IE-AAS-GC). The Raman data corroborate these results, since they also indicate the presence of crystalline phases for bulk samples heat-treated above 575 °C, as observed in Figure 4.

From the empirical calculations based on the amorphous bands present in Figure 7, where the area of the amorphous band in the AAS glass represents 100% amorphous and therefore 0% crystalline, and calculating the correspondent area for the AAS-GC result, the crystalline percentage can be estimated as 17%. The amorphous band area of the IE-AAS-GC sample indicates that the ion exchange process decreases the crystalline fraction from 17% to 14%. In fact, the ion exchange process modifies both the amorphous and crystalline phases, and an increase in the amorphous content is observed. This effect has been reported in the literature [8,18,29,30].

Table 4 shows the evolution of hardness over the various ion exchange treatments applied to the GC samples, and it is possible to conclude that there is a cumulative increase in the Vickers hardness of the ion-exchanged AAS-GC compared to the AAS-GC and AAS glass. By analyzing these results, the hardness for the IE-AAS-GC samples reached a peak at 12 h at both temperatures and then decreased. The decrease in hardness, more accentuated at higher temperatures and/or longer ion exchange times, is due to a viscoelastic relaxation phenomenon combined with a surface structural relaxation phenomenon [11,31]. Based on these results, the ion exchange condition that greatly enhances Vickers hardness is 12 h at 450 °C, but the analysis must again be complemented with the fracture toughness results.

The K_C values for ion-exchanged AAS-GC (Table 5) were difficult to obtain since most indentations are irregular and not compliant with the norm (c/a ratio above 2.5). Although only the result for 450 °C for 12 h (1.86 ± 0.1 MPa.m^1/2) followed the norm, all indentations required higher loads and/or times than the ones for AAS and AAS-GC. This makes it hard to compare the results, but it does indicate a stronger resistance of the ion-exchanged AAS-GC samples.

Table 7. Comparison of properties obtained for the AAS glass and the ion-exchanged AAS glass at 450 °C for 12 h [19], with the data obtained in this work for the AAS glass ceramic (575 °C for 1 h), and the AAS glass ceramic ion-exchanged at 450 °C for 12 h.

	AAS Glass [19]	IE–AAS Glass (450 °C—12 h) [19]	AAS GC (575 °C—1 h)	IE–AAS–GC (450 °C—12 h)
ρ [g/cm3] (±0.01)	2.49	2.49	2.46	2.46
T [%] (500–550 nm)	91	91	90	88
HV0.2	620 ± 10	716 ± 13	710 ± 14	773 ± 23
E [GPa]	87 ± 1	87 ± 1	85 ± 1	86 ± 1
G [GPa]	35 ± 1	35 ± 1	34 ± 1	35 ± 1
ν (±0.01)	0.23	0.23	0.24	0.23
KC [MPa.m1/2]	1.7 ± 0.1	2.2 ± 0.1	1.6 ± 0.1	2.0 ± 0.2 **
σf [MPa]	310 ± 84	468 ± 169	201 ± 92	362 ± 84
Weibull modulus	4.27/4.02 *	2.47/3.20 *	2.89/2.32 *	4.85/5.06 *

* The first value indicates the m_Wiebull obtained by linear regression, and the second value is obtained by the maximum likelihood method (ASTM, C1239-06 [24]). ** The value obtained for these samples required a higher load condition (196 N, 15 s).

presents a compilation of the mechanical properties determined for the different treatments performed on the glass composition studied: AAS glass and AAS glass ion-exchanged at 450 °C for 12 h (IE-AAS) [19], AAS glass ceramic (AAS-GS) heat-treated at 575 °C for 1 h, and the ion-exchanged AAS glass ceramic (IE-AAS-GC) at 450 °C for 12 h. As observed, the elastic constants (E; G; ν) are constant regardless of the treatments, while density and transmission in the visible region present small variations. On the other hand, Vickers hardness, indentation fracture strength, and fracture strength are strongly dependent on the treatment. From Table 7, one can conclude that the thermal and chemical treatments present cumulative results regarding hardness. In fact, the Vickers hardness goes from 620 ± 10 in the AAS glass to 710 ± 14 in the AAS-GC (575 °C for 1 h) and to 773 ± 23 for IE-AAS-GC (450 °C for 12 h), corresponding to cumulative hardness increases of ~15% and ~25%, respectively.

Regarding the K_C values obtained in similar conditions (49 N and 15 s), the ion-exchanged AAS glass presents the highest value (2.2 ± 0.1 MPa.m^1/2). Although the ion-exchanged AAS-GC required a much higher load to obtain a measurable indentation, the measured K_C presents a similar value (2.0 ± 0.2 MPa.m^1/2), indicating that the chemical treatment has the highest effect on the fracture toughness. Similar conclusions can be drawn from the fracture strength results, since only the ion exchange process enabled the increase in σ_f. In fact, the experimental fracture strength of the AAS glass decreased ~35% after the heat treatment. Since the average K_C values of the AAS-GC samples are statistically equal to the parent glass values, the only possible explanation is that the crystalline phase caused tensile internal stresses, which lowered the experimental values, instead of inducing compression stresses, thus enhancing the glass properties, as in the case of the ion exchange treatment.

The calculated Weibull modulus indicates a lower dispersion of the results for the ion-exchanged AAS–GC, and the size of the critical defects (estimated from Equation (2)) is ~10 um, which is of the order of the layer depth yielded by ion exchange for a 12 h treatment [19].

5. Conclusions

An aluminosilicate glass composition was heat-treated under different conditions in order to promote transparent glass ceramics. These glass ceramics were studied in terms of their transmission in the visible region, Vickers hardness, fracture strength, and indentation fracture toughness. The glass ceramic samples obtained by heat treatment at 575 °C for 1 h presented the highest increase in Vickers hardness while maintaining a visible transmittance and indentation fracture toughness similar to the parent glass (~90%). However, a decrease in fracture strength was observed. These glass ceramic samples were then submitted to ion exchange treatments in order to study the evolution of the mechanical properties for different time–temperature conditions.

Hardness is the property that undergoes a stronger effect, since ceramization and ion exchange, alone, yielded an increase in hardness of around 15% each. However, when the two treatments are combined, there is a cumulative increase in hardness of the order of 25%, reaching a maximum of 773 ± 23 HV.

It was also concluded that the percentage of crystallinity in these glass ceramics after the thermal treatment of 575 °C for 1 h reaches 17%, decreasing to 14% after ion exchange, where the nanocrystals are of the order of ~27 nm.

The indentation fracture toughness (K_C) of the ion-exchanged samples increased, both for the AAS glass and the AAS glass ceramic, but there was no strong difference between them, possibly due to the introduction of triaxial stresses during the heat treatment that cause the glass ceramics to fracture at lower tensions.