Lish Wng,Jinghu Jing,Hun Liu,Bssiouny Sleh,Aibin M

aCollege of Mechanics and Materials,Hohai University,Nanjing 210098,China

b Suqian Institute,Hohai University,Suqian 223800,China

Received 26 June 2019;received in revised form 20 August 2019;accepted 14 December 2019 Available online 30 June 2020

Abstract Mg–Y–Zn alloys with long period stacking ordered(LPSO)structure have received much attention recently and exhibit great potential in applications such as automotive,aerospace and in bio-medical fields.This paper aimed to investigate the effect of different phase constitution of LPSO structures on corrosion rate of bio-medical Mg–Y–Zn alloys.The results showed that as-cast Mg98.5Y1Zn0.5 alloys containing only 18R structure exhibited the highest corrosion resistance with the corrosion rate of 2.78mm/year.The precipitation of 14H lamellas within α-Mg grains during solid solution treatment introduced the crystallographic orientation corrosion by accelerating micro-galvanic corrosion.The increase of 18R/14H interfaces deteriorated the corrosion resistance,and the grain boundaries also suffered from severe electrochemical dissolution.This work suggested that Mg–Y–Zn alloys with single LPSO structure(either 18R or 14H)exhibited better corrosion resistance than alloys with co-existence 18R and 14H LPSO structures.

Keywords:Corrosion behavior;Long period stacking ordered phase;Magnesium alloys;Solution treatment;Bio-materials.

1.Introduction

Magnesium alloys(Mg)have been targeted for a variety application in automotive,aerospace and bio-medical implants because of their superior properties over other alloys[1–4].A good understanding on the corrosion mechanism and the corrosion behavior of Mg alloys in various conditions including in vivo environment is essential to develop high performance Mg alloys and to extend their application field[5–10].

Mg is an instinct reactive metal which has a high tendency to be eroded in aqueous solutions.For years much efforts have been dedicated to overcome this intrinsic defect of Mg alloys,such as surface film modification and microstructural optimization[11–15].Recently it is found that alloying with an appropriate proportion of rare earth element(RE)and transition metal elements(TM,such as Zn,Cu,Ni,Al and Co)leads to the formation of novel second phases with long period stacking ordered(LPSO)structure[5,11–18].These alloys usually possess excellent mechanical properties,corrosion resistance as well as high biocompatibility,showing great application potential in biodegradable orthopedic implants[1,19–29].The yield stress of the as-extruded Mg89Zn4Y7alloy with LPSO phase could reach to 480MPa[30].In biological environment,the mechanical properties of 18R-containing Mg–8Y–1Er–2Zn alloy still maintained at a relatively high level with the tensile yield stress of 216MPa,ultimate tensile strength of 286MPa and elongation of 6.8%,respectively,after immersed in simulated body fluid(SBF)for 240h[31].It is widely accepted that the excellent mechanical property of these alloys resulted from the presence of LPSO structures,which displayed high hardness,high modulus,and could strengthen the alloys via load transfer mechanism,kinking band strengthening,and fine grain strengthening[20,32–37].Moreover,some researches showed that the LPSO phase played a significant role in enhancing the corrosion resistance of Mg alloys[2,19–24].Table 1 lists the collected corrosion data of Mg and its alloys under various processing conditions.In comparison to conventional bio-Mg alloys like WE43,ZK60 and ZX60,the Mg96.83Y2Zn1Zr0.17with 18RLPSO phase showed a lower degradation rate[38],which was mainly ascribed to the barrier effect of LPSO structure.In addition,Mg-Y-Zn alloys with LPSO structures possessed good biocompatible properties.Zn is an essential element in human body and plays an indispensable role in human health[39].As one of the most important LPSO-forming elements,Y element shows weak toxicity with a tolerant daily intake of 4.2mg,and can be binding to globulin[40,41],which has low effects on viability of macrophages and the inflammatory reaction[42,43].Therefore,it can be concluded that the Mg-Y-Zn alloy system combines excellent mechanical characteristics,corrosion resistance and good biocompatibility,suitable to be employed as biomaterials.

Table 1Corrosion data of Mg and alloys under various conditions.

According to recent researches[8,44–47],the types,morphology and distributions of LPSO phases have great influences on the corrosion properties of LPSO-containing Mg alloys.Up to now,several LPSO structures have been reported in Mg-RE-TM alloy system,including 10H-type,12H-type,14H-type,15R-type,18R-type,21R-type,and 24R-type LPSO phases,among which the 18R and 14H are the most common types[18,44–48].The detailed stacking sequences and atomic structures of several LPSO structures are summarized in Table 2.Zhao et al.[38]found that Mg96.83Y2Zn1Zr0.17alloy with 18R LPSO phase possessed a better corrosion re-sistance compared with LPSO-free Mg alloys.Similar results were also reported in 14H-containing alloys.The corrosion rate of Mg–11.3Gd–2.5Zn–0.7Zr alloy with 14H-type LPSO structure was two times lower than their LPSO-free counterparts[37].Hence,both 18R and 14H LPSO phases showed great influence on the corrosion behavior of Mg alloys,and it is necessary to understand the difference between roles of 18R-LPSO and 14H-LPSO phases in corrosion characteristic.Unfortunately,little research focused on the corrosion behaviors of Mg alloys with 18R and 14H phases contained in the same alloy.Therefore,by tuning the contents of 18R and 14H LPSO structures in Mg98.5Y1Zn0.5(at%)alloy via solid solution treatments,the current research investigated the corrosion behaviors of Mg-Y-Zn alloys with different volume fractions of 18R and 14H phases,and the corrosion mechanisms of 18R and 14H structures onα-Mg matrix were elucidated.

Table 2Stacking sequence and lattice parameters of various LPSO phase in Mg-TMRE alloys.

2.Experimental procedure

2.1.Fabrication process

The Mg98.5Y1Zn0.5(at%)alloy with moderate content LPSO phase was selected in this study.The alloy was prepared by melting high purity Mg,Zn metals and Mg–30%Y master alloy at 710°C under a mixed protective atmosphere of CO2and SF6.After homogeneously mechanical stirring,the molten metals were maintained at 710°C for 20min,and then poured into a pre-heated(200°C)steel mold.The actual chemical composition of the alloy was analyzed to be Mg98.61Y0.96Zn0.43by inductively coupled plasma analyzer(ICP).The as-cast Mg–Y–Zn alloy was cut into test coupons with a dimension of 10mm×10mm×2mm.Solid solution treatment at 500°C for 4h,8h and 12h were conducted on as-cast alloy to tune the content of 18R and 14H phases(the samples were abbreviated as ss-4h,ss-8h,and ss-12h,respectively).During solution treatment,the samples were coated with Al foil and embedded in a crucible filled with graphite to protect the material.The electrochemical specimens were embedded in an epoxy resin with 1 cm2exposing area as working electrode surfaces.The other end was linked to an electrical connection with a copper wire.Before electrochemical testing,the electrochemical samples were grinded to 2000 grit by SiC abrasive paper and then polished with a diamond polishing spray(1μm).

2.2.Microstructural characterization

The microstructure of as-cast and solution-treated Mg–Y–Zn alloys was characterized by an optical microscopy(OM,Olympus BHM).The specimens were mechanically grinded by SiC papers(from 120 grift to 2000 git),polished with diamond polishing(1μm),and then etched with an alcohol solution containing 4% nitric acid(volume fraction).To measure the volume fraction of LPSO phases,the software Image Pro-Plus 6.0 was employed,and at least three OM images were analyzed for each alloy state.As the 14H phase exhibited lamellar shape with different length and width withinα-Mg grains,its volume fraction is hard to be measured,and therefore we mainly measured the content of 18R phase in this work.Phase identification of the alloy was performed with an X-ray diffractometer(XRD,Smartlab 9kw)at a scanning angle of 10°?85° and a scanning speed of 10(°)/min with a copper target.To identify different LPSO structures,TEM observation was carried out by a FEI Tecnai-G2thermo-emission transmission electron microscope at an accelerating voltage of 200kV.TEM Samples were mechanically polished first,and then thinned by the twin-jet electro polishing method.

Table 3Composition of the simulated body fluid in this study[35].

2.3.Corrosion behavior

The corrosion behaviors of the as-cast and solution-treated alloys were examined by the hydrogen evolution test and electrochemical analyses.The device used for hydrogen collection and volume measurement was made by a set of glass apparatus including beaker,burette and funnel.Installed samples were suspended vertically in solution by fishing line[59].The hydrogen evolution test was performed in simulated body fluid(SBF,its composition was listed in Table 3)at 37±0.5°C in water bath.The indoor temperature of the laboratory was controlled at 25±1°C by air conditioning.In the relevant literature,CO2-bicarbonate buffered Hank’s solution was used to maintain the pH similar to the pH in the body[52,60,61].As for this work,the precipitation could increase the pH of the solution during the hydrogen immersion.So,the SBF was renewed every 24h to ensure a stable pH value.The sample surface area ratio(cm2)to the SBF solution volume(mL)was established at 1/20.200g/L CrO3+10g/L AgNO3solution was used to clean samples for the removal of surface corrosion products after immersion.The corrosion ratePAH(mm y?1)and Pw(mm y?1)was obtained by hydrogen immersion rate and weight loss methods evaluated by Eqs.(1)and(2),respectively,whereis total volume of hydrogen evolved for the total immersion time,(mg)means the specimen weight loss,A(cm2)is the exposed surface area of the specimen,andt(day)is the immersion time[16,59].

Scanning electron microscope(SEM,Sirion)was used to characterize the surface morphology of the eroded alloys.In order to study the electrical corrosion behavior,the electrochemical workstation(CHI660C)was employed.A standard three-electrode method was used,with Pt as counter electrode,SCE as reference electrode,and the specimen as the working electrode.Open circle potential test(OCP)was carried out for 4000s.Electrochemical Impedance Spectroscopy(EIS)was examined after OCP test at the OCP potential.The scan frequency of EIS was ranged from 0.1Hz to 100kHz.The EIS spectra was fitted using the ZSimpWin software.The scan range of dynamic polarity(PDP)test was from?250mV to 500mV versus OCP at a scan rate of 1mV/s.The specimens were measured three times to ensure reproductivity.The instantaneous corrosion rate,Pi(mm y?1),was determined from the corrosion current density,icorr(mA cm2)by[62–64]:

Fig.1.XRD pattern of as-cast Mg98.5Y1Zn0.5 alloy.

3.Results

3.1.Microstructure evolution

Fig.1 presents the XRD patterns of as-cast Mg98.5Y1Zn0.5alloy.It can be seen that the as-cast Mg98.5Y1Zn0.5alloy was mainly composed ofα-Mg and Mg12YZn-type(LPSO)phase.Fig.2 shows the microstructures of as-cast and solutiontreated alloys.seen from Fig.2a,the discontinuous blocky shaped secondary phases were uniformly distributed at Mg grain boundaries of as-cast Mg98.5Y1Zn0.5alloy.Fig.3a shows the TEM observation of second phase in as-cast alloy,and the corresponding selected area electron diffraction(SAED)pattern demonstrated that the Mg12YZn-type phase exhibited a 18R LPSO structure[55].After 4 h solution treatment(Fig.2b),the 18R block phases were transformed into lamellar structure within theα-Mg grains gradually.Fig.3b shows the TEM image of lamellar phase and its corresponding SAED pattern,which proved the existence of 14H LPSO structure.Table 4 lists the variation of volume fraction of 18R LPSO structure in Mg98.5Y1Zn0.5alloys with different solution treatments.The as-cast alloy contained 9.7% 18R phase,and it decreased to 7.5% after 4h solution treatment.When the solution time increases to 8h,the 18R continued to be replaced by 14H structure,and at this time its volume fraction was approximate 2.9%.With increased solution time,the fine lamellar 14H phase continued to precipitate withinα-Mg grains.As shown in Fig.2d,after 12 h solution treatment,the 18R LPSO phase almost disappeared completely(with its volume fraction0.76%).At this situation,14H LPSO phase was distributed uniformly within the matrix.Owing to the unique orientation relationships between 14H andα-Mg(shown in Fig.3b),the newly precipitated 14H lamellas are parallel with each other in the same grain,but their orientations are different between different grains,as illustrated in the Fig.2e.

Table 4volume fractions of 18R LPSO phase in different alloys.

3.2.Corrosion behaviors

Fig.4 shows the average hydrogen evolution rate and the corrosion rate of as-cast and solution-treated Mg98.5Y1Zn0.5alloys after immersed in 37°C SBF for 120h.It can be seen from Fig.4a that all the target samples showed gradual increases in hydrogen evolution during the first 40h’s immersion.With prolonged immersion time,the slop of ss-8h plot grew rapidly,suggesting the poor corrosion property of ss-8h sample.The whole hydrogen evolution rate can be ranked as the following sequence:ss-8h>ss-12h>ss-4h>ascast alloy.Fig.4b shows the corresponding corrosion rate,PAH,evaluated from the total hydrogen evolution rate,and the corrosion ratePWcalculated from weight loss.There was a good agreement in the corrosion trends betweenPAHandPWvalues.For each specimen,the average corrosion rates measured from the total evolved hydrogen,PAH,were lower than the average corrosion rates measured from weight lossPW.This was mainly attributed to the partial dissolution of hydrogen in Mg metal as well as in the solution[16,62].The detailed data for the average corrosion rates ofPWandPAHis presented in Table 5.The as-cast Mg98.5Y1Zn0.5samplecontaining only 18R structure owned the lowest corrosion ratePWof 2.78mm/year,and the ss-8h sample had the highest corrosion rate(5.83mm/year).However,the average corrosion rate was declined when 18R LPSO phase were totally substituted by 14H structure.This could be attributed to the electrochemical homogeneity of the system.In addition,during solution treatment,the contents of the 18R/14H interfaces gradually increased.The sample after 8 h solution treatment possessed the largest amount of unstable 18R/14H interfaces,which may be responsible for the poor corrosion resistance of the specimen.While after 12h solution treatment,the 18R structure as well as the 18R/14H interface nearly disappeared.At this time,the corrosion rate decreased and the corrosion resistance was improved.Combining the microstructure and the hydrogen test results,it could be inferred that the introduction of 18R/14H interfaces may do harm to the corrosion resistance of Mg98.5Y1Zn0.5alloys.

Table 5Electrochemical parameters obtained from potentiodynamic polarization curves.For comparison,the corresponding instantaneous corrosion rate was also given based on hydrogen collection data and weight loss assessment.

Fig.2.Optical images of various Mg98.5Y1Zn0.5 alloys:(a)as-cast;(b)solution-treated for 4h;(c)solution-treated for 8h;(d)solution-treated for 12h and(e)its high-magnification image.

Fig.3.TEM images of(a)18R LPSO structure in as-cast alloy and(b)14H LPSO structure in ss-12h alloy.Insets of(a)and(b)are corresponding SAED patterns of 18R and 14H phases.

Fig.4.Hydrogen evolution volume and the corrosion rate of the Mg98.5Y1Zn0.5 alloys in as-cast and solution-treated conditions after immersed in 37 °C SBF for 120 h:(a)Hydrogen evolution volume;(b)Corrosion rate.

3.3.Corrosion morphology

Fig.5 shows the corrosion morphologies of the 18Rcontaining as-cast Mg98.5Y1Zn0.5alloy after removing the corrosion products.Immersed for 1h,some corrosion pits occurred on the surface(Fig.5a).Driven by the micro-galvanic cell made up ofα-Mg and LPSO particles,corrosion continued with the prolonging of immersion time.During the corrosion process,18R LPSO phase could be peeled off from the material with the cracking and shedding of the corrosion products and left some corrosion pits on the surface.Considering that the 18R phase was distributed both at grain boundaries and within the grains,the deep crystallographic type corrosion(Fig.5a)may be attributed to the larger 18R phase at the grain boundary falling off from the surface with the exfoliation of corrosion products.The immersion time increased,some corrosion pits grew deeper and bigger.The corrosion area became continuous and enlarged along the 18R/α-Mg interfaces(Fig.5b and c).After 120h immersion,some regions experienced very aggressive attack,and long deep corrosion grooves are formed around the LPSO phase on the surface(Fig.5c).Similar phenomenon was also reported in Mg-Gd-Zn alloys with 18R LPSO structure[25],where theα-Mg matrix was preferentially corroded and corrosion penetrated progressively along the residual LPSO phase.

Fig.6 displays the corrosion morphologies of 4h solutiontreated Mg98.5Y1Zn0.5alloy after removing corrosion product.Sample in this situation contained both 18R and 14H LPSO structure.After immersion for 1h,some corrosion pits were observed initiating on the surface,especially around LPSO phase(as can be seen from Fig.6a and b).When immersion time was prolonged to 24h,the corrosion extended along a particular direction,probably following a preferred crystallographic orientation which resulted in the larger corrosion pit on the surface.The reason for this special corrosion morphology was mainly due to the distribution of lamellar 14H structure,which was precipitated along the basal planes ofα-Mg grains.Actually,for Mg and its alloys,the electrochemical performance has been considered to exhibit significant dependence on the crystallographic orientation.According to Song and Liu et.al[65,66],the influence of crystallographic orientation on the corrosion behavior may be related to the binding energy and surface energy.In the present work,solution treatment lead to the precipitation of lamellar 14H structure,which accelerate the micro-galvanic corrosion along the 14H/α-Mg interfaces(the length direction of 14H lamellas)and finally resulted in the crystallographic orientated corrosion.Moreover,the obvious deep corrosion grooves detected on the surface can be attributed to the unstable 18R/14H interface as illustrated in Fig.6c and d.

Fig.5.SEM morphology of the as-cast Mg98.5Y1Zn0.5 alloy after immersed in 37°C SBF solution for different times:(a)1h;(b)24h;(c)120h.

Fig.6.SEM morphology of the ss-4h Mg98.5Y1Zn0.5 alloy after immersed in 37°C SBF solution for different times:(a)and(b)1h;(c)and(d)24h.

Fig.7 shows the corrosion morphologies of ss-12h Mg98.5Y1Zn0.5alloys with 14H structure.Unlike other samples,the 14H-containing Mg98.5Y1Zn0.5alloy presented an intact surface and no obvious corrosion pits were observed at low magnifications(Fig.7a and c).Immersed for 1h,some shallow corrosion grooves were formed on the interface(Fig.7b).Prolonging immersion time,corrosion grew along a certain orientation within the same grain,resulting in parallel corrosion pits,but the orientation was quite different in different grains.The corrosion morphology at grain boundary is observed at high magnification(Fig.7d).The intermetallic particles with fine-scale plate-like shape arranged parallel with each other were 14H structure.The grain boundary between different orientated 14H clusters experience severe electrochemical dissolution(Fig.7b–e).Overall,the LPSO/α-Mg micro-galvanic corrosion cell facilitated the erosion ofα-Mg phase along the boundaries of lamellar 14H precipitates with the parallel corrosion pits left behind.

Fig.7.SEM morphology of the Mg98.5Y1Zn0.5 alloys with 14H structure after immersed in 37°C SBF solution for different times:(a)and(b)1h;(c)and(d)24h;(e)24h at high magnification.

3.4.Electrochemical measurements

Fig.8 shows the potentiodynamic polarization(PDP)curves of as-cast and solution-treated Mg98.5Y1Zn0.5alloys.The detailed electrochemical parameters calculated from PDP plots are summarized in Table 5.It can be seen that the polarization curves of all four specimens were not symmetrical between their anodic and cathodic branches.The plateaus appeared in the anodic regions.In the anodic regions,the current density of as-cast and ss-4h alloys increased slowly with increasing polarization before the pitting potential(Epit.,as marked on the curves),and afterEpit.the current density increased rapidly.This indicates the present of partially protective films on the surface the alloys.There was an active dissolution rate in the anodic curve of ss-8h alloy.The current density increased rapidly with the polarization potential,which suggests that the corrosion products in this case almost presented no protective effect.In the cathode polarization branches the cathodic current densities of the four samples were very different under the same potential.At the same cathodic polarization potential,the 18R containing sample had the lowest corrosion current density.Table 5 also includes the values of the corrosion rate evaluated from the polarization curves,Pi,and the corrosion rate evaluated form hydrogen evolution ratePAHand weight lossPw.In all conditions,Piwas much lower thanPAHand weight lossPw.The lack of consistency amongPi,PAH,Pwmay result from the fact that,the electrochemical measurements do not reflect the entire corrosion process.Since polarization curves are often measured soon after specimen immersion in solution.For most Mg alloys containing multi phases,as demonstrated by Atrens and Song et.al,the corrosion rate may be increase with immersion time during the period in which the surface film breaks down and replaced by a more stable one,in that case the initial corrosion rate is less than that of the steady state.The lessPivalue may also relate to the partially electrochemical controlled corrosion process[13–16].The as-cast Mg98.5Y1Zn0.5sample with only 18R LPSO structure possessed the lowest corrosion current density(Icorr)of 4.11μAcm?2.This could be ascribed to the effect of the corrosion product film.In the initial stage,the production of Mg(OH)2promoted the formation of a protective passive film on the surface,resulting in the low corrosion rate of the as-cast alloy.Further solution treatment led to the phase transformation from 18R to 14H,which in turn affected the corrosion behavior.The appearance of 14H structure in the electro-system increased the corrosion current density.The ss-8h samples owned the highest current density(10.64μA cm2).

Fig.8.Potentiodynamic polarization curves of as-cast and solution-treated Mg98.5Y1Zn0.5 alloys.

EIS measurements measured at the equivalent potential of for as-cast and solution-treated alloys are presented as both Nyquist and Bode plots in Fig.9.The Nyquist plots of all the samples exhibited a clear capacitive arc at the high frequency region shown in Fig.9a.The high-frequency capacitive loop could be attributed to the electric double layer and the charge transfer process at the electrode/electrolyte interface[1,67–69].The higher increase size of high frequency loop reflects the stabilization of surface layer resulting in a decrease in the active dissolving surface,leading to a significantly higher impedance[70–72].The capacitive loops in the high-frequency domain for as-cast alloy exhibited the highest diameter sequenced with ss-4h samples compared with other three alloys which indicate the high corrosion resistance property of as-cast alloys.The ss-8h alloy showed distinct inductive features at the lower frequencies.According to Song et.al the induction characteristics in Nyquist plots can be related to the breakdown of surface film and the initiation of dissolution of Mg alloys[16,69].The decrease in corrosion resistance of ss-8h samples was also observed from the impedance values read from the Bode plots at the low frequencies.

The EIS plots of Mg98.5Y1Zn0.5alloys in all conditions can be equivalent to the circuit shown in Fig.9d.Considering the different electrochemical characters between as-cast and solution-treated samples,two equivalent circuits are used to fit the curves.The fitting results are listed in Table 6.Rsolrefers to the solution resistance.RctandCPEdlconnected in parallel represent the high frequency capacitive loop.CPEdlis a constant phase element used to present an unperfect capacity of the electric double layer owing to the nonhomogeneity in the electrochemical system.Rctmeans the charge transfer resistance.TheCPEhas a frequency independent constant phase shift and can be defined by two values,Qdlandndl.Ifndlis equal to 1,theCPEis identical to a capacitor,ifndlis equal to 0,theCPEis just a resistor.TheRfandCPEf(defined byQfandnf)show the film resistance and film capacity,respectively.Since the oxide films formed on the surface of the alloys during the corrosion process are not always perfect[72].The parallel circuit elements(CPEfR)refers to the unperfect capacity of surface film.For ss-8h samples,there exists obvious inductive loop in Nyquist plots,soLandRLconnected in parallel to display the low frequency inductance area.Lrepresents the inductance indicating the initiation of corrosion(the low frequency inductance region).

According to the fitting results in Table 6,the as-cast alloy showed the highestRctandRfvalues followed by ss-4h sample.The ss-8h sample had the lowestRctandRfvalues.This indicates that samples with 18R structure have the lowest metal dissolution rate and the highest film protection property.The electrochemical results have the same trend toward the hydrogen immersion test.The lowRctandRfvalues for ss-8h may be related to the less stable system caused by the increased 18R/14H interfaces.

4.Discussion

Based on above results,the corrosion behaviors of different LPSO-containing Mg98.5Y1Zn0.5alloys differ greatly between each other.The reason for this phenomenon may be related to the electrochemical homogeneity,and the different structures and morphologies of LPSO structures in as-cast and solution-treated alloys[73].The solution treatment stimulated transforms from blocky 18R to lamellar 14H structure(Fig.2).Thus,the differences of two LPSO phases must affects the corrosion resistance of as-cast and solution-treated Mg98.5Y1Zn0.5alloys in SBF solution.Due to the large potential difference betweenα-Mg matrix and LPSO phase,micro galvanic couple forms,which facilitates the corrosion process described below.

Fig.9.EIS data(scatter plot)and model fit(solid lines)of as-cast and solution-treated Mg98.5Y1Zn0.5 alloys and their equivalent circuit:(a)Nyquist plots;(b)Bode plots:|Z|-frequency plots;(c)Bode plots:phase angle-frequency plots;(d)equivalent circuit.

Table 6Fitting results of EIS spectra.

We assume that the electrical reactions are rare when the region is covered by oxidation–reduction products.The current density on the exposed sample can be defined asIcor.For potentials close toEcorr,the corrosion current calculated by the cathodic half-reaction[74]

I0,cis the cathode phase exchange current that exhibits the LPSO structure electrical activity.αcmeans the cathodic electrochemical charge transfer coefficient.The n’is the number of electrons in the rate-determining step.βcrefers to the Tafel slop of cathodic branch.Ecoris the corrosion potential of couple betweenα-Mg and LPSO phase.Eeq,cis the equilibrium potential for the cathodic half-reaction.WhenEcoris approximate toEeq,c,the corrosion density of the electro-system is equivalent with cathodic current density.Since the composition of the as-cast and heat treated Mg98.5Y1Zn0.5alloys are the same,theI0,cchanges because of the different LPSO phase in as-cast and solution-treated samples.For as-cast specimen with 18R structure,Icoris the lowest thus the cathodic reaction occurred hardly on the only 18R containing samples.For 12h heat treated samples with only 14H structureIcoris smaller than ss-8h Mg98.5Y1Zn0.5alloys.This indicates that the cathodic reaction occurred on LPSO structures with both 18R and 14H is much more rapid.

Fig.10.Schematic diagrams of the corrosion mechanisms of Mg98.5Y1Zn0.5 alloys with 18R structure:(a)3-D corrosion morphology of the corroded surface;(b)galvanic corrosion cell;(c)corrosion path.

Detailed schematic diagrams for the corrosion process are presented in Figs.10 and 11 to interpret the corrosion mechanisms of Mg98.5Y1Zn0.5alloys with different LPSO structures.Considering the large potential difference between the Mg matrix and the LPSO phase,a micro galvanic couple will form to facilitate the corrosion process.The second phases act as cathodes,while the Mg matrix worked as anode during the corrosion process,resulting in the preferential initial corrosion sites along the LPSO/α-Mg interface(Figs.10b and 11b).For 18R-containing Mg98.5Y1Zn0.5alloys,the Mg(OH)2produced by corrosion will fill the corroded metal sites with the prolongation of immersion time,but the existed Cl?in the solution have a strong destructive effect on the corrosion products,which helps to keep the local site active,and then continue the corrosion process[70].Compared to the interior matrix,it is much easier for the interface between theα-Mg and 18R phases to suffer from further erosion.Thus,the corrosion region proceeds around the LPSO phases with moreα-Mg dissolved(Fig.10c),which eventually leads to long-term serious corrosion.

As for Mg98.5Y1Zn0.5alloys with 14H structure,there is no obvious inhomogeneous corrosion on the surface.Initially,the micro-galvanic cell that formed betweenα-Mg and 14H is causing corrosion,and the corrosion has certain directivity owing to the unique orientated 14H lamellas.Corrosion then grows along a certain orientation in a stripe shape like way parallel with each other inside the same cluster.As immersion time extended,corrosion continues to corrode during the second phase as shown in Fig.11c.There is obvious grain dissolution at the boundary.For Mg alloys with 14H structure,the orientation of 14H phase is various among different grains.The different orientations between grains may form a micro-galvanic corrosion,which results in obvious matrix dissolution at grain boundaries.

The above results show that the overall corrosion resistance of the material decreased greatly with the increase of the transformation from 18R to 14H during solution treatment.This may be due to the introduced large density of 18R/14H interfaces in the phase transition process.The transition from 18R to 14H involves complex crystallographic changes,which may also be an important factor affecting their electrochemical behavior[56].The transformation from 18R to 14H structure can be related to ledge mechanism which involves the synergistic effect of the stacking faults,lateral glide Shockley partial dislocations and short-range movement of Y/Zn solute atoms[57,75–78].The 18R/14H interface contains the transition regions where 18R continuous transform to 14H.That is also somewhere high density of stacking faults and dislocations take place.After nucleation,the growth of 14H is mainly dominated by migration of Y/Zn solute atoms.The 18R/14H interface is the place that the transformation continues and contain the Y/Zn solute atoms enriched area.The enrichment in solution atoms caused high strain and stress around the enrichment region[78],which could properly destabilize the system when they were exposed to electro-environment and then may be responsible for the poor corrosion resistance of ss-8h specimens with a large number of 18R/14H interfaces.The 18R/14H interface has complex crystallographic characteristics,including the different orientation relationships between 18R,14H andα-Mg matrix,the accumulation and movement of dislocations at the 18R/14H interface,the accumulation of stacking faults as well as the enrichment of solute atoms in the transition zone of phase transition.The synergistic interaction of these factors may be responsible for the rapid dissolution of LPSO interface containing 18R/14H in electrochemical environment.

Fig.11.Schematic diagrams of the corrosion mechanisms of Mg98.5Y1Zn0.5 alloys with 14H structure:(a)3-D corrosion morphology of the corroded surface;(b)galvanic corrosion cell;(b)corrosion path.

5.Conclusions

Microstructure evolution and the corrosion behaviors of ascast and solution-treated Mg98.5Zn1Y0.5alloys with different contents of 18R and 14H structures in 37°C SBF solution were systematically investigated in this work.The following conclusions can be drawn:

1)The as-cast Mg98.5Zn1Y0.5alloy containing only 18R struc ture owned the lowest corrosion rate of 2.78mm/year,followed by 4h and 12h solid-solution-treated alloys,which were 3.93mm/year and 5.12mm/year,respectively.The 8h-treated sample had the highest corrosion rate of 5.83mm/year.

2)Micro-galvanic cell formed on the interface betweenα-Mg matrix and LPSO phase in all alloys resulted the initial corrosion and then fertilized the further corrosion processes.Moreover,the lamellar 14H particles accelerated the corrosion process through micro-galvanic corrosion and led to the crystallographic orientation corrosion.

3)The increase in the 18R/14H interface deteriorated the corrosion resistance of the solution-treated alloys.The complex crystallographic characteristics of 18R/14H interface led to high dissolution rate of grain boundary at the transition region.

Confllict of interest statement

The authors declare no conflict of interest.

Acknowledgment

The study was supported by the National Natural Science Foundation of China(Grant Nos.51774109 and 51979099),the Fundamental Research Funds for the Central Universities(Grant No.2018B48414 and 2018B690X14),Postgraduate Research & Practice Innovation Program of Jiangsu Province(Grant No.KYCX18_0570).The Key Research and Development Project of Jiangsu Province of China(Grant No.BE2017148),Postgraduate Education Reform Project of Jiangsu Province(JGLX19_027),Natural Science Foundation of China(Grant No.51979099).