Optimized sintering and mechanical properties of Y-TZP ceramics for dental restorations by adding lithium disilicate glass ceramics

The novel dental ceramics can be fabricated at lower temperatures when sol-gel derived lithium disilicate glass ceramics (LDGC) was used as an additive for yttria stabilized tetragonal zirconia polycrystalline (Y-TZP) ceramics. The effect of LDGC on the sintering, mechanical, and translucent properties of Y-TZP ceramics was investigated in the present study. The results showed that the LDGC additive effectively improved the densification of Y-TZP at 1100 °C, which was much lower than the sintering temperature for pure Y-TZP. When sintered at 1100 °C, the Y-TZP with 1 wt% LDGC reached a relative density of 95.45%, and prossessed a flexural strength of 482.4 MPa and a fracture toughness of 5.94 MPa·m1/2. Moreover, its translucency was also improved. While, the addition of LDGC could result in an escape of yttrium atoms from the grain lattice of zirconia, which induced the tetragonal-monoclinic transformation of zirconia and abnormal growth of monoclinic grains. The escaped yttrium atoms diffused into the intergranular glass phase. The results indicated that the novel Y-TZP-LDGC ceramics has a great potential to be used for all-ceramic restorations.

Zirconia was initially introduced as a biomaterial for the total hip replacement in 1988 [10]. At present, zirconia plays an important role in all-ceramic restoration system. It is well-known that zirconia has much higher fracture toughness than other ceramics used in dentistry, owing to the effect of transformation-toughening phenomena associated with the martensitic transformation, i.e., tetragonal to monoclinic phase (t-m) [6,10]. The volume expansion ranging from 3% to 5% of zirconia grains after this phase transition induced by localized stresses can result in the significant compressive stresses on the propagating cracks. The hindered crack propagation can effectively increase the fracture toughness of Y-TZP [11,12]. Y-TZP restorations are usually manufactured via a solid-state sintering at the temperature above 1500 [13,14]. The manufacturing ℃ temperature is higher, and the product's cost is more expensive, which will inevitably limit its commercial applications [15]. The most common method to decrease the manufacturing temperature for ceramics is to use sintering additives. Actually, several researches have reported about the performance of additives on the sintering properties of Y-TZP. Those results showed that the densification of Y-TZP could be effectively promoted due to the viscous flow of glass phase during sintering, such as CaO-Al 2 O 3 -SiO 2 (CAS) and Li 2 O-Al 2 O 3 -SiO 2 (LAS) [13,[16][17][18][19][20].
Lithium disilicate glass ceramics (LDGC) is polycrystalline material consisted of lithium disilicate crystalline grains and glassy phases [21]. Due to the glassy matrix and the interlocking microstructures of the needle-like lithium disilicate crystalline grains, LDGC presents a good translucency while maintaining a high strength (> 350 MPa) and a fracture toughness (2.5-2.8 MPa·m 1/2 ) [22]. These properties have a decisive influence on the reliability and esthetics of a restoration product. Currently, LDGC has been widely used for dental restorations, including inlays, crowns, and bridges [23][24][25]. Therefore, using LDGC as a sintering additive to improve the sintering performance of Y-TZP is worth of an expectation. Moreover, the addition of glass ceramics with excellent translucency is also a potential way to improve the characterized chalk color of Y-TZP [8,21,26].
In the present study, LDGC fabricated by sol-gel method was introduced into Y-TZP to optimize its mechanical and optical properties. The influence of LDGC addition on the sintering behavior, microstructure, and mechanical properties of Y-TZP had been investigated thoroughly. The optical properties of LDGC added Y-TZP ceramics were also characterized. Y-TZP with 0 wt%, 1 wt%, 3 wt%, and 5 wt% of LDGC were prepared in this study, with sample codes of 100Z, 99Z1LD, 97Z3LD, and 95Z5LD, respectively. Y-TZP powders (TZ-3YBE, Tosoh Co., Japan) and pre-prepared LDGC powders were mixed by ball milling in ethanol for 3 h. After ball milling, the mixed powders were dried and screened through 200-mesh sieve, and then compacted into rectangular blocks with a dimension of 30 mm × 40 mm × 6 mm, followed by a cold isostatic pressing at 250 MPa. Finally, the compacts were sintered at 900-1100 for ℃ 2 h in air atmosphere with a heating rate of 5 /min. ℃

2 Characterizations
Density of the samples was measured by the Archimedes method with distilled water as the medium. The differential scanning calorimetry (TG/DSC) analysis was conducted by a simultaneous TG/DSC apparatus (STA409 PC, Netzsch, Germany) with a heating rate of 10 /min under nitrogen atmosphere (50 ℃ mL/min). The morphology and chemical analyses of the samples were studied by scanning electron microscope (SEM; S-4800N, Hitachi, Japan) and energy-dispersive X-ray spectroscope (EDS; Magellan 400, FEI, USA) after being polished and thermally etched at 1000 for ℃ 30 min. The grain size was measured from the SEM micrographs using Nano Measurer software according to the linear interception method. The number of measured grains is about 200. More detailed microstructural features were investigated by transmission electron microscope (TEM; JEM-2100F, JEOL, Japan). The phase analysis was conducted by X-ray diffractometer (XRD; D/MAX-RBX, Rigaku, Osaka, Japan) with Cu Kα radiation (λ = 0.15418 nm). Each group of the samples was tested for 3 times. The fraction of the monoclinic phase in ZrO 2 , X m , was calculated by using the Gravie and Nicholson method [27] The theoretical densities of the composites can be calculated with the rule-of-mixtures: where t  is the theoretical density of the composite ceramic, l  and z  are the theoretical densities of LDGC (2.5 g/cm 3 ) and Y-TZP (6.10 g/cm 3 ) [29], respectively, and l x is the weight percentage of the LDGC additive in the composite ceramic.

3 Mechanical properties
Both flexural strength and elastic modulus were tested on a universal testing machine (Instron 5566, Instron Co., USA), and measured by the three-point bending method with a loading rate of 0.5 mm/min and a support span of 20 mm. The samples were cut and polished into rectangular bars with the size of 3 mm × 4 mm × 26 mm. Six bars were tested for each group.
The flexural strength was calculated by where f  is the flexural strength, P is the breaking load, l is the length of the support span, and b and h are the width and thickness of the bar, respectively.
The fracture toughness was determined by singleedge-notched beam method with a loading rate of 0.05 mm/min and a sample span of 20 mm. This test was also carried out on a universal test instrument (Instron 5566, Instron Co., USA). The fracture toughness was calculated by where IC K is the fracture toughness, P is the load, l is the span length, a is the depth of the notch, and w and b are the width and thickness of the bar, respectively.
The Vickers hardness was determined by a Vickers indentation hardness tester (TUKON-2100B, Instron Co., USA). For each sample, 5 indentations with a load of 1 kg were measured on finely polished surface for 10 s.

1 Phase characterization
The fabricated LDGC powder was studied by means of SEM, XRD, and DSC techniques. Figure 1 shows the SEM micrograph, XRD pattern, and DSC profile of the as-prepared powder. The powder, showing particle aggregation in the SEM micrograph, consists of both amorphous phase and crystallization phase, which is characteristic of glass ceramics, where the diffraction www.springer.com/journal/40145 peaks of crystalline can be readily indexed to lithium disilicate (PDF #72-0102). The DSC profile shows that the melting point of LDGC is 980 , which means ℃ that LDGC powder will be melted into liquid phase when the sintering temperature is higher than 980 .
℃ XRD profiles of the Y-TZP-LDGC samples with different content of LDGC addition sintered at different temperatures are shown in Fig. 2. It is obvious that there coexist the dominant crystalline phase of tetragonal ZrO 2 (t-ZrO 2 ; PDF #80-0965) and the residual phase of monoclinic ZrO 2 (m-ZrO 2 ; PDF #86-1451). The fraction of m-ZrO 2 is calculated and listed in Table 1. Generally, the m-ZrO 2 fraction increases as a function of the LDGC content. This trend is most evident for the 1100 sintered samples. ℃ After sintering at 1100 , the m ℃ -ZrO 2 fractions are only 14.0 and 13.2 vol%, in the samples 100Z and 99Z1LD (0 and 1 wt% of LDGC, respectively). However, when the amount of LDGC addition is 3 wt%, the fraction of m-ZrO 2 increases to 32.7 vol% sharply. Finally, the m-ZrO 2 fraction reaches at 31.3 vol% in the sample with 5 wt% LDGC addition after sintering at 1100 . ℃
The results are shown in Fig. 3. The theoretical density of the Y-TZP-LDGC composites is calculated from the theoretical densities reported for tetragonal Y-TZP and LDGC. When the sintering temperature is 900 , ℃ the relative density of Y-TZP is 50.66%. However, with 5 wt% LDGC addition, the relative density increases linearly to 61.24%, as shown in Fig. 3(a). When sintered at 1000 , the rel ℃ ative density of Y-TZP without LDGC is 56.93%. With 3 wt% LDGC addition, the relative density of the composite rapidly   increases up to 85.94%. After sintered at 1100 , the ℃ relative density of samples greatly increases from 81.58% to 95.45% with only 1 wt% of LDGC addition. However, with a further increase of LDGC addition, the relative density gradually decreases, for example, 91.25% with 5 wt% LDGC addition.
The flexural strength increases from 37.0 to 76.6 MPa, the elastic modulus from 13.2 to 31.8 GPa, and the hardness from 0.6 to 1.5 GPa, for the samples sintered at 900 with LDGC addition increasing ℃ from 0 to 5 wt%. The flexural strength of samples with 5 wt% LDGC addition achieves a maximum value of 363.7 MPa, and the elastic modulus and hardness reach to 77.4 and 7.85 GPa, respectively, when the sintering temperature reaches at 1000 . While the flexural ℃ strengths are 483.8 and 482.4 MPa, when the LDGC additions are 0 and 1 wt% at 1100 , respectively. ℃ However, the flexural strength decreases with more of LDGC content. Consistent with flexural strength, the highest elastic modulus and hardness also reach at 81.9 and 11.13 GPa, respectively, when the LDGC addition is 1 wt%. When the LDGC content is further increasing, the elastic modulus and hardness gradually decreased. Furthermore, a similar tendency is observed in the fracture toughness at 1100 as shown in Table 2. The ℃ sample 99Z1LD, with 1 wt% LDGC addition, presents the highest fracture toughness of 5.94 MPa·m 1/2 .
The fracture surfaces of samples sintered at 1100 ℃ were studied by SEM as shown in Fig. 4. In Fig. 4(a), it could be found that there are pores at the triplejunction among ZrO 2 grains. In sample 99Z1LD, the pores among ZrO 2 grains disappeared, which means 99Z1LD has a higher relative density than 100Z. In samples 97Z3LD and 95Z5LD, macro pores can be observed on the fracture surfaces. The existence of these pores reduces the relative density and has a negative effect on the mechanical properties of the samples.
It is noted that the addition of LDGC contributes to the densification of Y-TZP. The melt-out effect of added LDGC at grain boundaries can facilitate the atom diffusion and promote the densification of samples through liquid sintering. In addition, the increment of temperature can reduce the viscosity of glass phase, which can also promote the atom diffusion and thus  facilitate the densification. Therefore, with the increment of LDGC addition and sintering temperature, the relative density of the samples gradually increased. However, when the LDGC addition is more than 3 wt%, the relative density of the samples sintered at 1100 will ℃ be decreased. This is induced by the pores within the ceramics (Figs. 4(c) and 4(d)) and the volumetric expansion caused by the t-m transformation of ZrO 2 grains [18,30].
When the sintering temperatures are 900 and 1000 , ℃ the relative density of Y-TZP-LDGC samples increases with the increment of LDGC content, and the flexural strength, elastic modulus, and hardness of the samples increase accordingly. After sintering at 1100 , ℃ the sample with 1 wt% LDGC addition has the highest values of relative density, elastic modulus, hardness, and fracture toughness at the same time. The above results show that the flexural strength, elastic modulus, and hardness of Y-TZP-LDGC samples follow the same trend and strongly correlate with the variation of relative density, which indicate that the density is a dominate factor for the mechanical properties of Y-TZP-LDGC samples. It is well-known that the lower density of the ceramic indicates more internal pore defects inside (Fig. 4). The pore defects can cause the stress concentration and reduce the cross-sectional area of the material, which is unfavorable for the mechanical properties [11,12]. In addition, the volume expansion caused by t-m transformation may result in pores and micro-cracking, and thus the increasing m-ZrO 2 content has adverse effects on the mechanical properties. Moreover, the excessive intergranular glass phase may negatively affect the mechanical properties of Y-TZP, since the mechanical properties of LDGC are lower than that of Y-TZP. Therefore, after sintering at 1100 , although the density of 97Z3LD and ℃ 95Z5LD is higher than that of 100Z, the flexural strength of them is lower than that of 100Z [15,31].
For zirconia-based materials, stress-induced t-m transformation is the primary toughening-reinforcing mechanism. Therefore, the content of t-ZrO 2 plays an important role in the fracture toughness of Y-TZP-LDGC samples. In the present work, the density of samples is increased with the addition of LDGC, but the content of t-ZrO 2 will be reduced. For the samples sintered at 1100 , the relative density of 99Z1LD is ℃ higher than that of 100Z, while its m-ZrO 2 content is close to that of 100Z, and thus the fracture toughness of 99Z1LD is higher than that of 100Z. Although the relative density both of 97Z3LD and 95Z5LD is higher than that of 100Z, the fracture toughness of them is lower than that of 100Z, which may be caused by the higher m-ZrO 2 content. Figure 5 shows the SEM micrographs of the original surfaces of samples sintered at 1100 . In ℃ Fig. 5(a), the sample 100Z has a uniform grain size and some pores can be observed on its surface (indicated by yellow open circles). With 1 wt% LDGC addition, pores disappear on the sample surface. When the LDGC addition increases to 3 wt%, a thin glass phase appears at the grain boundaries (by the yellow arrow). As the LDGC addition further increases to 5 wt%, zirconia grains are gradually wrapped by the glass phase, and the thickness of the glass phase at the grain boundary is increased obviously. These results indicate that the LDGC gradually melts and fills the gaps among the ZrO 2 grains during the sintering process. Figure 6 shows SEM micrographs of the polished and thermal etched surfaces of Y-TZP-LDGC samples sintered at 1100 . It is clear in Fig. 6(a) that the size ℃ of ZrO 2 grains is uniform, and there are some intergranular pores in the sample. When 1 wt% LDGC is added, the intergranular pores disappear, without obvious changes in ZrO 2 grain size ( Fig. 6(b)). When the LDGC addition is more than 3 wt%, it is remarkable that there shows a bimodal size distribution of ZrO 2 grains (Figs. 6(c) and 6(d)); smaller ones are at nano scales (about 100 nm), while the larger ones are around 1-4 μm as indicated by white circles. Besides, the average diameter of both larger and smaller ZrO 2 grains increases gradually with the increasing content of LDGC addition. The shape of smaller ZrO 2 grains turns accordingly from polygon into roundness. Table 3 summarizes the average grain size of the samples  sintered at 1100 . ℃ According to liquid phase sintering theory, liquid phase can wet the solid particles, with resulting in capillary pressure in the space between particles, which can promote the densification by rearranging the particles to denser packing and increasing the mass transfer rate through dissolution-reprecipitation mechanism [13,[31][32][33]. Therefore, the intergranular pores eliminate in sample 99Z1LD and the relative density of Y-TZP increases with a small amount of LDGC addition. Furthermore, the increased mass transfer rate is conducive to grain growth. Thus, the average grain size of samples is increased as shown in Table 3.

3 Microstructural characterization
In fact, the LDGC additive mostly distributes at grain boundaries during sintering process. Therefore, the thickness of glass phase grain boundary increases with more addition of LDGC. This thick glass phase grain boundary would hinder the mutual contacts between grains and subsequently increase the diffusion path of grain boundary, resulting in limited mass transfer and low grain-growth rate. Therefore, when LDGC addition is more than 1 wt%, further increase in LDGC content leads to a decrease in relative density. Additionally, the similar result is also observed by Tekeli in SiO 2 /8YSCZ (8 mol% yttria-stabilized cubic zirconia) composites, in which the overdose SiO 2 addition has been proved to retard the densification of 8YSCZ by hindering the grain growth [34,35].
The detailed microstructure on micron and nano scale was analyzed by TEM imaging, electron diffraction as well as chemical analysis. Figure 7 gives the TEM bright field micrographs of 99Z1LD and 97Z3LD samples sintered at 1100 . It can be concluded that ℃ the grain size of 99Z1LD sample is uniform and the phase of marked grain is demonstrated to be tetragonal ZrO 2 (Figs. 7(a) and 7(b)). However, two types of ZrO 2 grains with different sizes are present in sample 97Z3LD (Figs. 7(c) and 7(d)), which is consistent with the SEM results (Fig. 6(c)). Furthermore, the selected area electron diffraction (SAED) patterns have confirmed that the larger ZrO 2 grains are the monoclinic phase, while smaller ones are cubic (Figs. 7(c)-7(f)).
Elemental mapping results on polished surface of the 1100 sintered samples are shown in Fig. 8. Si ℃ distribution reveals the existence of LDGC addition. At this magnification, element distribution of Zr, Y, and Si looks homogeneous in the sample of 99Z1LD. However, for 97Z3LD and 95Z5LD samples, it is obvious that the distribution of all the elements is non-uniform, and larger size grains lacking of Y and Si elements (white dashed ellipse in Figs. 8(b) and 8(c)) are observed, which further confirm the formation of larger-size ZrO 2 . Elemental analysis was also performed to fully understand the distribution of glass addition at a higher magnification in TEM. The Si distribution in Fig. 8(d) indicates that the glass addition is mainly distributed along ZrO 2 grain boundaries. Meanwhile, the linear variation of element content in   Fig. 8(e) shows that the glass addition at the grain boundary riches in yttrium. Therefore, it can be concluded that the Y element may diffuse from the larger-size ZrO 2 into intergranular glass. The consequence is that the lack of Y in larger-size ZrO 2 will directly induce the phase transition from t-ZrO 2 to m-ZrO 2 . It is well-known that Y 2 O 3 is the stabilizing agent of the zirconia. The content and distribution of Y have a direct influence on the phase form of zirconia [36]. Thus, it is reasonable to suppose that the lack of www.springer.com/journal/40145 Y in larger-size ZrO 2 will directly induce the phase transition from t-ZrO 2 to m-ZrO 2 .
Several studies have shown that the introduction of bio-glass (such as 64S, CaO-P 2 O 5 -SiO 2 -MgO, and SiO 2 -Na 2 O-CaO-P 2 O 5 ) into Y-TZP can induce a t-m transformation and lead to an increase in the m-ZrO 2 fraction [30,31,37]. It was explained that this could be caused by the inhomogeneous spreading of liquid phase during the sintering process. In fact, it is difficult to achieve the uniform mixing of precursor powders due to the agglomeration of LDGC powders ( Fig. 1(a)). The recrystallization of the inhomogeneous intergranular glass phase and the mismatch in coefficient of thermal expansion between the glass phase and zirconia matrix will cause considerable localized stress and structure changes during cooling process, which may also induce the undesirable phase transition [18].
Although it has been reported for several times that the m-ZrO 2 fraction of Y-TZP increases when glasses used as the additive, the phenomenon of Y segregation has been rarely reported [13,16,18,30,31,35]. Yttrium silicates (Y-Si-O), or the Y 2 O 3 -SiO 2 binary system, including Y 2 Si 2 O 7 , Y 2 SiO 5 , and 2Y 2 O 3 ·3SiO 2 , have been widely investigated and used as high-temperature structural ceramics [38][39][40]. Y 2 SiO 5 can be obtained by calcining Y 2 O 3 and SiO 2 when lithium used as an additive [38,41]. In the Y 2 O 3 -SiO 2 binary system, the additive of lithium can lower the reaction temperature of Y 2 O 3 -SiO 2 by eutectic reactions. Apparently, the introduction of lithium is beneficial to the formation of Li-Si-Y-O liquid phase and can promote the reaction between Y 2 O 3 and SiO 2 . In addition, the previous work has shown that several yttrium silicates, including Y 4.67 (SiO 4 ) 3 O, Y 2 SiO 5 , and Y 2 Si 2 O 7 , are obtained in the Y 2 O 3 -SiO 2 system with a lithium additive, after a calcination at 1025 [41]. This might be the reason ℃ why Y element diffuses away from ZrO 2 grains to the intergranular lithium-containing glass phase.
The t-ZrO 2 fraction depends both on the grain size and Y content in samples at room temperature. For 3Y-TZP (3 mol% yttria-stabilized tetragonal zirconia) used in the present work, the critical grain size of t-ZrO 2 is about 1 μm, which means the grains larger than 1 μm would transfer to m-ZrO 2 during the cooling process [10]. This is in good agreement with the TEM results observed in the present work (Figs. 7(d) and 7(e)). Moreover, the Y segregation in the Y-TZP-LDGC samples should be noted. It is easy to find that there is more Y in smaller grains, and less in larger grains. Since the zirconia stabilized by higher Y content (about 8 mol%) can be cubic phase at room temperature [34,42], c-ZrO 2 grain is observed in the region of smaller grains with higher Y concentration (Figs. 7(d) and 7(f)) [34]. Due to the limited total content of Y in the starting powders, the content of c-ZrO 2 grains is too low to be detected by the XRD analysis.

4 Translucency
As can be seen from Fig. 9, the significant difference in translucency is identified among the Y-TZP-LDGC samples sintered at 1100 . The optical images show ℃ that the additive of LDGC has a significant impact on the translucency of the samples. Specifically, the addition of 1 wt% LDGC can greatly increase the translucency of the sample. However, with a further increase of LDGC, the translucency gradually decreases.
Limited translucency is a major drawback of Y-TZP and directly influences the appearance of dental restorations [8,43,44]. The main factor affected the translucency of Y-TZP is the light scattering, mainly at pores, grain boundaries, or second phases. In the present Y-TZP-LDGC system, the addition component LDGC performs an excellent translucency. According to the above results, the additive of LDGC can eliminate the pores in Y-TZP (as shown in Figs. 5 and 6). Therefore, the addition of LDGC improves the translucency of Y-TZP. However, when the addition is more than 1 wt%, the density of samples decreases dramatically, and the translucency will be correspondingly decreased.

Conclusions
The present work investigates the influence of LDGC addition on the sintering, mechanical, and translucent properties of Y-TZP ceramics. The results show that the addition of LDGC can significantly promote the densification of Y-TZP ceramics through liquid phase sintering at relatively lower temperatures. Meanwhile, the addition of LDGC can also cause the yttrium atoms to escape from the zirconia grain lattice to the intergranular glass phase, resulting in an increase of m-ZrO 2 fraction in Y-TZP ceramics and a decrease of mechanical properties at the same time. Moreover, the diffusion of Y atoms from zirconia grain lattice can also result in an abnormal growth of monoclinic grains. When sintered at a relatively low temperature of 1100 , Y ℃ -TZP added with 1 wt% LDGC would possess a flexural strength of 482.4 MPa and a fracture toughness of 5.94 MPa·m 1/2 . It also exhibits a better translucency, which is a promising material for all-ceramic dental restorations.