*2.3. Sintering Process: Thermal Cycles*

The green samples were debinded and pressureless solid state sintered (SSS) in flowing air (LINN Elektronik HT—1800 VAC, LINN HIGH THERM GMBH, Hirschbach, Germany). Samples were dewaxed with a cycle up to 800 ◦C (10 ◦C/h ramp) in flowing air and then pressureless sintered in flowing air in the range of 1450–1570 ◦C for different holding times (2–80 h) depending on the composition. The dewaxing and sintering steps for the 3D-printed green bodies were performed at 1550 ◦C for ATZ for 1 h after a debinding step performed at 800 ◦C.

In addition, the two-step sintering (TSS) process was also tested. In this case, T1 and T2 were in the ranges of 1400–1500 and 1350–1450 ◦C, respectively, with zirconia and ATZ.

#### *2.4. Physical, Microstructural and Mechanical Characterization*

Sintered density was determined by Archimedes' method.

Diffraction patterns were collected by using a Philips X-ray powder diffractometer with Bragg–Brentano geometry and Cu Kα radiation (40 kV and 35 mA) to identify the crystalline phases in the sintered samples and to evaluate the tetragonality of the tetragonal phase.

Viscosity measurements were performed using a Malvern Kinexus Pro+ rheometer (Kinexus pro+, Malvern Instruments, Ltd., Worcestershire, UK) at 25 ◦C with cone-plate geometry (4◦, 40 mm) in shear rate control from 0.1 to 300 s<sup>−</sup>1.

The microstructural analysis of both the surface and cross sections of sintered bodies was performed with a scanning electron microscope (SEM, LEO 438 VP).

The flexural strength was determined at room temperature with four-point bending tests (five tests for each composition). Samples, in the form of 2 × 2.5 × 25 mm bars, were prepared and tested in accordance with the standard ENV843-1:2004 (cross head speed of 0.5 mm/min and support span of 20 mm). Hardness (*Hv*) was determined by means of Vickers indentation with a load of 9.8 N, while fracture toughness (*KIC*) was determined by means of Vickers indentation with a load ranging from 9.8 to 98 N. To calculate fracture toughness, the formula proposed by Niihara [45] for Palmqvist cracks was used (Equation (1)):

$$K\_{IC} = \frac{0.035 \left(H\_{\overline{v}} a^{0.5}\right) \left(3 \frac{F}{H\_v}\right)^{0.4} \left(\frac{l}{d}\right)^{-0.5}}{3} \text{ for } 0.25 < \frac{l}{a} < 2.5 \tag{1}$$

where *a* is the indent half-diagonal, *E* is the Young's modulus, and *l* is the Palmqvist crack length.

#### **3. Results and Discussion**

#### *3.1. Stabilization of Zirconia: Variables That Influence Transformability*

3.1.1. Type of Stabilizer

Different zirconia-based materials were produced and characterized to study the effect of the type of the stabilizer on the t–m phase transformation. The dopants could be classified according to their oxidation state (Cu2+, Y+3, Ce4+, and Ta5+). More precisely, they are stabilizers of the tetragonal phase (Y2O3 and CeO2) and toughening oxides (Ta2O5 and CuO) [46].

Cations' valence and size affect the stabilization mechanism of the tetragonal phase [47–50], even if the correlation is not univocal, as suggested by Yoshimura et al. [51].

The phase composition and crystallographic parameters were evaluated for each mixture (lattice constants c and a and their c/a ratio, namely "tetragonality") of doped tetragonal zirconia. The values of fracture toughness and hardness were also determined (Table 2).


**Table 2.** Properties of zirconia-based materials doped with different stabilizers and toughening oxides. DR is the relative density, c/a is the tetragonality, KIC the fracture toughness, and HV the microhardness.

Yttrium oxide (Y2O3) is the most common stabilizer of the tetragonal phase, and Y-TZP is widely used due to its strong mechanical properties [26,52]. The addition of different amounts of yttria influenced the c/a ratio, which indicated the transformability of the available tetragonal phase. In comparing the values in Table 2 for the mixtures of Zr2Y and Zr3Y, it is clear that a higher yttria content (3 vs. 2 mol%) led to a greater stabilization of the tetragonal phase, which corresponded to a decrease in the c/a ratio (1.0159 vs. 1.0166) and a lower toughness (4.0 vs. 9.4 MPa m1/2).

Cerium oxide (CeO2) is another well-known stabilizer of zirconia, and ceria-doped zirconia exhibits very high values of fracture toughness [53]. The ZrCe sample in our study showed toughness value four times higher than that of Zr3Y (18.4 vs. 4.0 MPa m1/2, respectively) as indicated in Table 2. CeO2 is a stabilizer as Y2O3, but its c/a ratio is higher; this means that the tetragonal phase is less stabilized, so its transformation is easier, thus leading to an increase in fracture toughness. On the other hand, as described in the literature [54], CeO2 does not allow one to obtain high values of mechanical resistance due to its limited capability to contain grain growth during sintering. Indeed, ZrCe grains are wider (ca. 2.0 μm) than Y-TZP ones (ca. 0.5–0.8 μm) [27]. As the oxidation state is the same of Zr4+, Ce4+ does not generate oxygen vacancies inside the ZrO2 cell, so, in a humid environment, the t–m spontaneous transformation is not promoted and CeO2-stabilized zirconia shows significantly high resistance to LTD [17,54,55].

Tantalum oxide (Ta2O5) is known in the literature for its toughening effect when added to 3Y-TZP [56,57]. In our study, the addition of Ta2O5 led to a higher value of fracture toughness than 3Y-TZP (9.4 vs. 4.0 MPa m1/2, respectively), as shown in Table 2. The addition of Ta2O5 to 3Y-TZP increased the c/a ratio (1.0173 vs. 1.0159, respectively) such that the chemical driving force for the t–m transformation was enhanced, and this led to a higher value of fracture toughness. On the other hand, the stabilizing effect of Y2O3 was contrasted by the addition of Ta2O5, which is a toughening oxide that increases the t–m martensitic transformation temperature [50], resulting in a toughening effect.

Copper oxide (CuO) was also tested as toughening agent for Y-TZP. The results reported in Table 2 show that the addition of CuO only led to a slight increase in the fracture toughness of the 3Y-TZP (4.6 vs. 4.0 MPa m1/2, respectively). This result is in contrast with the results reported by Ramesh et al. [58], where a different Y-TZP powder was used.

After comparing the fracture toughness values (Table 2) as function of the tetragonality, a linear correlation was obtained, as shown in Figure 1. If the c/a ratio of the tetragonal phase was near 1 (i.e., the c/a value of the cubic phase), the tetragonal phase was more stable and hence the t–m transformation became more difficult and the fracture toughness decreased. On the contrary, if the c/a ratio of the tetragonal phase increased up to 1.022 (which is the b/a value of the monoclinic phase), the t–m transformation was favored and the fracture toughness increased.

**Figure 1.** Fracture toughness vs. tetragonality of the different stabilizers in zirconia-based materials.

#### 3.1.2. Stabilizer Content

The effect of different contents of stabilizer (Y2O3) was studied in ZTA composites with a 60/40 alumina/zirconia weight ratio. The results reported in Table 3 show that the fracture toughness reached a maximum value of 6.2 MPam1/2 with the lowest amount of stabilizer (60402Y). The same results were previously observed in ZTA composites with a 50/50 alumina/zirconia weight ratio, as reported in Table 3 [59].

**Table 3.** Properties of ZTA materials doped with different amounts of stabilizer. DR is the relative density, c/a is the tetragonality, KIC the fracture toughness, HV the microhardness, and MOR is the four-point flexural strength.


<sup>1</sup> Reprinted with permission from ref. [59]. Copyright © 2021 Elsevier Ltd.

This behavior can be explained by the analysis of the variation of the tetragonal phase amount and the tetragonality with stabilizer content. In fact, 60403Y had a lower amount of tetragonal phase (less than 80%) and a lower tetragonality than those of 60402Y. This means that a lower quantity of tetragonal phase was available to the toughening t–m transformation in the 60403Y composite. Furthermore, in the same sample, the lower tetragonality enhanced the stability of the tetragonal phase, which caused a decrease in the fracture toughness. These observations are also in line with the study of Yoshimura et al. [60], which reported the dependence of the c/a ratio on stabilizer content.

#### 3.1.3. Critical Grain Size

3Y-TZP was sintered in six different conditions in order to highlight the effect of the grain size variation on tetragonality and, consequently, fracture toughness. The experimental results are reported in the Table 4.

After increasing the sintering time to 60 h at 1550 ◦C, the fracture toughness and grain size increased up to maximum values of 7.7 MPa m1/2 and 1.19 μm, respectively (Figure 2). Furthermore, a strong dependence between the fracture toughness and tetragonality was observed at the microstructural level. Indeed, with the increase in sintering time, tetragonality increased, i.e., the tetragonal cell instability grew. This instability, caused by the distortion of the cell, promoted the t–m transformation and a consequent increase in fracture toughness.


**Table 4.** Properties of zirconia-based (3Y-TZP) samples sintered with different thermal cycles. DR is the relative density, Dm the average grain size, c/a is the tetragonality, KIC the fracture toughness, and HV the microhardness.

**Figure 2.** Fracture toughness vs. average grain size of the different sintered zirconia-based materials.

The sample sintered at 1550 ◦C for 80 h was characterized by the lowest sintered density due to the formation of the monoclinic phase, and it showed many cracks. In fact, XRD analysis confirmed that all the samples were mainly constituted by the tetragonal phase with traces of the cubic phase, while the sample sintered at 80 h showed an increase in monoclinic phase content.

According to these data, the critical grain size can be estimated to be equal or greater than 1.19 μm for this 3Y-TZP material. This value is in agreement with the critical grain observed by Lange [32].

#### *3.2. Parameters That Influence Mechanical Properties*

The relationships between the microstructure and mechanical properties of Y-TZP ceramics have been extensively studied over the past four decades, and different effects have been identified.

It was demonstrated that the toughening effect, related to the t–m transformation mechanism in Y-TZP ceramics, is promoted by larger grain sizes [12,32–36]. On the other hand, some mechanical properties, including flexural strength, are known to be enhanced by fine microstructures [36,61,62].

Indeed, as grain size coarsens, the critical defect enlarges, thus leading to a strength decrease [63]. According to the Griffith (Equation (2)), strength (*σR*), fracture toughness (*KIC*), and failure origin size (*c*) are strictly connected and their control is necessary to obtain reliable structural ceramic materials.

$$
\sigma\_R \sim \frac{K\_{IC}}{\sqrt{\pi c}} \tag{2}
$$

Unfortunately, the best conditions (composition, grain size, transformability, etc.) to reach ceramic strength in zirconia-based materials are not the same for maximizing

fracture toughness, so the reliability of these ceramics comes from the compromise of these two properties [64].

Hardness is also influenced by microstructure [65]. Generally, hardness is strictly related to density, but no univocal correlation between hardness and grain size has been proven. The hardness values of 3Y-TZP do not show the influence of the grain dimension in the submicrometric range [66].

A typical method to obtain ceramics with fine microstructures and improved mechanical properties (flexural strength) is based on the application of innovative sintering processes that limit grain growth. Among the best known sintering methods to refine ceramic microstructures, the spark plasma sintering (SPS) [67,68] and microwave sintering (MWS) [69] methods are the most efficient.

A simple and cost-effective method for industrial applications to obtain near full dense ceramics with controlled grain growth is TSS (two-step sintering) [70], in which the sample is first heated to a higher temperature to achieve an intermediate density and then cooled down and held at a lower temperature until it is fully dense. This sintering method has been successfully applied for ZTA composites [71,72].

The effect of TSS on the 3Y-TZP and ZTA samples (Table 5) was studied and compared to that of classic SSS. In the case of 3Y-TZP, TSS showed an advantageous effect on grain size (almost halved), as shown in Figure 3a,b. However, TSS seemed to have no effect on the fracture toughness. This was probably due to two opposite and concomitant effects of TSS that compensate for each other. The grain size refinement contrasted with the toughening effect achieved when the grain size approached the critical value. On the other hand, the tetragonal phase obtained with the TSS was more transformable, as evidenced by the slight increase in the tetragonality. It is probable that the longer holding time at the higher temperature promoted the migration of the stabilizer (Y3+) [52]; hence, the yttria concentration within the tetragonal phase decreased and enhanced transformability.

**Table 5.** Properties of 3Y-TZP and ZTA materials sintered with the single step (SSS) or two-step cycles (TSS). DR is the relative density, KIC the fracture toughness, HV the microhardness, MOR is the four-point flexural strength, Dm is the average grain size (A refers to alumina and Z to zirconia grains), and c/a is the tetragonality.


Again, the flexural strength values were very similar despite the halved grain size. It is probable that the grain refinement obtained with TSS did not contribute to a decrease in critical defect size. In fact, as observed by Xiong et al. [73], the TSS method could yield the formation of thermodynamically stable large pores, thus showing its limit in eliminating last residual porosity (1–2%). The effects of grain size refinement and critical defect dimension compensate for each other, thus leaving the strength value unaltered (as also described by Trunec [62]).

In the case of the ZTA composites, the TSS method effectively limited grain growth (Figure 3c,d). Comparing two samples with the same stabilizer content, the grain size refinement resulted in a lower toughness, probably due to the average grain dimension being too far from the critical grain size. The strength values of the 60403Y samples were found to be similar, likely because the increase in the critical defect size was not sufficiently compensated for by the refinement of the microstructure, as suggested by Trunec [62]. For the 60402Y samples, dynamic pore coalescence occurred in the second step of TSS, which did not aid the elimination of residual porosity and had detrimental effects on bending strength [73].

**Figure 3.** SEM micrographs of zirconia samples 3Y-TZP SSS (**a**) and 3Y-TZP TSS (**b**), as well as of ZTA samples 60403Y SSS (**c**) and 60403Y TSS (**d**).

Finally, the bending strength was also influenced by the powder preparation technique. Using the freeze-drying technique to dry the slurry, the production of a homogeneous granulate without aggregates was achieved (Figure 4). This granulation process strongly influences the quality of a green and sintered body [74]; in our study, higher values of bending strength were obtained (872 ± 47 MPa for 60402Y TSS-FD; see Table 5).

**Figure 4.** SEM micrograph of 60402Y powder prepared with freeze-dry granulation.

### *3.3. New Manufacturing Techniques: 3D Printing*

The production of ceramic components via the DLP technique is strictly connected to the availability of a suitable ceramic slurry. Nowadays, the most important producers of vat polymerization printers commercialize feedstocks for their 3D printer models with limited possibility to access to other resins available on the market. Another problem for the AM of ceramics with the DLP technique is the low disposability of printable slurries filled with desired ceramic powders.

For the preparation of new resin–ceramic powder mixtures, one of the main problems related to the addition of a high content of ceramic powder to the photopolymeric resin is the increase in the viscosity of the mixture. This drawback was solved here by wisely selecting monomers with different functionalities and molecular weights. The shear thickening behavior that is commonly observed in high solid loaded suspensions was reduced by the use of an appropriate surfactant and a zirconia powder with a lower surface area (7 ± 2 m2/g). In this way, high content ceramic photocurable resins (see Materials and Methods section) with low viscosity, suitable for the DLP printing process, were prepared.

The shear viscosity for two ATZ resins is reported in Figure 5.

**Figure 5.** Photocurable slurry viscosity.

Prototypal dental endosseous implants were obtained via the DLP technique with the developed ZTA resin (Figure 6), which was sintered up to 1550 ◦C for 1 h and reached a final density of 96.8%.

**Figure 6.** Endosseous dental implant in ATZ printed with the DLP technique.

More complex shapes, as the lattice structure shown in Figure 7, were successfully printed with a final relative density of 98%. Layer-by-layer deposition is highlighted in Figure 8. SEM observations revealed a regular lattice structure profile, where the overlapping layers and their homogeneity in thickness were clearly visible. The slicing value was set to 50 μm and fell to 35 μm after sintering shrinkage. Nevertheless, the layer adhesion could be further enhanced to completely avoid the delamination defects partially present in these items.

**Figure 7.** Lattice structure in ATZ printed by the DLP technique.

**Figure 8.** Micrograph of the lattice structure profile.

These preliminary outcomes highlighted the possibility to develop resins with the required ceramic material and the feasibility to print ceramic materials with low cost and widely available DLP printers.
