Microstructure and mechanical and corrosion properties of hot-extruded Mg-Zn-Ca-(Mn) biodegradable alloys

V.E.Bzhenov ,A.V.Li ,A.A.Komissrov ,A.V.Koltygin ,S.A.Tvolzhnskii ,V.A.Butin ,O.O.Voropev,A.M.Mukhmetshin,A.A.Tokr

a Casting Department,National University of Science and Technology“MISiS”,Leninskiy pr.4,Moscow 119049,Russia

b Laboratory of Hybrid Nanostructured Materials,National University of Science and Technology“MISiS”,Leninskiy pr.4,Moscow 119049,Russia

c Department of Metallurgy Steel,New Production Technologies and Protection of Metals,National University of Science and Technology“MISiS”,Leninskiy pr.4,Moscow 119049,Russia

Abstract Biodegradable Mg-based implants are widely used in clinical applications because they exhibit mechanical properties comparable to those of human bone and require no revision surgery for their removal.Among Mg-based alloys,Mg-Zn-Ca-(Mn) alloys have been extensively investigated for medical applications because the constituent elements of these alloys,Mg,Zn,Ca,and Mn,are present in human tissues as nutrient elements.In this study,we investigated the effect of the hot extrusion temperature on the microstructure,mechanical properties,and biodegradation rate of Mg-Zn-Ca-(Mn) alloys.The results showed that the addition of Mn and a decrease in the extrusion temperature resulted in grain refinemen followed by an increase in the strength and a decrease in the elongation at fracture of the alloys.The alloys showed different mechanical properties along the directions parallel and perpendicular to the extrusion direction.The corrosion test of the alloys in the Hanks’ solution revealed that the addition of Mn significantl reduced the corrosion rate of the alloys.The Mg-2 wt% Zn-0.7 wt% Ca-1 wt% Mn alloy hot-extruded at 300°C with an ultimate tensile strength of 278MPa,an yield strength of 229MPa,an elongation at fracture of 10%,and a corrosion rate of 0.3mm/year was found to be suitable for orthopedic implants.

Keywords: Biodegradable Mg alloy;Mg-Zn-Ca-(Mn);Hot extrusion;Mechanical properties;Corrosion rate.

1.Introduction

Mg alloys are used as load-bearing orthopedic temporary implants because their density and Young’s modulus are similar to those of the cortical bone and they can prevent the stress-shielding effect [1,2].It is known that Mg increases the bone density by regulating the Ca transport [3].During implant degradation,the release of Mg ions can increase the bone strength.In addition,Mg enhances the adhesion of osteoblastic cells and promotes osteogenesis [4-6].These properties of Mg cause bone remodeling around the implant during in-vivo biodegradation tests [7].

Extensive studies have been carried out to develop Mg alloys with both excellent mechanical properties and corrosion performance.Mg-Zn-Ca-(Mn) alloys are promising for biodegradable applications as all their constituents are human tissue nutrients [5,8-10].These alloys show low grade 0-1 cytotoxicity in accordance with ISO 10993-5:1999 standard[8,11,12].Hence,Mg-Zn-Ca-(Mn) alloys exhibit a biosafety level suitable for cellular applications.

As-cast Mg-Zn-Ca-(Mn) alloys exhibit poor mechanical properties[12,13].Heat treatment decreases the corrosion rate of these alloys by partial or full dissolution of the secondary phases and improves their mechanical properties by precipitation hardening.However,the effect of heat treatment on the mechanical properties of Mg-Zn-Ca-(Mn) alloys is not significan [14-17].

Various methods such as hot extrusion,equal channel angular pressing,and hot rolling are used to improve the mechanical properties of alloys [8,18-22].It is well known that the grain size significantl affects the mechanical properties of Mg alloys because of their high Hall-Petch strengthening coefficien (～300 MPa·μm1/2)[9].In contrast,these alloys show a Zn solid-solution hardening of only 20MPa/wt% [23].The effect of deformation processing on the corrosion properties of Mg-Zn-Ca-(Mn) alloys is unclear [8,24].

In Mg-Zn-Ca alloys,an increase in the Zn and Ca contents leads to a moderate decrease in the grain size[13,25-27].As compared to these elements,Mn is more effective in the grain refinemen of Mg-Zn-Ca alloys under both the as-cast and deformation-processed conditions,and hence improves the mechanical properties of the alloys[13-15,28,29].The effect of Mn on the extruded alloy grain size depends on the formation ofα-Mn precipitates that hinder the grain growth during the extrusion process [20].The addition of Mn also improves the corrosion properties of Mg-Zn-Ca alloys.This can be attributed to the formation of a Mn oxide film which acts as a barrier against chloride ion penetration and the formation of intermetallic phases with harmful cathodic impurities like Fe [5,29].

At low extrusion temperatures and speeds,the bimodal structure with fin dynamic recrystallized(DRXed)and coarse textured unDRXed grains with strong basal texture are formed in Mg-Zn-Ca alloys.With an increase in the extrusion temperature and speed,the fully DRXed structure with coarse grains is obtained and the basal texture degrades [6,30,31].Therefore,with an increase in the extrusion temperature and speed,the strength of Mg-Zn-Ca alloys decreases,while the elongation at fracture increases.The addition of Mn shows the same effect on the grain structure and mechanical properties of alloys as that shown by the decrease in the extrusion temperature [20,28].Low-alloyed Mg-Zn-Ca-(Mn) alloys can be extruded at speeds up to 1m/s [30,32],but for alloys with a large fraction of secondary phases,lower extrusion speeds(for example 2mm/s [21]) are required.

Despite the numerous studies carried out on Mg-Zn-Ca-(Mn) alloys,the effects of the constituent elements and extrusion conditions on their properties are not clear yet.The aim of the study was to investigate the effect of the phase composition and extrusion temperature on the mechanical and corrosion properties of Mg-Zn-Ca-(Mn) alloys and to evaluate their potential for application in orthopedic implants.

2.Materials and methods

The polythermal sections of the Mg-Zn-Ca-(Mn) phase diagrams and Mg-Zn-Ca solidus surface projection were calculated according to the calculation of phase diagram(CALPHAD) method using Thermo-Calc software [33].Furthermore,the Scheil-Gulliver solidificatio of the alloys was calculated [34].The thermodynamic database TCMG4(TCS Mg-based Alloys Database) was used [35].

High-purity Mg (99.98 wt% purity),Zn (99.995 wt%),Mn (99.8 wt%),and Ca (99.8 wt%) were used as the rawmaterials for the preparation of the alloys.To minimize the element loss during melting,Mg-14 wt% Ca and Mg-2.7 wt% Mn master alloys were prepared.Before melting,the raw materials were ground using an angle grinder with a cup brush to prevent the contamination of the resulting alloys by oxides.Melting was carried out in a resistance furnace with a graphite crucible under an Ar+2 vol% SF6protective atmosphere.The resulting melt was purged with Ar at 760°C for 3min and was then held at this temperature for 10min before pouring it into the mold.Cylindrical ingots with a diameter of 60mm and a length of 200mm were cast into an aluminum alloy permanent mold preheated to 150°C.The alloys with 2 or 4 wt% Zn,0.7 wt% Ca,and 1 wt% Mn were prepared.The chemical compositions of the prepared alloys were analyzed using energy dispersive X-ray spectroscopy (EDS) on as-cast metallographic sections with an area of 1 mm2and are listed in Table 1.

Table 1 Chemical compositions of the prepared alloys.

In order to relieve the residual stresses and homogenize the alloys,the as-cast ingots were annealed at 350°C for 6h.Then,the parts with shrinkage cavities were cut and the ingots were machined to cylindrical billets with a height of 145mm and a diameter of 50mm.Prior to the hot extrusion process,the billet and die were preheated to 220 or 300°C,respectively.The hot extrusion of the alloys was carried out using the direct extrusion method at a ram speed of 2mm/s and an extrusion ratio of 6.Finally,the cylindrical extruded bars with a diameter of 20mm and a length of～1m were obtained.

The as-cast and hot-extruded alloy specimens were ground and polished for metallographic observations.The microstructural examination of the extruded bars was carried out in the directions parallel and perpendicular to the extrusion direction.For the microstructure and EDS analyses,a Tescan Vega SBH3 scanning electron microscope (SEM) equipped with an EDS system (Oxford Instruments (UK)) was used.The phase volume fractions of the as-cast samples were analyzed by determining the area occupied by each phase in the SEM image using“Tescan”software.The grain sizes of the alloy samples in the as-cast and hot-extruded conditions were determined according to the linear intercept method using optical microscopy (OM) (Carl Zeiss Axio Observer D1m).The grains in the alloy microstructures were examined by etching the samples for 5s using an etchant with the following composition:11g picric acid,11mL acetic acid,and 100mL ethanol.

Fig.1.Tensile test specimens:(a) standard cylindrical;(b) small fla plate.Numbers show the dimensions in mm.

The mechanical properties of the alloys were investigated on a universal testing machine (INSTRON 5569) equipped with an advanced video extensometer.Tensile tests were carried out on two types of specimens:the standard cylindrical specimens obtained by extruded bar lathe machining(Fig.1a) and small fla plate tensile test specimens produced by wire cutting the hot-extruded bars(Fig.1b).Three standard cylindrical specimens were tested for each alloy and extrusion temperature.A video extensometer was used and the tensile properties along only the extrusion direction were investigated.The small fla plate specimens were cut along the extrusion direction and perpendicular to the extrusion direction.The compression test was carried out on 3mm×3mm×6mm small cuboid specimens that were cut along the directions both parallel and perpendicular to the extrusion direction prepared by wire cutting of the extruded bars.The optimum deviation range for the small tensile and compression specimens was obtained by repeating the test on 6-12 specimens.

In-vitro immersion corrosion testing was carried out on disks with a diameter of 18mm and a height of 5mm(surface area=～8 cm2).These disks were prepared by wire cutting the hot-extruded bars.Finally,the disks were ground using 320 grit abrasive SiC paper.For each alloy and extrusion temperature,three disks were cut.During the immersion test,the disk samples were immersed in 400mL Hanks’ solution(PanEco,Russian Federation)at 37°C for 192h.The optimum solution volume to sample surface area ratio of 50mL/cm2was used for the corrosion rate measurements [36,37].The hydrogen evolution capability of the corroding samples was evaluated as a function of time.In addition,1mL of the evolved H2was converted into the weight loss of 1mg of the samples according to Ref.[36,38-39],and the average corrosion rate in mm/year was calculated according to ASTM Standard [40].The variation in the Hanks’ solution pH value was measured using a pH meter (HI83141 Hanna Instruments(USA)).

The electrochemical corrosion measurements of the alloys were carried out using a potentiostat/galvanostat/frequency response analyzer corrosion system (IPC Pro-MF) under the same corrosive medium and conditions as those used for the immersion corrosion test.A three-electrode system with the alloy samples as the working electrodes with an exposure area of 1 cm2was used.Platinum and saturated silver/silver chloride (Ag/AgCl) electrodes were used as the counter and reference electrodes,respectively.Prior to the electrochemical analysis,the alloy samples were immersed in a 0.3 wt%HNO3aqueous solution for 2s followed by rinsing in distilled water.Potentiodynamic polarization experiments were carried out from the cathodic region at -2300mV to the anodic region at -1000mV at a scan rate of 1mV/s.Three curves were obtained for each alloy sample.The corrosion current density and corrosion potential of the alloys were determined by Tafel fitting Using the corrosion current density values,the corrosion rate of the alloys could be calculated [41].

3.Results and discussion

3.1.Phase composition of the Mg-Zn-Ca-(Mn) alloys

Fig.2 shows the polythermal sections of the Mg-Zn-Ca and Mg-Zn-Ca-Mn phase diagrams calculated using Thermo-Calc software.The liquidus and solidus lines are marked in red and blue,respectively.After the solidification the Mg alloys with up to 10 wt% Zn,3 wt% Ca,and 1 wt% Mn showedα-Mg,Mg2Ca,Ca2Mg6Zn3,α-Mn,and MgZn as the main phases.With an increase in the Ca and Zn contents,the liquidus temperature of the alloys remained almost constant;however,the solidus temperature decreased significantl .For example,at the Zn contents>7 wt%,the solidus temperature was 295°C because of the formation of the low-melting MgZn phase.A wide freezing range is undesirable for alloys as it promotes the shrinkage porosity and generates hot tears in the ingot structure during solidification In the case of the alloys with>1.5 wt% Mn,the primary solidificatio of theα-Mn phase occurred,and a further increase in the Mn content resulted in an increase in the liquidus temperature of the alloys,thus increasing the melting and pouring temperatures and melt ignition probability of the alloys.

Fig.3 shows the Mg-rich corner of the Mg-Zn-Ca system solidus surface projection calculated using Thermo-Calc software.Depending on the composition of the alloys,their solidificatio resulted in the formation of the Mg2Ca or Ca2Mg6Zn3phases,or both of these phases.In most of the cases,the experimentally investigated phase composition of alloys is consistent with the calculated solidus surface projection [13,17,20,42].However,the alloys with high Zn contents showed low fractions of binary MgxZnyphases because of non-equilibrium solidificatio [15,43].In addition,it should be noted that the alloys with Zn>0.5 wt% showed theα-Mg,Ca2Mg6Zn3,and Mg2Ca phases [25,27,44].This is not consistent with the CALPHAD calculation results,which indicated that the Ca2Mg6Zn3phase solidifie at the Zn contents higher than 1.5 wt%.The Mg-2 w t% Zn-0.7 wt% Ca and Mg-4 wt% Zn-0.7 wt% Ca alloys with the compositions ofα-Mg+Ca2Mg6Zn3+Mg2Ca andα-Mg+Ca2Mg6Zn3,respectively,were chosen for further investigation because of their moderate secondary phase fractions and short freezing ranges.In the case of the Mg-4 wt% Zn-0.7 wt% Ca alloy,the formation of the MgZn phase through the Scheil-Gulliver solidificatio pathway was also possible.The alloys with 1 wt%Mn,which showed an additionalα-Mn phase,were also analyzed.

Fig.2.Polythermal sections of the Mg-Zn-Ca (a-f) and Mg-Zn-Ca-Mn (g-i) phase diagrams.

3.2.Microstructures of the as-cast Mg-Zn-Ca-(Mn) alloys

Fig.4 shows the microstructures of the as-cast Mg-Zn-Ca-(Mn) alloys.The ternary Mg-Zn-Ca alloy with 2 wt%Zn showed theα-Mg,Mg2Ca,and Ca2Mg6Zn3phases.An increase in the Zn content to 4 wt% resulted in the formation of theα-Mg and Ca2Mg6Zn3phases only.The microstructure of the quaternary alloys with Mn was similar to that of the ternary alloys without Mn,and the only difference was the presence a small fraction ofα-Mn phase precipitates in the former.This indicates that the phase composition of the as-cast alloys was consistent with that calculated using Thermo-Calc software.The grain size of the as-cast alloys is also shown in Fig.4.It can be clearly observed that the grain size of the alloys without Mn was nearly two times larger than that of the alloys with Mn.This Mn-induced grain refinemen can be attributed to the enhanced solute supercooling [14,29,45-47].

Fig.3.Calculated solidus surface projection of the Mg-Zn-Ca system at the Mg corner and phase composition plot.The numbers in red and blue indicate the fractions of the Mg2Ca and Ca2Mg6Zn3 phases,respectively in wt%.

Fig.4.Microstructure of the as-cast Mg-Zn-Ca-(Mn) alloys:(a) MgZn2Ca0.7;(b) MgZn4Ca0.7;(c) MgZn2Ca0.7Mn1;(d) MgZn4Ca0.7Mn1.

The Mg2Ca and Ca2Mg6Zn3phase fractions of the as-cast Mg-Zn-Ca-(Mn) alloys are shown in Fig.5.The Mg2Ca and Ca2Mg6Zn3phase fractions of the MgZn2Ca0.7 alloy were 1.1 and 1.4 vol%,respectively.With the addition of Mn,the Ca2Mg6Zn3phase fraction of the MgZn2Ca0.7 alloy increased slightly to～2 vol%.The alloys with 4 wt%Zn showed 2.8 vol% of the Ca2Mg6Zn3phase irrespective of the presence of Mn.The Ca2Mg6Zn3phase fraction obtained experimentally was consistent with that calculated using Thermo-Calc.However,in the case of the Mg2Ca phase,the amount calculated using Thermo-Calc was lower than that obtained experimentally.This confirm the discrepancy between the calculated surface projection and real alloy phase composition obtained in Ref.[25,27,44],as discussed in Section 3.1.

The elemental composition of theα-Mg solid solution of the as-cast Mg-Zn-Ca-(Mn) alloys obtained via EDS analysis is shown in Fig.6.As expected,the Zn content of theα-Mg solid solution increased with an increase in the Zn content of the alloys.This is because of the high solubility of Zn in Mg.The Mn content ofα-Mg was close to that of the alloys.This suggests that Mn almost completely dissolved inα-Mg and only a small amount of Mn was present in theα-Mn phase.In the MgZn4Ca0.7Mn1 alloy,α-Mg consisted of approximately 3.3 wt% of Zn,Mn,and Ca,which provided the maximum solution strengthening.However,the contribution of solution strengthening on the mechanical properties of magnesium alloys is not significan [9,23].Moreover,the high content of alloying elements,especially Mn,inα-Mg improves the corrosion resistance of alloys [5,29].

3.3.Microstructure of the hot-extruded Mg-Zn-Ca-(Mn)alloys

The SEM microstructures of the Mg-Zn-Ca-(Mn) alloys extruded at 300°C along the direction parallel to the extrusion direction are shown in Fig.7.The alloys extruded at 220°C showed the same microstructure,which is not shown here.In the extruded alloys,the Mg2Ca,Ca2Mg6Zn3,andα-Mn phases were elongated towards the extrusion direction as compared to the as-cast alloys.The difference in the secondary phase fractions of the hot-extruded alloys with 2 and 4 wt% Zn remained the same as that in the as-cast condition.The insets of Fig.7 show the microstructures of the extruded alloys at high magnifications The secondary phases in the hot-extruded alloys were significantl fragmented.In addition,small precipitates of secondary phases could be observed inα-Mg.These precipitates were formed during extrusion via a stress-induced precipitation mechanism.

The OM images of the Mg-Zn-Ca-(Mn) alloys hotextruded at 220 or 300°C along the directions parallel and perpendicular to the extrusion direction are shown in Fig.8.These alloys mainly consisted of fin DRXed grains with a small amount of large unDRXed grains elongated along the extrusion direction.The unDRXed grains were almost absent in the ternary alloys (without Mn) extruded at 300°C.In the case of the quaternary alloys (with Mn),unDRXed grains were observed after extrusion at 220 and 300°C.These results are consistent with those reported previously for extruded Mg-Zn-Ca-(Mn) alloys [6,30,31].

The average grain size of the Mg-Zn-Ca-(Mn) alloys extruded at different temperatures is shown in Fig.9a.At low extrusion temperatures,the alloys showed fin grains.For example,the alloys extruded at 220 and 300°C showed the grain sizes of 3-4.5 and 4.5-7.5μm,respectively.The average grain size of the ternary Mg-Zn-Ca alloy extruded at 220°C was nearly two times smaller than that of the ternary alloy extruded at 300°C.However,in the case of the quaternary Mg-Zn-Ca-Mn alloys,the extrusion temperature showed little effect on the average grain size of the alloys.Consequently,the effect of Mn addition on the grain size of the Mg-Zn-Ca alloys was the same as that of the decrease in the extrusion temperature.

The grain size distribution of the Mg-Zn-Ca-(Mn) alloys extruded at 220 and 300°C is shown in Fig.9b.The obtained grain size distribution data could be best fitte with lognormal distribution.At the extrusion temperature of 220°C,the ternary and quaternary alloys showed the same grain size distribution curve.At the extrusion temperature of 300°C,the ternary alloys showed an increase in the peak grain size and a decrease in the peak frequency.This indicates that in the alloys without Mn,the grain size range expanded with an increase in the extrusion temperature.These results suggest that Mn addition inhibited the grain growth during the recrystallization.In this study,the Mg2Ca,Ca2Mg6Zn3,andα-Mn phases acted as the particle-simulated nucleation sites[20,28].In addition,these phases can dynamically precipitate during extrusion and act as pinning obstacles in the growth of DRXed grains via the Zener drag effect [32].

3.4.Mechanical properties of the hot-extruded Mg-Zn-Ca-(Mn) alloys

The engineering stress-strain curves obtained for the Mg-Zn-Ca-(Mn) alloys after hot extrusion are shown in Fig.10.Large standard cylindrical specimens were used for the tensile test.It should be noted that all the alloys showed the same slope for the elastic portion in the stress-strain curve.

The tensile properties of the cylindrical Mg-Zn-Ca-(Mn)alloy specimens hot-extruded at 220 or 300°C (Fig.1a) are shown in Fig.11a-c.Except for the MgZn4Ca0.7 alloy,all the alloys extruded at 220°C showed high (up to 350MPa)tensile yield strength (TYS) and ultimate tensile strength(UTS),but low elongation at fracture(El)<5%.On the other hand,the alloys extruded at 300°C showed low TYS and UTS (<300MPa),but high El (>5%) as compared to those extruded at 220°C.The MgZn2Ca0.7 alloy hot-extruded at 300°C showed the highest El of 19%.An increase in the Zn content from 2 to 4 wt% showed little effect on the tensile properties of the alloys.The quaternary alloys with Mn showed higher TYS and UTS and lower El than the ternary Mg-Zn-Ca alloys.

The tensile properties of the alloys depended significantl on their grain structure.The quaternary Mg-Zn-Ca-Mn alloys extruded at 220°C showed the minimum average grain size,which contributed to their highest strength.However,the large unDRXed grains present in their microstructure reduced their El [6,30,31].This indicates that the grains had a strong basal texture,and only the basal slip was active during the deformation.On the other hand,the ternary Mg-Zn-Ca alloys extruded at 300°C showed a fully DRXed structure,which increased their elongation at fracture because of the weak basal texture and activation of the non-basal slip systems.However,the large grain size of these alloys reduced their strength.This indicates that the quantity of solute atoms inα-Mg and precipitation strengthening showed lesser effect on the tensile properties of the Mg-Zn-Ca-(Mn) alloys than that shown by the grain structure.These results are consistent with those reported previously [28,31,42,43,48].It has been reported that after deformation processing,alloys with a TYS of 300-350MPa exhibit an El of up to 10-15%,but alloys with a TYS 200-250MPa show an El of up to 20-25% [30].In this study,only the TYS and El values of the MgZn4Ca0.7 alloy were close to the maximum values.

Fig.5.Phase fractions of the as-cast Mg-Zn-Ca-(Mn) alloys.

Fig.6.Zn,Mn,and Ca contents of the α-Mg solid solution of the as-cast Mg-Zn-Ca-(Mn) alloys.

Loads can be applied to bone implants in different directions,promoting the formation of complex stress-strain distribution patterns.In addition,the stress-strain pattern can change during implant degradation.Hence,the anisotropy of implant material properties should be investigated,especially for hot-extruded magnesium alloys.In order to compare the tensile properties of the extruded Mg-Zn-Ca-(Mn) alloys along the extrusion direction and the direction perpendicular to the extrusion direction,their tensile tests were carried out on small fla plate specimens (Fig.1b) and the results are shown in Fig.11d-f.For the alloys extruded at both 220 and 300°C,the tensile properties of the specimens cut perpendicular to the extrusion direction were inferior to those cut along the extrusion direction.The difference between the TYS and UTS values obtained along the directions parallel and perpendicular to the extrusion directions was up to 150MPa.An increase in the extrusion temperature to 300°C reduced the anisotropy of the tensile properties of the alloys because of the formation of a structure with a large amount of DRXed grains.Hence,the anisotropy of mechanical properties should be considered while designing Mg-Zn-Ca-(Mn) alloy bone implants.

Limb implants are generally subjected to compression loading.Hence,it is important to investigate their compressive properties too.The compressive mechanical properties of the small cuboid specimens of the Mg-Zn-Ca-(Mn) alloys extruded at 220 or 300°C are shown in Fig.11g and h.With an increase in the extrusion temperature,the compression yield strength (CYS) and compressive strength (CS) of the alloys decreased.This trend is the same as that shown by the TYS and UTS of the investigated alloys.In addition,the CYS and CS of the quaternary Mg-Zn-Ca-Mn alloys were higher than those of the ternary alloys.It should be noted that the CYS of the ternary Mg-Zn-Ca alloys was nearly the same for the specimens cut parallel and perpendicular to the extrusion direction and no anisotropy was observed.As expected,the CS of the alloys extruded at 220°C (>500MPa) was higher than the UTS for all the alloys.The CS for all the alloys was 2-3 times higher than the maximal CS of cortical bone [49].This reduced the implant volume.Deformed magnesium alloys generally exhibit yield asymmetry (CYS/TYS).The quaternary alloys with Mn showed high yield asymmetry.This is consistent with the results reported previously [28].The lowest yield asymmetry was observed for the MgZn2Ca0.7 alloy(CYS/TYS～0.9) at both the extrusion temperatures.

3.5.Corrosion properties of the hot-extruded Mg-Zn-Ca-(Mn) alloys

Fig.7.Microstructures of the (a) MgZn2Ca0.7 and (b) MgZn4Ca0.7 alloys hot-extruded at 300°C.

Fig.8.Microstructures of the Mg-Zn-Ca-(Mn) alloys along the directions parallel and perpendicular to the extrusion direction after hot extrusion at 220 or 300°C.The yellow arrows show the extrusion direction.

Fig.9.(a) Average grain size and (b) grain size distribution of the Mg-Zn-Ca-(Mn) alloys hot-extruded at 220 or 300°C.

Fig.10.Engineering stress-strain curves obtained after the tensile test of the Mg-Zn-Ca-(Mn) alloys extruded at 220 or 300°C.

The corrosion rate of biodegradable alloys is important because fast degradation causes callus formation,impedes good connectivity of osteocytes with the implant surface,and promotes quick degradation of implant integrity [50].An implant should maintain its mechanical integrity during the 12-18-week period for partial or full replacement by natural tissue[2].At low corrosion rates the release of hydrogen is unproblematic for human tissues [24,50].The amount of hydrogen released during the immersion corrosion test of the Mg-Zn-Ca-(Mn) alloys extruded at 220 or 300°C in the Hanks’ solution is shown in Fig.12.It can be observed that the hydrogen released over time was not constant and increased after 100h of immersion.The possible reason of this behavior is the growth of a protective fil at the beginning of immersion and the formation of cracks on the fil surface because of an increase in stress with an increase in the fil thickness or fil spallation.Low corrosion rates of implant materials during the initial period after implantation are beneficia for medical applications.This is because during the firs week after the surgery the hydrogen evolution must be as low as possible to avoid inflammation Kraus et al.showed that the acceptable daily dosage of H2that can be carried away by rat tissues is 0.13mL for an implant with a surface area of 0.44 cm2[50].For an implant with a surface area of 1 cm2,the hydrogen volume that can be accepted by tissues is 0.295mL/cm2.The acceptable daily quantity of hydrogen that can be carried by rat tissues is shown in Fig.12 by the dashed line.All the investigated alloys,except MgZn4Ca0.7 released hydrogen in the acceptable range.The amounts of hydrogen released by the MgZn2Ca0.7 and MgZn4Ca0.7Mn1 alloys extruded at 220°C were close to the daily limit and could exceed it with an increase in the immersion duration.The maximum probability of inflammatio is observed in the early postoperative period;therefore,a slight excess in the content of released hydrogen one week after the surgery is acceptable.

Using the amount of hydrogen released (from Fig.12),the average corrosion rate of the alloys was calculated according to the standard practice [40].The corrosion rates of the alloys hot-extruded at 220 and 300°C are shown in Fig.13.The highest corrosion rates of 0.97 and 1.18mm/year were observed for the MgZn4Ca0.7 alloys extruded at 220 and 300°C,respectively.In contrast,the MgZn2Ca0.7Mn1 alloys extruded at 220 and 300°C showed the lowest corrosion rates(0.20 and 0.30mm/year,respectively).Considering the confidenc limit,no significan difference was observed in the corrosion rates of the alloys extruded at 220 or 300°C.This can be attributed to the slight difference in the grain sizes and phase compositions of the alloys.

On the basis of the results discussed thus far,it can be stated that Mn decreases the corrosion rate of alloys,but Zn increases it.With respect toα-Mg,the Mg2Ca phase acts as the anode,while the Ca2Mg6Zn3phase acts as the cathode[51-53].Thus,Mg2Ca dissolves preferentially in corrosion media as compared toα-Mg,but the Ca2Mg6Zn3phase temporarily blocks local corrosion [44].It was difficul to compare the corrosion rate of the alloys prepared in this study with those of the alloys reported previously because of their different impurity contents[54].However,it has been reported that among Mg-Zn-Ca-(Mn) alloys,alloys with～2 wt% Zn show the lowest corrosion rate [7,12,13,25-27,44].As discussed earlier,the microstructure of such alloys consist of both the Mg2Ca and Ca2Mg6Zn3phases.A possible explanation for this phenomenon was given by Zhang and Yang[25].The Mg2Ca phase dissolves during corrosion and the Ca content in the corrosion medium increases,accelerating the formation of the Ca-containing protection film

Fig.11.Mechanical properties of the Mg-Zn-Ca-(Mn) alloys evaluated for (a-c) large standard cylindrical tensile test specimens (Fig.1a),(d-f) small fla plate tensile test specimens cut along the directions parallel and perpendicular to the extrusion axis (Fig.1b) and (g-h) small compression test specimens cut along the directions parallel and perpendicular to the extrusion axis.Property:(a,d) TYS;(b,e) UTS;(c,f) El;(g) CYS;(h) CS.

Fig.12.Hydrogen evolution during 192h of immersion in the Hanks’ solution at 37°C for the Mg-Zn-Ca-(Mn) alloys extruded at:(a) 220°C;(b) 300°C.

Fig.13.Corrosion rates of the Mg-Zn-Ca-(Mn) alloys extruded at 220 and 300°C obtained by immersing the alloys in the Hanks’ solution for 192h at 37°C and by electrochemical corrosion testing at the same conditions.

The pH of the corrosive medium was measured during the immersion corrosion test (Fig.14).The pH of the Hanks’solution before the corrosion test was 7.4.The pH of the Hank’s solution increased to～9 at the end of the corrosion test of the ternary MgZn2Ca0.7 and MgZn4Ca0.7 alloys extruded at 220 and 300°C.This increase in the pH of the solution can be attributed to the release of (OH)-hydroxide ions into the corrosion medium [36].During the corrosion of quaternary alloys,the pH of the corrosive medium did not change in comparison with ternary alloys.For example,when the MgZn4Ca0.7Mn1 alloy extruded at 220°C was immersed in a corrosive medium,the pH of the solution changed in the range of 7.9-7.1.It is well-known that the corrosion rate of an alloy is related to the change in the pH of the corrosive medium [11,13,26,37,51].In this study,an increase in the corrosion medium pH was observed during the corrosion of the ternary Mg-Zn-Ca alloys with high corrosion rates.In the quaternary alloys with low corrosion rates the pH of the corrosive medium was close to the initial pH of the Hanks’solution.The slight decrease in the pH of the corrosive media observed during the corrosion testing of alloys with Mn is probably due to an increase in the self-dissolution of the alloys as a result of the negative difference effect induced by the presence of a thick fil of Mg-based corrosion products.Some researchers explain the negative differential effect by the influenc of monovalent magnesium ions [55-59].In any case,the presence of additional (divalent) magnesium ions near the metal-electrolyte interface with the current fl w should lower the pH of the corrosion medium,thus increasing the rate of self-dissolution.The decrease in the pH of the corrosion medium can be explained by the fact that the productshould be kept constant in the presence of solid Mg(OH)2.A high alkaline pH causes a negative effect on the healing of muscle and bone tissues,leading to hemolysis [37,60].On the other hand,the near neutral pH of the corrosion media for the quaternary Mg-Zn-Ca-Mn alloys may result in improved cell viability and proliferation [61].

The polarization curves of the extruded Mg-Zn-Ca-(Mn)alloys immersed in the Hanks’ solution at 37°C are shown in Fig.15.The polarization curves for the same alloys extruded at different temperatures were almost the same.The ternary Mg-Zn-Ca alloys showed the same corrosion potential of approximately -1.50V.The corrosion potential (-1.44V) of the quaternary alloys with Mn was more positive than that of the ternary alloys.Considering the ±0.02V error bars,it can be stated that the Zn content did not affect the corrosion potential of the alloys.It has been reported that an increase in the Zn content of Mg-Zn-Ca alloys causes a negative shift in their corrosion potential [13,27,44].Nevertheless,the addition of Mn to the Mg-Zn-Ca alloys caused a positive shift of+0.06V in their corrosion potential.This is consistent with the results reported previously [29].The addition of Mn not only resulted in an increase in cathodic reaction kinetics,but also retarded the anodic kinetics over a wide potential range.In this case,the anodic reaction,representing the dissolution of Mg,was kinetically easier for the ternary Mg-Zn-Ca alloys than for the quaternary Mg-Zn-Ca-(Mn) alloys.As a result,the ternary alloys showed higher corrosion rate than the quaternary alloys.

Fig.14.Change in the pH of the Hank’s solution after the immersion of the Mg-Zn-Ca-(Mn) alloys extruded at:(a) 220 and (b) 300°C for 192h at 37°C.

Fig.15.Polarization curves of the Mg-Zn-Ca-(Mn) alloys extruded at:(a) 220°C;(b) 300°C obtained during immersion in the Hanks’ solution at 37°C.

The corrosion rates of the Mg-Zn-Ca-(Mn) alloys calculated using their corrosion current densities obtained via electrochemical tests are shown in Fig.13.The corrosion rates obtained from the electrochemical tests were 1.5-7 times higher than those obtained from the immersion corrosion test.Prior to the electrochemical tests,the surface of the alloys was activated using a HNO3aqueous solution.The higher electrochemical corrosion rates can be attributed to the absence of the corrosion product fil in the electrochemical corrosion test samples.In the case of the immersion corrosion test,the corrosion process was kinetically weakened by the shielding effect of the corrosion products on a significan part of the sample surface.On the basis of the electrochemical corrosion test results,it can be stated that the extrusion temperature showed no effect on the corrosion rate of the alloys.Moreover,the addition of Mn decreased the corrosion rate of the Mg-Zn-Ca alloys.This is consistent with the immersion corrosion test results.The ternary and quaternary alloys with 4 wt% Zn showed close corrosion rates than their counterparts with 2 wt% Zn.The electrochemical corrosion test is a kind of express test and represents only a snapshot of the alloy corrosion behavior at the moment when it is performed.In addition,the corrosion rates obtained from the electrochemical test are affected by the negative difference effect [36,50,62].Therefore,we carried out immersion tests to evaluate the biodegradation rate of the investigated alloys.

Fig.16.(a,b,c) Microstructure and (d,e,f) EDS profile of alloy/corrosion product layer boundary of the (a,b,d,e) MgZn2Ca0.7Mn1 and (c,f) MgZn4Ca0.7 alloys hot extruded at 220°C after immersion in the Hanks’ solution at 37°C.

The Hanks’ solution contains most of the physiological inorganic salts at concentrations similar to those in the human body [36].This indicates that the corrosion products formed on the surface of the alloys during the in-vitro immersion corrosion test were almost the same as those formed under the in-vivo conditions.Fig.16a,16b,16d,and 16e shows the microstructures and EDS profile of the corrosion product layers formed on the MgZn2Ca0.7Mn1 alloy extruded at 220°C (with the lowest corrosion rate,0.2mm/year) after immersion in the Hanks’ solution at 37°C for 192h.Most of the sample surface was coated with a 2.5-μm thick layer mainly composed of O (64 at%) and 12,12,and 11 at% of Mg,Ca,and P,respectively (Figs.16a and 16d).After the corrosion of the Mg-Zn-Ca-(Mn)alloys in salt bath solutions,hydroxyapatite Ca10(PO4)6(OH)2[13,24,26,51],Mg3Ca3(PO4)4[37,63],Mg apatite [64] or other phosphates such as Ca4O(PO4)2,CaP3(OH)·2H2O,and Ca3(PO4)2[65] were formed.The X-ray diffraction patterns of the corrosion products showed closely spaced hydroxyapatite and calcium phosphate peaks.Hence,it was difficul to properly investigate the phase compositions of the corrosion products [12].The obtained Ca/P ratio(～1)better fitte to the passivation Mg3Ca3(PO4)4layer,instead of hydroxyapatite [37,63].Witte et al.showed that the Mg-containing calcium phosphate layer on the surface of an implant contributes to high osteoblastic activity [66].The areas shown in Fig.16b and e were also observed on the surface of the MgZn2Ca0.7Mn1 alloy sample but in small fractions.On the basis of the contrast of the backscattered electrons,three layers could be identified The outer layer was the passivation Mg3Ca3(PO4)4phase formed initially.The inner layer close to the alloy surface consisted of 64 at%O and 33 at% Mg (as revealed by the EDS results) and was identifie to be the Mg(OH)2phase.It is known that the Mg(OH)2phase is soluble,nontoxic,enhances osteoblast activity,and decreases the number of osteoclasts during bone remodeling [11,67].Another layer was observed between the Mg3Ca3(PO4)4and Mg(OH)2layers.The composition of this layer was intermediate to those of the Mg3Ca3(PO4)4and Mg(OH)2layers.

Fig.16c and f shows the microstructure and EDS profil of the corrosion product layer formed after the immersion of the MgZn4Ca0.7 alloy extruded at 220°C (with a high corrosion rate of 0.97mm/year) for 192h in the Hanks’ solution at 37°C.Most of the sample surface was covered with deep corrosion cavities formed between the Ca2Mg6Zn3secondary phase,which acted as the cathode with respect toα-Mg.The EDS results revealed that the corrosion product layer mainly consisted of the Mg(OH)2phase with a few Mg3Ca3(PO4)4outer layer fragments.The absence of the Mg3Ca3(PO4)4layer can be attributed to its spallation during the growth of the thick Mg(OH)2layer.During the corrosion of Mg alloys in salt bath solutions,first the Mg(OH)2layer is formed,upon which a phosphate protective layer is formed [24,27].In this study,the Mg3Ca3(PO4)4phase protective layer was formed first

The analysis of the polarization curves confirme that the corrosion behavior of the Mg-Zn-Ca-(Mn) alloys was affected more by their corrosion product fil morphology and composition than the alloy phase composition.For example,the MgZn4Ca0.7 and MgZn4Ca0.7Mn1 alloys showed the same volume fraction of the Ca2Mg6Zn3phase (2.8 vol%),which acted as the cathodic structural component.However,the corrosion rates of these alloys were significantl different.In the MgZn2Ca0.7Mn1 and MgZn4Ca0.7Mn1 alloys,Mn promoted the formation of protective film in the form of conversion coating by interacting with the phosphates in the Hanks’ solution,which significantl inhibited the anodic process of electrochemical corrosion.This was confirme by the positive shift of the corrosion potential in the polarization curves of the alloys (Fig.15).The inhibition of the anodic process occurred either via the anodic electrochemical reaction Mg→Mg2++2e-(overvoltage) or by the diffusion of Mg2+cations into the solution (concentration polarization).The exact controlling process could not be identifie in this study and needs further investigation.The decrease in the penetration depth of chloride ions confirme that the addition of Mn improved the stability of the surface corrosion product fil by forming MnO2and MnO oxides [29].The possible reason why Mn was not detected in the EDS profile in this work is the low thickness of the Mn oxide layer.

4.Conclusions

On the basis of the CALPHAD calculation results,four Mg-Zn-Ca-(Mn) alloys were prepared.The mechanical and corrosion properties of the hot-extruded alloys were investigated.The following results were obtained:

(1) The as-cast Mg-2 wt% Zn-0.7 wt% Ca alloy consisted of theα-Mg,Mg2Ca,and Ca2Mg6Zn3phases.With an increase in the Zn content to 4 wt%,the Mg2Ca phase disappeared.The addition of 1 wt% Mn to this alloy did not cause significan changes in its microstructure.This is because most of the Mn was present in theα-Mg phase and very fewα-Mn phase precipitates were observed.The addition of Mn also promoted grain refinemen in the as-cast condition.

(2) The hot extrusion of the alloys resulted in the fragmentation of the Mg2Ca and Ca2Mg6Zn3phases.The alloys extruded at 220°C showed a bimodal macrostructure with both fin DRXed and coarse elongated unDRXed grains.With an increase in the extrusion temperature to 300°C,the proportion of the DRXed grains and their mean size increased.The addition of Mn to the Mg-Zn-Ca alloys decreased the size of the DRXed grains and increased the number of the elongated unDRXed grains present in the alloys.

(3) The decrease in the extrusion temperature and the addition of Mn to the Mg-Zn-Ca alloys increased the TYS,UTS,CYS,and CS of the alloys,while reducing their El.The tensile and compressive mechanical properties of the alloys along the extrusion direction were higher than those in the direction perpendicular to the extrusion direction.However,the difference was not significant

(4) The corrosion rate of the alloys in the Hanks’ solution was independent of the extrusion temperature.However,the addition of Mn decreased their corrosion rate.In the case of the quaternary alloys with Mn,the pH did not change during the immersion corrosion test.The alloys with 2 wt% Zn showed lower corrosion rates than the alloys with 4 wt% Zn.

(5) The Mg-2 wt% Zn-0.7 wt% Ca-1 wt% Mn alloy extruded at 300°C was found to be suitable for application in bone implants because of its excellent mechanical properties(UTS=278MPa;YS=229MPa;El=10%)and biocorrosion rate (0.3mm/year) [9,50].

Declaration of Competing Interest

None.

Acknowledgments

The authors gratefully acknowledge the financia support from the Ministry of Science and Higher Education of the Russian Federation in the framework of Increase Competitiveness Program of NUST ?MISiS? (№К2-2019-008) implemented by a governmental decree dated 16th of March 2013,N 211.