Investigation of the degradation rate of electron beam processed and friction stir processed biocompatible ZKX50 magnesium alloy

Fatemeh Iranshahi,Mohammad Bagher Nasiri,Fernando Gustavo Warchomicka,Christof Sommitsch

Joining and Forming,Institute of Materials Science,Graz University of Technology,Kopernikusgasse 24/I,Graz 8010,Austria

Abstract Together with the mechanical properties,the degradation rate is an important factor for biodegradable implants.The ZKX50 Mg alloy is a suitable candidate to be used as a biodegradable implant due to its favorable biocompatibility and mechanical properties.Current research investigates the degradation rate and corrosion behavior of the ZKX50 as a function of the microstructure constituents and their morphology.Since grain refinemen is the main strengthening mechanism for the ZKX50,the effect of the microstructure refinemen on the corrosion rate was studied by applying electron beam processing(EBP)and friction stir processing(FSP)on the ZKX50 cast alloy.To study the effect of the microstructure constituents and their morphology a subsequent solution heat treatment (HT) was applied to the processed samples.The results show that the EBP and FSP lead to a uniform and remarkably refine microstructure of the ZKX50 alloy and homogeneous distribution of the intermetallic phases.The results of electrochemical corrosion tests together with the microstructure characterization show that microgalvanic corrosion is the predominant mechanism that occurs between the Ca2Mg6Zn3 intermetallic phase and α-Mg matrix.According to the results attained through the electrochemical tests,the EBPed-HT ZKX50 alloy shows higher corrosion resistance compared to all other conditions immersed in 0.5 wt.% NaCl solution.The dissolution and spheroidizing of Ca2Mg6Zn3 particles during the solution heat treatment provides higher corrosion resistance mainly by decreasing the microgalvanic corrosion.The microstructure of the heat-treated samples does not show a significan grain coarsening which can degrade the enhancement of the mechanical properties achieved by the EBP and FSP.

Keywords: Magnesium alloys;Electron beam processing;Friction stir processing;Corrosion;Biodegradable metals.

1.Introduction

Magnesium is essential mineral nutrient for human health which plays important role in the strength of the bones,muscle function,and nervous system.Around 50% of magnesium in the body is found in the bones [1].The combination of high mechanical strength and fracture toughness,low density,similar elastic modulus to that of human bones,and great biocompatibility make magnesium and its alloys interesting candidates for the application of biodegradable implants.Due to the degradation characteristic of magnesium alloys,there is no need for implant removal surgery after the bone has healed,which reduces both further suffering of the patient and medical costs [2].

The selection of the alloying elements in the Mg alloys mainly aims at improving the mechanical properties for broad application as in the automotive and aviation industries [3-6].Nevertheless,alloying element selection for bio-application of Mg alloy should be based not only on the improvement of mechanical properties but also on the biocompatibility of those elements.Zinc is one of the most common alloying elements in magnesium alloys and one which does not show local or general toxicity[7,8].Calcium with excellent biocompatibility is an ideal element to add to the Mg-Zn system for bioapplication purposes.Moreover,calcium is a major component of human bone and can accelerate bone growth.Several research studies on Mg-Zn-Ca(ZX)ternary alloys demonstrated good mechanical properties of these alloys as well [9-11].Wang et.al.reported that the cooperative action of Zn and Ca solutes improves the ductility of the Mg-1.8Zn-0.2Ca alloy during the deformation [12].Beside Zn and Ca,Zr is also a nontoxic powerful grain-refinin agent for magnesium alloys.An appropriate amount of Zr also helps enhancement of the mechanical properties and improvement of corrosion resistance [13,14].Thus,the newly developed ZKX50 alloy in the Mg-Zn-Zr-Ca system is a proper candidate for bioapplication mainly due to the biocompatibility of its alloying elements [15],together with its high strength and ductility[16].

The solid solution,precipitation strengthening,and grain refinemen are the effective mechanisms to enhance the mechanical properties of Mg-Zn-Ca alloys [11,15].H?nzi et al.reported a significan increment of the yield strength and ductility of the ZKX50 alloy after extruding the as-cast material [16].Vargas et al.examined the feasibility of the further enhancement of the mechanical properties through severe plastic deformation (SPD) using friction stir processing (FSP)[17].Li et al.assessed the effects of Zn in the mechanical properties of the Mg-Zn-Ca alloys through grain refinement solid solution strengthening,and precipitation hardening [10].Moreover,the formation of the Mg-Zn precipitate in theα-Mg matrix deaccelerates the corrosion rate acting as an anode in the composed microgalvanic pill with the matrix [18].However,Zn as an alloying element can adversely affect the corrosion behavior of Mg-Zn-Ca alloys due to the microgalvanic effect of Ca2Mg6Zn3intermetallic phase (IMP) which acts as cathode compared to the matrix [10].Moreover,the microstructure refinemen for improving the mechanical properties can also affect the corrosion rate of the Mg alloy in both negative and positive directions [19].Despite the demands for corrosion resistance material for most of the applications,the corrosion and degradation rate of the Mg alloys make them favorable for use as biodegradable implants [20].The high degradation rate of magnesium alloys can be also an issue for the bioapplication of Mg alloys in the physiological environment.The hydrogen gas produced during the corrosion of Mg alloys can create gas pockets around the implant,which cause tissue irritation and decreases the mechanical integrity of the implant [21].Therefore,understanding the corrosion mechanism,the effects of different microstructure constituents,and their morphology on the corrosion rate are crucial to control the degradation rate for developing biodegradable implants.Accordingly,current research aims at the assessment of the corrosion behavior of the ZKX50 to control the degradation rate,understanding the corrosion mechanisms,and also preserving the balance between the corrosion rate and mechanical enhancement.

The characteristics of the microstructure,such as grain size,morphology,and size of the IMPs and their distribution and chemical segregation in the microstructure can play important role in the corrosion behavior of magnesium alloys[22].Up to now,different approaches such as protective coatings [23,24],developing new alloys [25],and surface modifi cation have been widely investigated to improve the corrosion resistance of magnesium alloys [26].Moreover,modificatio of the microstructure in terms of controlling the grain size,the volume fraction,size,and distribution of the secondary phases can be also a strategy to control the corrosion rate of magnesium alloys [27].The microstructure refinemen aimed at improving the mechanical properties of the Mg alloy can also affect the corrosion behavior of magnesium alloys.In addition to SPD techniques for grain refinement a fast solidificatio process also leads to microstructure refinemen[28].Rapid and fast solidificatio can be obtained by various techniques including laser surface melting [29],centrifugal atomization,splat quenching [30],electron beam melting[31].Investigation of the effect of laser melting on corrosion resistance of Mg alloys AZ31,AZ61,and WE43 shows the significantl improved corrosion resistance after treatment due to the grain refinemen and uniform redistribution of the intermetallic phases [32].Gao et al.also reported the enhancement of the corrosion resistance of AZ91HP after laser melting processing [33].Electron beam processing (EBP) as a newly developed technique for surface modificatio possesses superior advantages over laser and ion beams owing to its high efficien y,simplicity,and reliability [34].In this technique,a high-velocity electron beam collides with the material and melts it.By moving the electron beam forward,the melted and evaporated alloy fl ws from the front to the back of the keyhole.In this process,the achievement of a refine microstructure can generally be expected due to the high cooling rate during solidificatio [35].As reported by Liu et al.,the rapid solidificatio of the high current pulsed electron beam (HCPEB) treated Mg-4Sm resulted in better corrosion resistance due to the homogenous microstructure and chemical composition [36].Additionally,our previous study on corrosion behavior of electron beam processed AZ91 shows improvement of corrosion-resistant due to the formation of the fin microstructure with a connected network of supersaturatedα-eutectic and homogeneous distribution of eutecticβ-Mg17Al12in the microstructure [19].

Friction stir processing (FSP) is a solid-state technique for microstructure modificatio in materials.In this process,a cylindrical wear resistance tool consists of a shoulder and a pin that rotates and moves toward the surface of the material until the pin penetrates the material and the rotating shoulder is in contact with the surface of the workpiece.The friction generated between the rotating shoulder and the workpiece leads to localized heating which softens the material and causes material movement around the pin.Severe plastic deformation at elevated temperatures results in significan grain refinement the break-up of secondary phases,and homogeneity of the processed zone [37-39].

The present work aims to investigate the effect of the microstructure refinemen process on the corrosion behavior of biocompatible ZKX50 Mg-alloy by electron beam processing (EBP) and friction stir processing (FSP).To understand the corrosion mechanism in the ZKX50 alloy,the role of the microstructure constituents and their morphology in the corrosion process are studied through electrochemical corrosion tests applied on the as-cast,friction stir processed (FSPed),electron beam processed(EBPed),and heat-treated(HT)samples of the ZKX50 alloy.

2.Experimental methods

2.1.Material

The as-received material in this study was ZKX50 cast alloy,with the nominal composition (in wt.%) of Mg-5.2%Zn-0.2%Zr-0.13%Ca-0.09%Mn.The material was produced by direct chill casting and machined in 6 mm thick plates.The alloy used was selected based on the alloys developed by H?nzi et al.,which represents high strength and simultaneously high ductility [16].

2.2.Electron beam processing

Electron beam processing (EBP) using a Probeam EBG 45-150 K14 electron beam welding machine was applied on the 6 mm thick plates of as-cast ZKX50 alloy.A series of preliminary tests were implemented to attain the optimum parameters of the EBP.A defect-free molten zone with desired penetration depth suitable for corrosion test and acceptable surface quality were the main criteria to select the suitable parameters of the EBP as discussed in our previous work [19].The optimized parameters of the EBP are listed in Table 1.

Table 2 EDS point analysis of different phases (value of the elements is given in wt.%).

2.3.Friction stir processing

Friction stir processing (FSP) using MTS I-Stir BR4 friction stir welding machine has been successfully performed on the 6 mm thick plates of as-cast ZKX50 alloy to refin the microstructure.The used FSP was carried out with a rotational speed of 500 rpm,a travel speed of 200 mm/min,and a tilt-angle of the tool of 3°.The tool used for FSP was a cylindrical-shaped wear-resistant H13 steel consist of a shoulder with a diameter of 18 mm,a helically grooved pin with a diameter of 10.44 mm,and a pin length of 3.3 mm.The process was with constant force.After the process,the FSPed samples were cut from the center of the stir zone of FSPed alloy for microstructure characterization and corrosion tests.

2.4.Heat treatment

To evaluate the effect of the IMPs on the corrosion rate of the ZKX50 alloy by decreasing their volume fraction in the microstructure,a solution heat treatment at 340 °C for 24 h was planned to apply on the EBPed and FSPed samples followed by water quenching.To avoid the grain coarsening,the 340 °C as the temperature of the heat treatment was selected close to the solidus temperature (340 °C) of IMP as shown by the ternary phase diagram of the Mg-Zn-Ca [16].

2.5.Microstructural characterization

To investigate the microstructure of the alloy in different conditions,an optional surface of as-cast alloy and the crosssection of the processed and heat-treated samples were embedded in epoxy resin.The surface of interest of samples was ground and polished up to 0.04 μm.The polished samples were etched by an acetic-picral solution to characterize their microstructure by light optical microscopy (LOM) studies using Zeiss observer Zm1 optical microscope.In addition to the LOM,a scanning electron microscope equipped with the fiel emission gun (FE SEM) TESCAN MIRA3 XMU,and the EDAX EDX detector Super Octane were used to reveal more details on the microstructure of as-received and processed samples.The average grain size was determined in LOM images,using the linear intercept method according to the ASTM E112-12[40].The average volume fraction of intermetallic phases was calculated with image analyzing ImageJ.

2.6.Corrosion tests

The corrosion behavior of as-cast,EBPed,EBPed-HT,FSPed,and FSPed-HT samples was evaluated by electrochemical corrosion analysis,including open circuit potential (OCP),potentiodynamic polarization (PDP),and electrochemical impedance spectroscopy (EIS).An Autolab (PGSTAT128N) potentiostat/galvanostat was utilized for electrochemical corrosion tests.These measurements were performed at 22 ± 0.5 °C in a 0.5 wt.% NaCl electrolyte using a common three-electrode corrosion cell including a counter electrode (CE) made of platinum foil,a reference electrode(RE) made of saturated Ag/AgCl,and the specimen as a working electrode (WE) (with a surface ground up to P4000 SiC papers).The obtained results from the electrochemical corrosion tests were analyzed using NOVA1.11 software.

The potentiodynamic polarization tests were carried out with a scan rate of 1 mV/s in the potential range of±130 mV.The EIS analysis was performed in the frequency range of 0.01 Hz-100 kHz using the amplitude of the sinusoidal potential signal of 10 mV for immersion time intervals from 1 h to 1 week with respect to the OCP.

Fig.1.Macroscopic pictures of (a) the cross-section of the EBPed,(b) the surface appearance of the EBPed (c) the cross-section of the FSPed,and (d) the surface appearance of the FSPed ZKX50 alloy (BM is the base metal).

To evaluate the biodegradation rate of the ZKX50 alloy,a series of hydrogen evolution tests using a lab-made setup was also implemented in Hanks’ balanced salt solution (produced by Carl ROTH,Germany) under the physiological temperature of 37 ± 1 °C and pH of 7.4 ± 0.1.In the hydrogen evolution test,the ZKX50 alloy samples using the fishin line were hanging in the glass funnels connected to the burettes to collect the released hydrogen during the degradation of the samples.For each test,three samples with the same condition,funnels,and burettes plus one extra funnel connected to a burette without a sample as a reference were placed in a beaker fille with Hanks’ solution on a heater-stirrer plate.All the burettes were fille with the same solution to a specifi level.To avoid rapid evaporation of the solution during the immersion time the surface of the solution was covered using small plastic balls.Based on the degradation rate of each series of the samples the solution was renewed.The pH value was also constantly adjusted by flushin the solution with CO2.All samples were ground up to P4000 SiC papers,washed with ethanol,dried in hot air,and weighted before each test.After one week of immersion,the samples were removed from the solution,carefully cleaned with chromic acid,and reweighted.The hydrogen collected during the immersion time is used to give the corrosion rate over time and the difference in the weight of samples before and after the immersion indicates the mass loss.

The corrosion weight loss of the studied samples immersed in 0.5 wt.% NaCl and Hank’s solution was measured using a digital balance with the precision of 0.001 mg.

2.7.Microhardness test

To explore the homogeneity of the microstructure and analysis the mechanical strength of the processed and heat-treated samples,microhardness tests were carried out by applying an indentation load of 0.1 kg for 15 s using the Vickers Microhardness testing machine.The microhardness of samples was measured at least 20 times in several points with a minimum distance of 0.3 mm from each other [41].

3.Results and discussion

3.1.Macroscopic analysis of the processed samples

The macroscopic view of the cross-section and the surface of EBPed and FSPed ZKX50 alloy are presented in Fig.1.A complete penetration at the root with about 5 mm depth free of cracks and pores provoked by a fast cooling rate during solidificatio is observed in the cross-sectional overview of the EBPed alloy(Fig.1(a)).The surface of the EB processed area shows an acceptable appearance (Fig.1(b)) for further experiment preparation.The cross-sectional view of the FSPed samples in Fig.1(c) shows the basin-shaped nugget of the alloy after FSP,while Fig.1(d) represents the good quality appearance of the surface of the FSPed samples.

3.2.Microstructure characterization

Fig.2.LOM images of (a) as-cast,(b) EBPed,(c) FSPed (d) EBPed-HT and (e) FSPed-HT ZKX50.

Fig.2 shows the light optical microscopic (LOM) microstructure of the as-cast,EBPed,FSPed,EBPed-HT,and FSPed-HT samples.The microstructure of as-cast in Fig.2(a) showsα-Mg grains with an average grain size of about 70 μm.The strip-like intermetallic phases (IMP) distributed at the grain boundaries and granular IMP inside the grains are also observed in the LOM pictures.Fig.2 (b and c) clearly shows the decrease of grain size to about 10 and 1 μm after EBP and FSP,respectively,which indicates an efficien grain refinemen through EBP and FSP.Due to the rapid solidificatio during the EBP,the microstructure of the alloy is composed of fin equiaxial grains,and the secondary phases were rearranged at the new grain boundaries ofα-Mg.In the EB processed area shown in Fig.2(b),no evidence of micro defects such as cracks,pores,and segregation within the grains were observed.Fig.2(c)presents the microstructure of the nugget zone (NZ) of the FSPed sample with ultra-fin grain.Fig.2 (d and f) show the microstructure of the processed samples after the heat treatment.The pinning effect of the IMPs hinders the grain coarsening of the alloy during heat treatment.

To understand and analyze the microstructure of the ZKX50 alloy,characterizing the microstructure constituents in the as-cast condition is firs required.In several research studies,the structure of the intermetallic phases in the Mg-Zn-Ca alloying system has been studied [10,16,42,43].The microstructure constituent with lamellar morphology and high content of Mg-Zn-Ca are predominant intermetallic phases in Mg-Zn-Ca alloying system.The most-reported structure for these lamellar IMPs is Ca2Mg6Zn3however it is still disputed [10,16].VargasH?nzi et al.conducted a comprehensive study on the microstructure constituents of the Mg-Zn-Ca alloying system by experimental and CALPHAD method.According to this research,the Ca2Mg6Zn3is solidifie in a eutectic transformation which is stable up to～360 °C.The second-largest fraction of IMPs is the binary Mg-Zn phase that precipitates within theα-Mg matrix at temperatures below 240 °C .The precipitation of Mg-Zn phases through the age-hardening treatment leads to precipitation hardening in these alloys [7,17].

Fig.3.SEM micrographs of the as-cast ZKX50 (a) low magnificatio (b) high magnification

Fig.3 shows the microstructure of the ZKX50 alloy in the as-cast condition analysed by the backscattered electron(BSE) mode of FESEM.Fig.3 shows the agglomeration of the secondary phases in the grain boundaries with brighter contrast compared to the matrix.The brighter contrast of IMP indicates the presence of elements with a higher atomic number.A broad EDS analysis was conducted to analyze the chemical composition of the microstructure constituents especially the existing IMPs in the samples.Due to the small size of the particles and interference of the neighboring matrix,the accuracy of the EDS result is limited.The mean value and the deviation from the mean values of data obtained from the EDS point analysis are given in Table 2 for four distinct areas including the matrix,sharp edge,granular,and strip-like IMP.In agreement with previous studies[10,16,17] the granular IMPs within the grains and strip-like IMP in the grain boundaries have close chemical composition to the Ca2Mg6Zn3structure.According to the EDS point analysis,the sharp edge particles contain a rich amount of Zn,Zr,and Mn which are visible in Fig.3 with the brightest contrast.

As seen in Fig.4 (a and b),the microstructure constituents of the EBPed samples are similar to the as-cast condition.In contrast to the as-cast condition,a refine microstructure is observed in the EBPed samples,since the cooling rate is much higher in EBP compared to the casting process.This microstructure contains fine-graineα-Mg (darker area) surrounded by strip-like IMP in the grain boundaries,rich in Mg,Zn,and Ca.Compared to the as-cast condition,the strip-like IMP constructs a more connected network around theα-Mg matrix.

The SEM analysis of the EBPed-HT sample is presented in Fig.4(c) and (d).In contrast to the microstructure of the EBPed sample shown in Fig.4(a) and (b),the morphology of the Ca2Mg6Zn3secondary phases is more granular in the solution heat-treated EBPed sample.Moreover,Fig.4(c)and(d)show that the connected network of the Ca-Mg-Zn particles around theα-Mg grains is interrupted after solution heat treatment.According to the ternary phase diagram of the Ca-Mg-Zn system[16],the melting interval of Ca2Mg6Zn3secondary phases is from 340 to 360 °C.Thus,partially melting of the Ca2Mg6Zn3particles and dissolution of Ca2Mg6Zn3in the matrix are expected during the heat treatment of the ZKX50 alloy at 340 °C.The chemical homogenization of the matrix accelerates the dissolution of Ca2Mg6Zn3particles and thus decreases the volume fraction of the Ca2Mg6Zn3IMP.The estimation of the intermetallic phase fraction obtained from image analyzing is given in Table 3.To decrease the uncertainty in the estimation of the phase fraction of intermetallic phases,the same thresholding parameters were applied on several images taken from each sample.According to the estimations,the mean estimated fraction of IMPs of EBPed alloy decreased from～2.2 to～1.2% after heat treatment.A complete dissolution of thin parts of the Ca2Mg6Zn3network changes the morphology of Ca2Mg6Zn3from strip-like to more granular morphology in EBP-HT samples (Fig.4(c)and (d)).

Table 3 IMPs phase fraction of the ZKX50 alloy in different conditions.

Table 4 Microhardness (HV 0.1) of ZKX50 alloy samples in as-cast,EBPed,FSPed,EBPed-HT,and FSPed-HT condition.

Fig.4.SEM micrographs of (a,b) the EBPed ZKX50,(c,d) the EBPed-HT ZKX50 ((a,c) low magnificatio (b,d) high magnification)

Fig.5.SEM micrographs of (a,b) the FSPed ZKX50,(c,d) the FSPed-HT ZKX50 ((a,c) low magnificatio (b,d) high magnification)

Fig.5(a) and (b) shows the microstructure of the FSPed sample,indicating dynamically recrystallized refine equiaxedα-Mg grains with dispersed IMP phases all over the microstructure.The Ca2Mg6Zn3particles are distributed evenly in the matrix however a non-homogeneity is observed in the size of the Ca2Mg6Zn3particles in the microstructure of the FSPed samples shown in Fig.5(a) and (b).Since the Ca2Mg6Zn3is a fragile intermetallic phase,the applied severe force during FSP can break the particles and spread them by moving the spinning tool pin.Moreover,the experienced temperature during the FSP is sufficien for partial eutectic melting of the Ca2Mg6Zn3[17],thus,the Ca2Mg6Zn3particles are partially melted and resolidifie as larger particles than the mechanically dispersed particles.

The microstructure of FSPed samples after solution treatment at 340 °C for 24 h is shown in Fig.5(c) and (d).The EDX and SEM analyses reveal that the Mg-Zn-Ca secondary phases are partially dissolved in theα-Mg matrix without considerable grain coarsening due to the existence of intermetallic phase particles inhibiting the grain coarsening.As discussed above in the case of the EBPed-HT samples,the thin and small Ca2Mg6Zn3particles are dissolved in the matrix by chemical homogenization of the matrix during the heat treatment.As seen in Fig.5(c) and (d),the morphology of the remaining Ca2Mg6Zn3particles is more globular after solution heat treatment.According to Table 3,a decrease in IMPs phase fraction is also observed after heat treatment of FSP samples.

The results of the microhardness measurement are given in Table 4 which clearly shows the grain boundary strengthening through the grain refinemen for EBPed and FSPed samples.The FSPed sample with a grain size of about 1 μm shows the highest microhardness compared to the EBPed sample with a mean grain size of 10 μm and as-cast with a mean grain size of 70 μm.A reduction in hardness is observed for the EBPed and FSPed samples while insignifican grain growth has been observed after heat treatment.The observed hardness reduction for heat-treated samples can be attributed to the decrease in the phase fraction of the IMPs due to the partial dissolution of the Ca2Mg6Zn3in the matrix.

Table 5 Electrochemical parameters in potentiodynamic polarization calculated by Tafel extrapolation and corrosion rate calculated by mass loss.

3.3.Corrosion behavior

3.3.1.Open circuit potential (OCP) test

Fig.6.Open circuit potential in as-cast,EBPed,EBPed-HT,FSPed,and FSPed-HT conditions for 60 min in 0.5 wt.% NaCl.

The open-circuit potential (OCP) of the samples was carried out during 60 min of immersion in 0.5 wt.% NaCl for all conditions.Fig.6 illustrates the mean values of three-time measurements of the OCP variations versus time for ZKX50 alloy in as-cast,EBPed,FSPed,EBPed-HT,and FSPed-HT.As seen,EBPed-HT alloy possesses a higher OCP than those under the other conditions.The increment of OCP in the early stage of immersion is mostly attributed to the growth of the corrosion layer [44].As seen in Fig.6,the corrosion product of the ZKX50 alloy shows insignifican passivation in the early stage of the immersion,especially in the heat-treated condition.Fig.6 shows that a relatively stable OCP value is attained in the longer term of immersion due to the dynamic balance between the advance of corrosion and the deposit of corrosion products.The heat-treated samples show more positive OCP values,which agrees with the lower fraction of intermetallic phases in their microstructure.As mentioned in the introduction,the Ca2Mg6Zn3andα-Mg matrix compose a microgalvanic cell in which the matrix is the anode and Ca2Mg6Zn3particles are the cathode.Thus,a more negative potential is expected by increasing the phase fraction of the Ca2Mg6Zn3IMP.Moreover,as seen in Figs.4 (c) and (d)and 5 (c) and (d) compared to Figs.4 (a) and (b) and 5 (a)and (b),the morphology of the intermetallic phases changes from a strip-like shape to a globular shape due to the partial dissolution of the IMPs during the heat treatment.Thus,the decreasing of the contact area between the Ca2Mg6Zn3as cathode and matrix as anode contributes to the increment of the corrosion potential of the heat-treated alloys.The coarser microstructure of ZKX50 alloy in the as-cast condition with thicker intermetallic phases provides less contact area between the Ca2Mg6Zn3andα-Mg compared to the fin microstructure of EBPed and FSPed samples.Thus,the alloy in as-cast condition shows a higher corrosion potential than the processed samples.In the FSPed sample,the Ca2Mg6Zn3particles are mechanically redistributed not only in the grain boundaries but also within the grains.Therefore,a wider contact surface is expected for both FSPed and FSPed-HT samples than the EBPed and EBPed-HT,respectively.Accordingly,the higher corrosion potential of EBPed-HT samples is attributed to the reduced contact between theα-Mg matrix and Ca2Mg6Zn3phases compared to that of the FSPed-HT samples.

Fig.7.Potentiodynamic polarization curves of as-cast,EBPed,EBPed-HT,FSPed,and FSPed-HT ZKX50 alloys in 0.5 wt.% NaCl.

3.3.2.Potentiodynamic polarization (PDP) test

The potentiodynamic polarization curves of the as-cast,EBPed,FSPed,EBPed-HT,and FSPed-HT samples are shown in Fig.7.As seen in Fig.7,the heat-treated samples show an increase in the potential of the polarization curve toward the less negative values which is in agreement with the OCP test result shown in Fig.6.As discussed in the previous section,the corrosion potential of the samples increases by decreasing the phase fraction of the Ca2Mg6Zn3particles which are acting as a cathode in the microgalvanic pill composed of the IMP particles and matrix.Fig.7 shows also a decrease in the corrosion current density (icorr) for the heat-treated samples.In general,an improvement of the corrosion resistance of the material is related to a reduction of the icorr.The icorrand corrosion potential (Ecorr) of the alloy in all conditions were obtained by applying the Tafel extrapolation method on potentiodynamic polarization curves which are listed in Table 5.As given in Table 5,the current density of the heat-treated samples (～5-9 μA/cm2) is nearly seven times less than that of the processed samples (43-52 μA/cm2).The as-cast and FSPed samples show a similar current density while the calculated icorrfor the EBPed sample is higher than the as-cast condition,which is attributed to the wider contact surface between matrix and IMPs (Fig.4 (a and b)).

Using Eq.(1),the corrosion rate (mm/year) of the samples used in the PDP test was estimated according to the icorr(mA/cm2) calculated by Tafel extrapolation [45].The corrosion rate (Pi) estimated by the data extracted from Tafel extrapolation reflect mainly the short-term corrosion behavior [46,47].Thus,the metal weight loss measurement was also conducted on the samples immersed in the same solution used in the PDP test for one week to evaluate the overall weight loss (Pw).Using Eq.(2) [45],the weight loss (ΔW)(g/cm2/d) was measured,and accordingly,the corrosion rate(given in Table 5)was calculated for the samples immersed in 0.5 wt.% NaCl for one week.In agreement with the corrosion rate estimated by Tafel extrapolation (Pi),the ZKX50 alloy in EBPed-HT and FSPed-HT conditions show a lowerPwcompared to the other conditions.Among all studied samples,the EBPed sample shows the highest corrosion rate which can be attributed to the higher contact surface between the Ca2Mg6Zn3phase and surrounding matrix in its microstructure compared to other samples.

3.3.3.Electrochemical impedance spectroscopy (EIS) test

Electrochemical impedance spectroscopy (EIS) is an effi cient and continuous approach for evaluating the long-term corrosion rate and understanding the corrosion mechanism[48].Accordingly,the EIS test was carried out for as-cast,processed,and heat-treated ZKX50 samples immersed in 0.5 wt.% NaCl solution for periods of one hour,one day,and one week.

Fig.8 shows the Nyquist plots of EIS measurements in 1,24,and 168 h for all conditions.The Nyquist plots for all the samples show an inductive loop at low-frequencies and a capacitive loop at high and medium-frequencies.The highfrequency region of the Nyquist plot represents the charge transfer resistance of the oxide fil composed during the corrosion and the presence of a capacitive loop in high and medium frequency shows the resistance of the corrosion product [49].The low-frequency part of the Nyquist plot indicates the Faradaic processes in the metallic-oxide layer interface representing the rate of the anodic reaction [50,49].

Fig.8.Nyquist diagrams in as-cast,EBPed,EBPed-HT,FSPed,and FSPed-HT conditions of ZKX50 alloy after (a) one hour,(b) one day,and (c) one-week immersion in 0.5 wt.% NaCl (d) Example of measured and calculated Nyquist diagrams using Randles equivalent circuit for EIS analysis.

As seen in Fig.8,the radius of the capacitive loops for heat-treated EBPed and FSPed ZKX50 samples are greater than that in the other conditions for all other immersion times.The larger radius of the capacitive loop indicates higher corrosion resistivity of the composed hydroxide layer.It shows that the composed hydroxide layer is more protective when the contact surface of the Ca2Mg6Zn3and matrix is less in the microstructure.In the short-term measurement (Fig.8(a)),the processed and heat-treated samples with fine microstructure show a larger capacitive loop compared to the as-cast condition which indicates the higher resistivity of the composed hydroxide layer on the refine microstructure.The radius of the capacitive loop decreases by increasing the immersion time for all conditions;however,the decrease is much faster for processed and heat-treated samples than the as-cast sample.The capacitive loop in the EBPed state shows a more pronounced reduction compared to the others and it is the smallest one after 1-day (Fig.8(b)) and 1-week immersion(Fig.8(c)).The better capacitive behavior of the EBPed-HT and FSPed samples can be attributed to the dissolution of the intermetallic phases that can act as microgalvanic cells.

To quantitatively analyze the electrochemical behavior of the EIS measurements,the obtained result from the EIS test was evaluated using the Randles equivalent circuit Fig.8(d).The Randles equivalent circuit comprises a resistor and a capacitor connected in parallel,and the complex of the resistor and capacitor is connected to another resistor in series.The Rs,Rct,Cdlin Randles equivalent circuit represent,respectively the resistance of the electrolyte,charge transfer resistance of the metal,and impedance of the hydroxide layer[19].A constant phase element was used instead of capacitance to account for the non-ideal behavior of the system[51]obtained from the Randles equivalent circuit for ZKX50 alloys.Electrical double-layer capacitance shows the water/electrolytes uptake through pores/defects in the hydroxide fil composed on the surface of the sample [52,53].Thus,an increase of Cdlwith time implies the increase of porosities/defects in the hydroxide film As given in Table 6,significan differences are not observed in the estimated values of Cdlwhich shows that the composed hydroxide layers on the surface of samples in different conditions have similar porosity/defect characteristics.However,a continuous increment for Cdlis observed over time showing that the integrity of the hydroxide fil decreases,and the fraction of pores increases during the immersion.The free space in the hydroxide layer provides space tostore the ions and thus the capacitance increases by increasing the trapped electric charge in the pores.From Table 6,it can be noted that the Rctof the alloy in all conditions decreases with time,suggesting a decrement in the corrosion resistance of the hydroxide layer.The increment of the Cdltogether with the decrement of the Rctover time implies that the composed hydroxide layer on the ZKX50 alloy is not a passive layer to decelerate the corrosion process.However,the Rctof the heat-treated samples compared to as-cast and processed samples indicates a higher resistivity of the heat-treated samples which is attributed to the lower fraction of the Ca2Mg6Zn3particles in the microstructure.

Table 6 Simulated parameters of the electrochemical impedance spectroscopy (EIS) data of the investigated ZKX50 alloys after 1 h,24 h(1 day),and 168 h (1 week).

3.3.4.Hydrogen evolution in simulated body flui

An important factor for the biodegradable application of Mg alloys is the hydrogen evolution during the corrosion of the Mg alloy which can lead to the formation of the gas pockets around the injured tissues,hydrogen bubbles in the blood,and pH increment in body flui [21].The hydrogen evolution rate in the simulated body flui is thus required to evaluate the functionality and side effects of biodegradable implants.From the volume of the released hydrogen during the hydrogen evolution test,some detailed information about the corrosion process can also be extracted such as the dissolution rate of the Mg in the solution,the corrosion rate of a magnesium alloy,and the degree of alkalization in the corrosion media[21].The hydrogen evolution test was thus also employed to assess the degradation rate of the as-cast,EBPed,FSPed,and heat-treated samples in simulated body fluid The results of the hydrogen evolution test are shown in Fig.9.Since the evolution of 1 ml of hydrogen corresponds to the dissolution of 1 mg of the Mg,thus,the corrosion rate in SBF can be directly estimated by the rate of the hydrogen evolution.To compare the result,the weight losses of the samples in the hydrogen evolution test were measured before and after the test,and accordingly,the corrosion rate was calculated based on the weight loss in SBF.The average rate of the hydrogen evolution and calculated corrosion rate are given in Table 7 which show a good agreement with each other.Consistent with the weight loss measurement in 0.5 wt.% NaCl solution,the corrosion rate of the processed samples is higher than the as-cast condition and the heat-treated samples show the minimum hydrogen release and corrosion rate.The hydrogen evolution rate of the EBPed -HT samples is one-fift of the as-cast condition thus the produced subcutaneous bubbles of the planted implant made of EBPed samples will be much less and better tolerated by the body than that made of the as-cast sample [21].

Fig.9.Hydrogen evolution rate in Hank’s solution.

Fig.10.FESEM micrograph of the corroded surface of the as-cast sample ((a) low magnificatio (b) high magnification)

3.3.5.Corrosion surface

The corroded surface of the samples immersed in 0.5 wt.%NaCl solution was analyzed by FESEM after cleaning the surface with chromic acid.Fig.10 shows the micrographs of the corroded surface after 1 h immersion in the 0.5 wt.%NaCl solution for the as-cast sample.As shown by the micrographs,the microgalvanic corrosion starts from theα-Mg which are in contact with Ca2Mg6Zn3particles,and progresses into the interior of the matrix.

Fig.11 shows the corrosion surface of the EBPed,FSPed EBPed-HT,and FSPed-HT samples after 1 h immersion in the 0.5 wt.% NaCl.Similar but more progressive corrosion is observed in the processed samples compared to the ascast one (Fig.11 (a,c,and d)).The progressive corrosion in the processed samples implies the increment of the microgalvanic corrosion due to the increment of the contact surface between theα-Mg and Ca2Mg6Zn3particles.As discussed in section 0,the redistribution of the fin and thin Ca2Mg6Zn3particles in the refine microstructure is due to the rapid solidificatio in EBP and the partial melting and mechanical dispersing in FSP,which produces more contact surface between the Ca2Mg6Zn3particles and the matrix.As expected the microgalvanic corrosion is also seen in heat-treated samples Fig.11(e,f,g,and h).By decreasing the volume fraction of the Ca2Mg6Zn3IMPs with globular morphology,however,the microgalvanic corrosion decreases significantl compared to the processed and as-cast samples.

4.Conclusion

The electron beam processing and friction stir processing of the ZKX50 Mg alloys lead to a significan microstructure refinemen with a homogenous distribution of the Ca2Mg6Zn3secondary phases at the grain boundaries.The solution heattreatment of the processed samples at the temperature around the solidus temperature of Ca2Mg6Zn3partially dissolves the Ca2Mg6Zn3particles and changes their morphology from strip-shaped to globular.The microstructure characterization in this research shows that the phase fraction of IMPs decreases under solution heat treatment but the pinning effect of the remaining IMPs still avoids the grain coarsening of the microstructure.The results of electrochemical corrosion tests show that the corrosion rate significantl decreases through solution heat treatment,while the refine microstructure of the processed samples that is preserved can guarantee the preservation of the mechanical properties obtained by EBP and FSP.The hydrogen evolution test in SBF also shows a significan decrease in the quantities of released hydrogen during degradation of the EBPed-HT samples compared to the other conditions and the EBPed-HT ZKX50 alloy can thus be a suitable candidate for further research to produce biocompatible and biodegradable implants.