The initial material used was the commercial Ti₃AlC₂ powder (99% purity, 200 mesh) (Chengdu Haixin High-tech Co., Ltd., Chengdu, China). A bulk dense Ti₃AlC₂ block was subsequently prepared using the hot-pressing method, with a sintering temperature of 1300 ℃, a pressure of 30 MPa, and a holding time of 60 min. The resulting block achieved a density of 4.24 g/cm³ with 99% densification. Following this, the rectangular Ti₃AlC₂ block was shaped to dimensions of 3 × 4 × 36 mm³ using electrical discharge machining (EDM). Subsequently, it was polished to a diamond grit size of 1.0 μm, and its edges were chamfered. The edges must be rounded (with a radius of 0.1-0.3 mm) to prevent premature fracture caused by stress concentration at the edges during testing.

Ultrafast thermal shock tests were conducted using a laboratory ultrafast thermal shock apparatus (​Fig. 1​). This apparatus consists of two components: a heating control unit and a cooling environment. The heating control unit employs a 500-1100 kHz AC power supply (Dongguan Jinbenlai Electromechanical Equipment Co., Ltd., Dongguan, China) to regulate temperature. The heating signal is transmitted to the induction coil within the chamber, achieving sample heating through power modulation rather than direct temperature control. Additionally, the heating rate is regulated by the power supply unit and is adjusted according to the quenching temperature and sample variations. A temperature sensor (Raynger 3i Plus, Raytek Corporation, Santa Cruz, California, USA) continuously monitors the sample surface temperature during the heating process.

The cooling environment consists of a container filled with liquid nitrogen (-196 ℃). After holding the sample at a constant temperature, it is immediately dropped into the container to simulate an ultra-low temperature environment. The entire thermal shock process is controlled by a dedicated computer system, enabling precise adjustment of power output to rapidly achieve the preset quenching temperature ranges (520 ℃, 720 ℃, 1100 ℃, 1250 ℃, and 1400 ℃). These temperature ranges hold significant importance in prior studies of MAX ceramic thermal shock behavior ​[4-6,29]​. It is particularly noteworthy that samples may undergo significant phase composition and microstructural changes both before and after these temperature points ​[11,30]​. Because the duration of heating has a significant impact on the formation and thickness of the oxide layer, ​Table 1​ lists the total duration from the start of heating to the end of the holding period for each temperature point, with the holding time set at 10 s in all samples. To ensure result reliability and minimize experimental variability, quenching was performed on at least three samples at each temperature point.

The phase composition of the samples was analyzed using a X-ray diffraction (XRD) equipment (D8 ADVANCE A25X, Bruker, Karlsruhe, Germany) with Cu Kα1 radiation (λ = 0.154 nm, 40 kV, and 30 mA). The flexural strength of the samples was measured using a universal testing machine (YC-100KN, Oneyice Company, Ningbo, China) via the three-point bending method, with a crosshead speed of 0.5 mm/min and a span of 30 mm. Finally, the surface and cross-sectional microstructure were examined using the field emission scanning electron microscopy (FSEM) (Apreo 2 C, Thermo Scientific, Waltham, MA, USA).

At elevated temperatures, thermal shock in Ti₃AlC₂ ceramics is often accompanied by oxidation of aluminum (Al) and titanium (Ti), as well as decomposition of Ti₃AlC₂ once a certain temperature is reached ​[29-32]​. The degree of oxidation increases with rising temperature. To better understand the thermal shock behavior of Ti₃AlC₂ quenched in liquid nitrogen, ​Fig. 2​ shows the X-ray diffraction patterns of Ti₃AlC₂ samples rapidly heated to six different temperature points and then cooled in liquid nitrogen. The phase compositions of the sample surfaces after cooling at different temperatures are detailed in ​Table 2​.

Based on XRD results, it can be confirmed that no oxides appear on the surface of samples cooled with liquid nitrogen after heating to 720 ℃ or below (​Figs. 2​(a)-​2​(c)). This contrasts with previous thermal shock experiments on Ti₃AlC₂ ceramics, where aluminum oxides appeared on the surface at 710 ℃ even after heating in a low-oxygen environment. Samples heated in air and then water-cooled or air-cooled exhibited oxides as low as 270 ℃ ​[11]​. Therefore, considering the significant variation in quenching conditions, it is concluded that: after heating to 710 ℃, the samples indeed form oxides, but these are very thin and present in low concentrations. Subsequently, the extremely low cooling temperature of liquid nitrogen creates a thermal shock between the hot interior and cold exterior of the surface oxides, causing them to detach or peel away from the substrate possibly. However, this phenomenon ceases to occur when the sample is heated to temperatures above a certain threshold, as the increased oxide content renders liquid nitrogen cooling insufficient to strip away all oxide layers. Thus, when the sample was heated to 1100 ℃, oxide formation was detected on the sample surface, with the Al₂O₃ (104) peak and TiO₂ (110) peak being particularly prominent in ​Fig. 2​(d) ​[32]​.

As the temperature increased, the formation of Al₂O₃ and TiO₂ led to the further oxidation of Al₂TiO₅. At 1250 ℃, the Al₂TiO₅ (110) diffraction peak was detected on the sample surface, accompanied by a high-intensity peak of the high-temperature decomposition product TiC (200) (​Fig. 2​(e)). However, the substrate phase still persisted at this point, albeit with very low intensity. This indicates that only partial decomposition had occurred within the sample at this stage, rather than complete decomposition. Only when the temperature reached 1400 ℃ did the underlying Ti₃AlC₂ phase completely disappear ​[33-37]​. The sample surface was then dominated by oxides, with the Al₂TiO₅ (110) peak intensity peaking at the heating temperature. The decomposition also produced a significant amount of TiC, with the (111), (200), and (220) peaks all becoming distinctly prominent in ​Fig. 2​(f).

To further elucidate the mechanism of Ti₃AlC₂ ceramics during liquid nitrogen quenching, the surface and cross-sectional analysis of the samples using scanning electron microscopy was performed, as shown in ​Figs. 3 and 4​. At temperatures of 720 ℃ and below, scanning electron microscopy revealed slight surface imperfections on the samples (​Figs. 3​(a) and ​4​(a)). The absence of oxides was confirmed, indicating that the surface irregularities and defects observed (​Figs. 3​(b) and ​4​(b)) resulted from the combined effects of liquid nitrogen quenching and the subsequent exfoliation of oxides. Although these defects occur relatively infrequently at this temperature, their incidence increases as the quenching temperature rises. At 1100 ℃, a distinct oxide layer was observed on the sample surface (​Fig. 3​(c)). The oxide covered most of the substrate surface and exhibited a relatively continuous and dense distribution (​Fig. 4​(c)). The presence of this oxide layer demonstrates that it withstood the thermal shock of liquid nitrogen quenching without separating from the substrate ​[15]​. The intact substrate indicates that the precisely controlled oxidation level provides protection for the Ti₃AlC₂ ceramic. When heated to higher temperatures, the oxide layer directly interfaces with liquid nitrogen, preventing more severe damage to the substrate.

At 1250 ℃, the oxide content further increased, as confirmed by EDS analysis (​Fig. 5​). Concurrently, the oxide composition was not merely a simple combination of Al₂O₃ and TiO₂, but also featured the deep oxide phase Al₂TiO₅. These three oxide phases coexisted on the sample surface, consistent with the conclusions drawn from X-ray diffraction analysis. As the oxide layer thickens and expands, increased grain spalling and porosity become evident on the outermost surface (​Fig. 3​(d)). However, close examination of the cross-section reveals a relatively dense oxide layer still adhering to the substrate (​Fig. 4​(d)). This confirms that the outermost oxide layer primarily consists of coarse, spherical Al₂TiO₅ particles, resulting in a loose and porous outer structure. The innermost layer retains a relatively dense oxide layer (primarily composed of Al₂O₃ and TiO₂) that provides some cushioning against the impact of liquid nitrogen. In contrast, ​Figs. 3​(e) and ​4​(e) reveal more severe structural defects, with numerous distinct cracks appearing on the sample surface at 1400 ℃. Concurrently, oxide particles further enlarged. EDS analysis indicates that the oxide composition at this stage is predominantly dominated by the Al₂TiO₅ phase (​Fig. 6​). Furthermore, as shown by the surface scanning results in Supplementary file, after liquid nitrogen quenching at 1400 ℃, the Al and O content on the surface of the Ti₃AlC₂ ceramic significantly increased, concentrating on the sample surface, while the Ti content decreased and enriched in the lower layer. Scanning electron microscopy images reveal an extremely porous and loose oxide layer that offers no protection against liquid nitrogen impact, leading to direct damage to the substrate. According to the X-ray diffraction results, indicating that Ti₃AlC₂ has completely decomposed, transforming the substrate from the Ti₃AlC₂ phase to the TiC phase. The TiC phase exhibits brittleness, and combined with the destructive effects of liquid nitrogen, the sample interior also shows degradation, appearing porous and incomplete ​[23]​. It is inferred that during the quenching process, the oxide film plays a crucial role in controlling thermal decomposition and thermal stress, thereby significantly influencing thermal shock resistance. Within the temperature range of 1250 ℃, a dense and intact oxide layer provides effective protection. However, as temperature increases, the composition and microstructure of the oxide undergo changes, losing its original protective function while simultaneously compromising the integrity of the overall structure.

To determine the residual bending of samples after liquid nitrogen quenching, all specimens underwent the three-point bending test. The obtained data are shown in ​Fig. 7​, and the strength loss rate was calculated by comparing the initial strength value (581 ± 18 MPa), as listed in ​Table 3​. As shown in the figure, the flexural strength exhibits a parabolic trend, initially increasing to a peak value and subsequently declining as the quenching temperature rises. At temperatures below 720 ℃, the integrity of the sample surface is compromised due to impact damage, and without the protective layer of oxidation, the flexural strength shows a slight decrease. Within this lower temperature range, the strength loss rate reaches a maximum of 4.5%. Notably, compared to previous experiments, the Ti₃AlC₂ sample exhibited a flexural strength of only 160 MPa after being heated to 270 ℃ in air and then quenched in water ​[11]​. This bending strength is significantly lower than the 520 ℃ value obtained under liquid nitrogen conditions. However, it must be noted that liquid nitrogen possesses a higher cooling rate than water, generating greater tensile stress within the sample. This stress can cause damage to both the matrix and surface, thereby hindering strength development. The XRD diffraction pattern of the sample cooled in water after heating at 270 ℃ revealed the presence of oxides, thus validating the prior hypothesis. During the liquid nitrogen impact test, the defects and irregularities observed on the sample surface at 520 ℃ were also traces of oxide detachment. Originally present on the sample surface, the oxides were too sparse and discontinuous to form a continuous layer. However, they continued to provide a protective effect before detachment, allowing the sample strength to remain relatively stable without significant degradation.

Similarly, the oxide content of the 720 ℃ sample increased with rising temperature, but it still flaked off upon impact, indicating that it was insufficiently stable to remain intact. When the temperature reaches 1100 ℃, the continuous and dense oxide layer has a thickness of approximately1.2 μm, as measured from the cross-sectional SEM image (​Fig. 4​(c)). At this point, the protective function of the oxide film is optimal, completely withstanding the rapid impact of liquid nitrogen. This prevents direct contact between the Ti₃AlC₂ substrate and liquid nitrogen, thereby avoiding damage from the liquid nitrogen impact. Moreover, the thermal expansion coefficient of the oxide layer is lower than that of Ti₃AlC₂, which partially counteracts the tensile stress induced by liquid nitrogen quenching by generating compressive stress on the substrate. The flexural strength of the sample slightly increased to 590.8 MPa, with a strength loss rate of −1.6%. When temperatures exceed 1100 ℃, on the one hand, the oxide particles on the sample surface become coarse and porous. Liquid nitrogen directly contacts the matrix through microcracks and microvoids, negatively impacting strength; On the other hand, Al₂TiO₅ begins to form. Since Al₂TiO₅ has a higher thermal expansion coefficient than Ti₃AlC₂, the strength drops to 485.5 MPa at 1250 ℃ ​[11]​. The oxide layer exhibits a bilayer structure. The outer loose layer (composed mainly of coarse Al₂TiO₅ grains) is about 3-5 μm thick, while the inner dense layer remains approximately 1 μm thick, as determined from cross-sectional SEM observation (​Fig. 4​(d)). The total oxide thickness is roughly 4-6 μm. The presence of the thick, porous outer layer initiates microcracks and reduces the protective effect. Despite this, the strength loss rate is only 16.6%. In contrast, the strength loss rate of the sample heated in air and quenched in water reaches as high as 94.9% at 1250 ℃. At 1400 ℃, the oxide layer becomes entirely porous and loose, with a thickness exceeding 20 μm. Revealing numerous defects and liquid nitrogen penetration into the matrix, further compromising surface integrity. Additionally, the complete decomposition of Ti₃AlC₂ into brittle TiC accelerates structural deterioration. Concurrently, oxidation intensifies, with increased Al₂TiO₅ formation leading to a further reduction in strength to 193.3 MPa.

Overall, when Ti₃AlC₂ samples are heated in air and quenched in liquid nitrogen, a dense oxide layer forms on the surface. This layer effectively protects the substrate, significantly reduces the impact of thermal gradients and thermal stresses, and maintains high strength values even under more abrupt temperature changes, demonstrating high application reliability. To further prove this, the Weibull analysis is done, which is a widely adopted statistical method in reliability applications for ceramics. A higher Weibull modulus indicates a more concentrated strength distribution and better reliability. To further determine the reliability of Ti₃AlC₂ under liquid nitrogen impact, the Weibull modulus for bending strength at elevated temperatures was calculated and subjected to reliability validation. The formula for the Weibull distribution is as follows ​[38]​:

F_i = i − 0.5/N represents the failure probability of the i-th graded sample, where i denotes the rank of the strength data, N is the total number of test samples, m is the Weibull modulus, σ_i is the measured strength, and σ₀ is the proportionality constant. ​Fig. 8​ shows the Weibull distribution curve of the bending strength distribution for Ti₃AlC₂ samples quenched in liquid nitrogen, with representative experimental temperatures of 1100 ℃ and 1250 ℃ selected for investigation. The results indicate that the Weibull modulus of the Ti₃AlC₂ sample heated to 1250 ℃ is 14.3, slightly lower than that of the polished Ti₃AlC₂ sample (17.6), as listed in ​Table 4​. And the Weibull modulus of the sample heated to 1100 ℃ was 19.2, indicating a higher value than that of the polished sample. The reason lies in the fact that at 1250 ℃, deep oxidation and defects within the oxide layer negatively impact the reliability of the test specimen. However, since the innermost layer directly adjacent to the substrate still possesses a relatively dense oxide layer, it can provide some cushioning effect against the impact of liquid nitrogen, thereby mitigating the negative effects to a significant extent. In contrast, at 1100 ℃, the surface of the Ti₃AlC₂ sample is covered by a continuous, dense oxide film that adheres tightly to the substrate. This film has minimal impact on reliability and provides a protective effect, resulting in high reliability. The Weibull modulus of Ti₃AlC₂ ceramics was compared with other commonly used structural ceramics. Results indicate that despite surface cracking during quenching, Ti₃AlC₂ ceramics exhibit a high Weibull modulus significantly exceeding that of traditional ceramics such as Al₂O₃ (1.5), maintaining exceptional reliability ​[38,39]​.

To show the flexural strength grades of samples following liquid nitrogen quenching, we have identified and summarized the residual flexural strengths of Ti₃AlC₂ obtained previously in various quenching media. This primarily covers two research: thermal shock in air, as reported by Hu et al., and thermal shock in argon, as reported by Liu et al. ​[11,40]​. The detailed strengths and strength loss rates are shown in ​Tables 5 and 6​.

A systematic comparison of this liquid nitrogen quenching process reveals the following key observations: water quenching resulted in significant strength loss at all temperatures (typically >50%, reaching as high as 94.9% when heated to 1250 ℃ in air, with direct fracture at 1400 ℃). In contrast, air-cooling or argon quenching preserves the dense oxide layer intact and induces compressive stress, resulting in increased strength at multiple temperature points, for example, the strength after argon quenching at 1040 ℃ reaches 647 MPa, representing an increase of 11.2%, demonstrates the highest reliability. Liquid nitrogen quenching exhibits unique temperature-dependent behavior: In the range of 520-720 ℃, the thin oxide film spalls off due to the cryogenic impact, leading to a slight decrease in strength. At 1100 ℃, a continuous and dense Al₂O₃-TiO₂ oxide layer successfully withstands the extreme cold shock and imposes compressive stress, raising the residual strength to 590.8 MPa. At 1250 ℃, the outer part of the oxide layer transforms into loose Al₂TiO₅, causing a strength loss of 16.6%. At 1400 ℃, the matrix decomposes completely into TiC, and the strength drops precipitously with a loss of 66.8%, although the sample does not fracture. It can be seen that the integrity of the oxide layer is the key factor governing thermal shock resistance. Regardless of the medium, provided that a continuous and dense oxide layer can form on the surface and is prevented from peeling off by slow cooling or an appropriate temperature gradient, Ti₃AlC₂ will exhibit excellent thermal shock resistance and may even demonstrate an exceptional increase in strength. Although the ultra-fast cooling rate of liquid nitrogen generates much higher thermal stresses than water quenching, air cooling, or argon quenching, Ti₃AlC₂ still achieves strength enhancement at 1100 ℃. Upon contact with the hot sample surface, liquid nitrogen vaporizes instantly and slows down the heat exchange rate. More importantly, nitrogen is an inert gas and does not chemically react with Ti₃AlC₂ or the oxide layer. Consequently, the continuous and dense Al₂O₃-TiO₂ oxide layer formed at 1100 ℃ remains intact, and its protective effect is sufficient to counterbalance the extremely large tensile stress. Moreover, the strength loss at 1250 ℃ is far lower than that of water quenching at the same temperature. Therefore, Ti₃AlC₂ possesses outstanding service reliability under cryogenic thermal shock in liquid nitrogen, with performance significantly superior to water quenching particularly in the temperature range of 1100-1250℃.

The quenching temperature directly influences the thermal shock behavior of materials. For liquid nitrogen quenching, higher cooling rates generate greater internal tensile stresses, adversely affecting the strength of Ti₃AlC₂ specimens and reducing their thermal shock resistance. To elucidate the thermal shock failure mechanism during liquid nitrogen quenching, the tensile stresses generated were quantitatively calculated. These tensile stresses can be calculated using the following formula ​[41]​:

where α is the thermal expansion coefficient, E is the elastic modulus, ΔT is the temperature difference during quenching, and ν is Poisson's ratio. For Ti₃AlC₂, the values of α, E, and ν are: α = 9.0 × 10⁻⁶ K⁻¹, E = 297 GPa, ν = 0.2. The quenching medium temperature is −196 ℃, with quenching temperatures of 520 ℃, 720 ℃, 1100 ℃, 1250 ℃, and 1400 ℃, corresponding to ΔT values of 716 ℃, 916 ℃, 1296 ℃, 1446 ℃, and 1596 ℃, respectively. The calculated tensile stresses are 2.4 GPa, 3.1 GPa, 4.3 GPa, 4.8 GPa, and 5.3 GPa. This indicates that surface tensile stress increases with rising temperature, leading to a reduction in residual bending strength. This finding is consistent with the results of the bending strength test. Scanning electron microscopy images reveal a thin oxide film coating the sample matrix, with defects observed in the Ti₃AlC₂ matrix. The oxide film imposes compressive stress on the substrate, partially mitigating direct contact with liquid nitrogen and providing some protection, resulting in an exceptional increase in strength at 1100 ℃. These calculated thermal stresses far exceed the theoretical strength of Ti₃AlC₂, indicating that the material’s survival under such extreme conditions relies heavily on the protective effect of the oxide layer.

Additionally, GB/T 37246-2018 (ICS: 81.060.30) Standard for Thermal Shock Resistance of Ceramic Materials specifies a method for evaluating the performance of fine ceramic materials under rapid temperature change conditions. By applying this method to quantitatively calculate the maximum allowable temperature difference ΔTc involved in liquid nitrogen impact testing, it provides a generalized argument from another perspective regarding the reliable temperature range of Ti₃AlC₂ at elevated temperatures. The ∆T_C can be calculated using the following formula:

∆T_C: Maximum allowable temperature difference; ∆T₁ and ∆T₂: Temperature differences between test temperatures T₁ and T₂ and the cooling medium (-196℃), where the bending strength decay rates at adjacent test temperatures T₁ and T₂ are less than 10% and between 10% and 20%, respectively. ​ ${\sigma }_{1}^{*}$ and ${\sigma }_{2}^{*}$: Bending strength loss rates at test temperatures T₁ and T₂, respectively. At T₁ (1100 ℃), the corresponding strength is 590.8 MPa, with ${\sigma }_{1}^{*}=-1.6\%$ at T₂ (1250 ℃), the strength is 485.5 MPa, with ${\sigma }_{2}^{*}=-16.6\%$. ∆T₁= 1296 ℃, and ∆T₂=1446 ℃. Substituting these values yields ∆T_C= 1392℃. Thus, 1392 ℃ represents the highest reliable temperature for Ti₃AlC₂ under liquid nitrogen thermal shock conditions, providing new validation for the Weibull analysis conclusions indicating high reliability at 1100 ℃ and 1250 ℃.