The Influence of Hydroxyapatite and Alumina Particles on the Mechanical Properties and Corrosion Behavior of Mg-Zn Hybrid Composites for Implants

Considering the necessity for a biodegradable implant alloy with good biocompatibility and mechanical strength, dual ceramic particles of HAP and Al2O3 were added to Mg-Zn alloy to produce a new hybrid composite using powder metallurgy. The paper reports the mechanical and corrosion behaviour of Mg-Zn/HAP/Al2O3 hybrid composites containing variable wt.% HAP and Al2O3 with 15 wt.% total ceramic content. The powders of Mg, Zn, Al2O3 and HAP were milled in a high-energy ball mill, and then compacted under 400 MPa and sintered at 300 °C. Density and compression strength increased with increasing Al2O3 content. HAP facilitated weight gain in Hanks balanced salt solution due to deposition of an apatite layer which promoted anodic behaviour with higher corrosion resistance. A hybrid composite of Mg alloy with 5 wt.% Al2O3 and 10 wt.% HAP displayed 153 MPa compressive strength, 1.37 mm/year corrosion resistance and bioactivity with a CA:P ratio of 1:1.55 and appears to be the most promising biodegradable implant material tested.


Introduction
Stainless steel (SS), Co-Cr-Mo, and titanium (Ti) alloys are traditionally used as biomedical alloys owing to their excellent mechanical properties [1], which enable them to play a crucial role in load-bearing implants for the replacement or repair of damaged bones. But they are not bio-degradable and their elastic modulus is higher than that of bone, causing stress shielding and bone absorption. They also release particles or toxic ions into the human body, which causes chronic inflammation, reducing biocompatibility, or tissue loss. In addition, the significant incompatibility of natural bone and implant material may lead to stress, lowering implant stability [2]. A biodegradable implant that decomposes in the body can solve these problems as it will not cause long-lasting physical irritation and furthermore avoids the second surgery to remove the implants [3].
Magnesium (Mg) and its alloys have become attractive candidates for a temporary implant material that avoids the necessity for a secondary operation to eliminate the implant material during healing. They are biodegradable, inherently biocompatible and possess low density and high mechanical properties [4]. Moreover, the modulus of elasticity of magnesium alloy is about 40-45 GPa, which is closer to human bone compared to stainless steel and titanium alloys. Mg also is crucial for health, safe and can be excreted by the kidney [1,5].
However, the main drawback of magnesium and its alloy is their fast degradation rate or corrosion rate in the physiological environment. The fast degradation of magnesium alloy not only results in the implant losing its mechanical integrity before the injured tissues have sufficient time to heal but also leads to serious hydrogen evolution and alkalization of body fluids [6]. Therefore, in order to use Mg as an effective implant, the corrosion rate of Mg needed to be slowed down by alloying the magnesium with non-toxic elements such as Ca, Zn or Zr [7] to form novel magnesium alloys. In this work, Zn was used as an alloying ingredients because Zn can enhance the mechanical and corrosion properties of Mg [7].
One possible way to improve Mg's mechanical properties and biocompatibility would be to reinforce the magnesium alloy with bioceramic. Calcium phosphate ceramic has been widely used for hard tissue replacements. One type of bioceramic, HAP (Ca 10 (PO 4 ) 6 (OH) 2 [8] has been widely used due its outstanding biocompatibility, good bonding with bone tissues and also its chemical composition, which is very closely similar to the biological apatite existing in human hard tissues. HAP also shows very low solubility as related to beta tricalcium phosphate (β-TCP) in body fluid [9] and, of the phosphate groups, it alone has the feature of biocompatibility. HAP can support bone growth and joints in bone structure without dissolving or breaking down. However, the successful application of HAP ceramic in load-bearing areas is limited because HAP does not have good mechanical strength and has low bioactive property due to its low resorbability [10].
Witte et al. [11] reported the potential of AZ91D magnesium matrix composite reinforced with HAP particles for load-bearing applications. However, HAP is brittle by nature, resulting in a drop in mechanical properties, as reported by Khalil et al. [12] for Mg-HAP and by Soon et al. [13] and Salleh et al. [14] for Mg-Zn/HAP. On the other hand, alumina (Al 2 O 3 ) bioceramics have been used widely as components in orthopaedics and dental application since the 1990s. They have been used in dental fixtures and aesthetics (dental restoration) owing to their outstanding biocompatibility and mechanical properties. Furthermore, Al 2 O 3 is also known to be an excellent bioinert ceramic [15] and has therefore become one of the most widely recommend reinforcement materials for HAP bioceramics [14]. Alumina is a hard ceramic material with high elastic modulus, strong ionic bond and oxygen-rich stoichiometry, which makes it chemically bioinert and stable in the human body [16].
Considering the needs of non-toxic and biodegradable alloy with bioactivity and good mechanical strength, this work investigates a Mg-Zn hybrid composite containing dual ceramic particles of HAP and Al 2 O 3 . Specifically, this paper reports the effect of HAP and Al 2 O 3 content in varying ratios in Mg-Zn composite on the microstructure, mechanical properties and corrosion behaviour in simulated body fluid (SBF).

Materials and Methods
Mg contains 98.5% purity (Merck, Darmstadt, Germany), meanwhile Zn is more than 99.9% pure (Alfa Aesar, Lancashire, England), and HAP (Sigma-Aldrich, Burlington, VT, USA) and Al 2 O 3 (Fluka, Sigma-Aldrich, Burlington, VT, USA) are more than 90.0% pure, respectively. The particle size of Mg, Zn, HAP and Al 2 O 3 powder is 288.74, 97.64, 2.87 and 10.44 µm respectively. Mg-Zn/HAP/Al 2 O 3 composites were produced with HAP and Al 2 O 3 content of 0, 5, 10 and 15 wt.% and 15, 10, 5 and 0 wt.%, respectively, for a total of 15 wt.% ceramic content. Mg and Zn content in the binary alloy matrix was 94 wt.% and 6 wt.%, respectively. Powders were mixed and milled in a Fritch planetary ball mill at 220 rpm for 2 h. The milled and homogenous powders were uniaxially cold pressed using 400 MPa for 2 min and sintered at 300 • C under argon gas flow at 10 • C/min for 1 h.
The sample of sintered composite was observed under SUPRA 35VP field emission scanning electron microscopy (FESEM) using backscattered and secondary electron (BE and SE) modes. Energy-dispersive X-ray (EDX) was used for compositional analysis. Phase identification on the sintered sample was performed using X-ray diffraction (XRD).
The density of the sintered sample was measured and tested using Archimedes' principle. The samples were immersed in water placed on the Sartorius electronic analytical balance by using four decimals of accuracy. Five readings were recorded and taken for each sample, and the average density was calculated. Meanwhile, for microhardness measurements, Vickers indenter at a load of 300 gf and dwell time of 10 s were conducted on the sample composites. Ten readings were calculated from each sample. A compression test was accomplished at room temperature using the ASTM E9-89a standard test methods.
Most bioactive materials tend to bond to living bone through the formation of an apatite layer that forms on the surface. However, in this study, Hanks Balanced Salt Solution (HBSS) was chosen due to its commercial availability and its continuing use in biomimetic experiments. It is useful to observe the apatite-forming ability on a material's surface in HBSS to predict the in vivo bonding ability of the material. The mineralization of the apatite layer on the composite was followed according to ASTM-G31-72 [17] in Hanks solution [18]. For the immersion of samples in HBSS, the samples were then ground using SiC emery paper starting from 360 to 2000 grit, and then the samples were polished using 1, 0.3 and 0.05 µm alumina paste. The polished specimens were put under warm airflow for a few minutes. Samples with diameter of 10 mm and thickness of 4 mm were measured and balanced using a four-decimal electronic analytical balance (Sartorius) before the samples were immersed in the HBSS. Before immersion, the volume of HBSS was calculated according to Equation (1): where V s is the volume of HBSS in (mL) necessary to be put into the falcon tube, while S a refers to the surface area (mm 2 ) of pellets. Meanwhile, the sintered samples were in the cylindrical disc shape, and S a was calculated according to Equation (2): where r (radius) and t (thickness) were both in mm unit of the sample. The total amount of the HBSS calculated as per Equation (1) was then poured into falcon tube, and the samples were submerged at the bottom of a tube and then were put into a water bath heated at 37 • C. After 2 h of immersion, the samples were taken from the water bath heated and were removed from the solution of HBSS. Then, the samples were put in a container and cleaned with dilute chromic acid (a mixture of CrO 3 and AgNO 3 ). The function of dilute chromic acid is to remove the corrosion layer by dissolving Mg(OH) 2 then rinsing with ethanol, followed by rinsing with de-ionized water. The samples then were put in an oven for 24 h at 70 • C. An average of five measurements was taken for each composition. Finally, the weight of the dried samples (m f ) after the corrosion test was measured, and a percentage change in mass at different times was calculated using the weight loss (W L ) Equation (3) for an in vivo bioactivity test.
where m o is the sample weight before the immersion test. The corrosion rates were calculated by the weight loss according to the following Equation (4) [19]: where ∆W is the weight loss in gram (g); A is the sample area exposed to the solution in centimetre (cm 2 ), and t is the exposure time in hour (h). A potentiodynamic polarization test was carried out using an Autolab (PGSTAT-302N) device. A standard three-electrode system was employed, comprising a specimen as working electrode, a platinum plate as a counter electrode and a saturated calomel electrode (SCE) as a reference electrode. Linear polarization tests were done with a scanning rate of 0.5 mv/s and were repeated five times for each composite. The corrosion current densities and the corrosion potentials were directly derived from the linear polarization plots by Tafels extrapolation in which the E corr and I corr were acquired from the intersection of the extrapolated I red and I ox Tafel lines.

XRD Analysis
Based on Figure 1, the XRD pattern of Mg-Zn alloy shows the absence of Zn peaks in the diffractogram. In the XRD pattern of as-milled Mg-Zn, Zn peaks were still observed in the diffraction angles of 38.90 • , 43.09 • and 54.22 • . As the compact was sintered at 300 • C, the Zn peaks disappeared, and the diffraction angles of sintered Mg-Zn were shifted to the higher diffraction angles, indicating the expansion of the Mg lattice as the Zn was being completely solid-solved into the lattice in sintering process. as working electrode, a platinum plate as a counter electrode and a saturated calomel electrode (SCE) as a reference electrode. Linear polarization tests were done with a scanning rate of 0.5 mv/s and were repeated five times for each composite. The corrosion current densities and the corrosion potentials were directly derived from the linear polarization plots by Tafels extrapolation in which the Ecorr and I corr were acquired from the intersection of the extrapolated Ired and Iox Tafel lines.

XRD Analysis
Based on Figure 1, the XRD pattern of Mg-Zn alloy shows the absence of Zn peaks in the diffractogram. In the XRD pattern of as-milled Mg-Zn, Zn peaks were still observed in the diffraction angles of 38.90°, 43.09° and 54.22°. As the compact was sintered at 300 °C, the Zn peaks disappeared, and the diffraction angles of sintered Mg-Zn were shifted to the higher diffraction angles, indicating the expansion of the Mg lattice as the Zn was being completely solid-solved into the lattice in sintering process.        Table 1 presents the average grain size of the Mg matrix. The grain size changed considerably for composites consisting of 0 wt.% and 5 wt.% Al 2 O 3 (180.03 µm and 368.40 µm, respectively). However, the grain size developed significantly larger with the addition of 10 and 15 wt.% Al 2 O 3 . Furthermore, the HAP and Al 2 O 3 particles are located typically at the grain boundaries of the matrix. The presence of ceramic particles at the grain boundary appears to decrease with increasing presence of Al 2 O 3 particles. Both situations can be explained by the large difference in the particle size of the HAP and Al 2 O 3 powders, i.e., 2.87 µm and 10.44 µm, respectively. Composites with high HAP content resulted in the refinement of the Mg grain size as HAP is finer, so more particles of HAP are available to resist Mg grain growth by the pinning of boundaries. Thus, composites with 0 and 5 wt.% Al 2 O 3 , which 15 and 10 wt.% HAP, respectively, have a finer Mg grain.    SEM for all composites is shown in Figure 4 for overall microstructure uniformity, especially the dispersion of the ceramic particles, either HAP or Al 2 Table 2 presents the density and porosity of Mg-Zn/HAP/Al2O3 composites with various Al2O3 contents. Addition of Al2O3 powder resulted in a slight increment in density ranging from 1.846 g/cm 3 to 1.947 g/cm 3 . This finding suggests that the higher density Al2O3 particles ensure that Al2O3 plays a key role in controlling the density of the composites. The density of Al2O3 is 3.97 g/cm 3 while the density of pure Mg is 1.736 g/cm 3 . Thus, adding more Al2O3 caused in a rise in the density of the composite in accordance with the trend of theoretical density estimated by the composite rule of mixtures (ROM). The increase in relative density with increasing Al2O3 content may also be due to the coarser Al2O3 particles better supporting the densification of the composite during powder pressing and sintering than did the fine HAP particles. Therefore, composites with 15 wt.% Al2O3 showed the lowest porosity among the various compositions.   Table 2 presents the density and porosity of Mg-Zn/HAP/Al 2 O 3 composites with various Al 2 O 3 contents. Addition of Al 2 O 3 powder resulted in a slight increment in density ranging from 1.846 g/cm 3 to 1.947 g/cm 3 . This finding suggests that the higher density Al 2 O 3 particles ensure that Al 2 O 3 plays a key role in controlling the density of the composites. The density of Al 2 O 3 is 3.97 g/cm 3 while the density of pure Mg is 1.736 g/cm 3 . Thus, adding more Al 2 O 3 caused in a rise in the density of the composite in accordance with the trend of theoretical density estimated by the composite rule of mixtures (ROM). The increase in relative density with increasing Al 2 O 3 content may also be due to the coarser Al 2 O 3 particles better supporting the densification of the composite during powder pressing and sintering than did the fine HAP particles. Therefore, composites with 15 wt.% Al 2 O 3 showed the lowest porosity among the various compositions. Stress-strain curves for representative samples with different composition of alumina are displayed in Figure 5 Figure 7 presents the change in the weight loss of Mg-Zn/HAP/Al2O3 composites in HBSS after 2 h immersion time. Composites with 10 wt.% Al2O3 and less gained in weight, while composites with 15 wt.% Al2O3 lost weight. The amount of weight loss was influenced significantly by the addition of Al2O3, particularly at 15 wt.% Al2O3. Increasing Al2O3 content increased weight loss.

Immersion Test
Weight loss due to Mg-Zn matrix degradation is possible as Mg has low resistance to corrosion [22]. However, since the immersion test was conducted in HBSS, the HAP in the composites appear to have encouraged the deposition of the calcium phosphate layer that caused a weight gain of the composites. It is believed that the weight gains for 0, 5 and 10 wt.% Al2O3 with composites was due to the presence of multiple protective layers contributed from thin Mg(OH)2 and the thick apatite layer that was promoted by the presence of HAP in the composites.
On the other hand, composites with 15 wt.% Al2O3 without HAP showed the highest weight loss at 2.582%. The presence of high Al2O3 content in the absence of HAP in the Mg matrix did not lead to the creation of an apatite layer at the surface because the Al2O3 is bioinert, inhibiting the development of an apatite layer. As a bioinert material, Al2O3 can only remain stable under biological conditions and does not substantially chemically react in the body or exchange electrons with any material it comes in contact with [6]. The formation and creation of an only-Mg(OH)2 layer on the surface of the 15 wt.% Al2O3 composite could not be effective in protecting the magnesium alloy matrix from corrosion.   Weight loss due to Mg-Zn matrix degradation is possible as Mg has low resistance to corrosion [22]. However, since the immersion test was conducted in HBSS, the HAP in the composites appear to have encouraged the deposition of the calcium phosphate layer that caused a weight gain of the composites. It is believed that the weight gains for 0, 5 and 10 wt.% Al 2 O 3 with composites was due to the presence of multiple protective layers contributed from thin Mg(OH) 2 and the thick apatite layer that was promoted by the presence of HAP in the composites.

Immersion Test
On the other hand, composites with 15 wt.% Al 2 O 3 without HAP showed the highest weight loss at 2.582%. The presence of high Al 2 O 3 content in the absence of HAP in the Mg matrix did not lead to the creation of an apatite layer at the surface because the Al 2 O 3 is bioinert, inhibiting the development of an apatite layer. As a bioinert material, Al 2 O 3 can only remain stable under biological conditions and does not substantially chemically react in the body or exchange electrons with any material it comes in contact with [6]. The formation and creation of an only-Mg(OH) 2 layer on the surface of the 15 wt.% Al 2 O 3 composite could not be effective in protecting the magnesium alloy matrix from corrosion. Figure 8 presents the macroscopic presence of the corroded surface after immersion in HBSS, with its surface degraded owing to the attack of Cl − ions in HBSS solution. Localized pitting was observed for in the samples, covering the entire surface during the course of immersion. The Mg(OH) 2 layer and possibly the white calcium phosphate layer known as the apatite layer were deposited on the surface of all composites. Results of this research are in agreement with Zhang et al. [22] and Bakhsheshi-Rad et al. [23], who discovered both MgOH 2 and apatite layers in Mg-Zn alloy and Mg-Ca-Zn, correspondingly.
In general, this observation is consistent with weight loss measurement data wherein a higher content of Al 2 O 3 was shown to increase the tendency to weight loss owing to the dissolution of the Mg-Zn alloy matrix. The super-protective Mg(OH) 2 layer and the white calcium phosphate in the protective layer inhibited excessive degradation in composites with HAP. Therefore, we conclude that composites with alumina alone could not prevent the surface from corroding in high Cl − ions medium because Al 2 O 3 barely interacted with HBSS solution to form calcium phosphate.   [22] and Bakhsheshi-Rad et al. [23], who discovered both MgOH2 and apatite layers in Mg-Zn alloy and Mg-Ca-Zn, correspondingly.
In general, this observation is consistent with weight loss measurement data wherein a higher content of Al2O3 was shown to increase the tendency to weight loss owing to the dissolution of the Mg-Zn alloy matrix. The super-protective Mg(OH)2 layer and the white calcium phosphate in the protective layer inhibited excessive degradation in composites with HAP. Therefore, we conclude that composites with alumina alone could not prevent the surface from corroding in high Cl − ions medium because Al2O3 barely interacted with HBSS solution to form calcium phosphate.

Hour (H)
Al2O3 Content (wt.%) 0 5 10 15 2  Figure 9 shows the morphology of composites under SEM after immersion in HBSS with insert images at higher magnification. The corrosion product was formed on the surface, and the deposited layer covering the samples exhibited a network of cracks and pits with different depths and sizes. The pits and cracks were observed in all composites, but the size and formation of cracks increased with the addition of high Al2O3 content.  Figure 9 shows the morphology of composites under SEM after immersion in HBSS with insert images at higher magnification. The corrosion product was formed on the surface, and the deposited layer covering the samples exhibited a network of cracks and pits with different depths and sizes. The pits and cracks were observed in all composites, but the size and formation of cracks increased with the addition of high Al 2 O 3 content. The cracks increased the contact area between the corrosive solutions, thus accelerating the corrosion of the alloy. The crack creation has been attributed to the water loss of the surface shrinkage and corrosion products surface shrinkage, as proposed by Bakhsheshi-Rad et al. [23]. The SEM images are in good agreement with this proposal (result of weight loss), where higher Al 2 O 3 content (at the expense of HAP content) contributed to higher weight loss, while lower Al 2 O 3 content (associated with higher HAP content) contributed to weight gain.
The cracks increased the contact area between the corrosive solutions, thus accelerating the corrosion of the alloy. The crack creation has been attributed to the water loss of the surface shrinkage and corrosion products surface shrinkage, as proposed by Bakhsheshi-Rad et al. [23]. The SEM images are in good agreement with this proposal (result of weight loss), where higher Al2O3 content (at the expense of HAP content) contributed to higher weight loss, while lower Al2O3 content (associated with higher HAP content) contributed to weight gain. EDX analysis was used to examine the composition of the deposited layer or corrosion product on the composites. Figure 10 and Table 3 show the presence of magnesium (Mg) and oxygen (O), as well as a small amount of phosphorus (P), calcium (Ca) and aluminium (Al) as the constituents of the corrosion products. The presence of Mg and O in the corrosion products demonstrated the possibility of the formation of magnesium hydroxide Mg(OH)2, which acted as the protective barrier against further degradation in the aggressive corrosion medium. The EDX analysis of the 0 wt.% Al2O3 (with 15 wt.% HAP) indicated the presence of Ca and P functioning as a sponge-like deposit that was found to cover almost all regions on the sample surface with the amount and Ca:P ratio at the highest value of all the composites. This analysis proposes that calcium-deficient HAP was placed on the substrate, as also discovered by . EDX reveals that for 0 wt.% Al2O3, the Ca/P molar ratio was 2.32, which suggests better bioactivity than the other composites. EDX analysis was used to examine the composition of the deposited layer or corrosion product on the composites. Figure 10 and Table 3 show the presence of magnesium (Mg) and oxygen (O), as well as a small amount of phosphorus (P), calcium (Ca) and aluminium (Al) as the constituents of the corrosion products. The presence of Mg and O in the corrosion products demonstrated the possibility of the formation of magnesium hydroxide Mg(OH) 2 , which acted as the protective barrier against further degradation in the aggressive corrosion medium. The EDX analysis of the 0 wt.% Al 2 O 3 (with 15 wt.% HAP) indicated the presence of Ca and P functioning as a sponge-like deposit that was found to cover almost all regions on the sample surface with the amount and Ca:P ratio at the highest value of all the composites. This analysis proposes that calcium-deficient HAP was placed on the substrate, as also discovered by [13][14][15][16][17][18][19][20][21][22][23][24][25][26][27][28][29][30][31][32]. EDX reveals that for 0 wt.% Al 2 O 3 , the Ca/P molar ratio was 2.32, which suggests better bioactivity than the other composites.   Figure 11 shows the Tafel plot for Mg-Zn/HAP/Al 2 O 3 composites with different compositions of Al 2 O 3 . In overall, the anodic polarization curves have been attributed to the dissolution of Mg, leading to the formation of Mg 2+ , while the cathodic polarization curves have been associated with the reduction of water [24]. Data derived from the intersection of the anodic and cathodic Tafel line extrapolations curves are corrosion potential (E corr ) and corrosion current (I corr ) and corrosion rate, as summarized in Table 4. Typically, in a potentiodynamic curve, the more positive E corr and the lower I corr correspond to lower corrosion rate, or in other words, corrosion resistance is higher. It was found that pure Mg (as a controlled sample) showed the worst corrosion resistance of all the composites with the value of the E corr (−1.675 V), I corr (0.530 × 10 −6 Acm −2 ) and corrosion rate (3.24 mm/year).    Composites having more positive E corr with higher I corr implies a higher corrosion rate, for two reasons. First, Zn as an alloying element of Mg matrix has a stabilizing effect on the protective film formed on the Mg alloy [22]. It has been informed that Zn can increase the corrosion potential of Mg alloy, thus reducing the corrosion rate [15]. Secondly, the additional ceramic particles with a total amount of 15 wt.% of the total composite composition also can reduced the corrosion rate of the composite.
When coupling the active magnesium with a moderately noble material, galvanic corrosion is the primary prospect that may accelerate corrosion of Mg. Instead, corrosion resistance was found to be improved with the addition of ceramic content into the Mg matrix, indicating that the Mg-Zn/HAP/Al 2 O 3 composite was the least susceptible to corrosion. Previous work by Ghasali et al. [25] found that the conductivity of the inert materials plays a important role in inhibiting galvanic corrosion between the Al 2 O 3 and Mg matrix since Al 2 O 3 is an insulator. In this research, both HAP and Al 2 O 3 are insulating ceramic particles, which may result in disruptive electric charges because of the influence of an electric field. This means that the electric charged cannot flow freely into the composite or very little electric current will flow through it under the influence of an electric field. Thus, it is expected to have lower E corr , I corr and corrosion rate than the magnesium alloy.
E corr , I corr and the corrosion rate of the 0 wt.% Al 2 O 3 composite (with the addition of 15 wt.% HAP) in Table 4 were lower than those of the 15 wt.% Al 2 O 3 composite (with 0 wt.% HAP). This decrease could be related to the higher HAP content in the 0 wt.% Al 2 O 3 composite, which known as the hydroxyapatite nuclei, and the hydroxyapatite will grow unexpectedly, consuming the Ca 2+ and PO 4 3− [22]. Furthermore, the presence in the corrosion products of high amounts of Mg and O in the EDX analysis shows the formation of magnesium hydroxide Mg(OH) 2 , also known as brucite [26]. Brucite forms when the samples are immersed in HBSS solution, and the MgO in the outer layers reacts with the corrosive solution and converts into insoluble Mg(OH) 2 . However, the aggressive Cl − in the HBSS solution transforms Mg(OH) 2 into a more soluble MgCl 2 [27] . The breakdown of Mg(OH) 2 decreases the size of the protected area, consequently promoting further dissolution of the sample. The multiple protection effects offered by the corrosion products such as Mg(OH) 2 and hydroxyapatite at the surface may be the reasons for the slower corrosion rate observed for the 0 wt.% Al 2 O 3 composite. Witte et al. [28] and Ye et al. [29] also have reported that protective layers on the composite such as Mg(OH) 2 and HAP composed of multiple corrosion products displayed better corrosion resistance than the single Mg(OH) 2 layer only.
The more negative value of E corr (−1.648 V), and higher values of I corr (0.474 × 10 −6 A /cm 2 ) and corrosion rate (2.90 mm/year) for the 15 wt.% Al 2 O 3 containing composite suggests that this composite has poor corrosion resistance. This low corrosion resistance results from the failure to form protective apatite crystals or any form of apatite layer at the surface to prevent the magnesium matrix from corrosion.
However, Al 2 O 3 is bioinert with poor surface bioactivity, which results in a weak bonebonding ability and provides no favourable surface for developing biological adherent interfaces with bone [21]. In other words, Al 2 O 3 is incapable of forming and developing any apatite crystal on its surface under any conditions. Therefore, since the 15 wt.% Al 2 O 3 composite did not contain HAP, it was unable to initiate other calcium phosphate and apatite when immersed in the HBSS solution. Without HAP particles, the composite had hardly any interaction with the HBSS solution that would allow the formation of calcium phosphate, because Al-OH has no affinity for calcium and phosphate.
Apparently, composites with a lower amount of Al 2 O 3 shifted the Tafel plot to a more positive value of E corr and lower value of I corr , indicating a lower corrosion rate. Thus, composites without Al 2 O 3 have the best resistance to corrosion in HBSS, as E corr is the most positive value (−1.604 V) and I corr is the lowest (0.252 × 10 −6 A/cm 2 ) found in this study. This is attributed to the fact that the composite without any Al 2 O 3 is the composite with the highest amount of HAP. HAP particles initiated the formation of a thicker protective HAP layer in addition to the Mg(OH) 2 layer at the composite surface [30], thus increasing the corrosion resistance of the magnesium alloy matrix.
Furthermore, the reduction in the corrosion rate of the composite for 0 and 5 wt.% Al 2 O 3 can be related to the refinement of the microstructure, which was affected by the addition of HAP and Al 2 O 3 into the Mg matrix. Thus, it can be concluded that fine grain composites of 0 and 5 wt.% Al 2 O 3 show better corrosion resistance than do coarse grain composites and hence, have the lowest corrosion rate. This trend is in agreement with other research Waizy et al. [31] and Aung & Zhou [32], which found that the materials with smaller grain size can provide more grain boundaries that can act as a physical barrier that prevents degradation due to corrosion.

Conclusions
The effect of HAP and Al 2 O 3 ratio on the physical, mechanical, corrosion and bioactivity properties of magnesium alloy hybrid composite metallurgy was investigated. HAP and Al 2 O 3 composition was designed from 0 to 15 wt.% and 15 wt.% to 0 wt.% with constant Mg-Zn alloy matrix composition (85 wt.%) to elucidate the properties of new hybrid composites.
Compression strength was increased from 126.48 MPa to 244.20 MPa, respectively with the increase of Al 2 O 3 content presenting the significant improvement of the strengthening of the soft Mg-Zn alloy provided by the Al 2 O 3 hard particles. Meanwhile the addition of 15 wt.% HAP (with 0 wt.% Al 2 O 3 ) in a polarization test showed that E corr was shifted to the most positive value, while I corr and the corrosion rate was lowest. Besides, the highest weight gain was observed, which is correlated to good bioactivity in forming an apatite layer due to the presence of the maximum amount of HAP in this single composite.
Considering the requirement of both high strength and good bioactivity properties, the most suitable composition is achieved by a Mg-Zn alloy matrix hybrid composite consisting of dual ceramic particles of 5 wt.% Al 2 O 3 and 10 wt.% HAP with 153 MPa compressive strength. The immersion test for corrosion showed that this composition exhibited 0.328% weight gain, 1.37 mm/year corrosion rate and a 1:1.55 Ca:P ratio for bioactivity.
This work highlights the significant of adding dual ceramic particles (bioactive HAP and strong Al 2 O 3 particles) to a Mg-Zn alloy-based hybrid composite in comparison to a single ceramic addition. Improvement in mechanical, corrosion and bioactivity properties of the new hybrid composite could contribute to the replacement of cortical and cancellous bones with a stronger and bioactive Mg-Zn alloy for biodegradable implant application.