1.   Introduction

Due to the increasing requirement of lightweight products in the industrial, including aerospace, automotive, and electronics industries, the demand for lightweight magnesium components is increasing [1–4]. However, to meet the demands of some applications, such as motor pulleys and engine blocks, ceramic particles must be introduced into the Mg matrix to enhance the matrix stiffness and wear resistance [5].

Recently, a novel ternary nanolayered MAX phase (M_n+1AX_n, where M is an early transition metal, A belongs to group IIIA or IVA, X is C and/or N, and n = 1−3) that exhibits metal and ceramic-like properties gained attention [6–7]. Different from traditional reinforcements, such as TiC, SiC, B₄C, graphite, and Al₂O₃ [8–11], Al elements can diffuse out from MAX phases at a certain temperature due to the low migration energy of Al in Al-contained MAX phases [12–13]. This can produce a strong interface between Al-contained MAX phases and the Mg matrix. For example, Yu et al. [14] and Amini et al. [15] experimentally observed the formation of a robust amorphous interfacial layer and nanosize Mg grains among Ti₂AlC particles. Anasori et al. [16] found that 20vol% Ti₂AlC-reinforced Mg composite obtained by power metallurgy could dissipate 30% mechanical energy during each compressive load at 250 MPa. In addition, Yu et al. [17–18] found that the high damping capacity and superior self-lubricating capacity of Ti₂AlC were attributed to its high dislocation density and graphite-like layered structure. However, the distribution of MAX phases in the Mg matrix is discontinuous in all the composites fabricated by semi-solid stir casting and powder metallurgy reported in the literature [14–15].

There has been no relevant research on the three-dimensional (3D) continuous network magnesium composite reinforced by MAX phases. According to the literature [19–20], the 3D continuous network that is composed of a ceramic skeleton and metal skeleton exhibits high load-bearing and thermal shock resistance performance. The metal skeleton transfers and disperses stress, whereas the ceramic skeleton efficiently improves the stiffness of the metal matrix. The interlocking effect of this structure endows the material’s excellent damage tolerance and low risk of failure [21–22]. Investigations reveal that co-continuous ceramic–metal composites prepared by the infiltration method exhibit excellent mechanical properties [21,23–25]. For example, co-continuous TiC_x/Cu–Cu₄Ti composites produced by infiltrating the melting Cu into TiC_0.5 porous ceramic preforms exhibited an excellent thermal shock resistance and good mechanical properties [21]. Wang et al. [23] reported that co-continuous Ti₃AlC₂/Al composites produced by pressureless infiltration could maintain their good mechanical properties at relatively high temperatures. Furthermore, Yu et al. [26] reported that mechanical properties of Ti₂AlN_x can be modified by controlling the vacancy of N. The elastic modulus and intrinsic hardness of Ti₂AlN increase with the increased N content. This suggests that the regulation of the mechanical performance of the Ti₂AlN/Mg composite is possible by controlling the N content in Ti₂AlN instead of regulating the Ti₂AlN content in the Mg matrix.

On the basis of the abovementioned points, the in-situ formed porous Ti₂AlN_x (x = 0.9, 1) preforms were infiltrated by pure magnesium in this work. To investigate the relationship between their mechanical properties and their microstructure, 2D and 3D characterizations were conducted.

2.   Experimental

2.1   Fabrication process

As shown in Fig. 1, starting powders of Ti, Al, and TiN (purity ≥99%, particle size 300 mesh) were mixed with molar ratios of 2Ti : 1.05Al : 0.9N and 2Ti : 1.05Al : 1.0N to prepare Ti₂AlN_0.9 and Ti₂AlN preforms. The mixed powders were first compressed into a green body under an axial pressure of 40 MPa. Thereafter, the green body was put into a graphite die coated with BN and sintered in a vacuum sintering furnace at 1400°C in 20 min under Ar gas. Note that the powder mixtures used in this study contain an excess of 5at% of Al to compensate the loss of aluminum by preferential out-diffusion during the sintering process [26]. The obtained porous Ti₂AlN_x preforms that were sandwiched between two Mg cylinder blocks were then heated to 750°C in 90 min in an Al₂O₃ crucible under Ar gas, which was followed by the furnace cooling to room temperature.

Figure  1.  Fabrication procedure of Ti₂AlN_x–Mg composites.

下载: 全尺寸图片 幻灯片

2.2   Microstructure characterization

Phase identification was performed by X-ray diffraction (XRD) using a Bruker (Karlsruhe, Germany) D8 diffractometer with Cu K_α radiation. Wavelength dispersive X-ray spectroscopy (WDS, CAMECA SX100) was used to determine the chemical composition of Ti₂AlN_x. The microstructural observation was performed by scanning electron microscopy (SEM, Merlin). Synchrotron X-ray micro-tomography experiments were conducted at the BL13W1 beamline with an X-ray energy of 36 keV at the Shanghai Synchrotron Radiation Facility to reconstruct the 3D real structure. The exposure time was set as 1 s using the Hamamatsu Flash 4.0 camera. The 4X lens was selected and the voxel size was 1.625 μm³. The data was reconstructed by a phase retrieval algorithm in the PITRE software and a 3D microstructure was acquired using the AVISO software. Based on 3D reconstruction, Ti₂AlN_x volumes were extracted and summarized in Table 1. In addition, densities of the composites were measured using Archimedes’ principle.

Table  1.  Material characterization results obtained from composites

Material	Volume fraction of Ti2AlNx	Bulk density / (g·cm−3)	Relative density / %
Ti2AlN0.9–Mg	55%	3.06 ± 0.02	97.03
Ti2AlN1.0–Mg	54%	3.09 ± 0.02	98.79

下载: 导出CSV 

| 显示表格

2.3   Mechanical test

Uniaxial compressive tests were performed on specimens with a diameter of 5 mm and a height of 8 mm on a universal servo-hydraulic mechanical testing machine with a strain rate of 0.5 mm/min at room temperature in air. Microhardness measurements were carried out using a Vickers microindenter at loads of 10, 30, 50, 100, and 200 N and were held for 15 s. Each test was repeated five times for each composite to evaluate its mechanical properties.

3.   Results and discussion

3.1   Microstructure of Ti₂AlN preforms and their composites

The backscattered 2D images of Ti₂AlN_0.9 and Ti₂AlN_1.0 preforms are shown in Fig. 2(a–d). Both preforms are characterized by a uniform distribution of pores. In enlarged areas (Fig. 2(b) and (d)), the grain size of the Ti₂AlN_0.9 preform is bigger than that of Ti₂AlN. This phenomenon was also reported in our previous study of the Ti₂AlN MAX phase [26], in which the grain size of Ti₂AlN decreased with the increasing N content in starting powders. Furthermore, the absence of contrast in the images obtained under the backscattered mode suggests that the fabricated performs are pure. Fig. 2(e–h) presents the microstructure of Ti₂AlN_0.9/Mg and Ti₂AlN/Mg composites. It is evident that light-gray Ti₂AlN_x phases and dark-gray Mg matrix are uniformly interlaced. Moreover, XRD analysis in Fig. 3 reveals that no other detectable trace appears except Ti₂AlN and Mg. The sharp and clear interface in backscatter electron micrographs (Fig. 2(e–h)) confirms that no chemical reaction happened between Ti₂AlN_x and Mg.

Figure  2.  Backscattered (BSD) scanning micrographs of Ti₂AlN preforms and its Mg composites: (a, b) Ti₂AlN_0.9 preform, (c, d) Ti₂AlN_1.0 preform, (e, f) Ti₂AlN_0.9/Mg composite, and (g, h) Ti₂AlN_1.0/Mg composites.

下载: 全尺寸图片 幻灯片

Figure  3.  X-ray diffraction (XRD) diffractograms of Ti₂AlN_0.9/Mg and Ti₂AlN_1.0/Mg composites.

下载: 全尺寸图片 幻灯片

Ti₂AlN_x skeletons (Fig. 4(a) and (c)) and the corresponding infiltrated Mg (Fig. 4(b) and (d)) were extracted from the 3D reconstruction to reveal the 3D information of Ti₂AlN_0.9/Mg and Ti₂AlN/Mg composites. They are clearly characterized by a 3D co-continuous structure without fragmentation. Marked areas in Fig. 4(a–d) were respectively enlarged. The enlarged areas in Fig. 4(a) and (c) indicate that holes in Ti₂AlN_x preforms are interconnected. Thus, the melt Mg could facilitate the infiltration of Ti₂AlN_x preforms and form the three-dimensionally interconnected Mg network found in Fig. 4(b–d).

Figure  4.  3D reconstructed images of Ti₂AlN_x/Mg composites: (a) Ti₂AlN_0.9 skeleton, (b) Mg in Ti₂AlN_0.9/Mg composite, (c) Ti₂AlN_1.0 skeleton, and (d) Mg in Ti₂AlN_1.0/Mg composite.

下载: 全尺寸图片 幻灯片

3.2   Mechanical properties

Table 2 summarizes the compressive strength and microhardness of Ti₂AlN_x/Mg composites. The compressive yield strength and microhardness of Ti₂AlN_0.9/Mg are 353 MPa and 1.12 GPa, respectively, which are 8.55% and 6.67% lower than those obtained from Ti₂AlN/Mg, respectively (compressive yield strength = 386 MPa, microhardness = 1.20 GPa). However, the ultimate compressive strength of Ti₂AlN_0.9/Mg (395 MPa) is 3.66% lower than that of Ti₂AlN/Mg (410 MPa). The yield strength is known to be proportional to Vickers hardness (V_H) for most materials [27–29]. These two composites have almost the same Ti₂AlN_x volume and density with only a difference in the N content and grain of Ti₂AlN_x. Our previous study about bulk Ti₂AlN_x revealed that the N deficiency in Ti₂AlN_x leads to the reduction of the intrinsic hardness [26]. The hardness obtained from nanoindentation decreases from 9.7 GPa of Ti₂AlN_1.0 to 8.1 GPa of Ti₂AlN_0.9. Furthermore, the grain size effect is found in MAX phases [30–32]. For example, the V_H values of coarse-grained (35 μm) and fine-grained (2 μm) Cr₂AlC are 3.5 and 6.4 GPa [30]. In this work, the grain sizes of Ti₂AlN_0.9 and Ti₂AlN (Fig. 2) are 12 and 6 μm, respectively. Here, the reduction of the compressive yield strength and microhardness in Ti₂AlN_x/Mg should be attributed to the grain size and N vacancy in Ti₂AlN. However, the 3D network ceramic and metal skeletons could efficiently transfer the load between each other. Due to this cooperative effect, the grain size and N vacancy effect are weakened in the ultimate compressive strength between Ti₂AlN_0.9 and Ti₂AlN. In addition, results show that the Ti₂AlN ternary nanolayered MAX phase-reinforced magnesium matrix composites exhibit higher compressive yield strength and ultimate compressive strength than the reported binary nitride reinforcement, such as the AlN/Mg composite with 93 and 313 MPa, respectively [33].

Table  2.  Compressive properties of different composites, including the 0.2% compressive yield strength (σ_0.2), compressive strength (σ_f), compressive fracture strain (ε_f), and Vickers hardness (V_H)

Material	σ0.2 / MPa	σf / MPa	εf / %	VH / GPa
Ti2AlN0.9–Mg	353	395 ± 7	14.4	1.12
Ti2AlN1.0–Mg	386	410 ± 8	13.1	1.20

下载: 导出CSV 

| 显示表格

Fig. 5 presents the indents of Ti₂AlN_x/Mg composites obtained under 50 N. No cracks propagated at the corners of indents for both composites. In the enlarged images, cracks appeared in Ti₂AlN_x MAX phases rather than at the interface. This means that the strong interfacial bonding strength and the plastic Mg matrix can block the crack propagation in this continuous skeleton structure. In contrast, more cracks were found in Ti₂AlN/Mg, as shown in Fig. 5(c) and (d). It is known that hardness and ductility are often mutually exclusive [34]. Our previous work shows that the elastic modulus and intrinsic hardness of substoichiometric Ti₂AlN_0.9 are 268 and 8.1 GPa, respectively, which are smaller than those of Ti₂AlN (elastic modulus = 278 GPa, intrinsic hardness = 9.7 GPa, respectively), obtained by nanoindentation [26]. Therefore, Ti₂AlN/Mg exhibits higher hardness and lower ductility, resulting in more cracks, which means the hardness and ductility of Ti₂AlN_x affect the mechanical behavior of composites.

Figure  5.  Microstructural morphologies of indents formed under 50 N: (a, b) Ti₂AlN_0.9/Mg composite; (c, d) Ti₂AlN_1.0/Mg composite.

下载: 全尺寸图片 幻灯片

For all composites, the fracture occurred 45° with respect to the compression loading axis. As shown in Fig. 6, the cylindrical side of Ti₂AlN/Mg after the compressive test was carefully observed. The enlarged area that is shown in Fig. 6(b) indicates that some tiny lines appeared in Ti₂AlN grains, far from the fracture line. However, Fig. 6(d) shows that no tiny line appeared in the untested Ti₂AlN/Mg composite. This difference suggests that Ti₂AlN_x/Mg specimens participated in deformation during the compressive test. Close to the fracture line, the enlarged area shown in Fig. 6(c) indicates that no interfacial decohesion happened between Ti₂AlN and the Mg matrix. Meanwhile, some Ti₂AlN particles fractured. This phenomenon is also found in our previous study about the Ti₂AlC/Mg composite because the interfacial bonding strength between Ti₂AlC and Mg is higher than the Ti–Al bonding strength in the Ti₂AlC MAX phase [14]. Electron energy-loss spectroscopy and band structure calculations reveal that there is only one very weak perturbation on the electronic structure when C is replaced by N in Ti₂AlC/N MAX phases [35]. The interface structure between Ti₂AlN and Mg should be similar to that of Ti₂AlC and Mg. Here, similar to Ti₂AlC/ Mg composites, no interfacial debonding occurred.

Figure  6.  Backscattered images of the cylindrical side of Ti₂AlN–Mg (a–c) after and (d) before the compressive test.

下载: 全尺寸图片 幻灯片

Fig. 7 indicates compressive fracture surfaces of specimens after compressive tests. For two composites, Fig. 7(a) and (d) reveals that the rough fracture surface was characterized by some tiny grooves. In the corresponding backscattered mode, Fig. 7(b) and (e) indicates that Ti₂AlN and Mg are uniformly distributed. There is no accumulation of MAX phases in the glide direction. This is different from composites from the semi-solid casting method and powder metallurgy, in which a seriously torn zone often appeared above the accumulated MAX phases in its reinforced metal composites [14,36]. The enlarged areas in Fig. 7(c) and (f) reveal that typical kink bands with very sharp radii of curvature or typical MAX phases delamination occurred in all Ti₂AlN grains. Thus, these results prove that both Ti₂AlN and Mg participated in the deformation during the compressive test due to their unique three-dimensional co-continuous network structure.

Figure  7.  Compressive fracture morphologies of (a–c) Ti₂AlN_0.9–Mg and (d–f) Ti₂AlN–Mg composites. (a) and (d) are in SEM mode and the rest are in BSE mode.

下载: 全尺寸图片 幻灯片

4.   Conclusion

Through non-pressure infiltration, 3D co-continuous magnesium composites reinforced by Ti₂AlN_x (x = 0.9, 1.0) were successfully fabricated. Uniformly distributed holes in Ti₂AlN_x preforms are interconnected without fragmentation. It is found that N vacancy and grain size effect in Ti₂AlN_x lead to the reduction of the mechanical properties of composites, especially the yield compressive strength (YCS) and microhardness. The YCS and microhardness of Ti₂AlN_0.9/Mg are 353 MPa and 1.12 GPa, respectively, which were 8.55% and 6.67% lower than those obtained from Ti₂AlN/Mg, respectively. Due to the continuous skeleton structure and strong interfacial bonding strength, the crack initiated in Ti₂AlN_x was blocked by the plastic Mg matrix. Moreover, the compressive fracture observation indicates that no interfacial decohesion occurred, and both Ti₂AlN and Mg efficiently participated in the deformation with the delamination of Ti₂AlN phases and severe plastic deformation of Mg. Therefore, this study demonstrates that it is possible to regulate the mechanical performance of Ti₂AlN/Mg composites by controlling the N vacancy in Ti₂AlN other than the reinforcement content of Ti₂AlN in the Mg matrix.

Acknowledgements

This work was financially supported by the National Natural Science Foundation of China (No. 52175284), the State Key Lab of Advanced Metals and Materials (No. 2021-ZD08), and the Beijing Government Funds for the Constructive Project of Central Universities (No. 353139535).

Conflict of Interest

The authors declare no potential conflict of interest.