Processing and Properties of Zirconia-Toughened Alumina Prepared by Gelcasting

Zirconia-toughened alumina (ZTA) using yttria-stabilised zirconia is a good option for ceramic-ceramic bearing couples for hip joint replacement. Gelcasting is a colloidal processing technique capable of producing complex products with a range of dimensions and materials by a relatively low-cost production process. Using gelcasting, ZTA samples were prepared, optimising the stages of fabrication, including slurry preparation with varying solid loadings, moulding and de-moulding, drying and sintering. Density, hardness, fracture toughness, flexural strength and grain size were observed relative to slurry solid loadings between 58 and 62 vol. %, as well as sintering temperatures of 1550 °C and 1650 °C. Optimal conditions found were plastic mould, 4000 g/mol PEG with 30 vol. % concentration, 61% solid loading and Ts = 1550 °C. ZTA samples of high density (maximum 99.1%), high hardness (maximum 1902 HV), high fracture toughness (maximum 5.43 MPa m1/2) and high flexural strength (maximum 618 MPa) were successfully prepared by gelcasting and pressureless sintering.


Introduction
Hip joint replacement surgery is a widely implemented solution for joint conditions with an estimated one million operations completed per year. The choice of material for the bearing couple, femoral ball and acetabular cup combination can affect the longevity of the hip joint replacement. The metal-polyethylene (PE) bearing couple is a commonly implanted combination, utilising mainly cobalt-chrome (Co-Cr) ball and PE cup components. The CoCr-PE couple has an average lifetime of 10-15 years, limited by corrosion, wear and ion release [1]. Ceramic-ceramic bearing couples display many times less wear than the alternatives and significantly minimise the risk of particle-induced osteolysis and peri-prosthetic fracture [2][3][4][5].
A D'Antonio et al. [6] compared metal-PE designs to an alumina-alumina bearing couple for identical total hip replacement (THR) device designs in an FDA-controlled, double-blinded study. The survival rate for the alumina to alumina device was 99.2% compared to 95.2% for CoCr-PE at a seven-year follow-up. Wear-related osteolysis was present in three CoCr-PE cases, whilst no osteolysis was reported for the alumina-alumina group.
In addition to superior wear properties, alumina possesses the capability to adsorb polar molecules, such as water and body fluids, thus promoting the formation of a liquid film as a lubrication system between the articulating ball and cup surfaces [1]. Furthermore, clinical studies have reported that ceramic-ceramic bearing couples have a low risk of dislocation [3,7].
However, ceramic materials often have a brittle nature. The majority of current ceramic-ceramic bearing couple manufacturers produce the components from pure alumina (Al2O3) ceramic. Alumina is one of the hardest materials and, thus, displays minimal surface wear; however, with a high hardness, there is a significant risk of brittle fracture.
Improving fracture toughness (4.40 MPa m 1/2 ) and flexural strength (282-551 MPa) of alumina would lessen the risk of brittle fracture and improve the efficiency and reliability of the device, contributing to improved patient outcomes. The addition of another ceramic zirconia (metastable yttrium-stabilised tetragonal ZrO2) to the alumina matrix to form zirconia-toughened alumina (ZTA) improves the fracture toughness via transformation toughening (6-12 MPa m 1/2 ).
Biomedical-grade yttria-stabilised zirconia (YSZ) is susceptible to low temperature degradation with thermal shock due to friction/stress occurring in the aqueous environment of the human body [8,9]. In fact, as determined by Willmann et al. [8], a zirconia-zirconia bearing couple, even in stabilised form, displays the lowest wear performance in comparison to CoCr-PE and alumina-alumina. YSZ is deemed unsuitable as hip replacement bearing couple material [1,[8][9][10].
The addition of YSZ to an alumina matrix to form ZTA optimises the hardness-fracture toughness-flexural strength combination of zirconia and alumina, deeming it suitable as the bearing couple material of hip replacements. High contents of alumina (80 wt. %) in ZTA and the high density (>99%) of samples also lessen the ageing phenomena [9][10][11].
The company CeramTec produces a product BIOLOX ® delta, a ZTA ball and cup system. The production of BIOLOX ® delta by CeramTec utilises hot isostatic pressing (HIP); a processing technique that combines pressing and sintering to achieve powder compaction and healing of voids. The drawbacks to the HIP process include high operating costs, highly specialised equipment and small production runs.
To improve accessibility, this study aims to develop a colloidal processing method capable of achieving high-quality ZTA components at a lower cost of production. Due to the small particle sizes associated with colloidal processing, there is a large contact area between the particles and the dispersing medium, and so, particle-to-particle force interactions in the suspension strongly influence the material behaviour. Using these force interactions, suspension behaviour can then be controlled and further fabrication can be economical to produce high-performance ceramics reliably [10].
Gelcasting is one type of colloidal processing techniques, as is powder injection moulding (PIM) and slip casting (SC). All techniques involve the following steps: (1) powder synthesis; (2) suspension preparation; (3) consolidation into desired shape; (4) removal of solvent phase of suspension; and (5) densification [12]. Gelcasting is a method that allows for moulding of ceramic powders by adopting concepts derived from traditional ceramics and polymer chemistry. Gelcasting produces a near-net shape gel network through the addition of an organic solution to a colloidal ceramic mixture, where in situ polymerisation immobilises the ceramic particles in a highly homogenous structure of a specified shape. The gelled parts undergo drying and densification, throughout which the homogenous structure is maintained.
The aim of this study was to optimise the gelcasting of ZTA observing solid loadings in slurry preparation, moulding type and de-moulding, solvent drying with osmotic drying versus air drying, pyrolysis and sintering at 1550 °C and 1650 °C. Characterisation of the mechanical properties, including density, hardness, fracture toughness, flexural strength and grainsize, is presented.

Materials and Methods
To fabricate ZTA, the gel casting method was used. Powders of alumina and yttria-stabilised zirconia are synthesized; using this powder, a suspension or slurry was prepared and then cast into the desired shape by moulding. In this study, the removal of the solvent phase of the slurry was done by liquid desiccant drying, and finally, pyrolysis and densification was done.

Slurry Preparation
Ceramic slurry suspensions of specific ceramic solid loading are produced from two separate components: the solid component and the liquid component. The liquid component (organic premix solution) was prepared by magnetically stirring monomers MAM and MBAM with dispersant (Dolapix) in de-ionised (DI) water for 20 min. The solid component is a mixture of the alumina and zirconia powders. The gelcasting slurry was prepared by ball-milling. The milling media used was zirconia balls of 5 and 10 mm. The ball-to-powder ratio was set to 2:5. Ceramic solid loadings of 58, 60, 61 and 62 vol. % were investigated.

Slurry Casting
Prior to removing the slurry from the ball mill, a 10 vol. % initiator (TEMED) solution is prepared by magnetic stirring for 20 min in DI water, and all moulds were pre-prepared with a parafilm base to allow easier removal.
The slurry was removed from the ball mill and sieved from the zirconia milling media to then be immediately de-aired. The initiator solution and catalyst were incorporated into the slurry during this step under a 4-min, 1-min and 30-s de-airing cycle. The slurry was then cast into the moulds, enclosed in a plastic bag with a small dish of water for hydration and placed into a conventional oven at 50-70 °C for 1 h for gelation to occur. The gelled parts remained in the oven in off-mode for natural cooling and were then de-moulded for solvent drying to be carried out.
Various mould designs and removal techniques were observed. Investigated mould materials include bees wax, paraffin wax and acrylonitrile butadiene styrene (ABS) (plastic), with and without the application of a mould release agent, WD-40. Wax mould removal was completed by gradual breakoff, as well as melting.

Liquid Desiccant Drying
Liquid desiccant drying of the gelled bodies was conducted in 3 subsequent phases: osmotic drying, air drying and oven drying. Phase 1, osmotic drying, was performed by completely immersing the green parts in liquid desiccant (PEG) solution for 20-25 h. For Phase 2, the parts were removed from the PEG, washed with DI water and left to dry in air for 2-3 days. The parts were then placed in the oven at 50 °C for 1 h for the final phase of drying.
Throughout each phase of drying, the weight of the green parts was measured to indicate when the next phase of the process should begin, i.e., when weight reduction of the parts becomes negligible, drying is considered to no longer be occurring efficiently, and thus, the next phase is initiated.
The investigated parameters were PEG molecular weight and PEG solution concentration. Experimental variation involved molecular weights of 400, 4000 and 20,000 g/mol in concentrations of 30 and 50 vol. % aqueous solutions.

Pyrolysis and Densification
After solvent removal, the green samples were placed in a sintering furnace for polymer binder removal under cycle with a hold at 600 °C for 2 h and up to 950 °C with a hold for 2 h. The parts were then machined and polished to the specific conditions required for various characterisation techniques and sintered for densification up to 1550 °C or 1650 °C with a hold for 2 h. Maximum sintering temperatures of 1550 and 1650 °C were investigated.

Density
Density was determined using helium pycnometry. Helium (He) gas was injected into the sample chamber and entered the pores and voids in the ceramic structures. The pressure change of the He gas in a calibrated volume then indicates the volume per unit weight of the samples. Results were converted to relative density (%) using theoretical density.

Mechanical Testing
Vickers hardness testing, fracture toughness and flexural strength testing were carried out. Vickers hardness (HV) testing using diamond was taken under a 10 kg load in 10 different areas for each tested sample and averaged. Samples tested under this technique were prepared by embedding in PolyFast (Struers, Cleveland, OH, USA) by a hot mounting machine and automated polishing with silicon carbide paper from Grade 320-Grade 4000. Polishing was carried out in manual mode.
Fracture toughness using diamond-like indentations impinged on the surface during hardness testing was analysed under optical microscopy (Zeiss, Oberkocken, Germany). Samples had red colouring applied to the surface to create a colour contrast and to allow a clear image of the indentation.
Indentation crack lengths were measured by optical microscopy and converted into fracture toughness values by the following equation for hard ceramics (in accordance with JIS R 1607) [13,14].
where KC is the fracture toughness (MPa m 1/2 ), P is the indentation load (kN), HV is the Vickers hardness, c is the crack length from indentation centre (mm) and E is the Young's modulus of the material (MPa). Flexural strength was tested using three-point bend testing, Instron Model 5948 Microtester (Norwood, MA, USA), with samples cut and polished to rectangular dimensions of 45 mm × 35 mm × 25 mm with a cross-head rate of 0.5 mm/min in accordance with ASTM C1161-13 [15]. An average of 5 specimens was used for 1650 °C and 3 specimens for 1550 °C.
where S is the flexural strength (MPa), P is the breaking force (N), L is the outer span, b is the specimen width and d is the specimen thickness.

Microstructural Analysis
Samples were analysed using Scanning Electron Microscopy (SEM), JEOL (Tokyo, Japan), to characterise grain size. Samples were polished after binder-burnout and prior to densification. Samples were polished with silicon carbide paper Grades 800-4000 to completely flat and parallel surfaces and then thermally etched to a temperature 50 °C less than the sintering temperature. Samples were coated with a layer of gold by sputter coating. Table 1 presents the results for moulding material and de-moulding techniques and the gelled ZTA samples after removal from the mould. ZTA appeared to adhere very strongly to both the bees wax and paraffin wax moulds. Producing moulds with dimensional accuracy with bees wax was difficult due to the lack of rigidity of the material and its adhesive surface nature. The ZTA parts that resulted from these moulds were extensively cracked and deformed. Thus, bees wax was ruled out for further investigation.

Moulding and De-Moulding
Paraffin wax in combination with the WD-40 release agent was used to minimise the adhesive properties of the mould surface. However, by break-off removal, the ZTA samples still adhered too strongly to the mould surface and experienced cracking and deformation during removal. Paraffin wax did have satisfactory rigidity and dimensional accuracy. Melting was done in an attempt to achieve mould removal by an alternative method to break-off. Using the low melting point of the wax (68 °C) and melting the mould off the gelled structure was attempted. Samples were heated to 70 °C, following gelation for the moulds to melt, and left in an off oven to cool. This removal technique produced almost crack-free ZTA samples with slight curvature. ABS (plastic) moulds were formed by building blocks to allow compartmental break-off for removal, as well as rigidity in structure and accurate versatility in dimensions. Slurry seeping and leakage, as well as adhesiveness all proved to be complications for this mould material. However, with the incorporation of a WD-40 layer over the mould surface, these problems were alleviated, and defect-free ZTA parts were obtained by break-off removal. Thus, this mould material and this removal technique were repeated for all remaining sample preparations of the experimental work.

Liquid Desiccant Drying
3.2.1. Air Drying vs. Osmotic Drying Figure 1 shows a sample weight reduction comparison between air drying and osmotic (PEG) drying. This comparison was conducted to confirm that the PEG method was a superior drying technique for gelcast ZTA samples. Results show that the overall amount of solvent removal is slightly greater by the PEG method than the air method, but more importantly, that the PEG method achieves a steadier rate of drying, which is critical to alleviating defect formation. Thus, further investigation involved optimising PEG drying parameters for maximum efficiency.  Figure 2 compares the PEG drying efficiencies of PEG 400, 4000 and 20,000, all at a 30% concentration aqueous solution. It was expected that the lowest molecular weight would be the least effective drying desiccant due to chain penetration into the porous structure and that the highest molecular weight would be the most effective. PEG 400 was confirmed to produce the lowest drying efficiency by a considerable amount, whereas PEG 4000 and 20,000 were similar, with PEG 4000 removing approximately 0.5% more solvent. This slight, yet reproduced difference may be attributed to the increase in osmotic pressure associated with larger polymer chains, which results in a smaller pressure gradient between the internal solvent and the external desiccant. This, in turn, may cause the rate and level of solvent diffusion out of the ZTA sample to be reduced.

Figure 2.
Sample weight reduction by PEG 400, 4000 and 20,000 at a 30% concentrated aqueous solution followed by air drying and oven drying.
In addition to this, if solvent drying is not adequate, the ceramic parts are expected to undergo extensive cracking during sintering cycles. Thus, PEG 400, 4000 and 20,000 drying efficiencies were qualitatively compared by observing the ZTA structures after densification, shown in Figure 3. Evidently, the structures dried with PEG 400 experienced very sudden solvent drying when placed in the sintering furnace and cracked into pieces, whereas PEG 4000 and 20,000 produced defect-free parts of satisfactory quality.   Figure 4 depicts the solvent removal capabilities of PEG 4000 and 20,000 in aqueous solutions of 30% and 50% concentrations. The 50% concentrated solutions proved ineffective for both molecular weights with less than a 1% weight reduction. The 30% solution however displayed satisfactory weight reduction in both cases and, thus, was used for all subsequent ZTA batch preparations. It can be deduced from these results that drying efficiency reduces with increasing PEG concentrations. This trend can be attributed to the viscosity increase in solution with higher PEG content, where viscosity is proportional to osmotic pressure. Thus, the increase in PEG concentration results in a smaller pressure gradient between the internal solvent and the external desiccant, causing the rate and level of solvent diffusion out of the ZTA sample to be reduced. Figure 5 shows relative densities obtained at solid loadings ranging from 58%-62% at sintering temperatures of 1550 °C and 1650 °C.

Density
Pycnometry was carried out for ZTA samples obtained from solid loadings of 58%, 60%, 61% and 62% and sintering at Ts = 1550 °C and 1650 °C. It was found from pycnometry measurements that density increased with solid loading for both 1550 °C and 1650 °C sintering temperatures, confirming the expected trend. The highest investigated solid loading of 62% achieved the highest relative densities (over 99%) for both sintering temperatures. A key detail to be extracted from the following data is that at a solid loading of 62%, and potentially higher, a sintering temperature of 1550 °C is capable of obtaining a similar density to that obtained at a 1650 °C sintering temperature. A lower sintering temperature is expected to result in a smaller grain size, which is desirable for mechanical performance, but risks a compromise in density. It can be deduced from these results that density has not, in fact, been compromised at T = 1550 °C and that it may well be possible to surpass a 99.8% density value with slightly higher solid loadings. Figure 6 displays the hardness values obtained by Vickers indentation for ZTA samples of 58%, 60%, 61% and 62% solid loadings sintered at Ts = 1550 °C and 1650 °C. Generally, hardness increases with increasing solid loading and increasing sintering temperature. However for samples with solid loadings of 62%, hardness is comparable for both sintering temperatures. The peak hardness of 1858 HV for Ts = 1550 °C was at 62% solid loading, whereas the peak hardness of 1902 HV for Ts = 1650 °C was at 61% solid loading.

Fracture Toughness
Optical microscopy was used to determine the fracture toughness of the ZTA samples of solid loadings of 58%, 60%, 61% and 62% sintered at Ts = 1550 °C and 1650 °C. The highest fracture toughness was 5.43 MPa m 1/2 and occurred at Ts = 1650 °C and 60% solid loading. It is expected that Kc would increase with decreasing solid loading, as increasing solid loading would likely result in increased hardness. Increasing sintering temperature would likely result in a larger grain size and, therefore, lower fracture toughness. Generally, these results were found and shown in Figure 7 with lower fracture toughness with higher solid loading. The effect of sintering temperature varied at the different solid loadings.  As shown in the optical micrographs in Figure 8, the crack path from the diamond hardness indentation was very fine. The application of red surface colouring assists the process. However, the judgement of where the crack tip is located was based on visual appearance and measured accordingly.

Microstructural Analysis
A small grain size for ceramics is the target for increased resistance to crack propagation, as a high incidence of grain boundaries provides maximal hindrance to the path of a crack. This need for as small a grain size as possible without density compromise is reinforced in Figure 10, which shows a crack travelling along the boundaries of small grains and directly through the larger grains.
Various ZTA samples were analysed by SEM for comparison of: (1) solid loading of 61%-62%; and (2) sintering temperatures of Ts = 1550 °C-1650 °C. SEM images, seen in Figure 11, show highly homogenous structures for all parameter variations with no appearance of zirconia aggregates or pores present.
Average alumina grain size was indicated by the grain distribution and displayed similar results for 61% and 62% solid loadings of 1.54 and 1.56 μm, respectively. Although these are average values, it can be seen in each SEM image that some grains are greater than 2 μm in one dimension. This can be attributed to the higher sintering temperature of 1650 °C, which allows greater particle fusion. The distribution calculations propose Ts = 1550 °C as the superior sintering temperature, achieving the smallest grain size of 1.37 μm.

Discussion
ZTA samples were prepared by gelcasting. The preparation method includes a slurry preparation, slurry casting in moulds, solvent drying, pyrolysis and densification. Slurries were prepared to solid loadings of 58-62 vol. %. In slurry casting, moulding and de-moulding were investigated and optimised with respect to specific dimensional requirements. The results are presented in Table 1 and show that ABS (plastic) with WD-40 release agent and break-off removal produced samples with defect-free gelled bodies with dimensional stability.
Results for solvent drying of green gelcast ZTA samples found that osmotic drying using polyethylene glycol (PEG) as the liquid desiccant is superior to air drying. Weight reduction measurements obtained for PEG 400 compared to PEG 4000 and PEG 20,000 suggest that the polymer chains were of small enough size to be able to penetrate into the pores of the gelled ceramic structures and to hinder the potential for solvent diffusion. Similar results were found by M. Trunec [16], where smaller PEG polymer chains penetrated the gelled bodies and reduced the dewatering.
Weight reduction measurements obtained for PEG 20,000 suggest that the problem of polymer chain diffusion into the porous ceramic structure, as encountered with PEG 400, was alleviated. However, diffusion levels were not as high as those obtained by PEG 4000. The mechanism of osmotic drying uses the maximum osmotic pressure gradient between the gelled structure and the surrounding liquid desiccant. Obtained weight reduction data suggest that the larger chain size of PEG 20,000 liquid desiccant induces a higher osmotic pressure, closer to that of the gelled structure, and, thus, reduces the pressure gradient between the two mediums. Similarly, measurements obtained for the 50 vol. % aqueous PEG solutions also suggest a reduced osmotic pressure gradient, an effect that may be attributed to the increase in viscosity of the solution associated with the higher concentration. This correlation was also reported by B. Michel et al. [17].
A 30 vol. % aqueous solution of PEG 4000 (molecular weight = 4000 g/mol) showed the highest solvent removal efficiency of the investigated parameters ( Figure 1). Thus, it can be deduced that PEG 4000 at a 30% concentration was a suitable compromise for providing an adequate osmotic pressure gradient to induce good levels of solvent diffusion and a large enough polymer chain size to prevent penetration into the pores of the structure. In addition to this, comparison of the sintered ZTA samples produced (Figure 3) confirmed the adequacy of PEG 4000 at a 30% concentration in producing high-quality brown structures. It can be seen in Figure 1 that the gelled ceramic structures experience approximately 30%-35% of total solvent drying in the air and oven drying stages carried out after osmotic drying.
After solvent drying, pyrolysis and densification was done. Looking at the range of solid loading between 58 and 62 vol. % for sintering at 1550 °C or 1650 °C, 62 vol. % and 1550 °C produced ZTA samples with the highest obtained density of 99.2%. These parameter values also produced an exceptional hardness of 1877 HV, equivalent to 18.41 GPa.
It was shown that a sintering temperature of 1550 °C was not only capable of achieving a density similar to that achieved by sintering at 1650 °C, but also a more favourable trend over the span of solid loading variations ( Figure 5). At Ts = 1650 °C, values for density plateaued at approximately 99% at solid loadings higher than 60 vol. %. However at Ts = 1550 °C, density measurements show a relatively steady increase over the span of solid loading variations, indicating that potentially, the maximum possible density had not been attained.
It is expected that a higher sintering temperature may result in a larger grain size and potentially have a detrimental impact on mechanical performance of the ZTA samples. It was shown that Ts = 1550 °C induces a smaller grain size than Ts = 1650 °C without compromising density ( Figure 11). Furthermore, particle fusion was evident at Ts = 1650 °C for both 61% and 62% solid loadings, whereas Ts = 1550 °C did not exhibit this.
Fracture toughness generally increased with decreasing solid loading (Figure 7). It was expected that samples with higher hardness would have lower fracture toughness. Samples with higher solid loadings generally had higher hardness ( Figure 6). It was also expected that for higher alumina content, hardness would be higher and fracture toughness would be lower. In this study, however, the alumina-to-zirconia ratio was fixed at 82.43 to 17.57 vol. % and solid loading was instead varied between 58 and 62 vol. %.
From microstructural analysis, there was a good degree of homogeneity of alumina and zirconia, which is thought to improve fracture toughness in a composite material. Furthermore the smaller grain size would reduce the crack energy and shorten the crack length. In most cases, except solid loadings of 60%, a smaller grain size and a lower sintering temperature attained higher fracture toughness. The crack lengths were generally equal in all directions from the diamond indentation, in part due to homogeneity and grain size.
The toughening mechanisms in the phase transformation of zirconia could have contributed to reducing crack propagation. Tulliani et al. determined a critical zirconia size in a ZTA composite of 1.2 μm, where lower than this value, the toughening mechanism of zirconia could be reasonably ruled out [18]. Therefore, as the finer grain size of approximately 1.37 μm was found in this study, it can be surmised that the toughening mechanism of zirconia is likely. Other studies that varied zirconia content of ZTA samples found that higher zirconia did result in higher fracture toughness and flexural strength, but lower hardness [11].
Flexural strength tended to reach a peak for Ts = 1650 °C at 61% solid loading, whilst for Ts = 1550 °C at a 60% solid loading and then decreased (Figure 9). Solid loading had a major effect on flexural strength, as did sintering temperature. With increasing of solid loading from 58%-61% the strength increased from 558-609 MPa for Ts = 1650 °C, and from 58%-60%, the strength increased from 549-574 MPa for Ts = 1550 °C. However, increasing solid loading above 60%-61% led to a drop in the flexural strength of the samples. Samples with higher hardness and density produced from 62% solid loading were more brittle, as shown with flexural strength testing, either due to less homogeneity throughout the sample, particle agglomeration or higher air entrapment from the viscosity of the slurry. Looking at Figure 11c, the homogeneity of the alumina and zirconia looks adequate, and air entrapment cannot be identified; however, there appears to be particle agglomeration of zirconia. This may have led to point defects and reduced flexural strength.
Increasing sintering temperature from Ts = 1550 °C to Ts = 1650 °C generally increased the flexural strength of the samples. The higher sintering temperature increased the grain size, as seen in Figure 11, and hardness, as seen in Figure 6. Although the higher hardness for Ts = 1550 °C reduced the flexural strength, the larger grainsize for Ts = 1650 °C allowed for higher flexural strength. This was only the case for samples with solid loadings between 58 and 61%. The flexural strength for samples with solid loadings of 62% did not follow this pattern. There was more variability for samples produced from solid loadings of 62% for both sintering temperatures when tested for flexural strength, and therefore, the slight difference between flexural strength at Ts = 1550 °C and Ts = 1650 °C is not significant. Liu et al. [19] found a similar effect of solid loadings on mechanical properties; however, the effect of sintering temperature was not tested.

Conclusions
ZTA samples of high density (maximum 99.1%), high hardness (maximum 1902 HV), high fracture toughness (maximum 5.43 MPa m 1/2 ) and high flexural strength (maximum 618 MPa) were successfully prepared by gelcasting and pressureless sintering. Prepared by gelcasting, the study optimised mould formulation for optimal de-moulding, osmotic drying, solid loading and sintering temperature, and a mechanical characterisation was carried out. Density, hardness, fracture toughness, flexural strength and grain size were all found to vary with solid loading and sintering temperature. Optimal conditions found were plastic mould, 4000 g/mol PEG with 30 vol. % concentration, 61% solid loading and Ts = 1550 °C. ZTA prepared by gelcasting has improved properties over alumina alone and is a strong contender for a hip joint replacement ceramic-ceramic bearing couple.