Functionally Graded Al2O3–CTZ Ceramics Fabricated by Spark Plasma Sintering

We studied the fabrication of functionally graded Al2O3–CeO2-stabilized-ZrO2 (CTZ) ceramics by spark plasma sintering. The ceramic composite exhibits a gradual change in terms of composition and porosity in the axial direction. The composition gradient was created by layering starting powders with different Al2O3 to CTZ ratios, whereas the porosity gradient was established with a large temperature difference, which was induced by an asymmetric graphite tool configuration during sintering. SEM investigations confirmed the development of a porosity gradient from the top toward the bottom side of the Al2O3–CTZ ceramic and the relative pore volume distributed in a wide range from 0.02 to 100 µm for the samples sintered in asymmetric configuration (ASY), while for the reference samples (STD), the size of pores was limited in the nanometer scale. The microhardness test exhibited a gradual change along the axis of the ASY samples, reaching 10 GPa difference between the two opposite sides of the Al2O3–CTZ ceramics without any sign of delamination or cracks between the layers. The flexural strength of the samples for both series showed an increasing tendency with higher sintering temperatures. However, the ASY samples achieved higher strength due to their lower total porosity and the newly formed elongated CeAl11O18 particles.


Introduction
Functionally graded materials (FGM) have gained increasing attention over recent decades due to the diverse fields of potential applications such as in engineering [1] and biomedicine [2,3]. By definition, FGMs [4] are characterized by a gradual transition of the composition and/or physical properties of the material along the bulk in one or more directions. It results in bulky materials having different behaviors in the various parts of the material tailored to the application. FGMs can be classified mainly into continuous and step-graded types depending on the interface conditions, but definite interfaces cannot be observed in either case. In this regard, the local concentration of stress, which may induce delamination within the material, can be avoided [5].
Plenty of examples can be found for gradient structures in nature including the human body. Because of this, intensive research is carried out to mimic such structures and develop functionally graded ceramics appropriate for various biomedical applications (e.g., bones and teeth). Currently, there are numerous potential materials for biomedical applications, including hydroxyapatite (HAp), bioglass [6,7], alumina, and zirconia [2]. However, all of them suffer from some shortcomings when used as a standalone monolithic material. Hap and bioglass have very poor mechanical properties, which limit their use in replacing bones. From that point of view, Al 2 O 3 and ZrO 2 provide potential benefits because of their porosity distribution [41]. In our previous study [42], we also fabricated alumina with gradient porosity by creating a temperature gradient within the sample during SPS, using asymmetric graphite mold configuration.
In this study, our goal is to develop an Al 2 O 3 -CZT ceramic composite with gradually varying porosity and chemical composition. Composition is varied through four layers, while the porosity gradient is developed by in situ generated thermal gradient during SPS. This structure may provide mechanical properties comparable with that of human bones along with improved biocompatibility and bioactive features. Al 2 O 3 -CZT with gradient porosity can be a promising functional material in biomedical applications, since it combines good mechanical behavior with rapid osteointegration of the cells.

Synthesis of Raw Nanoparticles
Ceria-stabilized (16 mol%) tetragonal ZrO 2 nanoparticles were prepared by hydrothermal synthesis. The tetragonal phase stabilization and mechanical properties of ZrO 2 ceramics are strongly affected by the CeO 2 content [43,44] and it was found that CTZ ceramics possess enhanced strength and toughness with CeO 2 content between 12-16 mol%. The precursors were zirconyle (IV) chloride hydrate (ZrOCl 2 ·H 2 O) (Szkarabeusz Kft., Budapest, Hungary, 99.0%) and cerium (III) nitrate hexahydrate (Ce(NO 3 ) 3 ·6H 2 O) (Sigma-Aldrich Kft. Budapest, Hungary; 99.0%). ZrOCl 2 ·H 2 O and Ce(NO 3 ) 3 ·6H 2 O were dissolved in 200 mL of distilled water for a CeO 2 to ZrO 2 molar ratio of 16:84. The Ce-Zr mixed salt solution was precipitated by adding 0.3 M ammonium hydroxide dropwise until the solution reached pH 10 and white gel was formed. The gel was filtered and washed with distilled water until chlorine ions (Cl − ) could not be detected with a 0.03 M silver nitrate (AgNO 3 ) solution. Finally, the gel was dried in an oven at 90 • C in air for 24 h, then calcined at 700 • C for 2 h. Nanosized Al 2 O 3 was prepared by hydrothermal synthesis similar to CTZ; this route of synthesis is detailed in our previous study [42].
The calcined powders were investigated by X-ray diffraction (XRD) analysis, which confirmed that CTZ powder contains tetragonal crystals fitted by the JCPDF 38-1437 reference, whereas Al 2 O 3 powder shows cubic crystal structured so-called γ-Al 2 O 3 (JCPDF 10-0425). The average particle sizes were 90 nm and 70 nm for Al 2 O 3 and CTZ powders, respectively, determined by scanning electron microscopy (SEM) analysis.

Preparation and Fabrication of 4-Layered Al 2 O 3 -CTZ FGMs
To provide a gradual change in the composition of the ceramic, we prepared four powder mixtures with varying Al 2 O 3 to CTZ ratios, as shown in Table 1. The composition of each layer was chosen based on the results of previous research [14,45]. The powder mixtures were homogenized in a Fritsch planetary ball mill with ethanol. Milling was performed in a Si 3 N 4 tank with Si 3 N 4 balls with a rotation rate of 500 rpm for 5 h. After milling, the powder was dried at 110 • C. The potential impurities stemming from the milling process were determined by EDXRF (energy dispersive X-ray fluorescent spectrometer) analysis method and silicon less than 0.5 wt% was also detected in the Al 2 O 3 -CTZ powder mixtures. The particular mixtures with different compositions were arranged in layers in the graphite sintering mold according to Figure 1. The amount of each mixture per layer was 2.8 g. Table 1. Phase composition of each layer.

Layer
Composition (wt%) tests are referred by their letters indicating the graphite position (STD or ASY) followed by the temperature (e.g., ASY1200 stands for a sample sintered at 1200 °C in asymmetric graphite configuration). In both cases, the local temperatures were monitored by two S-type thermocouples inserted into the inner side of the graphite mold. For the sake of reproducibility, each experiment was repeated three times. After the heat treatments, the average thickness of the samples was 6.3 ± 0.64 mm, whereas each layer was 1.58 ± 0.19 mm thick.  The layered powder mixture was consolidated with the use of SPS equipment (HD P5, manufactured by FCT System GmbH, Rauenstein, Germany) under vacuum at three different temperatures (1200 • C; 1250 • C; 1300 • C) controlled by the pyrometer of the SPS. The pyrometer measures the temperature of the graphite surface of the hole drilled in the upper graphite piston located a few mm above the sample surface as is shown in Figure 1.
The applied sintering conditions are shown in Table 2. The heating rate was 100 • C/min. The cooling rate was lowered to 30 • C/min to minimize thermo-mechanical stresses. The samples were fabricated in vacuum and a uniaxial pressure of 25 MPa was applied. The pulse cycle of the SPS machine was set to 12 ms on and 3 ms off; the characteristic time of each single pulse was 3.3 ms. In addition to creating layers with different compositions in the sample, we also generated a gradient structure in terms of porosity, with temperature difference. We employed an asymmetric graphite configuration, where the powder was placed in an asymmetric position with respect to the graphite mold. This generated a temperature gradient within the sample. Sintering tests were also performed in standard graphite configurations (powders placed in the middle of the graphite mold) serving as reference materials. A schematic illustration of the sample positions is shown in Figure 1. Beyond the sintering conditions, the nominations of the samples are also given in Table 2, where the particular tests are referred by their letters indicating the graphite position (STD or ASY) followed by the temperature (e.g., ASY1200 stands for a sample sintered at 1200 • C in asymmetric graphite configuration).
In both cases, the local temperatures were monitored by two S-type thermocouples inserted into the inner side of the graphite mold. For the sake of reproducibility, each experiment was repeated three times. After the heat treatments, the average thickness of the samples was 6.3 ± 0.64 mm, whereas each layer was 1.58 ± 0.19 mm thick.

Analysis Methods
The samples were characterized for microstructure and mechanical properties. The total pore volume and pore size distribution of the sintered material were measured by mercury intrusion porosimetry (Poremaster 60 GT by Quantacrome Instruments). Determination of pore size by mercury intrusion porosimetry is based on the non-wetting behavior of liquid mercury. Mercury does not penetrate into pores by capillary force but only by applying external pressure. The required pressure is inversely proportional to the pore size. The relation between the pore size and the applied pressure, assuming cylindrical pores, is expressed by the Washburn equation: where p stands for the absolute applied pressure, r for the pore radius, γ for mercury surface tension (≈0.48 N/m), and Θ represents the contact angle (≈140 • ). Due to the relatively long homogenization milling of the starting powder mixtures, the chemical composition of each batch was analyzed primarily for possible contamination. The analysis was performed by an energy dispersive X-ray fluorescence spectrometer (Thermo Scientific Niton Xl3t GOLDD+, Waltham, MA, USA) equipped with a Rh anode X-ray tube and energy-dispersive SDD detector, used at 15-50 kV and 30-120 µA, 120 s acquisition time.
We performed phase identification and determined structural characteristics of the as-prepared specimens by X-ray diffraction analysis (Philips PW 1830, Amsterdam, The Netherlands) using Cu Kα radiation, in the range of 10-70 • 2θ in step-scanning mode with a step length of 0.04 • for 1 s acquisition time per angle. The analysis was performed on the raw powders and on both surfaces of the sintered samples, as well. The fractured cross section of the samples was analyzed by SEM (FEI/ThermoFisher Apreo S LoVac, Waltham, MA, US) with the use of a secondary electron detector (SE) and backscattered electron detector (BSE), operated at 20 kV at various magnifications. Elemental analysis was also carried out on the samples by energy dispersion X-ray spectroscopy (EDX, Ametek Octane Elect Plus) during SEM measurements.
Among mechanical properties, hardness and flexural strength were determined. Hardness was measured by depth-sensing micro-indentation tests on the cut and polished surface of the samples, with a CSM2008 instrument (Peseux, Switzerland) equipped with a Vickers indenter tip. The load was applied for 15 s with a force of 500 mN. Hardness was calculated as the average of five indentation tests for all four sections of the sample. For each indentation, the load-penetration depth curve was automatically acquired and Vickers hardness was deduced from the unloading part of the load-depth curve. The flexural strength was measured with a three-point bending test. It was performed according to the AMMRC 85-21 standard (U.S. army standard test method for the flexural strength of high-performance ceramics at room temperature) with an Instron 5566 tester (Norwood, MA, USA) on three rectangular specimens.

Temperature Gradient during the SPS Process
Gradient properties were created in the ceramic sample in two ways: (1) making up the bulk from layers of different composition, which leads to a chemical gradient, as well as (2) establishing a temperature gradient in the ceramic body during sintering, which results in a porosity gradient. The temperature gradient was generated by the asymmetric graphite configuration of the sample during the SPS sintering process, similarly to our previous study [42]. In Figure 2a-c, we illustrated the temperatures at the top (T 1 ) and the bottom (T 2 ) of the samples, recorded by the thermocouples as a function of pyrometer temperature for different sintering temperatures. For comparison, the temperatures of the reference sample, which was sintered in standard (non-asymmetric) graphite configuration, were also recorded and are shown in Figure 2 as (2) establishing a temperature gradient in the ceramic body during sintering, which results in a porosity gradient. The temperature gradient was generated by the asymmetric graphite configuration of the sample during the SPS sintering process, similarly to our previous study [42]. In Figure 2a-c, we illustrated the temperatures at the top (T1) and the bottom (T2) of the samples, recorded by the thermocouples as a function of pyrometer temperature for different sintering temperatures. For comparison, the temperatures of the reference sample, which was sintered in standard (non-asymmetric) graphite configuration, were also recorded and are shown in Figure 2. The temperature difference (ΔT) between T1 and T2 (T2 − T1 = ΔT) as a function of sintering time for both configurations ( Figure  2d) is also illustrated. The temperatures registered by the thermocouples were invariably above the set temperature measured by the pyrometer of SPS. The differences between the real and set values were higher for the asymmetric graphite configuration and increased with sintering temperature. (The horizontal dash lines indicate the set temperatures in Figure 2a-c). In the standard graphite configurations, the temperatures at the top of the sample (T1) were typically 40-70 °C over the set temperatures, while in the asymmetric configurations, the difference was twice as much. The temperature measured at the bottom of the sample (T2) The temperatures registered by the thermocouples were invariably above the set temperature measured by the pyrometer of SPS. The differences between the real and set values were higher for the asymmetric graphite configuration and increased with sintering temperature. (The horizontal dash lines indicate the set temperatures in Figure 2a-c). In the standard graphite configurations, the temperatures at the top of the sample (T 1 ) were typically 40-70 • C over the set temperatures, while in the asymmetric configurations, the difference was twice as much. The temperature measured at the bottom of the sample (T 2 ) was even higher, with a deviation of about 170 • C for sintering temperatures of 1200 • C and 1250 • C, and 200 • C for 1300 • C. These large temperature differences between the top and bottom of the dies can be explained by the inhomogeneous distribution of current density for the asymmetric graphite configuration during the SPS sintering process, as illustrated in Figure 1. Inhomogeneous distribution of the current density can also occur at the contacting points of the particles [46,47], leading to temperature differences inside the particles during sintering, as well. Figure 2d shows ∆T generated between the top and bottom parts of the samples during the different sections of sintering. In the standard configuration, ∆T was not significant (the highest difference was 40 • C at a sintering temperature of 1300 • C) and alternates in the heating and sintering periods. However, in the asymmetric configuration, the temperature gradient was large and persisted throughout the sintering process. The largest ∆T was characteristic of the heating section reaching ca. 100 • C, probably due to the Joule heating effect, which is one of the typical heating mechanisms of SPS [48]. In the controlled cooling segments, the temperature difference decreased for both configurations, probably because there is less need for intensive heating, consequently, Joule-heating and current density decreased. However, for the samples made by asymmetric sintering, the detected temperature gradient remained until the end of the heat treatment process, whereas in the case of samples produced in the standard graphite configuration, ∆T approaches to 0 • C.
The porosity and mean pore diameter values of the Al 2 O 3 -CTZ FGM sintered in various sintering conditions are shown in Figure 3. The porosity of the samples varied from 5% to 44%, with a rather wide range of pore sizes between 0.03 and 100 µm.
at the contacting points of the particles [46,47], leading to temperature differences inside the particles during sintering, as well. Figure 2d shows ΔT generated between the top and bottom parts of the samples during the different sections of sintering. In the standard configuration, ΔT was not significant (the highest difference was 40 °C at a sintering temperature of 1300 °C) and alternates in the heating and sintering periods. However, in the asymmetric configuration, the temperature gradient was large and persisted throughout the sintering process. The largest ΔT was characteristic of the heating section reaching ca. 100 °C, probably due to the Joule heating effect, which is one of the typical heating mechanisms of SPS [48]. In the controlled cooling segments, the temperature difference decreased for both configurations, probably because there is less need for intensive heating, consequently, Joule-heating and current density decreased. However, for the samples made by asymmetric sintering, the detected temperature gradient remained until the end of the heat treatment process, whereas in the case of samples produced in the standard graphite configuration, ΔT approaches to 0 °C.
The porosity and mean pore diameter values of the Al2O3-CTZ FGM sintered in various sintering conditions are shown in Figure 3. The porosity of the samples varied from 5% to 44%, with a rather wide range of pore sizes between 0.03 and 100 µm. As shown in Figure 3, porosity decreases with increasing sintering temperatures for both configurations, but in the asymmetric graphite position, it reaches lower values than in the standard configuration, which could be expected considering the larger temperatures on both the top and bottom sides of the sample in the asymmetric position. For both series, porosity was greatly reduced to 5% and 8% for ASY1300 and STD1300 samples, respectively, even with a uniaxial pressure of only 25 MPa during the fabrication process. The mean pore diameter values of each sample followed a similar tendency-with increasing temperatures, mean pore size continuously decreased. The differences in the mean pore diameter between the two series were relatively small and became even smaller at higher sintering temperatures ( Figure 3). As shown in Figure 3, porosity decreases with increasing sintering temperatures for both configurations, but in the asymmetric graphite position, it reaches lower values than in the standard configuration, which could be expected considering the larger temperatures on both the top and bottom sides of the sample in the asymmetric position. For both series, porosity was greatly reduced to 5% and 8% for ASY1300 and STD1300 samples, respectively, even with a uniaxial pressure of only 25 MPa during the fabrication process. The mean pore diameter values of each sample followed a similar tendency-with increasing temperatures, mean pore size continuously decreased. The differences in the mean pore diameter between the two series were relatively small and became even smaller at higher sintering temperatures (Figure 3). Figure 4, illustrates the pore size distribution of the sintered bulk ceramics. The samples showed remarkable differences in pore size distribution depending on SPS die configuration. The majority of the pores are around 0.1 µm but the ratio of this size range with respect to total pore volume is by far the highest in the STD series. At the highest sintering temperature (1300 • C), most of the smallest pores are eliminated due to the densification process, as indicated by the decrease in the relative pore volume from 85% to below 10% for 0.1 µm pores. As a result, pores of greatly different sizes are available in the sample.
In the ASY series, pores of about 0.1 µm in size also dominate regardless of sintering temperature, but their proportion is far smaller compared with the STD series and pores in the range of 0.03-100 µm can be also detected in a significant ratio even at the lowest sintering temperature (Figure 4). Based on several studies [49,50], the ideal pore sizes of biologically suitable materials vary in a wide range, from a few nanometers to tens of micrometers; therefore, we can assert that the obtained porosity structure of ASY samples may be suitable for biological applications. However, at the highest temperature (ASY1300), the larger pores are already negligible due to the much higher real temperatures characteristic of the asymmetric configuration. At these temperatures, the sintering process is accelerated; therefore, in the asymmetric configuration, the attained close to 1500 • C is too a high sintering temperature for a porosity-graded material. the sample.
In the ASY series, pores of about 0.1 µm in size also dominate regardless of sintering temperature, but their proportion is far smaller compared with the STD series and pores in the range of 0.03-100 µm can be also detected in a significant ratio even at the lowest sintering temperature (Figure 4). Based on several studies [49,50], the ideal pore sizes of biologically suitable materials vary in a wide range, from a few nanometers to tens of micrometers; therefore, we can assert that the obtained porosity structure of ASY samples may be suitable for biological applications. However, at the highest temperature (ASY1300), the larger pores are already negligible due to the much higher real temperatures characteristic of the asymmetric configuration. At these temperatures, the sintering process is accelerated; therefore, in the asymmetric configuration, the attained close to 1500 °C is too a high sintering temperature for a porosity-graded material.

Microstructure Analysis of the Al2O3-CTZ FGMs
The phase composition of the green and sintered samples was analyzed by XRD. The analysis was carried out on the bottom side of the sintered samples, as well as on each

Microstructure Analysis of the Al 2 O 3 -CTZ FGMs
The phase composition of the green and sintered samples was analyzed by XRD. The analysis was carried out on the bottom side of the sintered samples, as well as on each layer. Figure 5a shows the XRD patterns of the initial Al 2 O 3 and CTZ powders. The XRD analysis confirmed that γ-Al 2 O 3 and tetragonal CTZ phases were obtained in good agreement with the JCPDS reference no. 00-010-0425 and no. 00-038-1437, respectively. Figure 5b,c shows the phase composition of the bottom side of the sample for both series, whereas Figure 5d illustrates the phase composition of each layer for sample ASY1200.
A comparison of the phase composition of the initial powders and the sintered samples reveals several changes. γ-Al 2 O 3 transformed to the more stable α-Al 2 O 3 phase (JCPDS no. 10-0173) in all samples, as the effect of the SPS heat treatment (Figure 5a-c). Furthermore, a small amount of monoclinic ZrO 2 phase (JCPDS no. 37-1484) was also detected in the samples, regardless of the applied sintering conditions and configurations. However, tetragonal CTZ remained the major phase in the composition, predicting improved mechanical properties of the samples [14].  [51]. We suppose that little CeO 2 remained in the initial powder, which was reduced to Ce 2 O 3 in the reductive atmosphere created by the high vacuum and the close contact of the powder mixture with the SPS graphite die. In sample STD1200, neither a CeAlO 3 nor a CeAl 11 O 18 phase was detected, but with increasing temperature, CeAlO 3 was identified in small amounts at 1250 • C. The CeAl 11 O 18 phase only appeared in the composition at high sintering temperatures (1300 • C), proving that it forms at high temperatures [52]. For the asymmetric graphite configuration, the CeAl 11 O 18 phase appeared in all the samples, and its amount increased with sintering temperature shown by the increasing intensity of its diffraction peaks. Layer XRD analysis (Figure 5d) also showed a CeAl 11 O 18 phase in the phase composition of each relevant layer. layer. Figure 5a shows the XRD patterns of the initial Al2O3 and CTZ powders. The XRD analysis confirmed that γ-Al2O3 and tetragonal CTZ phases were obtained in good agreement with the JCPDS reference no. 00-010-0425 and no. 00-038-1437, respectively. Figure  5b,c shows the phase composition of the bottom side of the sample for both series, whereas Figure 5d illustrates the phase composition of each layer for sample ASY1200. A comparison of the phase composition of the initial powders and the sintered samples reveals several changes. γ-Al2O3 transformed to the more stable α-Al2O3 phase (JCPDS no. 10-0173) in all samples, as the effect of the SPS heat treatment (Figure 5a-c). Furthermore, a small amount of monoclinic ZrO2 phase (JCPDS no. 37-1484) was also detected in the samples, regardless of the applied sintering conditions and configurations. However, tetragonal CTZ remained the major phase in the composition, predicting improved mechanical properties of the samples [14].
CeAl11O18 (JCPDS no. 48-0055) and CeAlO3 (JCPDS no. 81-1186) phases were also detected in minor amounts in certain samples. CeAlO3, supposing only as a transition phase in the formation of CeAl11O18, was present only in sample STD1250, while CeAl11O18 was present in STD1300 and in all samples of the ASY series. These phases were formed as a result of a solid-state reaction between the Ce2O3 and Al2O3 during the sintering process, based on previous observations [51]. We suppose that little CeO2 remained in the initial powder, which was reduced to Ce2O3 in the reductive atmosphere created by the high vacuum and the close contact of the powder mixture with the SPS graphite die. In sample STD1200, neither a CeAlO3 nor a CeAl11O18 phase was detected, but with increasing temperature, CeAlO3 was identified in small amounts at 1250 °C. The CeAl11O18 phase only appeared in the composition at high sintering temperatures (1300 °C), proving that it forms at high temperatures [52]. For the asymmetric graphite configuration, the CeAl11O18 The microstructure of the samples was investigated by SEM to determine the effects of (i) the standard and asymmetric graphite configurations and (ii) the applied sintering conditions on the microstructure of the samples. Figures 6 and 7 show the cross section of the fractured surfaces of the samples sintered under different temperatures, revealing the structure of the different layers. Figure 6 shows the cross section of the samples of the STD series. The microstructure unambiguously shows the porosity of the samples. The sample sintered at the lowest temperature (STD1200) shows a fine-grained but highly porous structure through all four layers without significant difference between the layers. However, notable changes can be observed with increasing sintering temperature. The porosity of the samples was significantly reduced along with grain coarsening. Grain size increased from 0.2 µm (STD1200) to 1-2 µm (STD1300). There are, however, no differences between the different layers.
In contrast, in samples sintered in the asymmetric configuration, a gradient structure can be revealed, as shown in Figure 7. This can be attributed to ∆T within the sample, which was as high as 100-120 • C. The microstructure changes gradually through the four layers, particularly in terms of porosity and grain size for all samples sintered at different temperatures. Even for sample ASY1200 sintered at the lowest temperature, significant differences are visible between the first and fourth layers. Whereas the first layer seemingly consists of 1-3 µm large grains with negligible nanometer-sized porosity, in the top part of the sample (fourth layer), 0.5-1 µm large pores are detectable among the 0.4-0.5 nm Al 2 O 3 grains. (The pores were marked with white circles in the figures.) The differences are even more noticeable for the ASY1300 sample. The top (fourth) layer consists of 0.1-0.2 nm Al 2 O 3 grains with large pores in the micrometer range, while the bottom (first) layer is almost pore-free containing 1 µm of large particles. SEM analysis also revealed that the pore structure comprises interconnected pores, particularly for the ASY1200 and ASY1250 samples.
We assume that the presence of silica impurity stems from homogenization of the starting mixtures by the Si 3 N 4 balls, which may act as a sintering aid and promote the densification of the composites. At the grain boundary regions, binary or ternary eutectic phases with lower melting points could be formed in the Al 2 O 3 -SiO 2 -ZrO 2 system, which contributed to the better sinterability of the composites, particularly for the ASY series having higher local temperatures at the bottom. shown by the increasing intensity of its diffraction peaks. Layer XRD analysis (Figure 5d) also showed a CeAl11O18 phase in the phase composition of each relevant layer. The microstructure of the samples was investigated by SEM to determine the effects of (i) the standard and asymmetric graphite configurations and (ii) the applied sintering conditions on the microstructure of the samples. Figures 6 and 7 show the cross section of the fractured surfaces of the samples sintered under different temperatures, revealing the structure of the different layers. Figure 6 shows the cross section of the samples of the STD series. The microstructure unambiguously shows the porosity of the samples. The sample sintered at the lowest temperature (STD1200) shows a fine-grained but highly porous structure through all four layers without significant difference between the layers. However, notable changes can be observed with increasing sintering temperature. The porosity of the samples was significantly reduced along with grain coarsening. Grain size increased from 0.2 µm (STD1200) to 1-2 µm (STD1300). There are, however, no differences between the different layers. In contrast, in samples sintered in the asymmetric configuration, a gradient structure can be revealed, as shown in Figure 7. This can be attributed to ΔT within the sample, which was as high as 100-120 °C. The microstructure changes gradually through the four layers, particularly in terms of porosity and grain size for all samples sintered at different temperatures. Even for sample ASY1200 sintered at the lowest temperature, significant phases with lower melting points could be formed in the Al2O3-SiO2-ZrO2 system, which contributed to the better sinterability of the composites, particularly for the ASY series having higher local temperatures at the bottom.
Consequently, the microstructural observations confirm that the asymmetric configuration contributed to a gradually increasing density towards the bottom part of the samples, resulting in reduced porosity in that region, while the top of the samples preserved their porous microstructure.  Consequently, the microstructural observations confirm that the asymmetric configuration contributed to a gradually increasing density towards the bottom part of the samples, resulting in reduced porosity in that region, while the top of the samples preserved their porous microstructure.
Additionally, several studies [35,36] point out the possible formation of a temperature gradient in the radial direction of the samples, but its magnitude strongly depends on the conductivity of the material. For non-conductive materials, only a temperature difference of 10-25 • C was observed. Our SEM studies found no evidence of a radial temperature gradient in the microstructure of the samples.
Elemental analysis was also performed through the cross-section of the samples by EDX. Figure 8 presents the distribution of Al, Zr, Ce, and O along the cross section of the ASY1250 sample, and its fracture surfaces shown by BSE-SEM. The Zr and Ce content gradually increased from the top (fourth layer) to the bottom (first layer) of the sample. At the same time, an even more pronounced decrease occurred in the Al content. The BSE images also reflect the gradual increase in the CTZ phase (bright parts) along the sample. High magnification images taken of the regions at the supposed boundary of the different layers are also shown in Figure 8. However, there are neither sudden changes nor definite interfaces between them. In other words, we cannot speak about different layers in the sintered samples but only one material with a gradually changing microstructure.
High magnification images taken of the regions at the supposed boundary of the different layers are also shown in Figure 8. However, there are neither sudden changes nor definite interfaces between them. In other words, we cannot speak about different layers in the sintered samples but only one material with a gradually changing microstructure.
High magnification microstructure analysis revealed elongated particles in the structure. (They are shown with black circles in Figure 9). It is assumed that those particles belong to the CeAl11O18, since Akin et al. [51] and Kern at al [40] reported similar observations; furthermore, it also had a beneficial effect on the mechanical properties of the studied ceramics.  High magnification microstructure analysis revealed elongated particles in the structure. (They are shown with black circles in Figure 9). It is assumed that those particles belong to the CeAl 11 O 18 , since Akin et al. [51] and Kern at al [40] reported similar observations; furthermore, it also had a beneficial effect on the mechanical properties of the studied ceramics.

Mechanical Properties of Al2O3-CTZ FGMs
The mechanical properties of the composites were determined with microhardness and by three-point bending tests. Vickers indentation tests were carried out on all four layers of the samples to investigate the differences between the top and bottom sides of the samples accurately. In Figure 10, we compared the mean hardness values of both the STD and ASY series. Vickers hardness (HV) reflects gradual changes in the microstructure of all the samples.

Mechanical Properties of Al 2 O 3 -CTZ FGMs
The mechanical properties of the composites were determined with microhardness and by three-point bending tests. Vickers indentation tests were carried out on all four layers of the samples to investigate the differences between the top and bottom sides of the samples accurately. In Figure 10, we compared the mean hardness values of both the STD and ASY series. Vickers hardness (HV) reflects gradual changes in the microstructure of all the samples.

Mechanical Properties of Al2O3-CTZ FGMs
The mechanical properties of the composites were determined with microhardness and by three-point bending tests. Vickers indentation tests were carried out on all four layers of the samples to investigate the differences between the top and bottom sides of the samples accurately. In Figure 10, we compared the mean hardness values of both the STD and ASY series. Vickers hardness (HV) reflects gradual changes in the microstructure of all the samples. As can be seen, increasing hardness with respect to increasing sintering temperature is characteristic of both series, which reflects the change in the porosity of the samples. Lower porosity results in greater hardness-this correlation between hardness and porosity is well known [53,54]. The hardness of the constituent layers with different compositions show a gradual trend at each sintering temperature. Interestingly, the greatest hardness is exhibited by the 50-50 wt% mixed layer, which contains the least Al2O3, even though, theoretically, Al2O3 is harder than CZT [42]. This phenomenon suggests that (i) hardness is basically determined by the microstructure and not only by the composition, and (ii) the porosity of the fourth layers is significantly higher than that of first layers. This As can be seen, increasing hardness with respect to increasing sintering temperature is characteristic of both series, which reflects the change in the porosity of the samples. Lower porosity results in greater hardness-this correlation between hardness and porosity is well known [53,54]. The hardness of the constituent layers with different compositions show a gradual trend at each sintering temperature. Interestingly, the greatest hardness is exhibited by the 50-50 wt% mixed layer, which contains the least Al 2 O 3 , even though, theoretically, Al 2 O 3 is harder than CZT [42]. This phenomenon suggests that (i) hardness is basically determined by the microstructure and not only by the composition, and (ii) the porosity of the fourth layers is significantly higher than that of first layers. This may seem reasonable in asymmetric configurations, considering the higher temperature at the bottom of the sample, as Figure 1 illustrates. However, for standard configurations, such temperature deviation between the upper and bottom parts of the samples does not occur. In these cases, the higher porosity (and lower hardness) of the alumina-rich layers can be explained by the fact that the complete sintering of pure alumina requires a higher temperature than zirconia-alumina mixed phases. This way, the composition of the layers also affects ultimate porosity, and in turn, the hardness of the sample. In the asymmetric configuration, this combined effect of composition and temperature gradient can result in evenly distributed gradual changes in the hardness of the ASY series. The hardness of each layer is considerably greater compared with the STD series both at 1200 • C and 1250 • C. It can be explained by the differences in the higher temperatures during sintering. At a sintering temperature of 1300 • C, the two configurations also differed in the achieved peak temperature, and the higher temperature may result in enhanced hardness for the ASY series, too. However, it was not the case. Only minor differences were observed between the hardness values of the particular layers of the two series. It is not surprising, however, if we consider that the overall porosity of both samples was almost identical. The higher temperature of the ASY series was still not enough for the pure alumina layer to reach full densification. However, creating fully densified FGM was not our goal.
The evenly graded microstructure that can be attained at 1250 • C or 1300 • C by asymmetric sintering provides a more homogeneous distribution of HV within the sample. The formation of internal stress can be avoided between the consecutive layers, and this in-creases the load capacity of the samples. This is also confirmed by the microstructure, which has a smooth transition of composition and porosity without any sign of delamination at the imaginary interfaces of the layers (Figure 8).
The flexural strength of FGM bodies was determined with three-point bending tests and the achieved values show a trend similar to HV and porosity. With increasing sintering temperature, flexural strength increases, too. Moreover, strength is further improved in asymmetric graphite configuration. The average strength values attained at different temperatures are illustrated in Figure 11 for both series.
tween the hardness values of the particular layers of the two series. It is not surprising, however, if we consider that the overall porosity of both samples was almost identical. The higher temperature of the ASY series was still not enough for the pure alumina layer to reach full densification. However, creating fully densified FGM was not our goal.
The evenly graded microstructure that can be attained at 1250 °C or 1300 °C by asymmetric sintering provides a more homogeneous distribution of HV within the sample. The formation of internal stress can be avoided between the consecutive layers, and this increases the load capacity of the samples. This is also confirmed by the microstructure, which has a smooth transition of composition and porosity without any sign of delamination at the imaginary interfaces of the layers (Figure 8).
The flexural strength of FGM bodies was determined with three-point bending tests and the achieved values show a trend similar to HV and porosity. With increasing sintering temperature, flexural strength increases, too. Moreover, strength is further improved in asymmetric graphite configuration. The average strength values attained at different temperatures are illustrated in Figure 11 for both series.  The STD1200 sample has significantly lower flexural strength than all the other samples. The flexural strength of the ASY1200 sample is almost five times higher. The main reason for this huge difference is presumably the high porosity of the STD1200 sample. The strong correlation between strength and porosity is also known [43], but the almost identical total porosity does not mean equal strength. Although STD1250 and ASY1200 have nearly the same porosity as STD1300 and ASY1300, they significantly differ in strength, which was higher for the ASY series. This can be explained with (i) the different porosity distribution and (ii) the presence of CeAl 11 O 18 phase. These particles have an elongated shape, as it can be observed in Figure 9, so they can provide a reinforcing effect, similarly to whiskers. The flexural strength of the FGM samples is less compared with the values of completely dense alumina, zirconia, or ZTA composites. However, similar values were reported by Kamyshnaya et al. [33] and Mangalaraja [55] for the ceramics they prepared with a composition and porosity identical to our samples. Moreover, the mechanical properties of the fabricated FGM composites also match those of natural bones [56,57], making the composite better suited for bone replacement applications.

Conclusions
We prepared functionally graded Al 2 O 3 -CTZ ceramics fabricated using spark plasma sintering. Owing to the thermal gradient generated during SPS consolidation, the obtained material exhibited a gradually changing composition and porosity in the axial direction. The following conclusions can be drawn:

•
In the asymmetric graphite configuration, there is a significant temperature difference between the top and bottom part of the samples. The difference increases with increas-ing sintering temperatures and reaches as much as 120 • C at a sintering temperature of 1300 • C. At the same time, the temperature at the top side was also invariably higher than the set temperature. • Mercury porosimetry and SEM analysis showed the formation of a gradient pore structure for the samples heat treated in an asymmetric configuration. Total porosity decreased with higher sintering temperatures, regardless of the type of graphite configuration. However, it was still lower in the asymmetric graphite configuration. At the same time, the relative pore volume was distributed in a wide range, and pores of up to 100 µm were detected in a significant ratio, as well. • SEM analysis also showed that porosity and pore size distribution is anisotropicdensification increased toward the bottom part of the samples compared with the reference material. Microstructural analysis at higher magnification confirmed a smooth transition of the material in terms of composition and porosity without any sign of delamination or cracks at the imaginary interfaces of the layers. XRD and SEM analysis revealed that the newly formed CeAl 11 O 18 phase in structure, which has an elongated appearance, presumably affected the mechanical properties of the FGM.

•
Microhardness and flexural test results also indicate the changes in the composition and porosity of the Al 2 O 3 -CTZ material. Hardness increased gradually along the axis, particularly for the asymmetric configuration samples. A difference of nearly 10 GP in hardness was achieved between the top and bottom part of the sample heat treated at 1300 • C. For these samples, hardness distribution was more homogenous within the sample, which prevents the formation of internal stresses inside the materials, therefore providing better load capacity.

•
The flexural strength of the samples for both series shows an increasing tendency with higher sintering temperatures. However, the samples sintered in asymmetric configuration achieved significantly higher strength due to the lower total porosity of the sample on one hand, and the newly formed elongated CeAl 11 O 18 particles on the other. The FGM materials produced possess better mechanical properties than natural bones.
In summary, the present results suggest that the functionally graded composite material of Al 2 O 3 -CTZ is very promising for biological applications because the high-porosity portion of the FGM allows fast bone ingrowth, while the low-porosity part can withstand high mechanical stresses.

Data Availability Statement:
The data that support the findings of this study are available from the corresponding author.