Ann Dokowsk ,Bogus?w Admzyk -Cie′slk ,Milen Korlnik ,Witold Chromi′nski ,Jiri Kusek ,Jku Cifti ,Driusz Ku,Jros?w Mizer

aFaculty of Materials Science and Engineering,Warsaw University of Technology,141 Woloska,St,Warsaw 02-507,Poland

b Department of Metals and Corrosion Engineering,University of Chemistry and Technology,Technicka 5,Prague 6 166 28,Czechia

c Institute of Materials Engineering,Silesian University of Technology,Krasi′nskiego 8,Katowice 40-019,Poland

Abstract The microstructure-dependent corrosion resistance of dual structured fine-graine Mg-7.5Li-3Al-1Zn has been investigated.The alloys were extruded using extrusion with a forward-backward rotating die (KoBo,a newly developed SPD method) at two different extrusion ratios.The fine-graine microstructures formed in the alloys were characterized,and the influenc of grain refinemen on corrosion resistance was analyzed.For fine-graine (α + β) Mg-Li alloys,a higher extrusion ratio led to more intensive grain refinement however,this relationship did not improve their corrosion resistance in a chloride-containing solution.The corrosion resistance of the alloys was mainly controlled by the refinemen of α(Mg) and β(Li),along with the distribution of second phases.The presence of MgLi2Al at grain boundaries facilitated their dissolution.

Keywords: Corrosion;Grain refinement Grain size distribution;Magnesium alloy.

1.Introduction

Dual-phase structured Mg-Li alloys have drawn extensive scientifi attention,and they have been widely investigated in terms of their microstructure and the improvement of their mechanical properties by microstructure refinemen and second-phase strengthening using plastic or thermal processing [1-9].Many attempts have been made to evaluate the mechanical properties of ultrafine-graine (α+β) Mg-Li,which have been fabricated using severe plastic deformation (SPD)methods,such as high-pressure torsion [5,10],double-change angular pressing [11],and equal-channel angular extrusion[12].The occurrence of superplasticity was reported for dualstructured Mg-Li alloys processed by equal-channel angular pressing [13,14],heavy rolling and nitrate bath annealing[15],or forging [16].Another method giving a possibility of significan grain refinemen is a combination of an extrusion with applied additional plastic deformation by reversible torsion of the die (KoBo) [17].The KoBo method allows a very high deformation to be obtained at room temperature and has been previously used for Mg-Li alloys [18,19].

There have been many trials to describe the corrosion resistance of dual-phase structured Mg-Li alloys;nevertheless,the associated corrosion mechanisms are quite poorly described in the literature.This phenomenon is due to the fact that the corrosion behavior of dual-structure Mg-Li is a complex process [20-24],and in addition to multiple microgalvanic processes caused by the presence of second-phase precipitation in the matrix [25,26],many other microstructure-dependent mechanisms proceed simultaneously [27,28].Pitting and filifor corrosion are the most observed types of corrosion in dual-structured Mg-Li alloys[22,23,29,30].It should be noted that the corrosion rate of dual-structured Mg-Li alloys is the highest when compared to singleα-structured and singleβ-structured Mg-Li [31].

Generally,the corrosion resistance of Mg-Li alloys increases with microstructure refinemen [31-34].It has been shown that grain refinemen resulting from extrusion at an extrusion ratio of 20:4:1 improved the corrosion resistance of Mg-9.29Li-0.88Ca and changed the form of corrosion from pitting to general corrosion for extruded alloys with a higher extrusion ratio [35].In addition,it has been demonstrated that grain refinemen caused by rolling supported oxide fil formation [33].Yang et al.[22] showed that in Mg-8Li-3Zn-Al in 3.5 wt.% NaCl solution,Cl-ions led to the initiation and development of corrosion pits.It has also been noted thatβ(Li),which is more active thanα(Mg) (since Li is a very active metal),played the major role in the corrosion behavior of Mg-Li alloys [23,36].Another study showed that although the Volta potential ofα(Mg) phases is much higher than that ofβ(Li) phases,pitting occurrence was preferentially related to theα(Mg) phases,but not the (β)Li phases [36].

Due to their hexagonal close-packed (hcp) crystal structure,which results in a limited number of slip systems and anisotropic deformation,Mg-alloys are hard to deform[37,38].Hence,the possibility of significan grain refinemen(to fin grains with dimensions below 10 μm) for these materials is limited,but may be accomplished under conditions of permanent destabilization of their microstructure by methods where a high deformation rate can be applied [39].Due to its additional reversible oscillating die at the end of its extrusion injection cylinder,extrusion with a forward-backward rotating die (KoBo) allows for the deformation at lower temperatures than used so far of materials that are hard to deform (i.e.Mg-based alloys),by the introduction of a large deformation until a viscoplastic material fl w phenomenon occurs in the material [40].

The KoBo is a newly developed technique among SPD methods,which may be used for the reduction of the dimensions of Mg-based alloys [41].Therefore,in this study the firs examination of microstructure-dependent corrosion in chloride-containing solution of two Mg-7.5Li-3Al-1Zn alloys extruded via a KoBo is described.

2.Methodology

2.1.Microstructural characterization

The firs step in this study was to characterize the phase compositions of the alloys.Atomic absorption spectrometry(AAS,GBC 932 Plus) was performed to calculate the Li concentration in both the analyzed materials.X-ray diffraction(XRD,Bruker D8 Advance),operating at 40 kV and 40 mA with Cu Kαradiation,was used to characterize the detailed phase composition of the alloys.The results were recorded by stepwise scanning 2θfrom 10° to 120°,with a step size of 0.02° and a count time of 10 s per step.A semi-quantitative analysis of the relative concentration of the phases present in the alloys was carried out by comparing the integrated intensities of the diffraction peaks from each of the known phases.

The microstructure of the alloys was observed using a high-resolution scanning electron microscope (SEM,Hitachi SU8000).The extruded wires were cut into 3 mm-thick slices,polished with 1200-grit and 2400-grit SiC papers,and subsequently surfaces perpendicular to the extrusion direction were polished with a low-energy Ar+ion beam milling system(Hitachi IM4000 Ion Milling System) for 6 h.Observations were carried out immediately after sample preparation.

A transmission electron microscope (TEM,JEOL JEM-1200EX) with an acceleration voltage of 120 kV was used to investigate the dislocation distribution and to characterize the details of the phases formed in the alloys.Samples in the form of thin foils were prepared using a Gatan Model 656 Dimple Grinder and a Gatan Model 691 Precision Ion Polishing System (PIPS) at 2.5 kV.The observations were carried out on cross-sections perpendicular to the extrusion direction.

2.2.Corrosion tests

Electrochemical measurements were carried out in naturally aerated,quiescent 0.1 M NaCl solution using an FAS1 Gamry potentiostat equipped with three electrodes:platinum as the counter electrode,saturated Ag/AgCl as the reference electrode,and the measured sample as the working electrode.The electrolyte was produced using analytical grade reagents and distilled water.The corrosion potential(Ecorr) was recorded for 1 h under open-circuit conditions.After immersion,potentiodynamic tests were conducted with a scan rate of 5 mV/s,starting 0.5 V below EOCPand finishin at 2 V vs.Ref.The polarization curves were fitte using Gamry Echem software in Tafel mode.The post-corrosion morphology of the samples was observed after 1 h immersion under open-circuit conditions using scanning electron microscopy (SEM,Hitachi SU8000).

To compare the corrosion resistance of the alloys,the corrosion rate (Vc) was calculated based on traditional mass loss measurements.Samples were polished up to 4000-grit SiC paper,ultrasonically cleaned in isopropanol and dried in the air.Prior to immersion,the surface area of the samples was measured,and the samples were weighted.Afterwards samples were immersed for 1 h in naturally aerated 0.1 M NaCl.After immersion,samples were dried in the air and corrosion products were removed by samples immersion in 7% solution of nitric acid (HNO3) for 15 s (as per [42,43]).The Vcwas calculated [44]:

where:Δm is a mass loss (g),t is a time of exposure (day),s is a surface area (m2).

Corrosion rate of the investigated materials were compared with the results obtained for traditionally extruded and annealed Mg-7.5Li-3Al-1Zn alloy (coarse-grained alloy described in [45]).

Fig.1.Extrusion with a forward-backward rotating die (KoBo).

To describe the role of microstructure in the corrosion resistance of both alloys,the corrosion products formed on the samples as well as surfaces with chemically removed corrosion products had been chemically removed,were observed using SEM.Cross-sectional observations were also made.

3.Deformation method and materials

The materials examined in this study were extruded using the KoBo method.As shown in Fig.1,during KoBo extrusion,the input material is punch-pressed through the matrix where the oscillations of the bilaterally rotating die occur.This specifi combination allows for extrusion under conditions of permanent microstructure destabilization by applying a high deformation rate and leads to the formation of a fine-graine microstructure (approx.＜10 μm) [40,46].The extrusion was performed under the following conditions:punch speed=0.2 mm/s,die rotation angle=8°,initial temperature of the billet=24 °C,and die oscillation frequency=5 Hz (±1 Hz).

In this study,two fine-graine Mg-7.5Li-3Al-1Zn(AZ31+7.5 wt.%Li) alloys extruded using the KoBo were investigated.The firs was extruded at an extrusion ratio of R1=40:4,giving a deformation degree ofλ1=100,while the second was extruded at an extrusion ratio of R2=40:2,givingλ2=400.

4.Results

4.1.Microstructure characterization

As Li is a chemically active metal,it was necessary to confir the Li concentration in the extruded rods.As listed in Table 1,the concentration of Li after extrusion to the fina dimension of ?4 mm slightly decreased from the initial amount of 7.5 wt.% to 7.0 wt.%,and to 5.7 wt.% after extrusion to ?2 mm.Similarly to their coarse-grained counterparts [45],both materials exhibited a dual-phased structure composed of light areas ofα(Mg) and dark areas ofβ(Li),Fig.2.The darkβ(Li) was irregularly distributed in the alloy extruded to ?4 mm,and its twisted shape (Fig.3a,b) may have been a result of the insufficien frequency of the die oscillation [41];however,this hypothesis needs to be further examined.As a result of the higher degree of deformation,a more uniform microstructure formed (Fig.2c,d).Coarse white Al,Mn-rich phases were precipitated within the matrix of both alloys (marked as P1 in Fig.2b,d).

Table 1 Li concentration obtained by AAS (wt.%).

Table 2 Semi-quantitative analysis obtained by XRD for the extruded Mg-7.5Li-3Al-1Zn.

Moreover,various nano-sized precipitates were formed in the materials,and as shown in the XRD patterns,they were identifie as Mg17Al12,AlLi,and MgLi2Al (Fig.3).To confir their presence,TEM observations were carried out.As shown in Fig.4a,small round Mg17Al12precipitates were randomly distributed in the grain interiors in the alloy extruded to ?4 mm,Fig.4a.Randomly distributed Mg17Al12precipitates,which were similar in shape,were also identifie in the alloy extruded to ?2 mm (Fig.4c).The presence of elongated MgLi2Al precipitates at the grain boundaries was also confirmed and a higher number of them were formed in the alloy extruded to ?2 mm (Fig.4d).Additionally,the existence of round,coarse precipitates is shown in Fig.4b.This kind of precipitates can be identifie as AlLi.Unfortunately,we were not able to perform SAD to confir its presence;however,based on our previous observations,we suppose that the observed shape and distribution can support the classifi cation of these precipitates as AlLi [45].Surprisingly,a few ternary precipitates of and MgLiAl2were detected in the alloy extruded to ?4 mm (Fig.4c).When comparing the TEM microstructure of the alloys extruded at the two extrusion ratios,it can be seen that a higher number of second phases precipitated in the alloy extruded to ?2 mm (Fig.5).These precipitates were formed aroundβ(Li),and they are clearly visible as tiny white threads surrounding the dark areas ofβ(Li) in Fig.2b and d.Despite the differing degrees of deformation,the relative concentrations ofα(Mg) toβ(Li) did not change significantl,and in the alloy extruded to ?4 mm the ratio was 34.1 to 63.2%,while being 27.2 to 71.2% for ?2 mm(Table 2).

4.2.Corrosion testing

Fig.2.SEM images showing microstructure of Mg-7.5Li-3Al-1Zn:(a) image showing overall microstructure of the alloy extruded to ?4 mm,(b) image showing distribution of coarse precipitates in the alloy extruded to ?4 mm,(c) image showing overall microstructure of the alloy extruded to ?2 mm,(d)image showing distribution of coarse precipitates in the alloy extruded to ?2 mm,(e) the representative EDX analyzes for precipitates labelled as P1.

Fig.3.XRD patterns obtained for the examined materials.

The electrochemical results for both the analyzed materials,performed in naturally aerated 0.1 M NaCl,are shown in Fig.6.The Ecorrevaluation during 1 h of immersion in naturally aerated 0.1 M NaCl for both analyzed materials is shown in Fig.6a.Depending on the deformation ratio,the Ecorrcurves exhibited different trends.Higher values of Ecorrwere recorded for Mg-7.5Li-3Al-1Zn extruded to ?2 mm.At the beginning of immersion,a rapid increase in Ecorrtook place,from the initial value of -1.53 V/Ref to -1.47 V/Ref.During the following 3 min,Ecorrslightly decreased,and then once again increased towards more positive values,reaching -1.45 V/Ref.Subsequently,an abrupt decline occurred to -1.52 V/Ref,the potential remaining around this value,in a nearly steady state,for the remainder of the experiment.Lower values of Ecorrin the analyzed solution,suggesting a poorer resistance to the corrosive environment due to a lower equilibrium potential that made the oxidation process easier [22,47],were registered for the Mg-7.5Li-3Al-1Zn extruded to ?4 mm.In this case,a rapid increase in Ecorroccurred at the start of the immersion,starting below-1.56 V/Ref,reaching a value of around -1.53 V/Ref after 5 min of immersion,and continuing at a near steady state just above -1.53 V/Ref for the remainder of the experiment.The observed initial rapid increase in the Ecorrvalues for the period of＜5 min of immersion is unique for Mg-based alloys,and is rather associated with “cathodic activation”than with the suppression of anodic kinetics [31,48].Also,the variations around the steady-state value of Ecorr,recorded for both alloys during the immersion,are attributed to the changes in anodic kinetics under the corrosion conditions[31,48,49].

Fig.4.TEM images in bright fiel and selected area diffraction (SAD) patterns obtained from the alloy extruded to ?4 mm (a and b),and the alloy extruded to ?2 mm (c and d).

Fig.5.TEM images of the microstructure of the alloy extruded to ?4 mm (a),and the alloy extruded to ?2 mm (b).

Fig.6.The results of electrochemical measurements performed in naturally aerated 0.1 M NaCl (a) Ecorr evaluation over 1 h,(b) potentiodynamic polarization curves.

Fig.6b shows the polarization curves for the Mg-7.5Li-3Al-1Zn alloys deformed to ?2 and ?4 mm,and the characteristic electrochemical parameters calculated from the Tafel slopes are displayed in Table 3.The cathodic branches showed a similar trend in their slopes;however,slightly higher cathodic current densities (ic) were reported for Mg-7.5Li-3Al-1Zn extruded to ?4 mm than for the material extruded to ?2 mm.This trend suggests that the rate of water reduction is higher in the case of Mg-7.5Li-3Al-1Zn extruded to ?4 mm,and may be supported by the presence of second phases that act as local cathodes [50,51].In the anodic branches,a typical infl xion point is visible,related to the breakdown of the protective oxide fil formed on the alloys’ surfaces [49,52,53].This breakdown potential is clearly visible for both alloys;however,Ebwas higher for the material extruded to ?2 mm than for the material extruded to ?4 mm (Eb=-1.24 V/Ref for Mg-7.5Li-3Al-1Zn extruded to ?4 mm,and Eb=-0.97 V/Ref for Mg-7.5Li-3Al-1Zn extruded to ?2 mm,respectively).The portion of the active area of the anodic slope,after reaching Ecorrand preceding the infl xion point (Eb),was shorter for the alloy extruded with a lower extrusion ratio (extruded to ?4 mm) than for the material extruded with a higher extrusion ratio (extruded to ?2 mm),as shown in Fig.7.The greater the differences between these two values,the higher the resistance to the localized corrosion of the respective alloy.Additionally,the alloy extruded at a lower extrusion ratio (?4 mm) presented lower corrosion current density than that extruded with a higher extrusion ratio (Table 3).

Fig.7.The difference between Ecorr and Eb calculated for both analyzed materials showing susceptibility to localized corrosion.

Table 3 The electrochemical parameters calculated from Tafel extrapolation based on results from potentiodynamic polarization tests recorded in 0.1 M NaCl.

The results for the measured mass loss rates are shown in Fig.8.The mass loss results represent the corrosion rate directly,and as presented,the extrusion of Mg-7.5Li-3Al-1Zn to ?4 mm made this alloy the most resistant to the 0.1 M NaCl.The corrosion rate after 1 h of immersion gave results of 0.02 g·day-1·cm-2,which was significantl lower that for coarse-grained Mg-7.5Li-3Al-1Zn (0.5 g·day-1·cm-2).The higher extrusion ratio increased the corrosion activity of the alloy extruded to ?2 mm,giving a corrosion rate of 0.07 g·day-1·cm-2making it still lower than that calculated for its coarse-grained counterpart but higher than that the one obtained for the KoBo extruded alloy to ?4 mm.This indicates,that among SPD extruded dual-structured Mg-Li alloys,higher grain refinemen is not capable of reducing the corrosion rate.

Fig.8.Corrosion rate calculated after 1 h of immersion in 0.1 M NaCl for conventionally extruded and annealed Mg-7.5Li-3Al-1Zn (coarse-grained),and for Mg-7.5Li-3Al-1Zn KoBo extruded to ?=4 mm and ?=2 mm.

4.3.Post-corrosion observations

Fig.9.SEM surface observations of Mg-7.5Li-3Al-1Zn extruded to ?4 mm,after 1 h immersion in naturally aerated 0.1 M NaCl:(a) image showing overall corrosion after immersion,(b) image showing localized corrosion,(c) cross-sectional observations,(d) EDX analysis performed at point A1 in panel c.

Fig.10.Cross-sectional observations of the corrosion products formed on the Mg-7.5Li-3Al-1Zn extruded to ?2 mm,after 1 h immersion in naturally aerated 0.1 M NaCl:(a) image showing overall corrosion after immersion,(b) image showing localized corrosion,(c) cross-sectional observations,(d) EDX analysis performed at point B1 in panel c.

SEM images of the surface and its chemical composition(EDX) were obtained after immersion of the samples for a period of 1 h.Figs.9 and 10 depict corrosion attack on the surface and cross-sectional observations of Mg-7.5Li-3Al-1Zn extruded to ?4 mm and to ?2 mm after immersion for 1 h in 0.1 M NaCl.The surface of both alloys suffered from an intense local corrosion attack aligned along some preferred paths (Figs.9a and 10a).The SEM images of the corroded surfaces of both alloys taken with higher magnificatio show that localized corrosion attack was intense and led to the strong dissolution of local areas (Figs.9b and 10b).The cross-sectional images of the samples after immersion under open-circuit conditions revealed that corrosion in Mg-7.5Li-3Al-1Zn extruded to ?4 mm spread apparently selectively,and propagated into the interior of the material (Fig.9c).Theα(Mg) phase had a relatively higher potential of -2.37 vs.SHE,whereasβ(Li) had a much lower potential of -3.41 V vs.SHE [30,54];therefore,in the case of dual-structured Mg-Li alloys,β(Li) was more prone to dissolve thanα(Mg).Based on this preference,it can be claimed that the extension of corrosion proceeded throughβ(Li),as marked in Fig.9c.In the case of the sample extruded with a higher extrusion ratio (to ?2 mm),corrosion spreading at the sample/solution interface led to the formation of a poorly protective layer with a width of＜3 μm and was also accompanied by selective and deep local damage to the material (Fig.10c).EDX analyzes shown in Figs.9d and 10d revealed that the corrosion products formed on the surfaces of both alloys were enriched in Mg and Al,suggesting that besides conglomerates with the needle-like crystals typical for brucite(shown in the inset in Fig.10b),other compounds enriched in Al were present,which may support better passivation of the Mg-Li alloys [55,56].Ni and Cu detected during EDX analysis were present because prior to the observations the samples were coated with Ni-Cu.

After the corrosion products were removed,it was clearly seen that the corroded areas could be identifie with grain boundaries.There are two reasons of this:microgalvanic corrosion would be expected to occur between MgLi2Al and the grain boundary,and the grain boundaries are more prone to serve as an anode with respect to the cathodic grain interior because of their higher distortion (Fig.11).As shown in Fig.11b,the intergranular corrosion extends to intragranular locations,and the attack led to the dissolution ofβ(Li).Clearly,more intensive corrosion occurred at the surface of the material deformed at a higher extrusion ratio.

5.Discussion

Fig.11.Images taken after corrosion products were removed:(a) Mg-7.5Li-3Al-1Zn extruded to ?4 mm,(b) Mg-7.5Li-3Al-1Zn extruded to ?2 mm.

Fig.12.The comparative images of the surfaces.

The results obtained in this study gave a new insight into the corrosion of dual-structured Mg-Li alloys,which contributed to a better understanding of this process.Our combined microstructural,electrochemical,and analytical study highlighted the differences between the spreading corrosion mechanisms and evaluated the role of various microstructural features in their propagation.It would be expected that higher grain refinemen would give a positive influenc on the corrosion resistance of metallic materials in alkaline and near neutral solutions [35,57].For AZ31B alloy,the effect of grain size played a significan role in the case of an untwined microstructure,for which the corrosion rate was higher when the grain size increased from 65 to 250 μm [58].Also,Li et al.[29] described the corrosion resistance of Mg-5Li alloy,which was improved due to more refine grains.The presented study shows that declaring a definit ve relationship between the grain size and corrosion resistance of metallic materials,especially dual-structured ones,may be an unjustifie interpretation.Grain refinemen via KoBo extrusion improved the corrosion resistance of dual-structured alloys when compared to their coarse-grained counterparts.Among group of fine-graine (α+β) Mg-Li alloys,a higher extrusion ratio led to more intensive grain refinement however,this relationship did not improve their corrosion resistance in chloridecontaining solution.This phenomenon is attributed to two critical factors.First of all,corrosion attack on both the analyzed alloys with fine-graine structure started at the grain boundaries where MgLi2Al was formed.The more refine structure possessed a higher number of grain boundaries with precipitated MgLi2Al where the corrosion may be initialized,and the reactions occurring at grain boundaries were not compensated by reactions leading to the formation of a protective layer.Secondly,an important findin is that the corrosion process is extended throughβ(Li),which is more active thanα(Mg).A microgalvanic coupling betweenβ(Li) andα(Mg)has been previously reported [31,59].In our previous study,we claimed that the corrosion resistance of dual-phased conventionally extruded Mg-7.5Li-3Al-1Zn alloys depends on theα(Mg) toβ(Li) area ratio,where one acts as a cathode and drives the local system to a state of equilibrium [45].This effect forces the second phase to actively corrode,and therefore the amount ofβ(Li) andα(Mg) in the case of coarsegrained dual-structured Mg-Li alloys plays a predominant role in their corrosion mechanisms.In the case of fine-graine dual-structured Mg-Li alloys,regardless of the degree of deformation,the ratio of the relative concentration ofα(Mg) toβ(Li)did not change significantl,and taking this into consideration,it is reasonable to assume that grain refinement and the results of this refinemen on both phases are major factors influencin the corrosion mechanisms in the examined alloys.

Fig.12 summarizes the microgalvanic corrosion mechanisms in the Mg-7.5Li-3Al-1Zn alloys extruded to ?4 and ?2 mm.The microgalvanic processes in dual-phase Mg-Li alloys are mainly controlled by the anodic to cathodic area ratio ofβ(Li)α(Mg) [45].Nevertheless,at the beginning of the process,in stage I,due to either the formation of a galvanic cell between the MgLi2Al precipitate and grain boundary,or a higher distortion of the grain boundary when compared to the grain interiors,the most preferential corrosion sites were grain boundaries (marked with the red arrow in Fig.12).Although at the beginning of immersion more sites of corrosion initiation were present on the alloy with higher grain refinement the predominant corrosion mechanism occurred during stage II,where microgalvanic corrosion proceeded betweenα(Mg) andβ(Li),and intergranular corrosion was supplemented by an intragranular process.In this case,microgalvanic corrosion was controlled by the area ratio of cathodic to anodic grain interiors.As commonly known for heterogeneous microstructures,the corrosion system was composed of half-cell electrodes,where the phase with the lower potential was anodic,and the phase with higher potential cathodic.Therefore,after reaching stage I,anodic dissolution ofβ(Li) was observed.Moreover,the higher refinemen of both phases resulted in a higher number of microgalvanically coupledα(Mg)/β(Li) grains with a lower area ratio than in the material where the grains were larger.In the latter case the number of microgalvanically coupledα(Mg)/β(Li)grains was lower,and the area ratio ofα(Mg)/β(Li) was higher.This trend influence the spatial distribution of the cathodic and anodic reactions,resulting in a higher degree of penetration for the alloy with the more refine grains.

6.Conclusion

In this study,through observing the microstructure and investigating the corrosion behavior of KoBo extruded Mg-7.5Li-3Al-1Zn alloys extruded to ?4 and ?2 mm,the effects ofα(Mg) andβ(Li) distribution,their relative ratio,and the role of second phases have been examined.The following conclusions can be drawn:

?The KoBo extrusion to the dimension of ?4 mm led to the formation of non-uniform fine-graine microstructure in dual-structured Mg-7.5Li-3Al-1Zn alloys.The increase in the extrusion ratio caused the formation of homogenous fine-graine dual-structured microstructure.

?In the case of fine-graine dual-structured Mg-Li alloys,the main roles in the corrosion mechanism were played by the refinemen ofα(Mg) andβ(Li) and the distribution of second phases.The presence of MgLi2Al at grain boundaries facilitated their dissolution.

?For fine-graine (α+β) Mg-Li alloys,a higher extrusion ratio led to more intensive grain refinement however,this relationship did not improve their corrosion resistance in chloride-containing solution.