1. Introduction
Oxide ceramics are typical structural materials with wide applications, mainly used in cutting tools, bearings, high temperature engine components, and knee replacement prostheses, owing to their excellent ambient and high temperature mechanical properties. However, due to their strong chemical bonding and low diffusion coefficient, it is difficult to fabricate the sophisticated parts through forming methods like metals [
1,
2,
3]. In 1986, Wakai and co-workers [
4] reported the superplasticity of polycrystalline ceramics for the first time. Since then, some fine-grain ceramics have been regarded as superplastic materials at elevated temperature. Thus, in recent years, superplastic formation (SPF; including forging, extrusion, bending, stretching, and gas bulging forming) has been an appropriate technology to manufacture ceramic parts with complex shapes.
ZrO
2 toughened Al
2O
3 ceramic (ZTA) is an outstanding structural ceramic due to its excellent mechanical properties [
5,
6] and good superplastic deformability. Superplasitc deformation and net-shape forming of ZTA ceramics have been investigated by many researchers [
7,
8,
9,
10,
11,
12]. Wakai et al. [
12] investigated the superplastic deformation of ZTA ceramic, and found that its limited elongation is about 200% during deformation at 1450 °C. Chen and co-workers [
10] reported that the dense ZTA samples can be deep-drawn to a dome height of at least 12 mm at a high punch rate of 0.6 mm/min at 1400 °C. In most oxide ceramics, the superplastic deformation always needs a lower strain rate of <10
−5 s
−1 and a higher deformation temperature of >1450 °C. Obviously, for industrial applications, the superplastic formation needs a higher efficiency (higher strain rate) and a lower production cost (lower deformation temperature) [
13]. This means that the superplasticity of oxide ceramics should be enhanced to accommodate the industrial demands.
To improve the superplasticity of oxide ceramics, attempts have made been using dopants [
14]. Doping TiO
2 and MgO into ZrO
2 can achieve a large strain of 206% at a higher strain rate of 1.2 × 10
−4 s
−1 and a relatively lower temperature of 1350 °C [
15]. Ti/Mn co-doped Al
2O
3 with 3 mol. % ZrO
2 can be shape-formed to 100% strain at a temperature as low as 1280 °C [
16]. In our previous studies [
17], it was found out that introducing 8 wt.% TiO
2 into the ZTA ceramics can increase the strain rate by two orders of magnitude. The dominant deformation mechanism of the TiO
2-doped ZTA is grain boundary sliding. In fact, for the deformation of the polycrystalline ceramics, when the grains slide over each other, stress concentration will occur at grain boundaries and triple junctions [
14]. Hence, other sufficient accommodation (e.g., diffusion, dislocation migration) is needed to inhibit cavity nucleation induced by stress concentration. During the high temperature superplastic deformation, diffusion is a momentous accommodation mechanism that relaxes the stress concentration. Generally, the deformed apparent activation energy can be used for evaluating the accommodated diffusion process [
18,
19]. Sakka [
15] calculated the apparent deformation activation energy in Ti/Mn-doped ZrO
2 ceramic as 411 kJ/mol, which is close to the grain growth activation energy (447 kJ/mol). Combined with the microstructure, they found that Ti and Mn dopants can improve the cations’ diffusion, grain growth, and dislocation recovery during deformation. Especially when liquid phase occurs along the grain boundaries due to doping, the diffusion process during deformation may be changed. In the Al
2O
3 and ZTA ceramics [
20,
21,
22] with (CuO-TiO
2-B2O
3-MgO) liquid phase, the grain boundary diffusion was enhanced significantly, which is in accordance with obvious alteration of the apparent deformation activation energy. Therefore, for the TiO
2-doped ZTA ceramics with outstanding advantages, further investigation of the relationship between the apparent deformation activation energy and grain boundary feature evolution during deformation is crucial to discuss the deformation accommodation mechanisms.
In the current study, the compressive deformation of the TiO2-doped ZTA ceramics at different temperatures of 1300–1400 °C were conducted to estimate their apparent deformation activation energy for determination of the respective diffusion path. Moreover, the grain boundary features and their evolution during the superplastic deformation were also investigated to clarify the accommodation deformation mechanisms.
2. Materials and Methods
The raw materials Al
2O
3 and ZrO
2 were synthesized at a mass ratio of 58:42 by heating ethanol-aqueous salt solutions [
23,
24]. The particle size of Al
2O
3 and ZrO
2 powders is approximately 50 nm. Commercial rutile TiO
2 (Nanjing Guanye Co. Ltd., Nanjing, China) purity ≥99.8%, particle size ≤ 80 nm) was selected as the dopant in this study (the doping contents are 0, 1, 4, and 8 wt.%, denoted as 0T, 1T, 4T, and 8T, respectively). Then, the mixed powders were sintered at 1400 °C to be used for investigation of compressive deformation. The grain size and grain shape were dependent on the TiO
2 content, as the reference reported [
17]. The detail methodology of sintering TiO
2-doped ZTA can be found in our previous study [
25]. To evaluate the stress exponent and apparent activation energy, the compression deformation was conducted at 1300 °C and 1350 °C by using the “stress-jump” method. The deformation behavior of TiO
2-doped ZTA ceramics at 1400 °C was used as reference [
17]. The deformation mold was made of high-strength graphite. When the deformation temperature reached the given value, the specimens were kept at the set temperature for 10 min before compressive testing. The temperature during deformation was measured using a platinum-rhodium thermocouple. The deformation test was conducted by using high temperature vacuum sintering furnaces. The deformation parameters in superplastic ceramics are calculated by the following constitutive equation:
where
is the strain rate,
A is a constant,
σ is the flow stress,
n is the stress exponent,
d is the grain size,
p is the grain size exponent,
Q is the activation energy,
R is the gas constant, and
T is the absolute temperature. The compressive strain
ε, the flow stress
σ, and the strain rate
were calculated by the directly measured load
P, immediate
Hx, initial height
H0, and initial end face area
A0, as a previous study [
17] reported.
The microstructure was characterized by means of scanning electron microscopy (SEM, Supra-55, Carl Zeiss Sigma NTS Gmbh, Jena, Germany). The SEM specimens were polished with a diamond paste of 1.5 μm and then thermally etched in air for 1 h at a 50 °C lower temperature than the deformation temperature of each specimen. The average grain sizes of the two phases were calculated using the linear intercept method: d = 1.56L, L is the intercept length. The grain boundary feature was characterized by using transmission electron microscope (TEM, Tecnai G220, FEI, Eindhoven, the Netherlands). The TEM thin foils were cut for Φ 3 mm disks from the different TiO2-doped ZTA ceramics. The foils were pre-thinned down to a few micrometers thickness by dimpling. A final thinning was performed by gentle ion milling.
3. Results and Discussion
The compressive deformations of the TiO
2-doped ZTA ceramics (0T, 1T, 4T, and 8T) were conducted in this study. To investigate the deformation behavior at different flow stress levels, the applied force was changed two or three times in one deformation testing at a given deformation temperature.
Figure 1 shows the strain rate with strain at different flow stress levels for the four TiO
2-doped ZTA ceramics conducted under 1300 °C. As can be seen in
Figure 1, when the applied force was just changed, the transition deformation occurred first. A moment later, a relatively steady-state deformation occurred subsequently, as shown in
Figure 1. In this study, the applied force was not changed until the occurrence of a short steady-state deformation. Considering that the cross sectional area of the specimens was gradually changed with deformation strain, the flow stress
σflow was approximately modified according to the following equation:
where
hx is the height of the specimen at a certain moment,
S0 and
h0 are the initial area of the end face and the initial height of the specimen, respectively. All flow stresses reported in this paper have been already modified by using the upper equation. It is shown that the measured strain rate was obviously related to the applied stress. For the comparative analysis of the deformation behaviors of 0T–8T at different temperatures, the relationships between flow stress and strain rate are plotted on a logarithmic scale in
Figure 2. In this plot, the flow stress and strain rate are only taken from the steady-state deformation stages. As can be seen in
Figure 2, with the decrease of the deformation temperature, the strain rate at the same flow stress decreased. In the previous studies, during the superplastic deformation at the higher temperature of 1400 °C, the grain growth is extremely slight. Thus, the influence of the grain growth during the deformation is not considered in the present study. In a word, the grain size
d, grain growth exponent
p, and the activation energy
Q in the same specimen are assumed as constants. Therefore, the stress exponent (
n value) is calculated from the slope lines according to
. In the present study, it is found that the
n values of the same material are almost equal at each deformed temperature. The values of stress exponents are in the range of 3.06–3.18 for 0T, 2.87–2.92 for 1T, 2.35–2.44 for 4T, and 2.03–2.16 for 8T, respectively. It is indicated that the deformation mechanism of the same material may be unchanged in the range of 1300–1400 °C.
The typical microstructures of the undoped and 8 wt.% TiO
2-doped ZTA ceramics after deformation at 1300 °C and 1350 °C are shown in
Figure 3. Compared with the undeformed [
25] and 1400 °C deformed specimens [
17], the average grain size and grain shape are almost unchanged for each material. For the undoped ZTA ceramic, no conspicuous texture of the elongated Al
2O
3 grains was formed, such as the similar material deformed at 1400 °C in the previous study [
17]. This may be because at the lower deformed temperatures of 1300 °C and 1350 °C, the strain rate of ZTA is extremely low, and the accumulated strain in the whole deformation is too small to form the conspicuous texture of the elongated Al
2O
3 grains. For the 8 wt.% TiO
2-doped ZTA ceramics, whether deformed at 1300 °C, 1350 °C, or 1400 °C, there is no obvious difference in grain size, grain shape, or grain boundary liquid phase. To understand the element composition of the liquid phase, the SEM-EDS (Energy Dispersive Spectrometer) results are shown in in
Figure 4. However, because the diameter of the region detected by SEM-EDS is larger than the liquid phase, it is difficult to examine the composition of the liquid phase accurately. Nevertheless, the SEM-EDS results can still be used for contrasting analysis of the element composition in different regions. As can be seen in
Figure 4, the regions of 1#, 2#, and 3# are located in the Al
2O
3 grain, ZrO
2 grain, and liquid phase, respectively. The Ti concentrations in Al
2O
3 and ZrO
2 are detected as 0.5 mol.% and 2 mol.%, which are basically in agreement with the limited solubility of TiO
2 in Al
2O
3 (<0.27 wt.%) and ZrO
2 (<16 wt.%). The higher Ti concentration of 6.9 mol.% in the liquid phase indicates that the liquid phase is rich in Ti element. In particular, the real value of Ti concentration in the liquid phase should be much higher than the measured value (6.9 mol.%), because the region 3# includes the unsolicited regions located at Al
2O
3 and ZrO
2 grains. Many Ti cations segregated on the grain boundary may be the reason of the formation of grain boundary liquid phase. Notably, in the previous studies, the lowest temperature of the Al
2O
3-ZrO
2-TiO
2 system is 1580–1610 °C, which is higher than the sintering and deformation temperature of 1300–1400 °C. Thus, the grain boundary liquid phases in this study are more likely to be generated owing to the solid-state phase transformation below the bulk eutectic temperature. Combined with the similar
n values for each individual material, the deformation mechanism is not altered with the temperature changing at 1300–1400 °C. Based on the microstructure evolution and
n values in the range of 2–3, the dominant deformation mechanism is grain boundary sliding. For the polycrystalline ceramics, when the grains slide over each other, stress concentration will occur at grain boundaries and triple junctions. If the strain rate is unduly fast, some cavities will nucleate and then form cracks at grain boundaries during superplastic deformation. Hence, grain boundary sliding must be accommodated by another process in order to keep the grains from separating. It has been proposed that rapid and sufficient dislocation motion or diffusion is necessary during deformation to prevent cavity formation [
14,
26]. Generally, the deformed apparent activation energy can be used to evaluate the accommodated process. Therefore, the alterations of the deformed apparent activation energy and grain boundary feature are discussed as follows.
Figure 5 shows the relationship between the strain rate and reciprocal absolute temperature for the ZTA ceramics doped with different amount of TiO
2. The slope of the straight line obtained from an Arrhenius type is used to determine the apparent deformation activation energy of the high temperature deformation. The values of the apparent activation energy in the undoped ZTA ceramic are calculated as 749 ± 35 kJ/mol (under 40 MPa) and 705 ± 29 kJ/mol (under 60 MPa), which are higher than those of pure Al
2O
3 (420–490 kJ/mol) [
27] and pure ZrO
2 (500–600 kJ/mol) [
28] ceramics. The higher
Q value in ZTA is probably related to the diffusion of species along the ZrO
2-Al
2O
3 interface (inter-diffusion phenomena between Al
2O
3 and ZrO
2 grains) [
29]. The sliding mobility of Al
2O
3-Al
2O
3 interfaces may be reduced by the segregation of Zr
4+ along these interfaces [
30]. The apparent deformation activation energy (705–749 kJ/mol) in the undoped ZTA ceramic is very close to the activation energy of sintering (700 ± 100 kJ/mol) [
31], superplasticity (740 kJ/mol) [
32], and grain growth (732 kJ/mol) [
32] in the typical ZTA ceramics, which are in agreement with those for grain boundary diffusion controlled by interface reaction reported in the ZTA ceramics with clean grain boundaries [
32].
For the ZTA doped with 1–8 wt.% TiO
2, the apparent activation energies (Q values) are calculated as
Q1T = 635–658 kJ/mol,
Q4T = 609–625 kJ/mol, and
Q8T = 589–605 kJ/mol, respectively. With the increase of TiO
2 concentration, the apparent activation energy of the TiO
2-doped ZTA ceramics decreased, which indicates that the interfacial diffusion during deformation may be changed by TiO
2 doping. Considering the occurrence of the glassy phase in the TiO
2-doped ZTA [
25], the alteration of the activation energy may be influenced by the grain boundary feature dependent on TiO
2 content. Generally, the alteration of grain boundary feature may affect the grain boundary transport rate. Harmer and his co-workers [
33,
34] summarized the grain boundary migration with different features for Al
2O
3 ceramics doped with various dopants. Their results indicated that the mass transport rate on the grain boundary increases with an increase of the width of the grain boundary phase [
33,
34]. In this study, the typical grain boundary features of the four deformed ceramics are shown in
Figure 6. Unlike the undoped ZTA ceramic with a clean grain boundary (
Figure 6a), grain boundary liquid phases with different features are found in the deformed TiO
2-doped ZTA ceramics, as shown in
Figure 6b–d. For the 1 wt.% TiO
2-doped ZTA ceramic, several thin liquid grain boundary phases (thickness is approximate 1–2 nm) can be observed at A-Z (meaning Al
2O
3-ZrO
2 interfaces), A-A, and Z-Z interfaces. Notably, not all grain boundaries are covered by the thin liquid phases. Some grain boundaries are still clean. For 4 wt.% and 8 wt.%TiO
2-doped ZTA ceramics, besides the thin grain boundary liquid phase, lots of thick grain boundary liquid phases (the thickness is 50–100 nm) (
Figure 6c,d) were also observed at grain boundaries and triple junctions. Therefore, with the increase of TiO
2 content, the grain boundary diffusion should be improved significantly due to the changing of the grain boundary feature, which is according to the decrease of the apparent activation energy and increase of the strain rate at a given temperature and stress during deformation.