Microstructural characterization of Ce-TZP/Al2O3 after degradation in water

ABSTRACT In this study, the long-term resistance of Ce-TZP/Al2O3 specimens to hydrothermal treatment is evaluated in terms of their microstructure. Treatment of the specimens is carried out in an autoclave with water at 134°C for up to 360 h. The biaxial strength and Weibull modulus of the specimens remains unchanged after 120 h of hydrothermal treatment, and their pseudo-plasticity is preserved. However, after 360 h of hydrothermal treatment the biaxial strength decreases by 12% and the Weibull modulus is also decreased. The transformation surface layer is observed through scanning electron microscopy (SEM) and transmission electron microscopy (TEM). The thickness of the transformation layer can be estimated only through TEM analysis, revealing minimal microcrack formation at the surface layer. For comparison, several Y-TZP specimens are also prepared. After 24 h of hydrothermal treatment, the monoclinic phase content in the Y-TZP specimen reached 60%, resulting in a 40% decrease in strength. This strength decrease can be attributed to microcrack formation in the surface layer.


Introduction
The mechanical properties of yttria-stabilized tetragonal zirconia (Y-TZP) are superior in comparison to other ceramics, making it a widely used material in various applications, such as optical connectors [1] and biomedical implants [2][3][4].The excellent mechanical properties of Y-TZP can be attributed to the presence of a metastable tetragonal (t) phase at room temperature.However, when this metastable phase comes in contact with water, it tends to transform into a monoclinic (m) phase [5,6].This phase transformation is accompanied by a 4% volume expansion [7,8], leading to the development of stresses and even microcracks.The formation of microcracks adversely affects the strength [9]; this is referred to as low-temperature degradation (LTD).
The applications of tetragonal zirconia polycrystal (TZP) ceramics are limited because of their susceptibility to degradation.For example, a large number of hip-joint prostheses were recalled in 2003 because of the LTD of Y-TZP femoral heads [6].Such degradation has also caused communication noise in optical connectors [1].Many approaches have been explored and implemented to address this issue.Some of these approaches include: (1) Introducing a second phase with high elastic modulus.In this approach, the volume expansion of zirconia during phase transformation is constrained by rigid surroundings [10].The use of a second phase with high elastic modulus, e.g., alumina, has been shown to be effective.However, this method may introduce internal stresses within the composite [5].(2) Reducing the sintering temperature.During the sintering of Y-TZP at elevated temperatures (>1450°C), yttrium ions may segregate to grain boundaries [11].Such segregations deplete the concentration of oxygen vacancies within ZrO 2 grains, resulting in higher m-phase content.Through the addition of a small amount (0.1-1 mol%) of the sintering additive, Al 2 O 3 , the sintering temperature can be reduced by more than 100°C [12] and the LTD resistance of Y-TZP can be improved.(3) Using other stabilizers.The LTD resistance of TZPs with alternative stabilizers has been evaluated.Ceria-stabilized zirconia (Ce-TZP) exhibits low degradation at low temperatures [13].However, the rapid grain growth of Ce-TZP during sintering results in lower strength [14].The grain growth at elevated temperatures can be inhibited through the addition of a second phase [15].Previous studies have demonstrated that the addition of 30 vol% Al 2 O 3 particles was effective in inhibiting the growth of ZrO 2 grains [16][17][18].
Among these approaches, the use of ceria as a stabilizer for TZP is a promising strategy [16][17][18].The charge of the Ce 4+ ion is the same as that of Zr 4+ , eliminating the need for additional oxygen vacancies to accommodate the incorporation of CeO 2 into ZrO 2 .
With a reduced number of oxygen vacancies, the transport of OH − or O 2− , resulting from the decomposition of water [6], is less likely.Previous studies have confirmed the LTD resistance of Ce-TZP/30%Al 2 O 3 specimens was indeed improved [17].Although the decrease in strength of Ce-TZP/Al 2 O 3 specimens after being soaked in water is minimal, phase transformation in a hydrothermal environment still occurs.More importantly, a detailed characterization of the microstructure of the transformation surface layer of Ce-TZP/Al 2 O 3 specimens has yet to be performed.In the present study, the long-term resistance of Ce-TZP/Al 2 O 3 specimens to hydrothermal degradation was evaluated and the microstructures of the transformation layers of the specimens were characterized.
The degradation of TZP depends heavily on the details of its processing technique [19][20][21][22][23].For example, the transition from a muffle furnace to a conveyor furnace significantly reduces the LTD resistance [6].In order to evaluate the degradation of Ce-TZP/Al 2 O 3 specimens in water, a comparison to the results for Y-TZP would be of interest.Therefore, some Y-TZP specimens were prepared as reference specimens.

Experimental
Previous studies [17] had suggested that the hydrothermal resistance of Ce-TZP/Al 2 O 3 was better than that of Y-TZP; nevertheless, the microstructure characterization on the degradation layer of Ce-TZP/Al 2 O 3 was still lacking.The following tests were conducted to evaluate the degradation resistance of Ce-TZP/Al 2 O 3 ; special attention was given to the microstructure characterization of surface layer.Apart from the Ce-TZP/Al 2 O 3 specimens, some reference Y-TZP specimens were also prepared for comparison purpose.

Ce-TZP/Al 2 O 3 specimen
The starting powder consisted of 70 vol% ceriastabilized zirconia (Ce-TZP) and 30 vol% alumina (Al 2 O 3 ) (Daiichi Kigenso Kagaku Kogyo Co., Japan); the CeO 2 content in Ce-TZP was 10 mol%.The powder was granulated by the manufacturer to facilitate the use of a die-pressing technique.Pre-firing was carried out at 1000°C for 1 h to remove the organic binder.The green compact was subsequently sintered at 1550°C for 2 h.The sintered discs were 19.1 mm in diameter and 1.8 mm in thickness.The density of sintered specimens was measured by a water displacement method.The average density value was obtained by measuring 10 specimen discs.The relative density of the specimen was calculated assuming the theoretical density of 5.56 g/cm 3 [24].
The cross-sections of the specimen were prepared by grinding and polishing, the thermal etching by heating at 1450°C for 0.5 h.The mean grain size was determined by lineal intercept method; at least 200 grains were counted.The grain size was then estimated by multiplying the average intercept length with a factor of 1.56 [25].The microstructure was examined by scanning electron microscopy (SEM, Nova NanoSEM 450, FEI Co., USA).Since the Ce-TZP/Al 2 O 3 specimen was electrical insulating, a thin Pt layer was coated onto the surface before SEM observation.The Pt coating could not penetrate into the pores or microcracks under surface, the electron charging could be detected if pores and/or microcracks were present in the surface layer.The charging layer was frequently taken as the layer of phase transformation [9].
The hydrothermal treatment was carried out in a Teflon vessel.The specimen discs were cleaned with acetone and alcohol, then immersed in deionized water.The autoclave was then placed in an oven at a temperature of 134°C for up to 360 h.The phase composition after hydrothermal treatment was analyzed by X-ray diffraction (XRD, TTR III, Rigaku Co., Japan) technique.The scanning was conducted in the 2θ range from 25° to 32° at a slow scanning rate of 0.24 °/min.Within this 2θ range, typical XRD peaks for t-ZrO 2 , m-ZrO 2 and Al 2 O 3 phase were all present; the amount of monoclinic (m) phase could therefore be estimated using the relation of Garvie and Nicholson [26].The detection depth of the XRD technique was around 5 μm.
The biaxial strength of the specimen discs was measured using a ring-on-ring jig.The diameter of the loading and supporting rings was 7 mm and 14 mm, respectively.The loading rate was 0.48 mm/min.There were 15 to 18 discs measured for each hydrothermal treatment condition.The scatter of strength values was expressed in terms of Weibull modulus [17].Apart from the SEM technique, the microstructure was also examined using transmission electron microscopy (TEM, Tecnai G2 F20, FEI Co., USA).For the TEM observation, a thin section was cut from surface region with a focused ion beam system (FIB, NanoLab 600i, FEI Co., USA).
In order to evaluate the stress-strain behavior of Ce-TZP/Al 2 O 3 specimen, a 4-point bending at a loading rate of 10 N/s was employed.The rectangular testing bars with the dimensions of 3 × 4x30 mm were prepared for the bending test.The preparation procedures of the testing bars can be found in another report [27].Compared to the specimen disc used for biaxial bending test, the volume under tensile stress in the rectangular bar was relatively larger.The stress (σ) and strain (ε) during 4-point bending were estimated using the following equations [28]: where P was the applied load; b the width (4 mm), d the thickness (3 mm) and d de the displacement of the loading pin.The upper span (l) was 10 mm and the lower span (L) was 30 mm.

Y-TZP specimen
Yttria-stabilized zirconia (Y-TZP, Tosoh Co., Japan) reference discs were also prepared using the same diepressing techniques.The Y 2 O 3 content in the Y-TZP was 3 mol%.There was a small amount of Al 2 O 3 (0.25 wt%) in the powder as reported by the manufacturer.The sintering temperature was 1450°C, the dwell time was 1 h.The techniques used for phase analysis, microstructure characterization and strength measurement were identical to those used for the Ce-TZP/Al 2 O 3 specimen discs.

Ce-TZP/Al 2 O 3 specimens
The sintering temperature for the Ce-TZP/Al 2 O 3 specimen used in this study was 1550°C.Our preliminary studies indicated a sintering temperature of 1450°C was sufficient for preparing dense Ce-TZP/Al 2 O 3 specimens with a relative density > 95%.A higher sintering temperature of 1550°C resulted in a higher density (>99%) and larger grain size (>1 μm).A specimen with higher density contains no open pores and a larger grain size induces a higher degree of phase transformation in Ce-TZP [29,30].These two microstructural characteristics facilitate the investigation of the hydrothermal degradation behavior of the Ce-TZP/Al 2 O 3 specimen within a reasonable time frame.
A representative SEM micrograph of a sintered Ce-TZP/Al 2 O 3 specimen is shown in Figure 1(a).The relative density of the sintered specimen discs was higher than 99%, as shown in Table 1.The bright particles in the SEM micrograph are ZrO 2 grains, with a mean size of 1.1 μm and the gray particles are Al 2 O 3 grains, with a mean size of 0.8 μm.
The microstructure of the sintered Ce-TZP/Al 2 O 3 specimens (before hydrothermal treatment) was also examined through TEM.A thin section collected using the FIB technique is shown in Figure 2(a).This section was taken from the surface to a depth of 5 μm into the specimen.Figure 2     The bending strength of the Ce-TZP/Al 2 O 3 specimen measured through biaxial loading is shown in Figure 7(a).The data were linearly fitted to determine the Weibull modulus, with R 2 values higher than 0.9.The Weibull modulus of the specimen before hydrothermal treatment was 7.2.After treatment in water for 120 h, the Weibull modulus remained almost constant at 8.8.The average biaxial strength value of the Ce-TZP/Al 2 O 3 specimen is shown in Figure 8.After hydrothermal treatment for 120 h, the strength decreased slightly by 4.3%, but the Weibull modulus was not affected.To assess the effect of the surface transformation layer on the stress -strain behavior of the Ce-TZP/Al 2 O 3 specimen, a loading -unloading-reloading technique was applied using rectangular testing bars (Figure 9).For the rectangular specimens before and after hydrothermal treatment for 120 h, a very small strain was observed when the stress reached 400 MPa during the first loading cycle.After unloading and reloading with a gradually increasing load, the nonrecoverable strain increased slightly.This indicates that (b) (a) the Ce-TZP/Al 2 O 3 specimen exhibits a pseudo-plastic behavior both before and after 120 h of hydrothermal treatment.
With an increase in the duration of hydrothermal treatment to 360 h, the strength of some specimens (located in the lower left corner of Figure 7(a)) was relatively low, resulting in a lower Weibull modulus (5.0).
The m-phase content increased slightly from 2% to 3% after hydrothermal treatment for 24 h, to 11% after 120 h, and to 28% after 360 h (Figure 4).The thickness of the transformation layer could not be estimated using SEM because of the lack of an electron charging layer.However, the TEM observation revealed the presence of twins as a result of the t-to-m phase transformation.The transformation layer was thicker than 12 μm after 360 h of hydrothermal treatment.More importantly, no microcracks were found, even after an exhaustive search of this layer.
As illustrated in Figure 5, estimating the thickness of the transformation layer in the Ce-TZP/Al 2 O 3 specimens through SEM was challenging because of the absence of microcracks.The transformation layer was defined through TEM analysis based on the presence of twins.However, the FIB technique used in this study could only reach a depth of 12 μm, thus hindering the estimation of the thickness of the entire transformation layer, which was considerably thick.
The resulting strength decrease in the Ce-TZP/Al 2 O 3 specimen and the change in Weibull modulus were minimal after hydrothermal treatment for 120 h (Figure 7).The stress -strain curve of the Ce-TZP/Al 2 O 3 specimen was similar to that of the as-sintered specimen (Figure 9), indicating that no additional cracks were introduced into the Ce-TZP/Al 2 O 3 specimen during hydrothermal treatment.
The Weibull modulus of the Ce-TZP/Al 2 O 3 specimen treated for 360 h was lower (Figure 7(a)).Although the value of the Weibull modulus depends strongly on the number of specimens tested [31], a decrease from 7.2 to 5.0 for a sample size of 15 to 18 specimens is significant.This implies that additional flaws are introduced into the surface layer after long-term exposure to a hydrothermal environment.

Y-TZP specimen
Yttria-stabilized zirconia (Y-TZP) discs were prepared as reference specimens using the same die-pressing, presintering, and sintering techniques used above.The microstructure of the specimen is shown in Figure 1(b).The sintered density was higher than 98% and the mean size of ZrO 2 grains was relatively small (0.6 μm, Table 1).Only the t-phase was detected in the sintered specimens (Figure 3(b)).After hydrothermal treatment for 5 h at 134°C, the content of the m-phase increased from 0% to 8%.Although the content of the m-phase was low, the biaxial strength decreased by 22% (Figure 7(b)).The content of the m-phase increased substantially after hydrothermal treatment (Figure 4).Furthermore, the intensity of ZrO 2 [101] t diffraction peak @ 2θ of ~ 30 degree in degraded Y-TZP was decreased and this peak was no longer symmetrical (Figure 3(b)).The m-phase content quickly increased to a saturated value of 75% after 72 h in a hydrothermal environment (Figure 4).A charging surface layer resulting from the presence of pores and/or microcracks was observed in the SEM micrograph (Figure 5(b)).From the thickness of the surface charging layer, the thickness of the transformation layer after 24 h of hydrothermal treatment was approximately 5 μm. Figure 7(b) shows the Weibull distribution for the Y-TZP specimens after 0 h, 5 h, and 24 h of hydrothermal treatment.The biaxial strength decreased with the increase in the content of the m-phase, although the Weibull modulus remained almost constant (4.1, 3.8, and 4.8, respectively).The decrease in strength indicated its dependence on the increase in m-phase content (Figure 8).
The surface layer of the specimen after 24 h of hydrothermal treatment was observed using TEM.A thin section was collected from the surface region of the Y-TZP specimen through the FIB technique (Figure 10(a)).Twins were observed within the ZrO 2 grains, and microcracks were observed at grain boundaries (Figure 10(b)), with most of them being parallel to the surface.
The LTD of Y-TZP has been extensively investigated in recent decades [6,[19][20][21][22]32,33]. According to the standard for dental implants, ISO 13,356, "Implants for surgery -Ceramic materials based on yttriastabilized tetragonal zirconia (Y-TZP)", the m-phase content should be lower than 25% after 5 h of hydrothermal treatment at 134°C.The reported 5-year survival rates for single crowns made of Y-TZP based on the standard clinical practice were higher than 91.2% [34].The Mehl -Avhrami -Johnson (MAJ) equation is used to estimate the lifetime of zirconia implants from in vitro data [6,32].Although such a prediction is of great importance, Lughi et al. suggested that the lifetime prediction is accurate only if the activation energy [101] t reflection was broader and shifted to its lefthand side with the increase in hydrothermal treatment time.Moreover, a left shoulder was observed in the t-ZrO 2 reflection.Previous studies suggested that one XRD peak of orthorhombic or rhombohedral was also close to 2θ of ~ 30 degree [35,36].The XRD pattern thus suggests that the tetragonal phase not only transformed to monoclinic phase, it may also transformed to other phase.Consequently, the activation energy for the t-to-m phase transformation is likely changed.When the hydrothermal treatment time was increased to 24 h, the m-phase content increased to 60% (Figure 4) and the biaxial strength decreased by 40% (Figure 7  attributed to the formation of microcracks [8,37].Microcracks were observed in the transformation surface layer (Figure 10), and they tended to parallel to the surface.Many twins were also observed within the ZrO 2 grains and aligned with each other.These microstructural characteristics of Y-TZP were also reported by Munoz-Tabares et al. [37,38].This demonstrates that the Y-TZP specimens prepared in this study could be regarded as reference specimens.

Ce-TZP/Al 2 O 3 vs. Y-TZP
A summary for the characteristics of Ce-TZP/Al 2 O 3 and Y-ZTP after degradation is shown in Table 2.A longer hydrothermal treatment time was used for Ce-TZP/Al 2 O 3 specimen; it would produce a thicker degradation surface layer to facilitate the subsequent microstructure observation.The Ce-TZP/Al 2 O 3 exhibits a better resistance to hydrothermal degradation in terms of time in water.The present study also suggests that the microstructure characterization on the transformation layer of TZP is challenging.When microcracks are formed in the surface layer of Y-TZP specimen, an electron charging layer can be observed.This technique was frequently applied to evaluate the aging resistance of Y-TZP [9].However, the technique was not appliable to Ce-TZP/Al 2 O 3 , the system with little microcracks in the surface region.A TEM technique is therefore needed to characterize the transformation layer.The formation of microcracks is rare in Ce-TZP/Al 2 O 3 (Figure 6), this finding is also supported by a lower strength decrease.
Figure 8 shows the biaxial strength of Ce-TZP/Al 2 O 3 and Y-TZP specimens as a function of the m-phase content.The strength of the Ce-TZP/Al 2 O 3 specimen was higher than that of the Y-TZP specimen for the same m-phase content.However, m-phase content is not the only factor affecting the strength [18] and the formation of microcracks plays an important role in hydrothermal resistance.
Two types of transformations, homogeneous and autocatalytic, have been reported for TZPs [39,40].The phase transformation in Y-TZP is typically homogeneous, such that the transformation often starts at one grain boundary and ends at another boundary within the same grain.Because volume expansion is associated with the phase transformation, stresses are induced by the transformation.These stresses are released partially through the formation of twins and partially through the generation of microcracks [29].Both twins and microcracks were observed in the degraded Y-TZP specimens (Figure 10).Y-TZP Time of hydrothermal treatment/h 360 24 m-phase content/% 28 (Figure 3a, Figure 4) 60 (Figure 3b, Figure 4) Remaining strength after degradation/% 88% (Figure 7a, Figure 8) 60% (Figure 7b, Figure 8) Surface charging layer (SEM)/μm ~1 (Figure 5a) 5 (Figure 5b) Surface transformation layer (TEM)/μm >12 (Figure 6) >3 (Figure 10) The phase transformation in Ce-TZP/Al 2 O 3 is typically autocatalytic [41,42].However, a direct observation of its transformation surface layer has not yet been made.TEM analysis conducted in this study provided insights into the transformation layer.The TEM micrograph shown in Figure 6 indicates that many small twins are formed within the ZrO 2 grain.In addition, most of these twins were not aligned with each other.
Similar to elongated alumina grains with an anisotropic thermal expansion coefficient, the formation of microcracks in the twins likely requires a critical size [43].Because many small twins were randomly oriented and not aligned with each (Figure 6(b)), the stresses associated with the randomly oriented twins did not accumulate in the same direction, preventing the formation of microcracks.The decrease in strength for the Ce-TZP/Al 2 O 3 specimens treated in water was minimal because no microcracks were formed.

Conclusions
In this study, the hydrothermal resistance of Ce-TZP/Al 2 O 3 specimens was evaluated, with Y-TZP specimens used as references.Although t-to-m phase transformation was observed in both TZP systems, the degrees of strength decrease in a hydrothermal environment varied.This discrepancy can be attributed to the presence of microcracks in the transformation surface layer.The findings of this study highlight the importance of microstructure characteristics in determining LTD resistance.If no microcrack is formed during the t-to-m phase transformation, the influence on strength is minimal.However, a TEM technique is needed to characterize the transformation surface layer with a limited number of microcracks.

Figure 6 (
a) shows a thin section collected from the surface of the Ce-TZP/Al 2 O 3 specimen to a depth of 12 μm after hydrothermal treatment for 360 h.Many twins were observed within the ZrO 2 grains (Figure 6(b)), with one relatively large ZrO 2 grain shown in Figure 6(c).Some twins originated from the grain boundary and aligned with each other.The twins in other ZrO 2 grains were small and randomly oriented (Figure 6(b)).The formation of twins was taken as evidence of phase transformation from t to m [30].The selected area electron diffraction pattern of the

Figure 2 .
Figure 2. (a) Thin section taken from the surface region of Ce-TZP/Al 2 O 3 specimen before hydrothermal treatment.(b) TEM micrograph of the area indicated with a square in (a).(c) Diffraction pattern of the ZrO 2 grain near surface (circled).

ZrO 2
grain is shown in Figure6(d), indicating the presence of the m phase.The thickness of the transformation layer, as estimated from TEM micrographs (Figure6(a)), was greater than 12 μm.Barely any microcracks were observed in this transformation layer.

Figure 3 .
Figure 3. XRD diffraction patterns of (a) Ce-TZP/Al 2 O 3 and (b) Y-TZP specimens before and after hydrothermal treatments for various times.

Figure 4 .Figure 5 .
Figure 4. Amount of m-phase after hydrothermal treatment for various times.

Figure 6 .
Figure 6.(a) Thin section taken from the surface region of Ce-TZP/Al 2 O 3 specimen after hydrothermal treatment for 360 h.(b) TEM micrograph in the surface region as indicated with a square in (a).(c) TEM micrograph at higher magnification of the ZrO 2 grain indicated with a square in (b).(d) Diffraction pattern of the ZrO 2 grain shown in (c).

Figure 8 .Figure 9 .
Figure 8. Biaxial strength of Ce-TZP/Al 2 O 3 and Y-TZP specimens as a function of m-phase content.

Figure 10 .
Figure 10.(a) Thin section taken from the surface region of Y-TZP specimen after hydrothermal treatment for 24 h.(b) TEM micrograph in the surface region as indicated with a square in (a).

Table 2 .
Hydrothermal resistance of the Ce-TZP/Al 2 O 3 and Y-TZP prepared in this study; corresponding data can be found in the figures indicated in brackets.Ce-TZP/Al 2 O 3

Table 1 .
Microstructural characteristics of Ce-TZP/Al 2 O 3 investigated in this study.Y-TZP specimens were also prepared for comparison.