Fabrication and characterization of friction stir-processed Mg-Zn-Ca biomaterials strengthened with MgO particles

: Magnesium alloy composites play an important role in biomaterials field. In this study, a novel Mg-Zn-Ca matrix composite was reinforced by adding 1.0 wt.% MgO nanoparticles via the high shear casting process. Hereafter, friction stir processing (FSP) was used to achieve a good dispersion of MgO particles and improve the mechanical properties of the composites. After the preparation of the novel composite materials, varied characterization and performance test methods have been selected for comparison. The results illustrate that through FSP, the corresponding microstructure and properties of as-cast MgO/Mg-Zn-Ca composites were significantly modified, and the best combination of the key parameters is 1200 rpm and 60 mm/min for rotational velocity and traveling speed, respectively. After the optimized FSP treatment, the grains size in FSP-processed composites were refined by 42%, to reach 1.04 μm. Due to the grain refinement and the redistribution of MgO particles, the hardness of the FSP-processed MgO/Mg-Zn-Ca composites were increased by 40%, to reach 101.2 HV. Further, it displayed excellent corrosion resistance as well as strength. Compared to the strengthening through grain refinement, the particle strengthening is more dominant based on the study. And meanwhile, the modified grains and added MgO particles are beneficial to the properties of the nugget zones.


Introduction
Magnesium (Mg) alloys have become one of the hot topics in many fields due to their adjustable mechanical behavior and corrosion resistance [1,2]. To further expand the application of Mg alloys, Mg metal matrix composites (Mg-MMCs) are being developed step by step, compared to conventional Mg alloys, Mg-MMCs have superior properties, such as higher strength, hardness and corrosion resistance [3,4]. In recent years, the reinforced Mg-MMCs with ceramic particles, for instance, hydroxyapatite (HA) or β-tricalcium phosphate (β-TCP), were found to significantly improve performances for these composites. In the work of Jaiswal et al. [5], it has improved the corrosion resistance, bioactivity, and mechanical strength of degradable Mg alloys through synthesizing heat treatment of hydroxyapatite-reinforced Mg-3wt% Zn-based composite materials. Liu et al. [6] have adopted the new melt shearing technique integrating with high pressure die casting to uniformly add nano-scale tricalcium phosphate (β-TCP) to Mg-3Zn-Ca composite materials, which have improved the hardness and tensile strength.
It is well accepted that ZrO2, TiO2 and MgO have good chemical stability, enabling them the suitable strengthening phases for the Mg matrix, as Lin et al. [7] have studied.
Goh et al. [8] have produced the MgO/Mg composites with improved strength and hardness. As a strengthening phase, MgO has not only improve the mechanical behavior of the material, but also the effect of increased corrosion resistance. Khalajabadi et al. [9] have used the powder metallurgy method to prepare Mg/HA/MgO nanocomposites, and the results show that with more MgO, the corrosion resistance turns better and better, which is due to thicker product film, and subsequent reduction of pitting corrosion. Lei et al. [10] have prepared MgO/Mg-Zn composites and found that the addition of MgO led to both grain refinement and second phase strengthening, and therefore mechanical properties and corrosion resistance are significantly improved. 4 However, Lin et al. [7] have found that irrespective of manufacturing methods, when the MgO content is about 0.5 wt.% or more, the phenomenon of agglomeration occurs, influencing negatively the mechanical behavior and corrosion properties of the composites. In this research, the application of severe plastic deformation (SPD) to improve the properties of the MgO/Mg-Zn-Ca composites has been proposed. SPD has been extensively applied to improve materials' properties, among them, FSP is one of the typical and essential SPD methods [11,12]. Due to the high heat input of FSP, it was confirmed to be effective to optimize the microstructure and properties of Mg-MMCs.
Vandana et al. [13] have incorporated nano-scaled HA into the pure Mg sheets through FSP, and the grains of the obtained composites in the stir zone are found to be refined from the initial size of 2000 μm to around 10 μm. Morisada et al. [14] have dispersed multi-walled carbon nanotubes (MWCNTs) into AZ31 using FSP. It was found that this method can refine the grains, and increase the microhardness by twice.
Sahraeinejad et al. [15] have incorporated nano-scaled B4C into the matrix of Al 5059, and found that the average microhardness within the stir zone has been increased from 85 HV to a maximum of 170 HV. Lee et al. [16] have added SiO2 nanoparticles to AZ61, the average grain size of prepared composite material is in the range of 0.5~2 μm, and the hardness of the material is almost twice of the matrix. Asadi et al. [17] have applied FSP to prepare a SiC/AZ91 composite layer, and 5 μm SiC particles for reinforcement and grain refinement. Based on the study, the stirring treatment is beneficial to refine the grains with increasing rotational and traveling speeds, which induces the strengthening. 5 Up to the present, a large number of composite materials have been prepared through FSP. However, most studies on performance improvement, focusing on the following points: (1) to rely on FSP to integrate reinforced particles into the metal matrix, (2) with combined actions of added particles as well as FSP, the grains are refined and the performance is improved accordingly. However, few studies have used FSP to improve the microstructure and performance of biomedical composites. In this study, the application of SPD to further improve the properties of the MgO/Mg-Zn-Ca biomedical composites through optimized powder metallurgy and casting is proposed.
Aiming at the first step of medical application, the properties of the casting composite, especially mechanical properties and corrosion resistance, are necessary to be improved to be fit for implanting the components in the human body. Therefore, the nanoparticle-reinforced Mg matrix and FSP-ed composite (MgO/Mg-Zn-Ca) have been fabricated and characterized in this study. The MgO/Mg-Zn-Ca was produced using casting and hot extrusion process, followed by FSP. The traveling speed and rotating velocity were changed accordingly, and the influences on the microstructure, mechanical behaviors and corrosion properties have been investigated to verify the effects of SPD.

Experimental procedure
The chemical composition of the MgO/Mg-Zn-Ca composite in this study is listed in Table 1. The Mg ingot, Zn ingot and Mg-Ca alloy, were melted in 690~700 o C under the protection of the mixed gas, i.e. SF6 and N2. After the MgO nanoparticles were brought into the molten metal, a high shear rotor-stator mixer was used to stir MgO 6 fully with the molten metal using a stirring velocity of 4000 rpm for 5 min. Then, the metal was cast into a mold at 680 o C to form ingots with a diameter of 60 mm. After that, the ingot was annealed at 300 o C for 20 h to achieve homogenization before final hot extrusion to the size of 1.5 mm × 30 mm (thickness × width). The ingots were then divided into billets with a length of 200 mm via wire-electrode cutting.

Zn
Ca MgO Mg The FSP was carried out using HT-JM 16 × 15/2 gantry friction stir processing machine. Fig. 1a shows the schematic diagram of the manufacturing process. The stirring tool was made of H13 die steel. The pin has an 8.0 mm shoulder diameter with a 2.5 mm tapered pin and a 1.0 mm length, shown in Fig. 1b. Based on the investigation, a tilt angle of 2.5° was machined, and the principal axis was rotated in a clockwise direction during the FSP. Furthermore, four combinations of the key parameters (pin rotational and traveling speeds) were selected, as listed in Table 2.  , which is an unaltered amount). Abbasi et al. [18] have used ω/ν to compare the influence of heat input when studying the mechanical behavior of friction stir welding on different areas of AZ31 magnesium alloy. Therefore, the heat inputs of different FSP processes can be compared by the values of ω/ν, listed in Table 2. Electrochemical corrosion test was performed at 37 ℃ in a Zennium electrochemical workstation, which consists of a glass beaker containing simulated body fluid (SBF) and a standard three-electrode system (saturated calomel electrode (SCE) -reference electrode, sample -working electrode, graphite -control electrode).
The surfaces of the tested samples were ground using silicon papers up to 3000#. After the open circuit voltage was recorded for 30 min, the potential dynamic polarization was performed using a scanning rate of 1 mV/s, and the voltage is from 0.5 V below the self-corrosion potential to 0.5 V above the self-corrosion potential.

Surface morphology after FSP
The friction heat that has been generated between the shoulder of the stirring tool

Microstructural analysis
As observed in Fig. 3a and b, the grains in the base materials (BM) are large and unevenly distributed. Here, the grain sizes were measured by the software of ImageJ.

11
The results of EDS ( Fig. 3c and d) Fig. 3f, and the average grain size is approximately 1.78 μm.
Goh et al. [8] have found that as an inhibitor, the addition of MgO can restrict grain growth. Fig. 3e shows the result of XRD analysis which was conducted to investigate the phases constituted during the FSP. The major diffraction peak is that of the α-Mg matrix. Weak diffraction peaks of MgO and precipitated phases (Ca2Mg6Zn3 and MgxZny alloy, including MgZn2, Mg2Zn3 and Mg4Zn7) are also observed. In this experiment, since the mass fraction of the added MgO is quite small, the diffraction peak of MgO in XRD is difficult to be inspected due to the power limit.   14 Referring to the distribution of MgO in Fig. 5, it can be observed that the MgO is more uniformly distributed in the Mg matrix at 1200 rpm-60 mm/min. More MgO clusters have been adverted in the metal matrix processed at 1200 rpm-80 mm/min. The presence of smaller and more uniformly dispersed particles in the matrix processed at 1200 rpm-60 mm/min is also related to the rotational situation. At 1200 rpm-60mm/min the rotational speed per unit distance is higher than that at 1200 rpm-80 mm/min. Therefore, for the same distance, the sample processed at 1200 rpm-60 mm/min will be subjected to more stirring. The distribution of the grain size in NZ at different processing parameters is shown in Fig. 6. When the parameters are 800 rpm-60 mm/min, 1000 rpm-60 mm/min, and 1200 rpm-60 mm/min, the average grain sizes were 1.08 μm, 1.13 μm, and 1.04 μm, respectively. The increasing stirring speed led to higher heat input, better fluidity of the plastic metal, and higher mechanical force. When the parameter is 1000 rpm-60 mm/min, the grain size is slightly larger than 800 rpm-60 mm/min and 1200 rpm-60 mm/min, which is related to two factors, i.e. the stirring speed and the distribution of inhibitors. Moreover, based on the above results, the effect of agitation is greater than heat input in refining the grain size.
The average grain sizes of the NZ at 1200 rpm-60 mm/min and 1200 rpm-80 mm/min are 1.04 μm and 1.01 μm, respectively. The rotational speed per unit length at 60 mm/min is higher than that at 80 mm/min, leading to more stirring for the same length. Meanwhile, higher rotational speed generates higher heat according to heat input, which is the direct reason for similar grain sizes of two procedures.   (Fig. 3e). The phase of MgxZny was reduced, which would improve the corrosion resistance. Song et al. [21] found that these phases may cause severe galvanic corrosion in the Mg matrix. Lu et al. [22] have shown that the Mg-3Zn phase present in the Mg-Zn-Ca alloy reduces the corrosion resistance. That is another reason for the NZ areas after the FSP owns higher corrosion resistance compared to the BM.
By comparing the XRD results in Fig. 7 and Fig. 3, it was found that FSP did not change the phase types in NZ. However, according to the previous analysis, FSP changes the grain size and particle distribution.

Electrochemical Testing
To study the general corrosion performance, i.e. pitting corrosion resistance, of the FSP-ed samples, the NZs with different parameters were electrochemically tested. The corrosion current (Icorr), corrosion potential (Ecorr), cathode Tafel slope (βc) and anode Tafel slope (βα) are listed in Table 3. In addition, according to the report of Bakhsheshi-Rad et al. [23], the polarization resistance (Rp) of the tested sample was calculated using It has a small corrosion driving force in the SBF solution, and the corrosion rate is slow. Using the same cathode potential, the electrode current of it has a larger cathode current density, and a more complete hydrogen evolution corrosion occurred, the anode side of all samples has a passivation tendency lower than the breakdown potential, which indicates the existence of a surface protective layer, this is mainly a dense Mg(OH)2 conservatory layer. When studying the passivation properties of Mg alloy, the corrosion behavior of Mg alloy. And Ambat et al. [26] pointed out that the corrosion layer of Mg and Mg alloy surfaces is magnesium hydroxide.
During the anodic polarization process, the current of the sample was increased with the potential shift and exhibited a strong discharge activity. This is because the MgO particles are finer and more uniform in the sample produced at 1200 rpm-60 mm/min, and Lin et al. [7] proposed that the MgO particles are in a positive potential position during the etching. Research by Ho et al. [27] shows that grain refinement can cause an increase in corrosion resistance. When the particles are uniformly dispersed, the potential is high and the distribution is uniform and has a small grain size, so uniform corrosion will occur, thereby improving corrosion resistance. In the microstructure with serious particle agglomeration, the agglomerated particles will lead to increased local misorientation difference, and Wang et al. [28] pointed out that high local misorientation difference will destroy corrosion resistance. And the high potential difference will lead to severe pitting, which is detrimental to the final pitting corrosion resistance of the samples.  After electrochemical testing, the corrosion morphology was characterized. Fig.9 shows the corrosion morphology of BM and different parameters. The chromic acid solution is used to remove the corrosion products to judge the corrosion degree of the 21 metal matrix. It can be seen from the figure that there are serious pitting-corrosion pits in the BM, while the area after FSP has a lot fewer pitting-corrosion pits. Because FSP refines the crystal grains, and makes the MgO particles fine and uniform. The study of Ralston et al. [29] showed that the corrosion resistance is directly related to the grain size. In the corrosion morphology of different parameters, when the parameter is 1200rpm-60mm/min, the degree of corrosion is the lightest, the corrosion pits are small and there are a lot of flat areas. The research of Singla et al. [30] showed that the distribution of reinforcing particles could reduce the corrosion degree of composites.
The distribution of MgO particles of 1200-60 is the most uniform, so the degree of corrosion is also the lightest.

Microhardness
Microhardness tests were executed to systematically understand the 22 microstructural evolution during the FSP. The results are presented in Fig. 10.
Irrespective of the processing parameters, the NZ has the highest hardness. Compared to the BM, the hardness of the NZ is approximately 35% higher due to grain size refinement as well as the presence of dispersed and fine MgO particles.
At the traveling velocity of 60 mm/min, the average hardness of the NZ was found to be 97.6 HV, 97.0 HV and 101.2 HV for 800 rpm, 1000 rpm and 1200 rpm, respectively. When the parameter is 1200 rpm-60 mm/min, the grain size is almost the smallest and the MgO particle distribution is the most uniform, leading to the highest hardness. Considering the influence of the different traveling velocities, the hardness of 1200 rpm-80 mm/min is 99.1 HV, which is lower than that of 1200 rpm-60 mm/min. Table 4 offers the information for the grain size, hardness and ω/ν value of the NZ.
Compared to 1200 rpm-80mm/min, although the grains are a little larger at 1200 rpm-60mm/min, the larger ω/ν results in a slightly higher hardness, which is related to the homogeneous distribution of the reinforcements in the NZ. Hence, it can be concluded that the reinforcing particles play a dominant role in strengthening the composite.  Considering sampling components from the NZ for medical application, for instance bone nail, the width of the NZ is a factor to control during the real processing.

Grain misorientation
From the above, the sample performance is optimized with 1200 rpm-60 mm/min. 25 The sample produced under that condition was selected for EBSD characterization to investigate the correlation between grain misorientation and mechanical performance in different districts of the sample. Fig. 11 shows pole figures in different districts of the sample. It can be judged that the BM has a certain texture, and the grain orientation of the NZ is overall perpendicular to the ND-TD plane (or parallel to WD). As a transition region, the influenced zone is in the intermediate state of transforming from the initial texture of the BM to the texture of the base plane {0001}. This is because the grain misorientation is different in the BM, and the orientation factor is different, which will lead to discontinuity in the strain tensor at the grain boundaries. This phenomenon can be interpreted by referring to the Sachs model, through research by Barnett et al. [32]. When Mg alloy is deformed at a temperature below 498 K, its plastic deformation the movement of the slip system, which has an adverse effect on the hardness. In NZ, a large number of bulging grain boundaries were eliminated and many LAGBs were generated, which is related to the dislocation pile-up from the results, and then improve the hardness accordingly. The LAGB ratio of the TMAZ is 0.42, more than that of the NZ, 0.35, however, there are many incomplete grain boundaries. Although the TMAZ is influenced by mechanical agitation and generated heat, the degree is not as much as that of the NZ because it is at the edge of the processed core. Therefore, the heat is not enough for recrystallization after the stirring, resulting in incomplete grains. This is also a reason for the lower hardness in the TMAZ compared to NZ.

Conclusions
In this research, the novel MgO/Mg-Zn-Ca matrix composite was created through optimized casting, and further friction stir processed to optimize the microstructure and properties. The conclusions are as follows.
(1) The key processing parameters of 1200 rpm-60 mm/min have created an FSPed sample with the best quality compared to the other procedures. The surface morphology is in good quality, and the reinforced particles are distributed uniformly.
(2) As for the FSP-ed sample using the appropriate solution, i.e. 1200 rpm-60 mm/min, the grain size is refined by 42%, and as a result, the hardness is improved by 40% through both fine-grain strengthening and nanoparticle strengthening. According to comparison, particle strengthening is dominant in the strengthening mechanism.
(3) Compared to the BM specimens, the NZ specimens are more susceptible to corrosion. The specimens produced under 1200 rpm-60 mm/min exhibited the highest corrosion potential and polarization resistance as well as the smallest corrosion current density. In others word, these specimens represent the best general corrosion resistance.
(4) In the NZ of the sample after the FSP under 1200 rpm-60 mm/min, {0001} basal texture and fully recrystallized grains were formed, which is beneficial to the properties.
In sum, the FSP is beneficial to the property improvement of the innovative composite by Mg-Zn-Ca and MgO nanoparticles, mainly mechanical properties and corrosion resistance. This is the first step for the research of the novel composite that will be applied in medical application, and it has been testified that from the points of mechanical and corrosion properties, the new-designed composite is acceptable.