Microstructure and Mechanical Properties of AZ31/ZrO2 Composites Prepared by Friction Stir Processing With High Rotation Speed

The nano-ZrO2-reinforced AZ31 alloy composites were fabricated by friction stir processing with three multi-passes and high rotation speeds. The AZ31/2.14 vol% ZrO2, AZ31/4.29 vol% ZrO2, and AZ31/6.43 vol% ZrO2 composites were designed. The fine microstructure and uniform dispersion of ZrO2 particles were observed in the AZ31/ZrO2 composites. The excellent interfacial bonding was observed between the Mg matrix and ZrO2. The hardness and tensile properties were enhanced after three cumulative friction stir processing passes. This was attributed to grain refinement and strengthening effects of ZrO2. The tensile properties and hardness of the AZ31/ZrO2 composite increased with the increase of volume fraction of the ZrO2 particles from 2.14 to 6.43 vol%.


INTRODUCTION
Mg alloys have some advantages, such as low density, high specific strength, and high damping capacity, which have been applied in automotive and aerospace industries (Suh et al., 2016;Zang et al., 2017;Zhang J. et al., 2017;Zhang L. C. et al., 2020). The Mg alloys and Mg-based composites with excellent mechanical properties have been developed to meet the large demand for light materials in automotive and aerospace industries Yu et al., 2018;Shahin et al., 2020). Recently, many studies were focused on the Mg-based composites by reinforcing Al 2 O 3 , ZrO 2 , SiC, graphene, graphene oxide, etc. fabricated by in situ synthesis process, semi-powder metallurgy, and friction stir processing (FSP) Meng et al., 2018;Zhang et al., 2018).
Among these techniques for fabrication of Mg-based composites, FSP has attracted extensive attention, which could fabricate Mg-based composites efficiently due to frictional heating and severe plastic deformation (Del Valle et al., 2015;Liang et al., 2017). At the same time, FSP could refine the grain size, homogenize the microstructure, and improve the dispersion of reinforcements (Chen et al., 2015;Arab and Marashi, 2019;Zhang and Chen, 2019). Navazani and Dehghani (2016) investigated the microstructure and mechanical properties of AZ31/ZrO 2 composites fabricated by FSP. The refinement of the microstructure and fine dispersion of ZrO 2 particles were obtained after FSP, and the mechanical properties were improved. In the case of FSPed Mg-based composites, the tool geometry and the process parameter optimization (like capping pass, multiple passes) were extensively investigated to improve the mechanical properties (Hashemi and Hussain, 2015;Khodabakhshi et al., 2017;Barati et al., 2019). The distribution of reinforcement and the interfacial bonding between matrix and reinforcement largely determines its mechanical properties Ajay Kumar et al., 2020). The previous researches have focused on FSPed Mg-based composites with the conventional rotation speeds (less than 2,000 rpm) (Hanas et al., 2018;Shang et al., 2019). High rotation speed FSP has advantages of fast heating rate, fewer process defects, a more stable process, good mechanical properties, and so on. The processing efficiency is high due to the high heat input during high rotation speed FSP . Liu et al. (2020) investigated the microstructure and mechanical properties of FSPed AZ31 alloy with high rotation speed, which shows excellent corrosion resistance and mechanical properties. The dispersion of β-Al 12 Mg 17 precipitates increased with increased processing speed. Thus, the FSP with high rotation speed could improve the dispersion of particles. It is meaningful to investigate the microstructure and mechanical properties of Mg-based composites by FSP with high rotation speed.
In this work, AZ31-matrix composites reinforced with a different volume fraction of ZrO 2 particles were fabricated by the FSP process with high rotation speed. The microstructure and mechanical properties of FSPed AZ31/ZrO 2 composites fabricated by different FSP passes and volume fraction of ZrO 2 particles were investigated. The effects of FSP process and ZrO 2 particles on mechanical properties were discussed.

MATERIALS AND METHODS
Commercial AZ31 alloy plates with the desired size of 150 × 80 mm and a thickness of 2 mm were used as the matrix. The chemical composition of AZ31 alloy sheets is shown in Table 1. Commercial nano-ZrO 2 with an average diameter of 50 nm, purity >99.99 wt%, were used as reinforcing particles, purchased from Shanghai Macklin Biochemical Co., Ltd.
The grooves with three different dimensions of 1.5 mm (depth) × 0.2 mm (width), 1.5 mm (depth) × 0.4 mm (width), and 1.5 mm (depth) × 0.6 mm (width) were machined in the middle of the AZ31 alloy plates. Then, the grooves filled the ZrO 2 particles. The volume fraction of the ZrO 2 particles were approximately 2.14, 4.29, and 6.43%. This was calculated based on the method in Balakrishnan et al. (2015). At first, a capping pass with a cylindrical pin-less tool with a shoulder diameter of 7 mm was carried out, which can encapsulate the ZrO 2 particles and prevent them from ejecting during FSP process. This capping pass was applied to improve the stability in the preplaced ZrO 2 particles in the groove, which involves closing the top surface of the groove. Then, FSP was performed with a tool with a shoulder diameter of 7 mm and a pin length of 1.9 mm at a high tool rotation rate of 9,000 rpm, a travel speed of 200 mm/min, and a tilt angle of 3 • . FSP with a different volume fraction of the ZrO 2 particles was carried out in one pass, two passes, and three passes, which is shown in Table 2. The temperature in the stir zone (SZ) was 383.8 ± 2 • C during FSP, which was corrected by a k-type thermocouple. An optical microscope (OM) observation was performed on the cross-section of the FSPed samples. OM was observed after etched by 4.2 g picric acid, 10 ml acetic acid, 70 ml ethanol, and 10 ml H 2 O for 10 s. Measuring of average grain size was carried out on OM images using Proimage software. The microstructure observation of ZrO 2 distribution in AZ31 matrix was performed by scanning electron microscopy (SEM, JSM-6610LV) and energy dispersive spectroscopy (EDS) at an accelerated voltage of 20 kV. The microstructure of ZrO 2 distribution in the AZ31 matrix was also observed by using JEM-2100F transmission electron microscope (TEM) at 200 kV. TEM specimens were collected from SZ and were twin-jet polished in a solution of nitric acid (10%), glycerin (30%), and methanol (60%) and, finally, ion-milled for 30 min. The phases of the AZ31 matrix and the composites were identified by an X-ray diffraction (XRD) analyzer (Bruker, D8 Advance) with a Cu target at a scanning angle of 20 • -90 • and a scanning speed of 4 • /min. Microhardness mapping was performed across the thickness section of these FSPed samples. Vickers microhardness was measured on the FSPed samples, which was done at 0.1 kg load and dwell time of 10 s. The measurements were carried out on a grid of five parallel lines for the composites. A total number of 145 points (5 × 29) were used with an indent spacing of 0.5 mm. Microhardness mapping was performed using Matlab software. Tensile specimens were electrical discharge machined from the SZ parallel to the FSP direction, and the dimensions of tensile specimens are shown in Figure 1. The tensile test was conducted

RESULTS AND DISCUSSION
Distribution of ZrO 2 Particles in FSPed AZ31/ZrO 2 Composites Figure 2 shows the SEM-EDS mapping of the SZ-center in the AZ31/ZrO 2 composite (Z3-1P specimen) after one FSP pass; five major elements O, Mg, Al, Zn, and Zr are present. The white phases with O and Zr elements were observed in the FSPed AZ31/ZrO 2 composite. It indicated that ZrO 2 clusters existed in the FSPed AZ31/ZrO 2 composite after one FSP pass. The SEM-EDS mapping of the SZ-center in the AZ31/ZrO 2 composite (Z3-2P specimen) after two cumulative FSP passes is shown in Figure 3; segregation of Zr and O was still observed. The ZrO 2 clusters were obviously observed after one FSP pass, and the segregation significantly reduced when two cumulative FSP passes were applied. Figure 4 shows the EDS elemental mapping analysis results from the microstructure of the FSPed AZ31/ZrO 2 composites (Z3-3P specimen) after three cumulative FSP passes. The EDS result (Point A: 50.37 at.% O, 36.19 at.% Mg, 3.24 at.% Al, 0.73, at.% Zn and 9.47 at.% Zr) suggested that the white particles were ZrO 2 particles. The SEM-EDS mapping also shows well-distributed ZrO 2 particles after three cumulative FSP passes, and it was observed to be free from clustering noticeably.
The agglomeration of ZrO 2 particles was easy to obtain due to insufficient plastic material flow with low heat input during a single FSP pass. The continuous clusters of ZrO 2 particles existed in the SZ of the Z3-1P specimen. The ZrO 2 particles did not distribute well in the matrix, and the agglomeration was unavoidable. The agglomeration of ZrO 2 particles reduced with increasing the FSP passes. The distribution of ZrO 2 particles was more uniform, with continuing the severe deformation up to three FSP passes. Vahedi et al. (2020) investigated the microstructure of FSPed AZ31 composites reinforced by micro-graphite and nano-graphene particles, and a similar  phenomenon was also observed. The well-distributed particles were obtained after the three FSP passes.
The distributions of ZrO 2 particles in the FSPed Z1-3P, Z2-3P, and Z3-3P specimens are shown in Figure 5. The SEM micrographs from SZ confirmed that there were no large agglomerations of ZrO 2 particles after three cumulative FSP passes with a different volume fraction of the ZrO 2 particles. The fraction of ZrO 2 particles in the SZ of the FSPed AZ31/ZrO 2 composites increased with the increase of the ZrO 2 addition, and more ZrO 2 particles were observed in the FSPed Z3-3P specimen. The better distributions of particles were observed in the FSPed Z1-3P, Z2-3P, and Z3-3P specimens. Figure 6 shows the XRD patterns of AZ31 alloy and FSPed AZ31/ZrO 2 composites, and the Mg, β-Al 12 Mg 17 phase were observed in AZ31 alloy. The Mg, β-Al 12 Mg 17 phase and ZrO 2 phase diffraction peaks were observed in the FSPed Z1-3P, Z2-3P, and Z3-3P specimens. β-Al 12 Mg 17 intermetallic is the most common and abundant in Mg-Al-Zn alloys. The results of XRD confirm the existence of ZrO 2 particles in the FSPed Z1-3P, Z2-3P, and Z3-3P specimens. Besides, it also suggested that no reaction occurred between the AZ31 alloy and ZrO 2 particles during the FSP. A similar phenomenon was also reported in Mazaheri et al. (2020), and the ZrO 2 particles were stable during the FSP. Figure 7a shows the OM of the base metal AZ31 alloy plates, and the initial microstructure of the base metal was fine equiaxed grains and some coarse grains. Several fine equiaxed grains were observed in the SZ-center in the FSPed AZ31/ZrO 2 composite (Z3-3P specimen); see Figure 7b. The grain size of the SZ-center of the FSPed Z3 specimen was refined due to the occurrence of dynamic recrystallization during the FSP. A distinct boundary existed between the SZ and thermomechanically affected zone (see Figure 7c). The grain size increased in the thermo-mechanically affected zone, and some coarse grains were observed (see Figure 7d). Figure 8 shows the OM of SZ-center of the FSPed AZ31/ZrO 2 composites (Z1-3P and Z2-3P specimens). The fine equiaxed grains were obtained in the SZ-center in the FSPed Z1-3P and Z2-3P specimens. The average grain sizes of SZ-center in the FSPed Z1-3P, Z2-3P, and Z3-3P were 8.46, 8.23, and 8.39 µm, respectively. A similar grain size was obtained in the SZ of FSPed AZ31/ZrO 2 composites with a different volume fraction of the ZrO 2 addition after three cumulative FSP passes. Figure 9a shows the TEM micrographs of homogeneously distributed ZrO 2 particles in the FSPed AZ31/ZrO 2 composite (Z3-3P specimen), and its selected area diffraction pattern is shown in Figure 9b. This also confirms the existence of ZrO2 particles in the Z3-3P specimen. The ZrO2 particles were uniformly dispersed in the Z3-3P specimen. Figure 9c shows the high magnification TEM micrographs, and the continuous and defect-free (micro-voids and cracks) interface was observed.     The distance of the layer-to-layer of the ZrO 2 particle is about 0.213 nm. Figure 9d shows the interfacial region between the ZrO 2 particle and the Mg matrix. The excellent interfacial bonding without any reaction product was observed. The wellbonding of AZ31-ZrO 2 is attributed to the lower heat generation and the severe plastic deformation during FSP.

Mechanical Properties of FSPed AZ31/ZrO 2 Composites
The microhardness mapping for the FSPed AZ31/ZrO 2 composites is shown in Figures 10A-C. The microhardness variations from SZ to base metal was pronounced, and the microhardness of SZ was the highest in FSPed AZ31/ZrO 2 composites compared with that in the base metal. The average microhardness values of SZ for Z1-3P specimen, Z2-3P specimen, and Z3-3P specimen were 74.02, 81.19, and 87.22 HV, respectively. The average microhardness of the AZ31 alloy was 60.14 HV. The microhardness of the FSPed AZ31/ZrO 2 composites improved compared with that of the AZ31 alloy. The average microhardness increased with the increase of ZrO 2 addition. At the same time, the region with higher hardness also increased with the increase of ZrO 2 addition, and the higher hardness within a larger area was observed in the Z3-3P specimen.
The engineering stress-strain curves of the AZ31 alloy and FSPed AZ31/ZrO 2 composites are shown in Figure 10D. The tensile properties of the AZ31 alloy and FSPed AZ31/ZrO 2 composites are shown in Table 3. The yield strength, tensile strength, and elongation of the FSPed AZ31/ZrO 2 composites (Z1-3P, Z2-3P, and Z3-3P specimens) increased to compare with that of the AZ31 alloy. The yield strength, tensile strength, and elongation of the FSPed AZ31/ZrO 2 composites increased with the increase of ZrO 2 addition. The FSP passes could affect the tensile properties of the FSPed AZ31/ZrO 2 composites, and the tensile properties of the FSPed Z3-1P, Z3-2P, and Z3-3P specimens increased with the increase of FSP passes. The yield strength, tensile strength, and elongation of the FSPed Z3-3P  specimen were 201 ± 3 and 270 ± 4 MPa and 19.6 ± 7%, respectively, which obtained the best tensile properties. The yield strength, tensile strength, and elongation improved by 17, 15, and 39%, respectively, compared with those of the AZ31 alloy. The hardness of FSPed AZ31/ZrO 2 composites increased compared with that of the base metal. This is attributed to the effects of FSP on the microstructure refinements and the direct strengthening from ZrO 2 particles on the metal matrix (Hu et al., 2016;Rashad et al., 2016;Naseer et al., 2019). The grain refinement results in a higher resistance to the motion of dislocations, which could increase the hardness and strength. Based on the Hall-Petch relationship , the finer grains could also improve the tensile properties of FSPed AZ31/ZrO 2 composites. At the same time, the ZrO 2 particles could act as obstacles against the dislocation movement (Zhang M. et al., 2020). Based on the microstructure of the Z1-3P, Z2-3P, and Z3-3P specimens, a similar grain size was observed in the SZ in these specimens. The larger area with higher microhardness was obtained in the SZ of the FSPed Z3-3P specimen. This indicated that the strengthening from reinforcement of ZrO 2 particles is predominant in the SZ of the FSPed specimens. The enhancement of strength can be explained by the Orowan strengthening mechanism (Xiang et al., 2017), which inhibits the dislocations from passing through the ZrO 2 particles. Figure 11 shows that a large number of dislocations generated around ZrO 2 particles. It suggested that the ZrO 2 particles played the pinning effect on the dislocations. The strength could increase due to the accumulation of dislocations.
The good bonding between matrix and reinforcement is also important for the mechanical properties of composites Ajay Kumar et al., 2020). The improvement of the mechanical properties of Mg alloy composite reinforced by ZrO 2 particles is mainly determined by the interfacial bonding between the Mg matrix and ZrO 2 particles. In this study, the good bonding between the Mg matrix and the ZrO 2 particles was obtained in the FSPed AZ31/ZrO 2 composites, which is beneficial to the mechanical properties. The effective stress transfer between the matrix and reinforcement particles was obtained due to the strong interfacial bonding. At the same time, the distribution of ZrO 2 particles also affects the mechanical properties of the composites. The dispersed ZrO 2 particles with less aggregation were required for good mechanical properties (Xiang et al., 2017). The uniform dispersion of ZrO 2 particles was obtained in AZ31/ZrO 2 composites fabricated by high rotation speed FSP after three cumulative FSP passes. Thus, the tensile properties of the FSPed AZ31/ZrO 2 composites could increase with the increase of volume fraction of the ZrO 2 particles. More ZrO 2 particles were added, and these could improve the mechanical properties. The FSP passes played a key role in the particle distribution and mechanical properties. In general, the void formation at the particles occurred due to particle decohension in the soft matrix, which resulted in the fracture (Babout et al., 2004;Pineau et al., 2016). Firstly, the void nucleation can be discussed by energy, and the formation of the crack should exceed or equal the surface energy of new surface creating, which can be expressed as follows: where E is the Young modulus, γ s is surface interfacial energy, and c is half of the void length, and the particle size can replace the void length. This indicated that a large particle needs lower stress to create a new surface, and the coarse particles were weakly bonded to the matrix. The interfacial voids were easily formed at large agglomerations of ZrO 2 particles in the FSPed Z3-1P specimen, and the fine round particles need higher stress to form the interfacial voids. The agglomerations of ZrO 2 particles reduced in the FSPed Z3-2P specimen, and the strength and elongation increased while increasing the FSP passes to two passes. The best strength and elongation of the FSPed Z3-3P specimen were obtained due to the better distributions of ZrO 2 particles. It suggested that the FSPed Z3-3P specimen needs higher strain to form the voids compared with the FSPed Z3-1P and Z3-2P specimen.

CONCLUSION
In this work, the AZ31/2.14 vol% ZrO 2 , AZ31/4.29 vol% ZrO 2 , and AZ31/6.43 vol% ZrO 2 composites have been successfully fabricated by FSP with one pass, two multi-passes, and three multi-passes and high rotation speed. The microstructure, distribution of ZrO 2 particles, and mechanical properties of the FSPed AZ31/ZrO 2 composites were investigated. The effects of FSP passes and ZrO 2 particles on the mechanical properties were discussed; the main conclusions generated are as follows: (1) The fine microstructure was observed in the AZ31/ZrO 2 composites after three multi-pass FSPs. The average grain sizes of SZ-center in the AZ31/2.14 vol% ZrO 2 , AZ31/4.29 vol% ZrO 2 , and AZ31/6.43 vol% ZrO 2 composites were 8.46, 8.23, and 8.39 µm, respectively. (2) The continuous clusters of ZrO 2 particles existed in the SZ of AZ31/6.43 vol% ZrO 2 composites after one FSP pass. The ZrO 2 particles did not distribute well in the matrix, and the agglomeration was unavoidable. The agglomeration of ZrO 2 particles reduced with the increase of FSP passes. The well-distributed ZrO 2 particles were obtained after three cumulative FSP passes. (3) The hardness and tensile properties of the AZ31/ZrO 2 composites after three cumulative FSP passes were enhanced with the increase of volume fraction of the ZrO 2 particles. The tensile properties of the AZ31/ZrO 2 composites increased with the rise of FSP passes. (4) The AZ31/6.43 vol% ZrO 2 composites after three cumulative FSP passes obtained the best tensile properties, which attributed to grain refinement from FSP, welldistributed ZrO 2 particles, and strong interfacial bonding between the Mg matrix and ZrO 2 particles.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation, to any qualified researcher.

AUTHOR CONTRIBUTIONS
QZ and XL conceived and designed the experiments. JZ, LW, and SC performed the experiments. YJ and SL analyzed the data. QZ and HC wrote the manuscript. All authors contributed to the article and approved the submitted version.

FUNDING
This research was funded by the Priority Academic Program Development of Jiangsu Higher Education Institutions.