Anisotropy oxidation behavior and mechanism of textured Ti₃(SiAl)C₂ ceramic

Abstract

Adjusting the grain growth orientation may be a feasible way to improve the oxidation resistance. In this work, textured Ti₃(SiAl)C₂ ceramic was successfully fabricated via the spark plasma sintering technique. The crystal orientation and microstructure were investigated by XRD combined with EBSD. The oxidation behavior of textured Ti₃(SiAl)C₂ was investigated at 1000–1300 °C for 10 h. The results showed that the oxidation was significantly anisotropic and the surface parallel to the compression direction exhibited better oxidation resistance below 1200 °C. The improved oxidation resistance was primarily attributed to the formation of passivated Al₂O₃ scale by rapid out-diffusion of Al elements on the orientated (hko) planes in textured Ti₃(SiAl)C₂. The formation of oxide scales was strongly dependent on the crystallographic orientation of Ti₃(SiAl)C₂.

Introduction

The ternary layered Ti₃SiC₂ is the most representative compound of M_n+1AX_n phases, where M is a transition metal, A is a group A element, X is C or N, and n = 1, 2, or 3. It possesses a hexagonal crystal structure, which consists of MX lamellae and A-A atomic layer arranged in alternating stacking order along the c-axis direction. Thus, the coexistence of covalent/ionic and metallic bonds in the chemical structure allows it to combine the properties of metals and ceramics, such as low density, easy machinability, high thermal and electrical conductivity, good damage tolerance, thermal shock resistance, as well as good high-temperatures oxidation resistance, etc^1,2,3. The combination of these properties makes Ti₃SiC₂ a candidate material for high-temperature structural applications.

Consequently, the high-temperature oxidation behavior of Ti₃SiC₂ has been extensively investigated in the last decade^4,5. For example, Sun et al.⁴ suggested that at 900–1100 °C, the oxide scales were composed of coarse-grained TiO₂ in the outer layer and fine-grained TiO₂ and SiO₂ in the inner layer. However, the oxidation resistance of Ti₃SiC₂ is not satisfactory above 1100 °C. Moreover, TiC always exists in Ti₃SiC₂ as impurities, which is detrimental to the high-temperature oxidation resistance. In this regard, Zhang et al.⁶ introduced Al as a sintering aid into the Ti-Si-C system, which effectively inhibited the formation of TiC and also improved the oxidation resistance of Ti₃SiC₂. The improvement in oxidation resistance is attributed to the high activity and diffusivity of Al, which inhibits the inward diffusion of oxygen by forming a passivated Al₂O₃ scale during high-temperature oxidation. Normally, the added Al is present in Ti₃SiC₂ as a solid solution, i.e., Ti₃(SiAl)C₂.

In recent years, textured ceramic materials have attracted much attention because of their special physical and mechanical properties^{7,8,9,10,11,12}. For example, Hu et al.⁸ successfully fabricated textured Nb₄AlC₃ using the strong magnetic field alignment (SMFA) method combined with spark plasma sintering (SPS). The flexural strength and fracture toughness were significantly improved. Moreover, Li et al.⁹ studied the oxidation resistance of textured Ti₂AlC and Ti₃AlC₂. Xu et al.¹⁰ fabricated textured Ti₃AlC₂ by SMFA combined with the SPS technique. They found that textured Ti₃AlC₂ exhibited significantly anisotropic oxidation behavior and the surface showed better oxidation resistance along the c-axis direction. For Ti₃(SiAl)C₂, it possesses a hexagonal crystal structure and anisotropic properties along the a- or b-axes and c- axis. Therefore, the oxidation behavior of Ti₃(SiAl)C₂ can be expected to follow the anisotropy of its crystal structure. However, the process of preparing textured MAX phases by the SMFA method is complex and costly. Hence, a simple and efficient preparation method, i.e., the SPS technique, was implemented to induce grain texture of the MAX phase by loading^13,14. The sintered MAX grains exhibit a clear preferential orientation and the obtained bulk material has mutually parallel basal planes, where a- and b-axes are randomly aligned in a plane orthogonal to c-axis and c-axis is neatly aligned.

However, most of the available works have focused on untextured Ti₃(SiAl)C₂, and to our knowledge, the oxidation behavior of textured Ti₃(SiAl)C₂ has not yet been reported. Therefore, this work aims to investigate the anisotropic microstructure and oxidation behavior of textured Ti₃(SiAl)C₂ fabricated by the SPS technique.

Results and discussion

Texture

Figure 1 presents XRD patterns collected from TTS and TSS samples. The diffraction peaks in all (00l) planes of TTS were stronger than those on TSS and standard PDF card. It suggested that (00l) crystallographic orientation (c-axis) was parallel to the compression direction, which indicates that the as-prepared Ti₃(SiAl)C₂ formed a textured structure. This texture is also verified by the diffraction peaks on TSS, which exhibited the strongest diffraction peak at $(10\bar 14)$, while on TTS, the strongest diffraction peak shifted to (0008), corresponding to (00l) planes. Based on XRD data, the Lotgering orientation factor was calculated. $f_{(00l)}$ was 0.62, which was higher than the value of Ti₃SiC₂ sintered by high-temperature deformation and slip-casted in the static magnetic field, but lower than the value of Ti₃SiC₂ slip-casted in the rotating magnetic field^15,16. In general, a larger value of $f_{(00l)}$ means a higher orientation along the compression direction. And $f_{(hk0)}$ was only 0.11, which was lower than the values of Ti₃SiC₂ slip-casted in the static and rotating magnetic fields¹⁶. This result suggests that the SPS technique allows alignment of the c-axis of the grains, but has limited control over the a- and b-axis orientations.

Fig. 1

XRD patterns of top and side surface of textured Ti₃(SiAl)C₂.

In addition, it can be seen that a small amount of Al₂O₃ (2θ = 37.77° and 43.35°) existed in the sample as an impurity. The presence of Al₂O₃ was unavoidable because it always existed in the prepared Ti₃(SiAl)C₂ powders.

Figure 2 shows SEM images of the fracture surfaces of textured Ti₃(SiAl)C₂. As shown in Fig. 2a, only transgranular fractures were observed and no lamellar features were present on the top surface, which is due to the weak bond between the base planes of Si or Al atoms and Ti atoms¹⁰. The crack propagates easily along the weak substrate plane. A typical lamellar structure can be observed on one side surface (Fig. 2b), where Ti₃(SiAl)C₂ grains showed a clear intercalation stacking along the c-axis. Such a tailored ceramic should have anisotropic properties along the top and side directions of the texture. This also demonstrates that SPS can prepare textured samples.

Fig. 2: SEM images of fractured surfaces of textured Ti₃(SiAl)C₂.

a TTS sample, b TSS sample.

To reveal the phase distribution and grain orientation, EBSD was conducted on TSS (the acquisition surface perpendicular to the compression direction) and TTS (the acquisition surface parallel to the compression direction) surface. The phase distribution and constructed orientation maps of the grains are shown in Fig. 3. By observing the phase distribution maps (Fig. 3a, b), the red grains were identified as Ti₃(SiAl)C₂ phase and the blue grains as Al₂O₃ phase, which is consistent with XRD results. As shown in Fig. 3c, d, different colors indicate different grain orientations on TSS and TTS. It is noticeable that on TSS, most of the grain-exposed crystal planes were (00l) planes. While on TSS, there are almost no (00l) planes exposed. On the contrary, $(01\bar 10)$ and $(1\bar 210)$ planes were the main grain-exposed crystal planes.

Fig. 3: EBSD characterization.

Phases distribution maps of TTS (a) and TSS (b) samples. Grain orientation and Inverse pole figure (IPF) of TTS plane (c) and TSS plane (d).

Note that, the volume fractions of the Al₂O₃ phase in TTS and TSS were 2.86% and 4%, respectively, as determined by EBSD. The presence of Al₂O₃ also inhibited the grain growth of Ti₃(SiAl)C₂ during sintering. Figure 4 presents the average grain size of Ti₃(SiAl)C₂ particles, which was statistically about 0.98 μm for TTS and about 0.89 μm for TSS. Besides, the grain size may also be caused by the different orientations. The large grain size of TTS leads to a decrease in the number of grains observed by EBSD.

Fig. 4: The grain size of textured Ti₃(SiAl)C₂.

a TTS sample, b TSS sample.

In addition, (0001), (1$\bar 2$10), and (01$\bar 1$0) pole figures derived from the stereographic projections of EBSD orientation data showed the texture orientation of TTS and TSS surfaces, as shown in Fig. 5. The highest texture intensities of TTS and TSS surfaces were 6.02 and 6.57 multiples of uniform density (MUD), respectively. These results of EBSD analysis are consistent with the orientation factors calculated from XRD data, which also demonstrates the high texture of the obtained textured Ti₃(SiAl)C₂ material.

Fig. 5: The pole figure of textured Ti₃(SiAl)C₂.

a TTS sample, b TSS sample.

Based on the above analysis, a preferential crystallographic orientation of the basal plane perpendicular to the compression direction was obtained in the as-prepared Ti₃(SiAl)C₂, thus providing the possibility to study its oxidation resistance.

Oxidation

To test the oxidation behavior of textured Ti₃(SiAl)C₂, the thermogravimetric analysis tests were conducted in air at 1000, 1100, 1200, and 1300 °C for 10 h, respectively.

Figure 6 displays the oxidation kinetics (weight change per unit area vs oxidation time) of TTS and TSS during the oxidation at different temperatures. The final mass gains at different temperatures are summarized in Table 1. The mass gain of both samples increased with increasing oxidation time. The oxidation gain curves showed a parabolic tendency below and at 1200 °C, confirming good oxidation resistance, which is consistent with the previously reported results⁴. When the temperature exceeded 1200 °C, the oxidation weight gain curves showed a linear tendency, indicating poor oxidation resistance. The TSS sample showed an overall better oxidation resistance compared to the TTS sample. After oxidation of 10 h, the weight gains of the TTS sample at 1000, 1100, 1200, and 1300 °C were 0.92, 1.60, 3.80, and 29.68 g m⁻², respectively. While for the TSS sample, the final weight gains were 0.54, 1.34, 3.74, and 35.43 g m⁻² at 1000, 1100, 1200, and 1300 °C, respectively. In addition, the weight gain ratios of TTS to TSS after 10 h were 1.70, 1.19, 1.02, and 0.84 at 1000, 1100, 1200, and 1300 °C, respectively. The lower the temperature, the larger the ratio. These results revealed that texturing can improve the oxidation resistance of Ti₃(SiAl)C₂.

Fig. 6

The weight gain per unit area as a function of time for textured Ti₃(SiAl)C₂ oxidized in the temperature range from 1000 to 1300 °C.

Table 1 Summary of the weight gain per unit area (g m⁻²) for TTS and TSS samples at different temperatures.

According to Fig. 6, the oxidation weight gain of textured Ti₃(SiAl)C₂ obeys a parabolic law at 1000–1200 °C. The corresponding square of weight gain per unit area as a function of oxidation time is shown in Fig. 7. Therefore, the correlation between the weight gain per unit area and oxidation time could be described as¹⁷:

$$\left( {\frac{{\Delta W}}{A}} \right)^2 = k_p\cdot t$$

(1)

where △W/A is the weight gain per unit area, k_p is the parabolic rate constant, and t is the oxidation time.

Fig. 7

Square of weight gain per unit surface area as a function of oxidation time for textured Ti₃(SiAl)C₂ oxidized at different temperatures.

When the temperature increased to 1300 °C, the oxidation weight gain of both samples deviated from the parabolic law and followed a linear law, indicating that the rate-controlling step became an oxidation reaction between Ti₃(SiAl)C₂ and O₂. Therefore, the correlation between the weight gain per unit area and oxidation time could be described as:

$$\frac{{\Delta W}}{A} = k_p\cdot t$$

(2)

The parabolic and linear rate constants for two samples oxidized at different temperatures are summarized in Table 2. After oxidation for 10 h, the oxidation rate constants for TTS samples were 2.4 × 10^–9 kg² m⁻⁴ s⁻¹, 8.5 × 10⁻⁹ kg² m⁻⁴ s⁻¹, and 4.6 × 10⁻⁸ kg² m⁻⁴ s⁻¹ at 1000 °C, 1100 °C, and 1200 °C, respectively. However, for TSS samples, the oxidation rate constants were 7.9 × 10⁻¹⁰ kg² m⁻⁴ s⁻¹, 5.8 × 10⁻⁹ kg² m⁻⁴ s⁻¹, and 4.4 × 10⁻⁸ kg² m⁻⁴ s⁻¹ at 1000, 1100, and 1200 °C, respectively. At the same temperature, based on the total test area occupied by the faces perpendicular to the c-axis (a pair faces) and parallel to the c-axis (two pair faces), their oxidation rate constants were calculated, and also summarized in Table 2.

Table 2 Summary of k_p (1000–1200 °C, parabolic, kg² m⁻⁴ s⁻¹) and k_l (1300 °C, linear, kg m⁻² s⁻¹) for Ti₃(SiAl)C₂.

The relationship between the parabolic rate constant (k_p, kg² m⁻⁴ s⁻¹) and temperature (T, K) could be described by the Arrhenius’ equation:

$$k_p = A\cdot {{{\mathrm{exp}}}}\left( - \frac{Q}{{RT}}\right)$$

(3)

where A is the pre-exponential factor, Q is the apparent or effective activation energy, R is the gas constant, and T is the absolute temperature. Take the logarithm on both sides of Eq. (3):

$$Ln\left( {k_p} \right) = Ln\left( A \right) - \frac{Q}{{RT}}$$

(4)

Arrhenius plots of the logarithm of k_p for TTS* and TSS* vs the inverse of the absolute temperature are shown in Fig. 8. The data point of 1300 °C was excluded from the activation energy calculation because their oxidation pattern was deviated from the parabolic law. The values of Q were obtained from the slope of the linear fit in Fig. 8. The results show that the Q values of TTS* and TSS* were 202 ± 30 kJ mol⁻¹, and 401 ± 17 kJ mol⁻¹, respectively. The Q value of TSS* is close to 350–380 kJ mol⁻¹, which is the activation energy value for oxygen diffusion at Al₂O₃ grain boundaries^18,19. Smialek et al.²⁰ studied the oxidation behavior of Al-containing MAX phases and revealed that oxygen diffusion at the grain boundaries of the Al₂O₃ scale was the primary oxidation mechanism of MAX phases. Therefore, textured Ti₃(SiAl)C₂ is no exception to this rule. Besides, Gao et al.²¹ reported that the Q value of purity Ti₃SiC₂ was 295 ± 20 kJ mol⁻¹ in the temperature range of 1150–1250 °C. A higher Q means that the oxidation rate constant of the material is more variable in the temperature range tested. However, at 1200 °C, the oxidation rate constants of the two samples were close, indicating that the lower the temperature, the greater the difference in the oxidation rate constants of the two samples. Based on the present investigation, it can be concluded that the oxidation resistance of TSS samples was superior to that of TTS samples below 1200 °C. Also, it can be inferred that the lower the temperature, the better the oxidation resistance of TSS sample.

Fig. 8

Arrhenius plot of the parabolic oxidation constants for the isothermal oxidation at 1273–1573 K for textured Ti₃(SiAl)C₂.

Figure 9 illustrates XRD patterns of TTS and TSS oxidized samples at 1000–1300 °C. The phase compositions of the two sample surfaces at different oxidation temperatures are summarized in Table 3. It can be seen that the main phase of the TTS sample was TiO₂ at 1000 °C, while the main phase of the TSS sample was Ti₃(SiAl)C₂, which indicates that the thickness of the oxide layer of the TSS sample was thinner than that of TTS sample, i.e., the oxidation resistance of TSS sample was superior to that of TTS sample. At and above 1100 °C, the main phase of both samples was TiO₂, while the Ti₃(SiAl)C₂ phase could only be detected at 1100 °C, which indicates that the samples underwent rapid oxidation when the temperature exceeded 1100 °C. Below 1200 °C, the Al₂O₃ phase could be observed in all samples. Previous work demonstrated that the formation of a passivated Al₂O₃ layer contributes to the oxidation resistance, and it explains well the very limited weight gain of the samples below 1200 °C (see Fig. 6)^6,10. While at 1300 °C, the diffraction peak of Al₂O₃ disappeared to be replaced by Al₂TiO₅ and the diffraction peak of TiO₂ (2θ = 36.1°) became significantly weaker. Al₂O₃ reacted with TiO₂ to form Al₂TiO₅, which led to the loss of the protective layer and a significant weight gain^10,22. In addition, the TSS samples showed higher oxidation rate constants, which was also mainly because the TSS samples generated more Al₂O₃ in the initial oxidation stage, and as the oxidation time increased, Al₂O₃ reacted with TiO₂ to form Al₂TiO₅.

Fig. 9

XRD patterns of scales grown on textured Ti₃(SiAl)C₂ at different temperatures.

Table 3 Phase composition on the surface of TTS and TSS samples after oxidation at different temperatures.

Figure 10 shows the surface morphologies of TTS and TSS oxidized samples at 1000–1300 °C. It can be seen that the grain size increased significantly with the increase in temperate. Figure 10a–d shows bright TiO₂ grains dispersed on small gray Al₂O₃ grains in the oxidized samples at 1000-1100 °C. As the temperature increased to 1200 °C, TiO₂ gradually grew from small grains to large elongated crystals and no Al₂O₃ grains were observed (Fig. 10e, f). At 1300 °C, TiO₂ grew from grains to large ridge-like structures. Al₂O₃ was substituted by Al₂TiO₅, which was dispersed on TiO₂ (Fig. 10g, h). The looser or porous TiO₂ and Al₂TiO₅ scales cannot prevent further diffusion of oxygen, resulting in a rapid increase in the weight of the sample¹⁰. In addition, below 1200 °C, the grain size of TTS oxidized samples was larger than that of TSS samples, which corresponded well with the oxidation weight gain curve (Fig. 6). At 1300 °C, the grain size of TTS oxidized samples was smaller than that of TSS samples, which may be related to the generation of Al₂TiO₅.

Fig. 10: SEM images of sample surfaces after the oxidation of TTS and TSS samples tested at different temperatures.

a TTS-1000 ℃, b TSS-1000 ℃, c TTS-1100 ℃, d TSS-1100 ℃, e TTS-1200 ℃, f TSS-1200 ℃, g TTS-1300 ℃, h TSS-1300 ℃.

To understand the oxidation mechanism of textured Ti₃(SiAl)C₂, the cross-sectional morphology of both samples after oxidation at 1000–1300 °C for 10 h was observed by SEM and EDS. Typical SEM images of the oxidized scales formed after oxidation at different temperatures and the corresponding EDS analysis are shown in Figs. 11 and 12, respectively. EDS images of the oxidized sample cross-sections at 1100 °C are similar to those at 1000 °C. By comparing the cross-sectional morphology of the two samples, the oxide layers of both samples showed the same multilayer structure and phase compositions at the same temperature. At 1000 °C and 1100 °C, a thin and stratified oxide layer was generated on the surface of the samples, which consisted of TiO₂ in the outer layer, Al₂O₃ in the middle layer, and a mixed layer of TiO₂ + SiO₂ in the inner layer. At 1200 °C, the oxide layer of the sample showed a more pronounced multilayer structure, which consisted of TiO₂ in the outer layer, an intermediate layer of Al₂O₃ interspersed inside the TiO₂ layer, and an inner layer of TiO₂ + SiO₂ mixed porous layer. While at 1300 °C, the oxide layer of the sample changed significantly, which consisted of an outer layer of discontinuous Al₂TiO₅, an intermediate layer of TiO₂, and an inner layer of TiO₂ + SiO₂ mixed layer.

Fig. 11: SEM images of cross sections of scales grown on TTS and TSS samples tested at different temperatures.

a TTS-1000 ℃, b TSS-1000 ℃, c TTS-1100 ℃, d TSS-1100 ℃, e TTS-1200 ℃, f TSS-1200 ℃, g TTS-1300 ℃, h TSS-1300 ℃.

Fig. 12

EDS mappings of cross sections of scales grown on TTS and TSS samples tested at different temperatures.

On the other hand, for TTS samples, the middle layer (Al₂O₃) of the oxide layers of TTS did not connect well with the outer layer (TiO₂) and the inner layer (TiO₂ + SiO₂ mixed layer) with the existence of faint cracks at 1000–1200 °C, which may be due to the lack of formation of a continuous Al₂O₃ passivation layer¹⁰. As a result, the inner oxide layer (TiO₂ + SiO₂ mixed layer) was significantly thicker compared to that of TSS, as shown in Figs. 11 and 12. The oxide layer thicknesses were 12 ± 0.5 μm, 19.8 ± 3.5 μm, and 33.8 ± 2.2 μm at 1000 °C, 1100 °C, and 1200 °C, respectively. For the TSS sample, at 1000–1200 °C, it can be seen that the Al₂O₃ layer was well bonded to the outer and inner layers, and no cracks were observed. The scale thicknesses were 7.9 ± 0.9 μm, 14.4 ± 1.2 μm, and 29.2 ± 2.4 μm at 1000 °C, 1100 °C, and 1200 °C, respectively.

At 1300 °C, no Al₂O₃ phase was connecting between the outer and inner layers, resulting in significant cracks separating the inner and outer layers in both samples. This is because Al rapidly diffuses outward at high temperatures to form Al₂O₃, which reacts with TiO₂ to form Al₂TiO₅, at the same time, loses its protective effect on the substrate, leading to the violent oxidation of the samples. In addition, the thickness of the oxide layer of the TTS sample was lower than that of the TSS sample, which is consistent with the results in Fig. 6. In general, we can observe that higher oxidation temperatures lead to more rapid oxidation and thicker oxide layers.

Oxidation mechanism of texture Ti₃(SiAl)C₂

Based on the above analysis, the oxidation process of Ti₃(SiAl)C₂ was controlled by the inward diffusion of O and the outward diffusion of C, Ti, and Al, while Si was relatively immobile, which is in agreement with the results of the previous studies^6,23. Therefore, the evolution of the microstructure of oxide scales and the atomistic model can be proposed to describe the formation mechanism of oxide scales on Ti₃(SiAl)C₂, as shown in Fig. 13.

Fig. 13

Schematic illustration of the evolution of the microstructure of oxide scales and an atomistic model describing the formation mechanism of oxide scale on Ti₃(SiAl)C₂.

The microstructure and growth kinetics of the different oxide scales are related to the crystallographic orientation of the crystals, as mentioned above, and the exposed grains on TTS correspond mainly to (00l) planes, whose planes were formed mainly by MX layers (TiC_x layers). But on TSS, most of the grains were (hko) planes, whose planes were formed by stacking MX layers and Si(Al)-Si(Al) atomic layers. As a result, TTS and TSS planes contain different Al contents. Moreover, Wang et al.²⁴ investigated the thermal stability of Ti₃AlC₂ powders at low oxygen partial pressure. The results showed that Al underwent selective oxidation on Ti₃AlC₂ and Ti₃AlC₂ powder was partially transformed into non-stoichiometric TiC_x and Al₂O₃, demonstrating the higher chemical activity of Al than other elements. For Ti₃(SiAl)C₂ material, the element Al can also be considered as an important element for the formation of the Al₂O₃ protective layer as a further oxidation-inhibiting layer.

For TTS samples, the (00l) plane was parallel to the oxide layer and the Si(Al)-Si(Al) atomic plane was perpendicular to the direction of Al diffusion, resulting in a slower diffusion of Al atoms from the (00l) plane to the oxide reaction layer, which had to penetrate the MX layer, thus making the growth of the Al₂O₃ layer more complicated and discontinuous. In contrast, for the TSS sample, the Si(Al)-Si(Al) atomic plane was parallel to the direction of Al diffusion, allowing rapid diffusion of Al atoms from (hko) plane to the oxide layer, which contributed to the formation of a continuous Al₂O₃ layer. Thus, the continuous Al₂O₃ layer can protect the TSS sample from further oxidation. Whereas, the low content of Al in the samples and the retarded diffusion of Al atoms in the TTS sample led to severe oxidation of the TTS sample.

From the above results, we can know that the element Al plays a very important role in the oxidation resistance of textured Ti₃(SiAl)C₂. Due to the low content of Al, the ease of its outward diffusion in TTS and TSS determines whether a continuous Al₂O₃ protective layer can be formed. Therefore, the texture affects the oxidation behavior of texture Ti₃(SiAl)C₂.

As for Ti₃SiC₂ without Al solid solution, however, its oxidation process was controlled by the inward diffusion of O and the outward diffusion of Ti and C, while Si was kept immobile. This process does not involve the ease of Si external diffusion. Therefore, we can infer that the effect of texture on the oxidation behavior of Ti₃SiC₂ is negligible.

Methods

Fabrication process

Textured Ti₃(SiAl)C₂ ceramic was prepared by a two-step sintering process. Firstly, Ti (~300 mesh, purity: 99.9%), Si (~300 mesh, purity: 99.9%), Al (~300 mesh, purity: 99.9%), and graphite (~300 mesh, purity: 99.99%) powders in a molar ration of 3:1:0.2:2 was used as the raw material for the fabrication of Ti₃(SiAl)C₂ powders. In this regard, a small amount of Si was substituted by Al to eliminate TiC impurities and improve its oxidation resistance^6,25,26. All the elemental powders were placed in an agate ball mill jar and mixed uniformly by planetary ball milling. After ball milling, the mixed powders were placed in a graphite mold and compacted under a pressure of 3 MPa, followed by pressureless sintering at 1450 °C for 1 h in a flowing argon atmosphere. Subsequently, the as-prepared Ti₃(SiAl)C₂ was pulverized and ball milled for obtaining fine-grained Ti₃(SiAl)C₂ powders. The milling speed and time were 300 rpm and 24 h, respectively. Finally, the fine-grained Ti₃(SiAl)C₂ powders were sintered at 1300 °C for 20 min in a vacuum under 40 MPa using an SPS furnace. The SPS curves for Ti₃(SiAl)C₂ are shown Fig. 14.

Fig. 14

SPS curves of textured Ti₃(SiAl)C₂.

Characterization

The phase compositions of the textured top surface (TTS), textured side surface (TSS), and oxide scales were characterized by XRD (Bruker D8, Germany). Cu-Kα radiation (k = 0.1542 nm) was adopted with a voltage of 40 kV, a current of 40 mA, and a scanning speed of 4◦/min. The degree of texture was evaluated by the Lotgering orientation factor, f_L, which can be expressed as:²⁷

$$f_L = \frac{{P - P_0}}{{1 - P_0}}$$

(5)

For a- and b-axes orientations, the values of P and P₀ correspond to the ratio $\mathop {\sum }\nolimits^ I_{\left( {hk0} \right)}/\mathop {\sum }\nolimits^ I_{\left( {hkl} \right)}$, and for c-axis orientation, the values of P and $P_0$ correspond to the ratio $\mathop {\sum }\nolimits^ I_{\left( {00l} \right)}/\mathop {\sum }\nolimits^ I_{\left( {hkl} \right)}$, where $\mathop {\sum }\nolimits^ I_{\left( {hko} \right)}$, $\mathop {\sum }\nolimits^ I_{\left( {hkl} \right)}$, and $\mathop {\sum }\nolimits^ I_{\left( {00l} \right)}$ were the sums of peak intensities of $\left( {hko} \right)$, $\left( {hkl} \right)$, and $\left( {00l} \right)$ planes, respectively. P is from the texture sample and P₀ is from the standard PDF card (74-0310) of Ti₃(SiAl)C₂.

The microstructure of the oxidized surfaces, cross-sections, and elemental distributions of textured Ti₃(SiAl)C₂ was observed using SEM (Supra 55, Zeiss, Oberkochen, Germany) equipped with EDS (Oxford Instruments, UK). The samples used for SEM analysis and EBSD (Supra 55, Zeiss, Oberkochen, Germany) observation were subjected to standard metallographic procedures, including cutting, grinding and polishing. In addition, the grain structure and the orientation map of the crystal texture were determined by EBSD. The scanning was performed in a hexagonal grid with a step size of 0.2 µm using an accelerating voltage of 20 kV. The sample was tilted by 70° during the EBSD acquisition. The grain boundaries are defined for misorientation angles greater than 5°. The data acquisition and processing were performed using Channel 5 software.

Oxidation tests

Before the oxidation test, the samples with dimensions of 8 × 8 × 2 mm³ were cut from the top and side of the textured Ti₃(SiAl)C₂ sample, respectively, and then sanded and polished. For the convenience of description, the top surface perpendicular to the compression direction is named TTS and the side surface parallel to the compression direction is named TSS. Regarding the oxidation test of the TTS sample, a pair of faces (8 × 8), perpendicular to the c-axis, accounted for 66.67% of the total area tested. While for the TSS sample, two pair faces (8 × 8, and 8 × 2), parallel to the c-axis, accounted for 83.3% of the total area tested.

The oxidation behavior of textured Ti₃(SiAl)C₂ was evaluated using vertical Setsys evolution microbalance (SETARAM, France) in flowing air at 1000–1300 °C for 10 h with a heating rate of 40 °C min⁻¹. The mass change versus oxidation time was recorded continuously with software.