Effect of Homogenization on Microstructure Characteristics, Corrosion and Biocompatibility of Mg-Zn-Mn-xCa Alloys

The corrosion behaviors of Mg-2Zn-0.2Mn-xCa (denoted as MZM-xCa alloys) in homogenization state have been investigated by immersion test and electrochemical techniques in a simulated physiological condition. The microstructure features were characterized using scanning electron microscopy (SEM), X-ray diffraction (XRD), transmission electron microscopy (TEM) and electron probe microanalysis (EPMA), and the corrosion mechanism was illustrated using atomic force microscope (AFM), X-ray photoelectron spectroscopy (XPS) and confocal laser scanning microscopy (CLSM). The electrochemical and immersion test verify the MZM-0.38% Ca owns the best corrosion performance with the corrosion rate of 6.27 mm/year. Furthermore, the film layer of MZM-0.38% Ca is more compact and denser than that of others. This improvement could be associated with the combined effects of the suitable content of Zn/Ca dissolving into the α-Mg matrix and the modification of Ca-containing compounds by heat-treatment. However, the morphologies were transformed from uniform corrosion to localized pitting corrosion with Ca further addition. It could be explained that the excessive Ca addition can strengthen the nucleation driving force for the second phase formation, and the large volumes fraction of micro-galvanic present interface sites accelerate the nucleation driving force for corrosion propagation. In addition, in vitro biocompatibility tests also show the MZM-0.38% Ca was safe to bone mesenchymal stem cells (BMSCs) and was promising to be utilized as implant materials.


Introduction
Recently, magnesium-based alloys and composites have a crucial role in implant materials applications [1][2][3][4][5][6], due to the moderate mechanical properties similar to bone tissue [7][8][9][10][11][12]. In addition, Mg 2+ is also the second most important cation for special physiological functions, which can activate the body enzymes, participate in the protein synthesis and muscle growth, and maintain the structure stability of nucleic acid [13]. Besides, Mg-based alloys could be gradually dissolved, absorbed, consumed by metabolic reactions, and then excreted with urinary system after the tissues heal [7,13,14]. Moreover, magnesium alloys have no systemic toxicity in human body and possess excellent biocompatibility with cells activity [15]. Thus, Mg and its alloys have been paid more attention to develop as the candidate biomaterials for clinical applications [1,2,6]. However, these Mg alloys were extremely susceptible to corrosion medium when they were exposed to the physiological condition, leading to the loss of mechanical integrity before the tissues have recovered completely [3]. As a result, how to accurately control degradation rate and explore the corrosion mechanism of biodegradable materials in vitro has become an important problem.

Electrochemical Measurement
According to ASTM G3-89 [42], the polarization measurements were measured in the Kokubo solution using an electrochemical workstation (Princeton Versa STAT 3F, AMETEK, Chicago, IL, USA). The standard three-electrode system consists of a saturated calomel electrode as the reference electrode, a platinum plate as the counter electrode and the as-prepared sample with an exposed area of 1 cm 2 as the working electrode. The composition of Kokubo solution is listed in Table 2. When the open circuit potential (OCP) reaches the steady-state for 30 min, the measurement was conducted with a scan rate of 1 mV/s. The corrosion potentials (E corr , V SCE ) and corrosion current densities (I corr , mA/cm 2 ) were calculated from the polarization curves by the Tafel extrapolation method. The corrosion rate (P i , mm/year) was related to corrosion current densities, according to the following equation [35]:

Immersion Test
In accordance with ASTM G31-72 [44], the immersion test was conducted in the Kokubo solution at 37 • C for 240 h, and the ratio of specimen surface area to solution volume was 1:30 cm 2 /mL. Prior to immersion tests, all samples were polished down to 0.25 µm and weighted. After the immersion test, the corrosion products were analyzed by SEM, XPS and XRD, respectively. Subsequently, according to the G1-03 (2011), the corrosion products were chemically removed using a solution containing 200 g/L CrO 3 and 10 g/L AgNO 3 , and then the corrosion morphologies were characterized by confocal laser scanning microscopy (CLSM) (OLS 4000, Olympus, Tokyo, Japan). The corrosion rate (CR) was determined using the following equation: where K is a constant (8.76 × 10 4 for rate unit of millimeter per year), W is mass loss (g), A is the sample area exposed to solution (cm 2 ), T is the time of exposed (h) and D is the density of materials (g·cm 3 ). Each specimen was tested three times under the identical conditions.

In Vitro Biocompatibility Assessment
Human bone marrow mesenchymal stem cells were adopted to evaluate the cytotoxicity of the investigated MZM-xCa alloys. The BMSCs were cultured in Dulbecco's modified Eagle's medium (DMEM, Gibco, Waltham, MA USA) supplemented with 10% fetal bovine serum (FBS, Gibco, Waltham, MA USA), 100 units/mL penicillin and 100 units/mL streptomycin in a cell incubator (humidified atmosphere with 5% CO 2 at 37 • C). The cytotoxicity tests were carried out by indirect contact. Polished samples were washed, dried in air and sterilized via Co60 γ ray radiation.
Extracts were prepared using DMEM serum free medium as the extraction medium with the surface area to extraction medium ratio 1.25 mL/cm 2 in a humidified atmosphere with 5% CO 2 at 37 • C for 72 h, according to ISO 10993-5:1999. The supernatant fluid was withdrawn and filtered to prepare the extraction medium, and subsequently preserved at 4 • C before the cytotoxicity test. The control groups involved the use of DMEM medium as negative controls. The BMSCs were seeded in 96-well plates at a density of 5000 cells/well and incubated for 24 h to allow attachment. Then, the medium was replaced with 100 µL extract or DMEM medium, and incubated for 1, 2 and 3 days, respectively. In addition, the relative growth rate (RGR) was calculated for all samples using the Cell Counting Kit-8 (CCK-8, Sigma-Aldrich, St. Louis, MO, USA). After each assay time point, 10 µL CCK-8 solution was added to each well and then it was incubated for 2 h. The optical density (OD) was measured at the wavelength of 450 nm using a microplate reader (iMARK, Bio-Rad, Hercules, CA, USA). The cell viabilities were expressed as relative growth rate (RGR) determined by RGR (%) = {(OD)sample/OD (negative control)} × 100%. All the data were presented as the mean RGR value ± standard deviation. Statistical analysis was conducted to evaluate the difference in cell viability by analysis of variance (ANOVA), where the statistical significance was defined as p < 0.05. Figure 1 reveals the microstructure evolution of as-homogenized MZM-xCa alloys with various Ca content addition. The grain size within Mg-2Zn-0.2Mn (Figure 1a) alloy was remarkably inhomogeneous, for which the smallest grain size was less than 40-50 µm and the largest size was more than 400 µm. Besides, the MZM has the equiaxed grains, and the average gain size was about 180 µm by the linear intercept methods. The grain size slightly reduced with the 0.38% Ca addition and only a small quantity of precipitation phases were detected in the matrix and triangular grain boundary (the inset of Figure 1b). When the Ca concentration further increased to 0.76% and 1.10%, the grain size of the two alloys significantly decreased and the average grain size was 87 µm and 60 µm, respectively. Some holes were observed in the matrix due to peeling of the precipitates. As for the reason of grain refinement, it was widely accepted that an alloy liquid consisting of a small amount of Ca atoms at the solid/liquid interface could form composition undercooling in the diffusion layer, in which the nucleation particles may be activated in the undercooling region, leading to forming more grain nucleation particles and refining the grain size [35]. Moreover, the Ca atoms, with low speed of diffusion, hindered the grain growth in the interfaces and limited the growth rate of crystal. Thus, the grain size of MZM-xCa alloys clearly refined (Figure 1d). (humidified atmosphere with 5% CO2 at 37 °C). The cytotoxicity tests were carried out by indirect contact. Polished samples were washed, dried in air and sterilized via Co60 γ ray radiation. Extracts were prepared using DMEM serum free medium as the extraction medium with the surface area to extraction medium ratio 1.25 mL/cm 2 in a humidified atmosphere with 5% CO2 at 37 °C for 72 h, according to ISO 10993-5:1999. The supernatant fluid was withdrawn and filtered to prepare the extraction medium, and subsequently preserved at 4 °C before the cytotoxicity test. The control groups involved the use of DMEM medium as negative controls. The BMSCs were seeded in 96-well plates at a density of 5000 cells/well and incubated for 24 h to allow attachment. Then, the medium was replaced with 100 μL extract or DMEM medium, and incubated for 1, 2 and 3 days, respectively. In addition, the relative growth rate (RGR) was calculated for all samples using the Cell Counting Kit-8 (CCK-8, Sigma-Aldrich, St. Louis, MO, USA). After each assay time point, 10 μL CCK-8 solution was added to each well and then it was incubated for 2 h. The optical density (OD) was measured at the wavelength of 450 nm using a microplate reader (iMARK, Bio-Rad, Hercules, CA, USA). The cell viabilities were expressed as relative growth rate (RGR) determined by RGR (%) = {(OD)sample/OD (negative control)} × 100%. All the data were presented as the mean RGR value ± standard deviation. Statistical analysis was conducted to evaluate the difference in cell viability by analysis of variance (ANOVA), where the statistical significance was defined as p < 0.05. Figure 1 reveals the microstructure evolution of as-homogenized MZM-xCa alloys with various Ca content addition. The grain size within Mg-2Zn-0.2Mn ( Figure 1a) alloy was remarkably inhomogeneous, for which the smallest grain size was less than 40-50 μm and the largest size was more than 400 μm. Besides, the MZM has the equiaxed grains, and the average gain size was about 180 μm by the linear intercept methods. The grain size slightly reduced with the 0.38% Ca addition and only a small quantity of precipitation phases were detected in the matrix and triangular grain boundary (the inset of Figure 1b). When the Ca concentration further increased to 0.76% and 1.10%, the grain size of the two alloys significantly decreased and the average grain size was 87 μm and 60 μm, respectively. Some holes were observed in the matrix due to peeling of the precipitates. As for the reason of grain refinement, it was widely accepted that an alloy liquid consisting of a small amount of Ca atoms at the solid/liquid interface could form composition undercooling in the diffusion layer, in which the nucleation particles may be activated in the undercooling region, leading to forming more grain nucleation particles and refining the grain size [35]. Moreover, the Ca atoms, with low speed of diffusion, hindered the grain growth in the interfaces and limited the growth rate of crystal. Thus, the grain size of MZM-xCa alloys clearly refined ( Figure 1d).   It can be seen that the distribution of Mn element was uniform along the matrix while Zn/Ca were gathered in localized areas. In addition, the volume fraction of phases (Mg 2 Ca + Ca 2 Mg 6 Zn 3 ) increased gradually as the Ca content increased, ranging from 0.26% to 1.93%, which was consistent with SEM morphologies. Figure 3 shows the XRD patterns of studied alloys and compared with as-cast state. It can be concluded that the as-cast MZM-xCa alloys ( Figure 3a) were mainly composed of α-Mg, Mg 2 Ca, MgZn, MgZn 2 and Ca 2 Mg 6 Zn 3 , and the homogenization MZM-xCa alloys ( Figure 3b) were composed of α-Mg, Mg 2 Ca and Ca 2 Mg 6 Zn 3 . The reason for the component discrepancy was mainly attributed to Zn/Ca dissolving into Mg substrate after homogenization treatment (Figure 3c). It can be seen that the distribution of Mn element was uniform along the matrix while Zn/Ca were gathered in localized areas. In addition, the volume fraction of phases (Mg2Ca + Ca2Mg6Zn3) increased gradually as the Ca content increased, ranging from 0.26% to 1.93%, which was consistent with SEM morphologies. Figure 3 shows the XRD patterns of studied alloys and compared with as-cast state. It can be concluded that the as-cast MZM-xCa alloys ( Figure 3a) were mainly composed of α-Mg, Mg2Ca, MgZn, MgZn2 and Ca2Mg6Zn3, and the homogenization MZM-xCa alloys ( Figure 3b) were composed of α-Mg, Mg2Ca and Ca2Mg6Zn3. The reason for the component discrepancy was mainly attributed to Zn/Ca dissolving into Mg substrate after homogenization treatment (Figure 3c).  It can be seen that the distribution of Mn element was uniform along the matrix while Zn/Ca were gathered in localized areas. In addition, the volume fraction of phases (Mg2Ca + Ca2Mg6Zn3) increased gradually as the Ca content increased, ranging from 0.26% to 1.93%, which was consistent with SEM morphologies. Figure 3 shows the XRD patterns of studied alloys and compared with as-cast state. It can be concluded that the as-cast MZM-xCa alloys ( Figure 3a) were mainly composed of α-Mg, Mg2Ca, MgZn, MgZn2 and Ca2Mg6Zn3, and the homogenization MZM-xCa alloys ( Figure 3b) were composed of α-Mg, Mg2Ca and Ca2Mg6Zn3. The reason for the component discrepancy was mainly attributed to Zn/Ca dissolving into Mg substrate after homogenization treatment (Figure 3c).   To ascertain the phase structure, TEM observation was further conducted. Figure 4 shows the TEM morphology of MZM-1.10% Ca alloys, and the random distribution of precipitation phases were detected. In addition, Figure 4a-c presents the different morphologies of precipitation phases, such as short-rod shape and granular shape. Figure 4d presents the HRTEM images and selected area diffraction pattern (SADP) of the precipitation (Figure 4c), in which the interplanar space was 0.84 nm. It was consistent with the value of Ca2Mg6Zn3 (100). Moreover, the elemental distribution of precipitates is presented in Figure 4e, which further verified the validity of TEM results. In a similar way, the precipitation of Mg2Ca was also identified, and the details are shown in Figure 4f-h. To ascertain the phase structure, TEM observation was further conducted. Figure 4 shows the TEM morphology of MZM-1.10% Ca alloys, and the random distribution of precipitation phases were detected. In addition, Figure 4a-c presents the different morphologies of precipitation phases, such as short-rod shape and granular shape. Figure 4d presents the HRTEM images and selected area diffraction pattern (SADP) of the precipitation (Figure 4c), in which the interplanar space was 0.84 nm. It was consistent with the value of Ca 2 Mg 6 Zn 3 (100). Moreover, the elemental distribution of precipitates is presented in Figure 4e, which further verified the validity of TEM results. In a similar way, the precipitation of Mg 2 Ca was also identified, and the details are shown in Figure 4f-h.   Figure 5 shows the electrochemical behaviors of homogenization MZM-x Ca alloys as a function of Ca content in the Kokubo solution at 37 • C, and the fitted results are summarized in Table 3. As indicated, as the Ca content increased, the MZM-0.38% Ca alloys shifted to a more positive potential (−1.57 V SCE ) compared to the MZM alloy. However, the sample of MZM-(0.76% Ca, 1.10% Ca) unexpectedly shifted to a negative potential (Figure 5a), in which the values were −1.63 V and −1.68 V, respectively. Based on the electrochemical theory, the E corr of samples depends mainly on the relative magnitude between the anodic and cathode reaction rates, which reflected the reaction tendency [27]. Besides, the corrosion current density (I corr ) was presented in a decreasing order: MZM-1.10 Ca < MZM-0.76 Ca < MZM < MZM-0.38 Ca. Lower current density means better corrosion resistance. Thus, it indicated that the MZM-0.38% Ca sample with 380 • C/24 h treatment possesses the noblest corrosion potential, smallest current density and best corrosion resistance. It was associated with the combined effects of the suitable content of Zn/Ca dissolving into the α-Mg matrix and the modification of Ca-containing compounds by heat-treatment. Thus, the corrosion resistance was improved. However, the excess Ca content present in the matrix could generate intermetallic and form micro-galvanic effect, which accelerated the Mg substrates dissolution.  Figure 5c. As known, the dimension size of the Nyquist plots was the important parameters to reflect the corrosion resistance. Namely, better corrosion behavior of the metal matrix is associated with the higher Z modulus at lower frequency, which is reciprocal to the correlation of corrosion rate for the alloys. In the case of bode phase plot (phase angle vs. frequency), it can be seen that the phase angles    Figure 6 presents the SEM micrographs of corrosion morphology of MZM-xCa in the Kokubo solution at 37 °C for 10 days. As can be seen, the substantial corrosion products were covered on the surface with cracks features, which was due to the dehydration when they removed from the Kokubo solution. In addition, a flat and smooth surface was obtained in sample of MZM, MZM-0.38 Ca and MZM-0.76 Ca. In contrast, a severe corrosion morphology with large size of pitting holes and multidefects features was observed in MZM-1.10 Ca. The composition of the products is summarized in the Figure 6 inset. It can be seen that the corrosion products were primarily composed of Mg, O, P, C and Ca elements. Furthermore, the elements distribution of products was shown in Figure 7. These elements such as P, O and Ca corresponded to each other, which indicated the presence of (Mg, Ca)3(PO4)2 (insoluble) in the products. To further study the composition of corrosion products, XRD and XPS were employed to determine the results, as shown in Figure 8. XRD results indicated the corrosion products were primary composed of Mg, Mg(OH)2 and a small quantity of hydroxyapatite (HA). As shown in Figure 8b, the XPS test identified the corrosion products consist of O − , OH − , CO3 2− ,   Figure 6 presents the SEM micrographs of corrosion morphology of MZM-xCa in the Kokubo solution at 37 • C for 10 days. As can be seen, the substantial corrosion products were covered on the surface with cracks features, which was due to the dehydration when they removed from the Kokubo solution. In addition, a flat and smooth surface was obtained in sample of MZM, MZM-0.38 Ca and MZM-0.76 Ca. In contrast, a severe corrosion morphology with large size of pitting holes and multi-defects features was observed in MZM-1.10 Ca. The composition of the products is summarized in the Figure 6 inset. It can be seen that the corrosion products were primarily composed of Mg, O, P, C and Ca elements. Furthermore, the elements distribution of products was shown in Figure 7. These elements such as P, O and Ca corresponded to each other, which indicated the presence of (Mg, Ca) 3 (PO 4 ) 2 (insoluble) in the products. To further study the composition of corrosion products, XRD and XPS were employed to determine the results, as shown in Figure 8. XRD results indicated the corrosion products were primary composed of Mg, Mg(OH) 2 and a small quantity of hydroxyapatite (HA). As shown in Figure 8b, the XPS test identified the corrosion products consist of O − , OH − , CO 3 2− , PO 3 2− , Mg 2+ and Ca 2+ . Thus, the results of EDS, XPS and XRD have a consistent conclusion, which was also in good agreement with previous studies [6,24,38,45].     Besides, the 3D corrosion morphologies of as-received MZM-xCa alloys after the products removal are shown in Figure 9, which were characterized by laser scanning confocal microscopy (CLSM). As indicated, the experiment alloys surface exhibit the different colors and corrosion profiles, which reflected and represented the different corrosion performance of as-received alloys in a simulated body fluid [45][46][47]. As for the MZM, the surface was slightly corroded and exhibited some shallow pitting holes. The MZM-0.38 Ca alloy revealed a smooth surface, and the fluctuation of profile curves was also relatively slight. With increasing Ca content, the volume fraction of the compounds remarkably increased, which can arise from micro-galvanic corrosion. The large size of pitting holes began to generate and propagate, as shown in Figure 9c,d. Furthermore, the data in Table 4 relate to the corrosion rate Pi. The results from both methods show that addition of 0.38 wt % Ca has the best corrosion resistance. As for the pH measurements (Figure 10), the MZM-0.38 Ca still possesses the low value of pH, indicating the lowest corrosion rate, which was consistent with electrochemical measurements and immersion test. Besides, the 3D corrosion morphologies of as-received MZM-xCa alloys after the products removal are shown in Figure 9, which were characterized by laser scanning confocal microscopy (CLSM). As indicated, the experiment alloys surface exhibit the different colors and corrosion profiles, which reflected and represented the different corrosion performance of as-received alloys in a simulated body fluid [45][46][47]. As for the MZM, the surface was slightly corroded and exhibited some shallow pitting holes. The MZM-0.38 Ca alloy revealed a smooth surface, and the fluctuation of profile curves was also relatively slight. With increasing Ca content, the volume fraction of the compounds remarkably increased, which can arise from micro-galvanic corrosion. The large size of pitting holes began to generate and propagate, as shown in Figure 9c,d. Furthermore, the data in Table 4 relate to the corrosion rate P i . The results from both methods show that addition of 0.38 wt % Ca has the best corrosion resistance. As for the pH measurements (Figure 10), the MZM-0.38 Ca still possesses the low value of pH, indicating the lowest corrosion rate, which was consistent with electrochemical measurements and immersion test. Besides, the 3D corrosion morphologies of as-received MZM-xCa alloys after the products removal are shown in Figure 9, which were characterized by laser scanning confocal microscopy (CLSM). As indicated, the experiment alloys surface exhibit the different colors and corrosion profiles, which reflected and represented the different corrosion performance of as-received alloys in a simulated body fluid [45][46][47]. As for the MZM, the surface was slightly corroded and exhibited some shallow pitting holes. The MZM-0.38 Ca alloy revealed a smooth surface, and the fluctuation of profile curves was also relatively slight. With increasing Ca content, the volume fraction of the compounds remarkably increased, which can arise from micro-galvanic corrosion. The large size of pitting holes began to generate and propagate, as shown in Figure 9c,d. Furthermore, the data in Table 4 relate to the corrosion rate Pi. The results from both methods show that addition of 0.38 wt % Ca has the best corrosion resistance. As for the pH measurements (Figure 10), the MZM-0.38 Ca still possesses the low value of pH, indicating the lowest corrosion rate, which was consistent with electrochemical measurements and immersion test.

Biocompatibility Assessment
Cell viability of BMSCs cultured in 100% extraction medium for one, two and three days is shown in Figure 11. It could be seen that, after one day of cultivation, the cells in extracts showed slightly lower viability than that of the control group (Figure 11a). After two and three days in culture of materials' extracts, the cells from MZM group exhibited still lower viability, but not a significant one, than the negative control group, and the cytotoxicity of these extracts was Grade 0-1 according to ISO 10993-5:1999, indicating that the MZM alloys did not induce toxicity to BMSCs. Nevertheless, as shown in Figure 11b-d, the cells in others extracts exhibited a relatively high viability compared

Biocompatibility Assessment
Cell viability of BMSCs cultured in 100% extraction medium for one, two and three days is shown in Figure 11. It could be seen that, after one day of cultivation, the cells in extracts showed slightly lower viability than that of the control group (Figure 11a). After two and three days in culture of materials' extracts, the cells from MZM group exhibited still lower viability, but not a significant one, than the negative control group, and the cytotoxicity of these extracts was Grade 0-1 according to ISO 10993-5:1999, indicating that the MZM alloys did not induce toxicity to BMSCs. Nevertheless, as shown in Figure 11b-d, the cells in others extracts exhibited a relatively high viability compared

Biocompatibility Assessment
Cell viability of BMSCs cultured in 100% extraction medium for one, two and three days is shown in Figure 11. It could be seen that, after one day of cultivation, the cells in extracts showed slightly lower viability than that of the control group (Figure 11a). After two and three days in culture of materials' extracts, the cells from MZM group exhibited still lower viability, but not a significant one, than the negative control group, and the cytotoxicity of these extracts was Grade 0-1 according to ISO 10993-5:1999, indicating that the MZM alloys did not induce toxicity to BMSCs. Nevertheless, as shown in Figure 11b-d, the cells in others extracts exhibited a relatively high viability compared to the former. It indicated that the extracts can offer a suitable environment for BMSCs proliferation. Among the four specimens, the MZM-0.38% Ca alloy exhibits the highest cell viability, reaching 108% after three-day cultivation. The better cell proliferation could be attributed to the synergetic effects of the release of suitable Mg 2+ , Ca 2+ , and Zn 2+ accompanied by a moderate degradation rate of the MZM-0.38% Ca specimen. In addition, the cells grown in different extracts exhibited a healthy, spindle-shaped morphology, which was similar to that of the negative control group. As a result, the results of cell viability and cell morphology observation (Figures 11 and 12) showed that as-homogenized MZM-xCa alloys did not induce toxicity to cells and met the biocompatibility requirements for implant materials. to the former. It indicated that the extracts can offer a suitable environment for BMSCs proliferation. Among the four specimens, the MZM-0.38% Ca alloy exhibits the highest cell viability, reaching 108% after three-day cultivation. The better cell proliferation could be attributed to the synergetic effects of the release of suitable Mg 2+ , Ca 2+ , and Zn 2+ accompanied by a moderate degradation rate of the MZM-0.38% Ca specimen. In addition, the cells grown in different extracts exhibited a healthy, spindleshaped morphology, which was similar to that of the negative control group. As a result, the results of cell viability and cell morphology observation (Figures 11 and 12) showed that as-homogenized MZM-xCa alloys did not induce toxicity to cells and met the biocompatibility requirements for implant materials.   to the former. It indicated that the extracts can offer a suitable environment for BMSCs proliferation. Among the four specimens, the MZM-0.38% Ca alloy exhibits the highest cell viability, reaching 108% after three-day cultivation. The better cell proliferation could be attributed to the synergetic effects of the release of suitable Mg 2+ , Ca 2+ , and Zn 2+ accompanied by a moderate degradation rate of the MZM-0.38% Ca specimen. In addition, the cells grown in different extracts exhibited a healthy, spindleshaped morphology, which was similar to that of the negative control group. As a result, the results of cell viability and cell morphology observation (Figures 11 and 12) showed that as-homogenized MZM-xCa alloys did not induce toxicity to cells and met the biocompatibility requirements for implant materials.   Aimed to further analyze the corrosion mechanism of MZM-xCa alloys in the Kokubo solution, the SKPFM were used to detect the surface Volta potential distributions. The second phase particles exhibited a brighter color (white color) than the surrounding matrix, indicating the difference in Volta potentials, as shown in Figure 13. In the case of MZM, MZM-0.38 Ca, almost no or trace of precipitatons were detected and exhibited a uniform distribution of Volta potential with relative noble value, corresponding to a better corrosion resistance [45,47]. As for the Figure 13d, the large size of compounds/phases could be observed, which caused the huge difference in Volta potentials between the matrix and phases. Moreover, with the Ca content increased, the large size and volume fractions of phases (Mg 2 Ca and Ca 2 Mg 6 Zn 3 ) were observed (Figure 1, Figure 2, and Figure 13), which gathered along the grain boundaries, and exhibited a dots-chains distribution. Furthermore, the morphologies were transformed from uniform corrosion to localized pitting corrosion with Ca further addition (Figures 6 and 9). It could be explained that the excessive Ca addition can strengthen the nucleation driving force for the formation of precipitations (Mg 2 Ca and Ca 2 Mg 6 Zn 3 ), and the large volumes fraction of micro-galvanic present interface sites accelerate the nucleation driving force for corrosion propagation. Aimed to further analyze the corrosion mechanism of MZM-xCa alloys in the Kokubo solution, the SKPFM were used to detect the surface Volta potential distributions. The second phase particles exhibited a brighter color (white color) than the surrounding matrix, indicating the difference in Volta potentials, as shown in Figure 13. In the case of MZM, MZM-0.38 Ca, almost no or trace of precipitatons were detected and exhibited a uniform distribution of Volta potential with relative noble value, corresponding to a better corrosion resistance [45,47]. As for the Figure 13d, the large size of compounds/phases could be observed, which caused the huge difference in Volta potentials between the matrix and phases. Moreover, with the Ca content increased, the large size and volume fractions of phases (Mg2Ca and Ca2Mg6Zn3) were observed (Figures 1, 2, and 13), which gathered along the grain boundaries, and exhibited a dots-chains distribution. Furthermore, the morphologies were transformed from uniform corrosion to localized pitting corrosion with Ca further addition (Figures 6 and 9). It could be explained that the excessive Ca addition can strengthen the nucleation driving force for the formation of precipitations (Mg2Ca and Ca2Mg6Zn3), and the large volumes fraction of micro-galvanic present interface sites accelerate the nucleation driving force for corrosion propagation. Moreover, the preferential corrosion areas were also observed at the interface positions, as shown in Figure 14. The schematic of corrosion mechanism is shown in Figure 15. When the fresh surface exposed to electrode containing corrosive ions, the Mg would immediately transfer into Mg 2+ and combine with OH − to deposit Mg(OH)2 accompanied by the evolution of hydrogen (Figure 15a, Equation (4)). Prolonged the immersion time, the film layer formation and rupture were under a dynamic balance (Figure 15b, Equations (3) and (4)). Besides, the more OH − ions present in the Kokubo solution convert H2PO4 − and HO4 2− to PO4 3− , and these PO4 3− ions can react with Ca 2+ and Mg 2+ in the electrolyte (Equations (6) and (7)) and result in the formation of calcium phosphate base apatite precipitations according to Equation (8). However, with further extension of the immersion time, the presence of abundant Cl − in Kokubo solution exerts the film layer more active and vulnerable to rupture (Figure 15c, Equation (4)). Besides, the defects, such as micro-cracks and vacancies, can provide diffusion channels for Cl − into film structure to react with MgO and Mg(OH)2, which accelerated the dissolution rate of the Mg substrate and caused the formation of localized corrosion (Figure 15d). Moreover, the preferential corrosion areas were also observed at the interface positions, as shown in Figure 14. The schematic of corrosion mechanism is shown in Figure 15. When the fresh surface exposed to electrode containing corrosive ions, the Mg would immediately transfer into Mg 2+ and combine with OH − to deposit Mg(OH) 2 accompanied by the evolution of hydrogen (Figure 15a, Equation (4)). Prolonged the immersion time, the film layer formation and rupture were under a dynamic balance (Figure 15b, Equations (3) and (4) in the electrolyte (Equations (6) and (7)) and result in the formation of calcium phosphate base apatite precipitations according to Equation (8). However, with further extension of the immersion time, the presence of abundant Cl − in Kokubo solution exerts the film layer more active and vulnerable to rupture (Figure 15c, Equation (4)). Besides, the defects, such as micro-cracks and vacancies, can provide diffusion channels for Cl − into film structure to react with MgO and Mg(OH) 2 , which accelerated the dissolution rate of the Mg substrate and caused the formation of localized corrosion (Figure 15d).     Figure 15. Schematic illustration of corrosion mechanism of as-studied alloys in the Kokubo solution. Figure 15. Schematic illustration of corrosion mechanism of as-studied alloys in the Kokubo solution.

Conclusions
(1) Electrochemical tests and immersion results both showed that MZM-0.38 Ca has the best corrosion resistance, and yields a uniform corrosion morphology on the surface. (2) The improvement of corrosion resistance could be attributed to the combined effects of the suitable content of Zn/Ca dissolving in α-Mg matrix and the modification of Ca-containing compounds by homogenization treatment. (3) With the increase of Ca content, nucleation driving forces for the phases formation were facilitated and the microstructure was refined. However, the large volume fraction of micro-galvanic presence between α-Mg and phase accelerates the micro-galvanic corrosion. (4) In vitro biocompatibility tests show the MZM-xCa alloys were safe to human bone marrow mesenchymal stem cells and were promising to be utilized as implant materials in the future.