Gelcast Zirconia Ceramics With High Strength and Simultaneously High Translucency for Dental Applications

Translucent zirconia represents a favourite material for monolithic ceramic dental restorations. However, materials approaches employed so far to improve the translucency of zirconia ceramics are accompanied by a significant decline in strength. Thus, we aimed to develop dental 3Y-TZP ceramics that can provide excellent strength and, simultaneously, enhanced translucency. In this investigation, machinable tetragonal zirconia ceramics based on fine mesostructured zirconia particles stabilized with 3 mol% of yttria and prepared by the gelcasting processing method were developed. Properties of sintered samples were characterised, namely: shrinkage, density, structure, surface roughness, hardness, biaxial strength, and total forward transmittance. Zirconia ceramics with an average biaxial strength of 1184 MPa and a total forward transmittance of 46.7% for a 0.5 mm thick sample at a wavelength of 600 nm were obtained. These ceramics exhibited homogeneous structure with grains sizes up to 620 nm and purely tetragonal phase composition. The developed ceramics provided a favourable combination of high translucency comparable even with the mixed cubic/tetragonal structure of a common 4Y-TZP, and very high strength that is achievable only in the pure tetragonal 3Y-TZP.


Introduction
The development of a manufacturing technology based on CAD/CAM milling of pre-sintered blanks in the shape of discs or blocks has enabled the production of all-ceramic full-contour restorations [1,2]. Among the ceramics used for CAD/CAM milling in dentistry, zirconia ceramics attract ever growing attention of dental professionals as well as patients. Tetragonal zirconia polycrystals stabilised with yttria (Y-TZP) are favourable because of their biocompatibility, chemical stability, high fracture toughness, and excellent flexural strength.
Unfortunately, the conventional tetragonal zirconia ceramics stabilised with 3 mol% of yttria (3Y-TZP) exhibit an opaque appearance, i.e. low optical translucency. As low translucency negatively affects the aesthetics of the restoration, the conventional zirconia ceramics have had to be veneered with more translucent glass-based materials known as veneering porcelain [3,4]. However, chipping and fracturing of the veneering porcelains have often been registered [5][6][7][8] and this shortcoming of veneered zirconia materials resulted in the introduction of and ongoing extensive investigation into monolithic Y-TZP restorations with improved translucency [9][10][11][12][13]. A measure of translucency is the total forward transmittance of light, i.e. the fraction of incident light intensity transmitted through the material in the forward direction [14]. The main reasons for limited translucency of tetragonal zirconia include high light reflection at the body surfaces, light absorption in the material, scattering at pores or inclusions, and scattering at grain boundaries of birefringent tetragonal zirconia [15,16].
Three methods for improving translucency have been reported [17][18][19]. The principle of all these methods consists in reducing the backscattering in the ceramics because other factors are difficult to control. The first method for reducing backscattering is based on the elimination of inclusions whose refractive index is different from the refractive index of the ceramic matrix. Not only impurities but also residual pores can be regarded as inclusions with a high scattering effect. Primarily, inclusions of a size similar to the wavelength of visible light significantly reduce translucency [14,16]. Thus, alumina, which is often used as a dopant for tetragonal zirconia ceramics [13,20], must be eliminated from translucent zirconia.
It was shown that an addition of 0.3 mol% of alumina to 3Y-TZP decreased the translucency of a 1 mm thick sample from ~38% to ~30% [20]. However, the absence of alumina makes sintering more difficult, decreases strength, and increases the propensity for low-temperature degradation [21][22][23]. The second method uses grain size adjustment in sintered tetragonal zirconia ceramics. Theoretical models predict that scattering at grain boundaries diminishes as the grain size becomes much lower than the wavelength of transmitted light [16,24,25]. On the other hand, in a very coarse-grained structure, the transmitting light interacts with fewer grain boundaries. Unfortunately, the optimum grain size for the best mechanical properties of common tetragonal 3Y-TZP usually lies between 300 and 600 nm, i.e. at grain sizes comparable with the wavelengths of visible light [26]. At smaller grain sizes, the transformation toughening effect decreases due to an "overstabilisation" of the tetragonal grains [26,27]. At larger grain sizes, the spontaneous transformation of tetragonal (t-ZrO 2 ) to monoclinic (m-ZrO 2 ) phase occurs [26,28,29]. In both cases, the mechanical properties are dramatically reduced. The third method increases translucency by decreasing the birefringence of zirconia ceramics. Zirconia ceramics stabilised with 4-5 mol% of yttria (4Y-TZP or 5Y-TZP) result in materials with a biphasic tetragonal/cubic structure with a varying content of the cubic phase of up to 50% [30][31][32]. The presence of the cubic phase (c-ZrO 2 ) in the tetragonal structure reduces the light scattering at grain boundaries and the tetragonal/cubic zirconia ceramics thus exhibit higher translucency compared with pure tetragonal zirconia ceramics. On the other hand, the c-ZrO 2 cannot undergo the transformation toughening effect, and the strength of tetragonal/cubic zirconia ceramics is lowered to 50-75% of the conventional high-strength tetragonal 3Y-TZP ceramics that can reach up to 1200 MPa in the 3-point bending [13,[33][34][35]. Recently, gradient multilayer pre-sintered discs have been developed with yttria content changing from 5 mol% at the top to 3 mol% at the bottom of the disc [36]. Although this disc design brings a better aesthetics of dental restorations, it exhibits a gradient in strength as well. The reliability of such structures is still questionable and need to be investigated in detail [37]. It follows from this short overview that an improvement of zirconia ceramics translucency always results in their strength deterioration. In applications of dental zirconia ceramics, dental professionals must always seek a compromise between translucency and strength. Other properties such as fracture toughness, hardness, abrasiveness, colouring, etc. can also be affected by the translucency improvement. Thus, an investigation into improving translucency without affecting other properties of zirconia ceramics, mainly strength, is still a challenging topic. In the case of multi-unit monolithic restorations with their demand for very high strength and reliability, the only way to increase translucency is the development of a pore-and impurity-free optimized tetragonal microstructure. Gelcasting is a colloidal processing method that can provide ceramic bodies with a more homogeneous and denser microstructure than can be obtained by uniaxial pressing or cold isostatic pressing, the common methods for the preparation of dental zirconia blanks [38]. The benefits of the gelcasting method arise from a regular and tight packing of well-dispersed particles [39,40]. However, the preparation of large machinable zirconia discs from nano-sized powders by the gelcasting method is a tricky process that needs to be investigated and carefully developed. Thus, in this investigation, we aimed to develop dental 3Y-TZP ceramics based on the gelcasting of zirconia nanoparticles that can provide excellent strength and, simultaneously, enhanced translucency. 6 were used as the linear monomer and cross-linker, respectively. The monomer ratio of the linear to the cross-linking monomer was 4:1. Polyethylene glycol 400 (8.07485, Sigma-Aldrich Chemie, Germany) as a drying additive (PEG) was used to improve the drying process of the gelled body. The content of individual components in suspension is listed in Table 1. The applied composition of the suspension and particularly the content of PEG were essential for defect-less production of large zirconia blanks and were based on an extensive investigation that goes beyond the scope of this article. Germany) with respect to pure monomers was added as a 2.5% water solution to initiate the polymerisation of dissolved monomers at room temperature [42,43]. After the addition of initiator, the ceramic suspension was cast into non-porous moulds and gelled under a nitrogen atmosphere to accomplish complete polymerisation during 2 h. Bodies in the shape of discs safely remove the drying additive (PEG) and other organic species from the particle compacts.

CNC milling and final sintering
A diamond-coated toroid mill (flat-end mill D-EPDR-2020-20-02 Epoch21, MMC Hitachi Tool Engineering Europe, Germany) with a corner radius of 0.2 mm and an end diameter of 2 mm was used for milling specimens for strength measurements from the pre-sintered ceramic discs (dia. 30 mm, thickness 6 mm). The milling parameters were as follows: depth of cut a p = 0.5 mm and width of cut a e = 0.8 mm for initial roughening, and a p = 0.1 mm and a e = 0.8 mm for finishing in the last step. Further details of milling can be found in our previous study [41]. The milled samples were sintered in an air atmosphere at a heating rate of 10 °C min -1 up to 800 °C, followed by a heating rate of 5 °C min -1 up to the final sintering temperature. Sintering temperatures of 1350, 1450, and 1550 °C with a 2 h dwell time were applied.

Evaluation methods
The specific surface area of the zirconia powder was measured by nitrogen adsorption according to the multipoint BET method (Autosorb iQ, Quantachrome, USA). Particle size distribution in the ceramic suspension was determined using dynamic light scattering (Zetasizer Nano ZS, Malvern, UK). The bulk densities of pre-sintered and sintered samples were determined in distilled water by the Archimedes method following the EN 623-2 standard [44], always for at least three samples. The relative densities were calculated using a theoretical density (t.d.) of 6.1 g cm -3 for t-ZrO 2 [45]. The densification curve was calculated from high-temperature dilatometer measurements (Linseis L75 Platinum edition, Linseis, Germany) by a procedure described in [46]. The presence of surface defects (such as pores, cracks or delamination) in sintered samples was determined using the penetration test described in [47]. were calculated from the maximum (fracture) force, disc thickness and the numerically calculated geometrical factor f for the given geometry and the combination of materials. The factor f was calculated and equalled to 2.05 for every group of specimens tested in our work.
More details of the biaxial strength measurement can be found elsewhere [50].  [53,54]. A profilometer tip radius of 2 µm was used. The data obtained by the profilometer were processed by the Gwyddion software [55].

Preparation and sintering behaviour of zirconia blanks
Images of the SZ zirconia powder used for the preparation of ceramic suspensions are shown in Fig. 1. Most of the primary particles were below 100 nm (Fig. 1a) and they were predominantly aggregated in bigger mesostructured particles of about 150-200 nm (as can be seen in the TEM image in Fig. 1b). These mesostructured particles consisted of individual crystals and exhibited mesoporous or dense structure [56]. This visual evaluation of the powder was supported by its high specific surface area of 11.8 m 2 g -1 , which corresponded to an average particle size of 83 nm. Fig. 2 shows particle size distribution in the ceramic suspension after the powder dispersion by ball milling.  It is shown that the pore size distribution in the particle compacts after binder removal with a minimum densification (800 °C/0 h) was very narrow, with the most frequent pore size of 60 nm, which was well below the size of the particles used. It follows from the comparison of particle and pore sizes that the particles in the gelcast bodies were homogeneously and regularly packed without any large interagglomerate pores. Moreover, the similar pore size distributions in bodies pre-sintered to higher temperatures (1000 °C/1 h, 1100 °C/1 h) demonstrated homogeneous sintering without preferential densification of denser particle domains. The preferential densification of denser domains is typical of inhomogeneously packed compacts prepared from agglomerated nanoparticles and is accompanied by pore opening in the first stage of sintering [57][58][59][60]. Fig. 3 shows the sintering behaviour of the zirconia sample pre-sintered at 1000 °C for 1 h. Bodies with this pre-sintering heat treatment exhibited the best machinability, as is discussed later. The gelcast zirconia bodies densified from an initial relative density of 54% (in the presintered state) to a relative density of 99% already at 1400 °C and reached almost full density at 1450 °C, with only negligible further densification during the dwell time. The pre-sintered machinable zirconia blanks densified extremely homogeneously. Table 2

Structure of sintered ceramics
The SZ samples sintered at temperatures in the range from 1350 to 1550 °C with a 2 h dwell time reached a relative density of 99% or more (see Table 3). Although the zirconia sample could be sintered to almost full density already at a temperature of 1350 °C, sintering at higher temperatures could still remove some residual pores (99.4 vs 99.8% t.d.). Well dispersed mesostructured zirconia powder coupled with the gelcasting method provided an easy-to-sinter bodies, as also confirmed by the dilatometric experiments. for a spontaneous transformation [26]. The grain size, density, and shrinkage after the sintering of all samples are summarised in Table 3. The XRD patterns of as-sintered ceramic surfaces are shown in Fig. 6 (observed data after background subtraction and K β elimination are presented). Only t-ZrO 2 was detected in all gelcast samples after sintering (the R wp parameters in the Rietveld analyses were ≤ 10%). The c-ZrO 2 often present in 3Y-TZP ceramics was not detected in our gelcast samples (as can be seen in Fig. 6b) and if present, its content would be negligible. It is also important to emphasize that even the sample with the average grain size of 620 nm (SZ 1550) had a fully tetragonal structure with no traces of m-ZrO 2 .

Mechanical and optical properties
The Vickers hardness (HV5) was measured for sintered ceramics and the average values are given in Table 4. The hardness decreased with increasing grain size from HV5 = 1511 for samples sintered at 1350 °C to HV5 = 1414 for samples sintered at 1550 °C. Fig. 7 shows the hardness of individual ceramics as a function of the inverse square root of the grain size.
Because all ceramics reached a relative density of over 99%, the porosity did not influence the hardness values, and the grain size and/or phase structure were the only parameters affecting these values. The hardness values in the graph in Fig. 7 could be reasonably fitted with a line (R 2 > 0.998), which means that the hardness dependence on grain size is governed by the Hall-Petch relationship [61][62][63][64]. Biaxial flexural strength was determined on the sintered ceramic discs with untreated surfaces after sintering and is summarised in Table 4. The Weibull plots for biaxial flexural strength of zirconia ceramics developed in this study are shown in Fig. 8  On the other hand, the maximum strength for conventional 3Y-TZP ceramics usually reaches its maximum at a smaller grain size [66,68,69] because the size of failure origins increases with microstructural coarsening [26,70]. However, it has been reported that zirconia nanoparticle compacts exhibit different grain growth kinetics during sintering compared with coarser (submicron) particulate compacts [26,71] and could be less sensitive to this effect.
where  0 is the characteristic Weibull strength, m is the Weibull modulus, and V eff is the effective loaded volume. The subscripts BT and B3B stand for bending tests (3-point [26,41] and overcame the strength of dental 3Y-TZP ceramics with improved translucency and even mostly surpassed the dental low translucent high-strength zirconia (i.e. opaque dental zirconia with alumina doping) [12,30,31,33,34,75,76]. Moreover, it must be noted that all the strength results in this study were obtained from tests performed on samples with no surface treatment after sintering (i.e. the samples were unground and unpolished). We believe that testing a body surface prepared in a similar way as the dental restorations are prepared (i.e. by milling the pre-sintered blanks followed by sintering) better simulate the mechanical behaviour of real products.
Total forward transmittance (translucency) in the wavelength range from 300 to 800 nm is shown in Fig. 9 for ceramics sintered at 1450 and 1550 °C. The sample sintered at 1350 °C was excluded from the transmittance experiments as a non-preferred material for dental applications due to its low strength and residual porosity. The transmittance in Fig. 9 is shown as a function of wavelength and sample thickness, where two nominal sample thicknesses, 0.5 mm and 1.0 mm, were used. As expected, the transmittance was lower for thicker samples. The transmittance curves steeply rose at wavelengths between 300 and 400 nm and then only slowly continued to increase with increasing wavelength. The SZ 1450 and SZ 1550 samples had quite a similar transmittance in visible light and they reached transmittances of 46.6 and 46.7% for a 0.5 mm thick sample, and 40.8 and 41.1% for a 1 mm thick sample at a wavelength of 600 nm, respectively (see Table 5). Although many studies have described the translucency of dental materials, comparing light transmittances across different studies is difficult or even impossible for three reasons. First, different parameters are used, such as Total Forward Transmittance (TFT), Translucency Parameter (TP) or Contrast Ratio (CR).
The second occasional problem is translucency misinterpretation in the case of TP or CR because these parameters describe the translucency correctly only at a transmittance of above 50% [78][79][80]. The third reason is the non-standardised conditions of transmittance measurement, resulting in a broad range of transmittance values. For example, one type of dental ceramic, very well known among dentists, was tested in four studies [78,79,81,82], where each study used the same sample thickness and the absolute measurement of total forward transmission. The translucency varied from 26% to 62%. The translucency comparison is therefore only valid within one study, where all samples are measured under the same conditions. For this reason, the translucency of the zirconia ceramics developed in our study was experimentally compared with translucency of commercial dental zirconia ceramics well established at the market. It is shown in Fig. 9 and summarized in Table 5 that commercial high-strength dental 3Y-TZP ceramics Ceramill ZI (Ceramill ZI White, Amann Girrbach, Austria), recommended for multi-unit bridges and full-arch screw retained restorations, exhibited substantially lower translucency in the whole spectrum of visible light.
The benefit of the gelcasting method can be also demonstrated on the translucency of alumina-doped 3Y-TZP ceramics prepared from standard 3Y-TZP powder (TZ-3YS-E, Tosoh, Japan) in previous study [41]. Even with this alumina-doped powder (0.25 wt% Al 2 O 3 ), the translucency of the sintered ceramics was close to the highly translucent Nacera Pearl 1 and well above the Ceramill ZI.  Amann Girbach, Austria) marked by producer as highly translucent zirconia. Ceramill Zolid HT+ contained a mixed cubic/tetragonal structure with 38% of cubic phase ( Fig. S1 and Table   S1 in the Supplementary material). However, it is well documented that increasing the content of c-ZrO 2 in the zirconia structure inevitably results in a strength and reliability reduction [13,34]. We have shown that eliminating alumina and other dopants, decreasing the residual porosity after sintering, and utilising a homogeneous coarse-grained structure (i.e. decreasing the number of backscattering centres) in 3Y-TZP ceramics leads to improved translucency without forfeiting excellent mechanical properties. The ceramics based on fine mesostructured 3Y-ZrO 2 particles and processed by the gelcasting method provided a favourable combination of high translucency comparable with a common cubic/tetragonal 4Y-TZP, and very high strength available only in the pure tetragonal 3Y-TZP.
The machinable zirconia discs and blocks were pre-sintered to a temperature of 1000 °C with a dwell time of 1 h. This heat treatment exhibited the most suitable combination of milling conditions and surface roughness. The surface roughness of sintered discs is given in Table 4.
A representative surface map of sintered sample is shown in Fig. 10. Typical milling traces were observed on the sample surface. It is interesting to note that this smooth milling surface (Ra=0.13 µm after sintering at 1550 °C) could be achieved with a relatively dense machinable ceramic (54.0 % t.d. after the pre-sintering at 1000 °C /1 h). The gel-cast bodies were densely packed and started to densify at this optimum pre-sintering temperature, but the mesostructured particles could be still easily removed during milling, providing a smooth surface. Using the presented processing method, blanks of typical dimensions used in dentistry (cylindrical block, large disc) could be prepared (Fig. 11). Fig. 12 shows the individual steps of 3Y-TZP crown production from the pre-sintered cylindrical block with a very smooth surface after milling and a glossy appearance after sintering.     Table 3 Density, grain size, and shrinkage of the zirconia bodies densified by different sintering temperatures. Table 4 Surface and mechanical characteristics of gel-cast samples after sintering. Table 5 Total forward transmittance (TFT) of sintered samples at a wavelength of 600 nm.