Simultaneous enhancement of mechanical properties and corrosion resistance of as-cast Mg-5Zn via microstructural modification by friction stir processing

Fei Long ,Gaoqiang Chen ,Mengran Zhou ,Qingyu Shi ,Qu Liu,∗

aState Key Laboratory of Tribology,Tsinghua University,Beijing 100084,China

b Department of Mechanical Engineering,Tsinghua University,Beijing 100084,China

Abstract Magnesium alloys are ideal lightweight materials;however,their applications are extremely limited due to their low strength,poor ductility,and weak corrosion resistance.In the present study,a friction stir processing (FSP) treatment was employed to optimize the mechanical properties and corrosion resistance of an as-cast Mg-5Zn alloy.The average grain size of the Mg-5Zn alloy was refined from 133.8 μm to 1.3 μm as a result of FSP.Along different directions,FSP exhibited the enhancement effects on different mechanical properties.Furthermore,according to the potentiodynamic polarization results,the corrosion current density at the free-corrosion potential of the FSPed sample,was 4.1 × 10-6 A/cm2 in 3.5 wt.% NaCl aqueous solution,which was significantly lower than that of the as-cast sample.Electrochemical impedance spectroscopy revealed that the polarization impedance,Rp,of the FSPed sample was 1534 Ω/cm2 in 3.5 wt.% NaCl aqueous solution,which was 71.4% greater than that of the as-cast sample.The corrosion morphology of the FSPed sample in 3.5 wt.% NaCl aqueous solution exhibited largely uniform corrosion,rather than severe localized corrosion characteristics,which further reduced the corrosion depth on the basis of reducing the corrosion current density.The results presented herein indicate that FSP is a viable technique for simultaneously improving the mechanical properties and corrosion resistance of the as-cast Mg-5Zn alloy.

Keywords: Friction stir processing;Magnesium alloy;Corrosion mode;Texture softening;Anisotropy in mechanical properties.

1.Introduction

Magnesium (Mg) accounts for approximately two percent by mass of the Earth’s crust,and it is the eighth most abundant element in the Earth’s crust.Owing to the advantages of high specific modulus and superior specific strength,it has been generally acknowledged as an ideal lightweight structural material.In particular,Mg has good mechanical properties,and is extensively used in the manufacturing of parts and instruments.With the current implementation of the green manufacturing philosophy and the development of lightweight innovations among automobiles,Mg alloys,which are nontoxic and harmless to humans and the environment,have emerged as one of the most promising metals.In addition,the density and elastic modulus of Mg are similar to human bone,and this metal exhibits good biocompatibility.As early as 1878,Witte [1] began using Mg wires to ligate blood vessels for hemostasis.In 2015,Zhao et al.[2] adopted pure Mg screws as implants for osteonecrosis treatment.However,the mechanical properties and corrosion resistance of pure Mg prevent further applications as biomedical materials and lightweight materials.Considering the insufficient mechanical properties of pure Mg,various researchers have investigated Mg alloying.However,many Mg alloys have low strength,poor ductility,and weak corrosion resistance,which significantly restrict their commercial application.

Extensive research efforts have been directed towards addressing the limitations of Mg alloys.Although researchers have developed ways to improve the materials’strengths(e.g.,solution strengthening,aging strengthening),these enhancements often have adverse consequences in term of ductility.For example,Gao et al.[3] dissolved Gd into a Mg alloy and found that as the Gd content increased,the ultimate tensile strength (UTS) also increased,whereas the elongation decreased.In another experiment,Liu et al.[4] aged an as-extruded Mg-1.5Gd-1.3Y-0.6Zn-0.6Mn alloy and observed that within 150 h of the aging treatment,the material’s elongation was inferior to that of the alloy in its original state.Thus,researchers have explored ways to simultaneously increase an alloy’s strength and ductility through fine grain processing [5–10].For example,Ma et al.[7] refined the ZE41A alloy grains to 1.5 μm via equal channel angular pressing(ECAP),thereby improving both its ductility and strength.Meanwhile,researchers have widely reported on the optimization of Mg alloys’ corrosion resistance.They concentrated on the development of coatings for these alloys to increase the corrosion resistance,and they achieved good results [11,12].However,the coating merely promoted corrosion resistance on the surface.All alloying elements could potentially form micro-galvanic couples between different phases through the generation of precipitates,which cause local pitting.For Mg alloys containing precipitates,once the coating is destroyed,large areas of pitting resulting from the precipitates and the subsequent failure of the workpiece have not been fundamentally inhibited;this has become a technical barrier that hinders further applications of Mg alloys.Therefore,it is crucial to develop methods for improving the corrosion resistance of Mg alloys through microstructural modifications [13–15].

Researchers have used several strategies to regulate and control the corrosion resistance of Mg alloys containing certain precipitates.Among these methods,grain refinement technology comprises one of the most typical approaches,as it enables simultaneous improvements in strength,ductility,and corrosion resistance.However,this method has inconsistent effects on corrosiveness;corrosion resistance is reduced in some cases,and it may be improved in other cases,while being accompanied by a variety of issues.For example,Song et al.[16]treated an as-cast AZ91D alloy with ECAP technology to obtain an ultrafine-grainedα-phase matrix and refinedβ-phase particles,which reduced the corrosion resistance of the alloy.Coy et al.[17] applied excimer laser cladding technology to treat the surface of a die-cast AZ91D Mg alloy and obtained a highly uniform and fine microstructure.Many precipitates were dissolved into the matrix of the AZ91D alloy,and the impedance of the treated alloy in 3.5 wt.% NaCl aqueous solution increased significantly.However,some hole defects observed in the processing area and microcracks between the overlapping layers increased the corrosion sensitivity from another perspective.Laleh and Kargar [18] applied a surface mechanical attrition treatment (SMAT) to refine the surface of an AZ91 Mg alloy.After the surface treatment with an appropriately-sized ball,the alloy’s corrosion resistance was optimized.It was proposed that this was associated with the re-dissolution of theβ-Mg17Al12phase.However,the maximum thickness of the surface layer obtained via SMAT was only 150 μm.

Recently,researchers have explored the optimization of Mg alloys’ properties via friction stir processing (FSP),which is a severe plastic deformation (SPD) technology based on friction stir welding (FSW).Compared with similar technologies(e.g.,ECAP and high-pressure torsion (HPT)),FSP involves relatively simple processing parameters,uncomplicated procedures,lower costs,and fewer restrictions on the shape and size of the workpieces.To date,many researchers have conducted studies to understand the effect of FSP/FSW on the microstructure and mechanical properties of Mg alloys [19–29].However,there are limited studies exploring the effect of FSP on the corrosion behavior of Mg alloys [30–35].Several studies [30–33] have demonstrated that the etch pits of Mg alloys processed by FSP exhibited significant reductions in pitting sizes.These reductions imply a wide range of potential applications.However,some alloys exhibited deteriorated corrosion resistance after FSP treatment.For example,the FSP treatment employed by Seifiyan et al.[34] reduced the electrode potential and increased theIcorr(the corrosion current density at the free-corrosion potential) of the AZ31 Mg alloy,that is,FSP deteriorated the alloy’s corrosion resistance.Liu et al.[35] conducted FSP on a Mg-9Li-1Zn alloy and found that theIcorrdecreased,while its electrode potential showed an abnormal decrease.In general,the unpredictable corrosion resistance of Mg alloys treated by FSP revealed a relatively complicated relationship that could not adequately be described by a universal law.

In the context of biomedical materials,aluminum and some rare-earth elements generate substances that are harmful to organisms during their degradation processes [36–41].As a result,the alloying elements available for biomedical Mg alloys are limited.Zinc is one of the most abundant elements in the human body,and it exhibits high biocompatibility [42].According to results presented by Li et al.[43],the addition of Zn enhanced the mechanical properties of a Mg alloy.Haferkamp et al.[44] reported that the corrosion rate of Mg alloys could be reduced by increasing the mass fraction of Zn in the Mg alloy.Thus,Mg-Zn alloys broaden the application prospects of Mg for biomedical materials.At the same time,the density of a Mg-5Zn alloy is only about 1.8/cm3,which makes it a nontoxic lightweight material [45–47].The maximum solid solubility of Zn inα-Mg is 1.6 wt.% at room temperature [48].The Mg-5Zn alloy is a typical Mg alloy that contains precipitates and has attracted significant attention from researchers.However,to date,the effects of FSP on Mg-5Zn have not been examined.In the as-cast Mg-5Zn alloy,both the grain and the precipitates are very coarse.Typically,the average grain size would be over 100 μm.Therefore,it is worth reporting the microstructural modification,such as the grain refinement and the precipitates breaking/dissolution,of the Mg-5Zn alloy by FSP and the subsequent enhancement of mechanical properties and corrosion resistance.

To this end,the present study evaluated the strength,elongation,and corrosion resistance of friction stir processed(FSPed)Mg-5Zn alloy and explored the mechanism governing the changes in properties based on microstructural analysis.

2.Experimental procedures

2.1. Sample preparation

The Mg-5Zn alloy was smelted according to the nominal composition presented in Table 1 and cast into a cuboid billet.A flat plate (thickness=5 mm) was cut from the billet for FSP.The pin used in the present study was a circular truncated cone with an M6 standard thread having a pitch of 1 mm.The bottom and top diameters of the pin were 6 mm and 4 mm respectively,and the length between the faces was 4.7 mm.The tool’s concave shoulder measured 14 mm in diameter.One-pass FSP was conducted at a tool rotation rate of 800 rpm,with a traverse speed of 40 mm/min (hereafter denoted as “800/40′′) and a tilt angle of 2.55° The stirring pin rotated in counter-clockwise direction.Tests of the mechanical and corrosion properties were performed on samples cut from the stir zone.

Table 1Nominal composition of cast alloy (wt.%).

2.2. Tensile tests

Tensile tests were performed on two types of specimens at room temperature: (1) a gauge with dimensions of 25 mm(length)×6 mm(width)×3 mm(thickness)cut from the stir zone parallel to the processing direction (PD),using a strain rate of 1 × 10-3s-1at room temperature;(2) a gauge with dimensions of 2 mm×2 mm×2 mm was machined perpendicular to the PD (i.e.,transverse direction (TD),as shown in Fig.1,where the gauge was completely within the stir zone),using a strain rate of 1 × 10-3s-1at room temperature.

Fig.1.Schematic diagram of tensile specimens along the transverse direction.

2.3. Potentiodynamic polarization tests

Potentiodynamic polarization (PDP) curves were recorded using a Zennium electrochemical system (Zahner,Germany)at room temperature (25 ± 0.5 °C) in 3.5 wt.% NaCl aqueous solution.The scanning rate was 1 mV/s,over the potential range from–1620 to–1020 mV.The cell comprised a working electrode (Mg alloy sample),a counter electrode (Pt electrode),and a reference electrode (saturated calomel electrode).Each measurement was performed after immersing the sample in 3.5 wt.% NaCl aqueous solution for 3600 s.

2.4. Electrochemical impedance spectroscopy

Electrochemical impedance spectroscopy (EIS) was obtained in 3.5 wt.% NaCl aqueous solution at room temperature (25 ± 0.5 °C) using a Zennium electrochemical system(Zahner,Germany).A three-electrode system was employed,using the Mg alloy as the working electrode,a Pt foil as the counter electrode,and a saturated calomel electrode as the reference electrode.EIS spectra were assessed at open circuit potential,with a frequency in the range of 10-2–105Hz and sinusoidal perturbation.The amplitude of the alternating current (AC) signal was 5 mV.The impedance measurement for each sample was carried out for 30 min after 1800s of immersion in 3.5 wt.% NaCl aqueous solution.The obtained data were analyzed and fitted using ZView software.

2.5. Microstructural characterization

An optical microscope (OM;Olympus) was used to examine the microstructure of the cross-section of the entire shoulder-effect area of the FSPed Mg-5Zn alloy.The grain size of as-cast Mg-5Zn alloy was statistically analyzed based on electron back-scattered diffraction (EBSD).The morphology and distribution of precipitates in the cross-section of all Mg alloy samples were observed by using a field emissionscanning electron microscope (FE-SEM;Zeiss) in the backscattered electron (BSE) mode.The present study also employed EBSD to measure the average grain size in the stir zone of the Mg alloy cross-section after FSP.In addition,the X-ray diffractometer,which was equipped with a scintillation detector and a rotating Cu target,was utilized to analyze the phase composition of all Mg alloys within an angular range of 20°to 90°The X-ray diffraction(XRD)results were analyzed by using Jade software.After immersion tests,the corrosion products of all Mg alloys were removed.Then,the corrosion morphology in the cross-section of the alloys was observed by SEM.The detailed procedure was as follows: after immersion in 3.5 wt.% NaCl aqueous solution for 24 h,the corrosion product film was removed from the samples by dipping the samples into a solution of 25 wt.% CrO3–1 wt.%AgNO3for 10 min.After removing the corrosion products,the samples were observed by SEM.

3.Results and discussion

3.1. Tensile properties

Mg alloys with hexagonal close-packed (HCP) structure often show strong anisotropy in mechanical behaviors after SPD.In order to evaluate the anisotropy of the FSPed Mg-5Zn alloy’s tensile properties along different directions,tensile tests were performed along the PD and TD,and the results are shown in Fig.2.These results indicated that the mechanical properties of the FSPed Mg-5Zn alloy exhibited significant anisotropy along the PD and TD.Along the PD,there was a slight increase (10.9%) in yield strength (YS).The UTS and elongation reached 214 MPa and 11%,corresponding to 21.6% and 61.5% improvements,respectively.Along the TD,the YS and UTS reached 152 MPa and 193 MPa,achieving enhancements of 65.2% and 9.7%,respectively.It should be noted that the elongation of the FSPed-TD sample was lower than that of the corresponding as-cast sample.

Fig.2.Mechanical properties of the as-cast and FSPed Mg-5Zn samples.

3.2. Electrochemical tests

Fig.3 shows the PDP curves of the as-cast and FSPed Mg-5Zn samples after 60 min of immersion in 3.5 wt.% NaCl aqueous solution.There were two reaction stages (AB and BC) identified on the PDP curves.The first stage,AB,repre sents the cathode reaction stage,and BC represents the anode reaction stage;overall,AB/BC represent different corrosion processes.The chemical reaction occurring during AB was mainly the hydrogen evolution reaction in the aqueous solution.In this case,H2O molecules accept electrons to generate OH–ions and release H2gas.Subsequently,the chemical reactions occurring during BC mainly included the anodic dissolution of Mg and the abnormal anodic hydrogen evolution,which is known as the negative difference effect (NDE).The potential corresponding to the turning point (B) is called the self-corrosion potential (Ecorr).Notably,when the applied potential or current density was sufficiently negative,the anodic dissolution of Mg was extremely low,almost diminished to zero.In this case,the cathodic reaction dominated,whereas the anodic reaction was very weak,as described by Song and Atrens[49].However,since corrosion under the self-corrosion potential was the result of the equilibrium between the anode and cathode reactions,the current magnitudes in the cathode and anode zones were not considered to reflect the actual corrosion rates.

Fig.3.PDP curves of the as-cast and FSPed Mg-5Zn after 60 min of immersion in 3.5 wt.% NaCl aqueous solution.

According to Song et al.[50],the anodic polarization behavior of the Mg alloy was complicated,and there was no Tafel zone.In contrast,in the cathode zone,the hydrogen evolution reaction was relatively simple,and there was a Tafel zone.The Tafel extrapolation of the cathodic branch of the PDP curve is commonly used to assess the corrosion rate of Mg,and numerous researchers have applied this method[51,52].

Owing to the complicated mechanism involved with the anode region,only a small portion of the curves in the anode region,and all curves in the cathode region,were considered in this study,as presented in Fig.3.

The free-corrosion potential,Ecorr,relative to the two states of the Mg-5Zn alloy,and the corrosion current density at the free-corrosion potential,Icorr,as calculated by a Tafel extrapolation of the cathodic branch of the PDP curve,are presented in Table 2.

Table 2Fitting results for the PDP curves of the Mg-5Zn alloy in different states in 3.5 wt.% NaCl aqueous solution.

According to Table 2,the free-corrosion potential of the Mg-5Zn alloy increased after FSP,and the corrosion current density decreased.This indicated that the corrosion resistance improved as a result of FSP.The current density of the Mg-5Zn alloy in the cathode reaction stage after FSP was one order of magnitude lower than that before FSP.This implied that FSP inhibited the cathode reaction kinetics of the Mg alloy,i.e.,FSP reduced the cathodic hydrogen evolution reaction.

After immersion in 3.5 wt.% NaCl aqueous solution for 30 min,EIS of the as-cast and FSPed Mg-5Zn alloys were obtained at the open circuit potential,and the results are presented as a Nyquist diagram (Fig.4a) and a Bode diagram(Fig.4b).These EIS data revealed that the Mg-5Zn alloys before and after FSP treatment had similar patterns along the middle-frequency and high-frequency regions,whereas there were significant differences in the low-frequency region.For example,Fig.4a shows a large semicircular capacitive loop in the high-frequency region,whereas a small capacitive loop emerged in the medium-frequency region.The large capacitive loop primarily reflected the charge transfer process,during which Mg lost electrons,thus generating Mg2+ions in the electric double-layer (EDL) at the interface between the alloy and the solution.The diameter of the capacitive loop is equal to the charge transfer of the test alloy.The small capacitive loop was associated with the corrosion product layer formed on the alloy surface during corrosion,and its diameter is equal to the impedance of the corrosion product layer.

Fig.4.EIS of the Mg-5Zn alloy in 3.5 wt.% NaCl aqueous solution: (a) Nyquist diagram;(b) Bode diagram.

It should be noted that in the low-frequency region,the ascast Mg-5Zn alloy exhibited an inductive loop;however,after FSP,the low-frequency inductive loop almost disappeared.According to some literature reports [30,53,54],an inductive loop in the low-frequency region is associated with the breakdown of the Mg alloy’s corrosion products and the transient intermediate products adsorbed on the new alloy surface.

Based on the EIS characteristics,the equivalent circuits(Fig.5)were constructed and used to quantitatively fit and analyze the electrochemical impedance response on the surface of the Mg-5Zn alloys immersed in 3.5 wt.% NaCl aqueous solution before and after FSP.The circuit in Fig.5a corresponds to the as-cast alloy,and that in Fig.5b corresponds to the alloy after FSP.In the diagrams,the physical meanings of each component in the equivalent circuits are given here:Rsrepresents the impedance of the solution;R1indicates the charge transfer impedance of the EDL at the interface of the corrosion products;CPE1signifies the EDL capacitance at the interface of the corrosion products [55];R2indicates the charge transfer impedance of the EDL at the interface between the alloy and the solution;CPE2represents the EDL capacitance at the EDL interface;L1is an inductive component that is used to simulate the changes caused by the expansion of the active anode region;R3represents the resistance relative to the local environmental changes near the anode.

Fig.5.Equivalent circuits describing the electrochemical impedance response on the surface of the Mg-5Zn alloy: (a) As-cast alloy;(b) FSPed alloy.

Since the electrochemical interface involved in the corrosion process typically does not exhibit ideal pure capacitance,a constant phase element (CPE) is commonly used to replace the capacitance element in the equivalent circuit.This component can be used to describe the inhomogeneity of the electrochemical reactions on the electrode surface,and its impedance value can be calculated using Eq.(1),

whereiis an imaginary number,ωis the angular frequency,Trepresents the magnitude of the CPE,andnis the exponential term of the CPE (usually between 0 and 1),which is used to evaluate the degree of compression of the impedance spectrum arc.Whenn=1,the CPE is a pure capacitor;whenn=0,the CPE is a pure resistor [56,57].

Since the Mg alloy represents a corrosion system controlled by activation polarization,the concentration polarization aspect can be ignored.In this case,the Faraday impedance of the Mg alloy can be considered equivalent to the polarization resistance,Rp,which therefore also represents the corrosion rate and resistance of the Mg alloy.When the frequency of the AC signal infinitely tends to zero,Rpcan be calculated using Eqs.(2) and (3) (Fig.5a,5b):

In general,a higher Rpindicates a more difficult charge transfer and a lower alloy corrosion rate of the alloy.Therefore,alloy corrosion resistance rates can be assessed by comparing the magnitudes of their Rpvalues.

The fitting results are presented in Table 3,and the Rpvalues are calculated using Eqs.(2) and (3).The obtained fitting results are presented in Fig.6.

Fig.6.Impedance of the as-cast and FSPed Mg-5Zn alloys in 3.5 wt.%NaCl aqueous solution.

Table 3Fitting results based on the EIS impedance spectra of the Mg-5Zn alloy.

According to Fig.6,the impedance of the Mg-5Zn alloy in 3.5 wt.% NaCl aqueous solution increased significantly after FSP,indicating that the corrosion resistance of the material was improved.

3.3. Microstructural analysis

In order to analyze the reason for the improvements in the strength,elongation,and corrosion resistance of the Mg-5Zn alloy after FSP,the microstructure of the Mg-5Zn alloy after FSP was examined.The cross-section of the entire shouldereffect area was cut and observed with an OM,as shown in Fig.7.

Fig.7.Microstructure of Mg-5Zn alloy after FSP treatment: (a) Cross-sectional micrograph of the FSPed region;(b) Micrograph of the heat affected zone;(c) Micrograph of the stir zone.

As can be seen in Fig.7,after FSP treatment,the grain size was significantly refined.

The grain parameters (GPs) and grain boundaries (GBs) of the as-cast Mg-5Zn alloy were observed using EBSD,and this material’s grain size distribution was analyzed.The results are presented in Fig.8,which indicates that the grains of the ascast Mg-5Zn alloy were coarse and unevenly distributed.The average grain size in the as-cast Mg-5Zn alloy was 133.8 μm,in particular,some grains even exceeded 300 μm in size.EBSD was also used to analyze the Mg-5Zn alloy after FSP.The inverse pole figure(IPF)map of the stir zone is presented in Fig.9a.

Fig.8.EBSD results for the as-cast Mg-5Zn alloy sample: (a) Grain parameter and grain boundary maps;(b) Grain size distribution.

Fig.9.EBSD results for the stir zone of the FSPed Mg-5Zn alloy sample: (a) IPF map;(b) Grain size distribution (ND=normal direction).

Compared with the as-cast alloy’s microstructure,the microstructure of Mg-5Zn after FSP revealed significant grain refinement in the stir zone,where the majority of the grains were equiaxed.The grains were uniformly dispersed throughout the microstructure,and almost all of the grains were smaller than 5 μm in diameter;the average grain size in the Mg-5Zn alloy after FSP was 1.3 μm,which is two orders of magnitude smaller than that in the as-cast Mg-5Zn alloy.

The nature of the texture in the FSPed Mg-5Zn alloy was confirmed based on pole figures (PFs) as showed in Figs.9 and 10.As can be seen from both the IPF in Fig.9 and the PFs in Fig.10,the stir zone of the FSPed Mg-Zn alloy had a strong (0001) texture.

Fig.10.Pole figures for the stir zone of the FSPed Mg-5Zn alloy.

Fig.11 shows the image quality (IQ) maps and grain boundary misorientation distribution of the FSPed Mg-5Zn alloy.It can be seen from Fig.11 that the FSPed Mg-5Zn alloy was dominated by high-angle grain boundaries(HAGBs),whereas low-angle grain boundaries(LAGBs),consisting mainly of dislocations and sub-grain boundaries,accounted for a relatively low percentage.

Fig.11.(a) Image quality maps and (b) grain boundary misorientation distribution of FSPed Mg-5Zn alloy.

Fig.12 presents the kernel average misorientation (KAM)map for the FSPed Mg-5Zn alloy.KAM maps can reflect the homogeneity degree in terms of plastic deformation;in a KAM map,higher values indicate higher dislocation density and higher strain[58].Based on Fig.12 and previous analysis,it can be known that although the grains of the FSPed Mg-5Zn alloy were significantly refined and the density of the grain boundaries increased significantly,the dislocation density was relatively low.

Fig.12.Kernel average misorientation map of the FSPed Mg-5Zn alloy.

Fig.13a shows the microstructure of the as-cast Mg-5Zn.In the BSE micrographs,the elements with higher atomic numbers appear bright.Therefore,the dark portions in the images represent theα-Mg matrix,while the bright portions represent the precipitates.It is clear from Fig.13a that the ascast Mg-5Zn alloy contained numerous dendritic precipitates distributed at the grain boundaries.

Fig.13.BSE images of the Mg-5Zn alloy: (a) as-cast;(b) FSPed.

Fig.13b presents a BSE micrograph of the Mg-5Zn alloy after FSP.Notably,the FSP treatment significantly altered the microstructure of the as-cast alloy.Large precipitates were broken down into refined and uniformly distributed precipitates.

The XRD patterns of the as-cast and FSPed Mg-5Zn are presented in Fig.14.The microstructure of this alloy mainly comprisedα-Mg,Mg7Zn3,and MgZn2,and the content of Mg-Zn compounds decreased significantly after FSP.These results indicated that FSP did not alter the phase composition of the Mg-5Zn alloy;furthermore,the precipitates in the Mg-5Zn alloy after FSP were re-dissolved,which is consistent with several literature reports [59–62].

Fig.14.XRD patterns of the as-cast and FSPed Mg-5Zn alloys.

Based on the results present in Figs.6–14,as a result of FSP,there was a significant grain refinement in the Mg-5Zn alloy,the breakage and uniform dispersion of the precipitates.Moreover,FSP led to a dissolution of partial precipitates and significant anisotropy (texture).All of these factors significantly influenced the properties of the Mg-5Zn alloy.

According to previous studies [63,64],the mechanical properties of FSPed alloys are mainly affected by the following three factors: fine grain strengthening,precipitate dissolution,and texture.Fine grain strengthening (Hall-Petch theory)often plays a positive role in enhancing the strength and elongation of alloys,whereas precipitates dissolution often has a negative effect on the YS of alloys and a positive effect on the elongation of alloys [65,66].The specific effect of texture is relatively complex.According to the Hall-Petch formula,the significant refinement of grains and precipitate dissolution by FSP has a significant effect in terms of improving the YS[67].However,for FSPed Mg alloys,the YS is often lower than that of the coarse-grained base metal,thus suggesting that the effect of the texture on the mechanical properties of FSPed Mg alloys is significant [35].The term “texture softening” is often used to explain the role of texture in affecting a material’s mechanical properties [32,68,69].

Considering the Hall-Petch formula,the impact of fine grain strengthening should be obvious.However,the YS along the PD only increased by 10.9%,while the YS along the TD increased by 65.2%.Meanwhile,it can be seen from both the IPF map in Fig.9a and the PF in Fig.10,the stir zone of the FSPed alloy had a strong (0001) texture.This indicated that the strong (0001) texture played a significant role in weakening the YS along the PD.Along both the PD and TD directions,the combined effects of fine grain strengthening and texture overcame the negative effect of precipitate dissolution,so the YSs are increased.In addition,FSP effectively broke and refined the large dendritic second phase,such that there was no large dendritic second phase present in the FSPed alloy structure;this enhanced the effect of dispersion strengthening.

For elongation of Mg alloys,it has been reported that FSP treatment could significantly enhance the elongation of as-cast Mg alloys [70].Ma et al.[70] attributed the enhancement of Mg alloys’ elongation by FSP to precipitate dissolution and improved microstructural homogeneity.From the structural changes in the FSPed Mg-5Zn alloy,the second phase of the FSPed Mg-5Zn alloy was broken,and its distribution was more refined and uniform.This partially inhibited the dislocation pile-up in the single second phase.In this case,the hindrance of this phase to the dislocation motion was weakened.Meanwhile,the stress concentration was dispersed,and the ductility of the material was enhanced.All of these factors were also observed by Orozco-Caballero et al.[71].However,we found that the elongation of the FSPed Mg-5Zn alloy along the PD was significantly higher (61.5%) than that of the as-cast alloy,while the elongation of the FSPed Mg-5Zn alloy along the TD was lower than that of the as-cast alloy.These results indicated that the texture had a significant effect on the elongation of FSPed Mg-5Zn.Based on the EBSD results,it is clear that the stir zone of the FSPed Mg-5Zn alloy has a strong (0001) texture.Since (0001) belongs to the basal slip system,we consider that the basal slip would be more easily activated along the PD.Regarding the anomaly that the elongation of FSPed-TD Mg-5Zn was lower than that of the as-cast sample,according to our analysis,there may be data with higher yield strength and lower elongation along the TD under the effect of (0001) strong texture compared to the data along the PD.In some cases,The elongation along the TD is even lower than that of the as-cast state.This is similar to the phenomenon observed by other researchers [72–74].

Fine grain strengthening played a major role in increasing the UTS;however,there was an anomaly that YSTD>YSPDbut UTSTD

As far as the corrosion resistance of Mg alloys is concerned,it is well known that grain size,dislocation density,texture,the degree of galvanic corrosion,and other factors all have a certain influence.Theoretically,as the grains become finer,more grain boundaries are created.A structure with a high density of grain boundaries makes the Mg alloy more susceptible to corrosion than a structure with a low density of grain boundaries,i.e.,it reduces the corrosion resistance of Mg alloys.However,other microstructural changes often coincide with changes in grain size and they may have more significant effects on the corrosion behavior of Mg alloys[81].In the present study,the as-cast Mg-5Zn alloy grains were refined by two orders of magnitude after FSP treatment,but the FSPed Mg-5Zn alloy still exhibited much improved corrosion resistance in 3.5 wt.% NaCl aqueous solution.This indicated that the effect of the grain size (grain boundary density) on the corrosion resistance was weak relative to the effect of other factors.

Numerous studies [82–86] have shown that the effect of the dislocation density on the corrosion resistance of Mg alloys is significant.Song [85] proposed that a large number of dislocations was introduced at the grain boundary of the AZ31 alloy by hot rolling,which adversely affected the corrosion resistance.Hamu et al.[86] also observed that multipass ECAP greatly increased the dislocation density and thus accelerated the corrosion of AZ31.With regard to the straininduced changes in dislocation density caused by FSP,it can be seen from Fig.11 that the grain boundaries of the alloy after FSP were mainly HAGBs.This indicated that recrystallization occurred relatively completely and the dislocation density was relatively low.The KAM diagram in Fig.12 also illustrated this phenomenon.The strain-induced change in dislocation density introduced by FSP did not significantly deteriorate the corrosion resistance of Mg alloys,as also suggested by Liu et al.[30].

At a fundamental level,the Mg alloy texture influences the corrosion resistance because the atomic density of adjacent crystal planes is significantly different [85].For HCP metal materials,the atomic density of the basal plane (0001) is the highest,followed by the prism planes {1120} and {0110}[85,87].The most closely packed crystal plane (0001) of Mg has the highest atomic density and binding energy.Therefore,orientation along plane (0001) could effectively inhibit the escape of metal ions from the metal lattice into the ambient solution.It indicated that the lower electrochemical dissolution rate of Mg alloys could be obtained by the formation of strong (0001) texture.[88].Figs.9 and 10 indicate that the FSPed Mg-5Zn alloy is dominated by the (0001) basal plane,so the texture had a positive effect on the corrosion resistance in the present study.Liu et al.[30] indicated that the corrosion resistance of Mg alloys prepared via FSP did not differ much in the cross-section or surface,which were both much higher than that of the as-cast alloy.However,the texture of the FSPed Mg-5Zn alloy was significant,which indicated that the texture was not a major influencing factor in terms of the corrosion resistance of the FSPed Mg-5Zn alloy.

For galvanic corrosion caused by precipitates,after FSP treatment,the precipitates (cathodic sites) in the Mg alloys which introduced anodic dissolution ofα-Mg by microgalvanic couple are significantly fragmented and partly dissolved,and the effect of galvanic corrosion will be weakened.Several studies[18,89]have shown that the dissolution of precipitates leads to an enhancement of corrosion resistance.It is generally accepted that such dissolution can reduce pitting corrosion.Since the FSP of Mg-5Zn dissolved precipitates,the corrosion resistance improved.The dissolution of the precipitates promoted by FSP was primarily based on SPD.As SPD occurs during FSP,the diffusion of alloy elements will be accelerated,and re-dissolution can occur at a lower temperature [70,90–94].Meanwhile,since subsequent cooling rate of Mg alloy after FSP is relatively high,the alloying elements can not be completely precipitated.Therefore,the formation of partial supersaturated solid solution was obtained.In addition,FSP resulted in the breakage and uniform dispersion of the precipitates.Since the corrosion mechanism of Mg alloys in 3.5 wt.% NaCl aqueous solution primarily involves microgalvanic corrosion between theα-Mg matrix and precipitates,the refinement of the cathodic precipitates can reduce the local area ratio of the cathode and anode [55],which reduces the current density of galvanic corrosion.Therefore,the effects of fragmentation and uniform distribution of the precipitates on the corrosion resistance are significant.In another study,Liu et al.[55] also indicated that significant refinement of the precipitates (cathodic sites) could weaken their accelerating effect on the corrosion of the Mg alloy.In the present study,it was hypothesized that the corrosion current in a certain region wasI,and the corresponding region of the corrosion pit wasS;thenI/Srepresents the intensity of charge transfer generated where the corrosion actually occurred,i.e.,the accelerating effect of galvanic corrosion.In this case,it is known thatIas-cast>IFSPed.As shown in Fig.15,Sas-castIFSPed/SFSPed.These results indicated that the intensity of charge transfer in the actual corroded area of the FSPed Mg alloy was reduced.Therefore,the accelerating effect of the cathodic sites,i.e.,precipitates,on the corrosion of the Mg alloy was weakened by FSP treatment.

Fig.15.SEM micrographs of Mg-5Zn alloys after immersion in 3.5 wt.% NaCl aqueous solution for 24 h and removal of the corrosion products: (a) As-cast;(b) FSPed.

Fig.15 presents the SEM micrographs of the as-cast and FSPed Mg-5Zn alloys after immersion in 3.5 wt.% NaCl aqueous solution for 24 h and removal of the corrosion products.The Mg-5Zn alloys before and after FSP exhibited significant differences in terms of their morphology after corrosion.Specifically,the number of corrosion pits decreased after FSP and the pits changed their shape to gullies and became shallower.Additionally,the corrosion form changed from primarily severe local pitting corrosion to largely uniform corrosion.

According to the Nyquist diagram,the low-frequency inductive loop of the FSPed sample almost disappeared.This may be caused by the lack of massive desorption of the corrosion product film.Considering the morphology of the removed corrosion products,the corrosion of the FSPed sample was more uniform than that of the as-cast alloy.This indicated that the corrosion product film of the FSPed sample was more uniform,less permeable,and less desorbed than that of as-cast sample.To some extent,this may also explain why the corrosion was somewhat inhibited.

On the other hand,apart from the weakened galvanic corrosion,another reason why FSP may have improved the alloy’s corrosion resistance is by changing the corrosion form.

Mg alloys was more susceptible to corrosion at the grain boundaries and in the second phases.For coarse grains,after the grain separation caused by corrosion,the corrosion cracks were more likely to form relatively large corrosion pits,as reported by Liu et al.[35].The precipitates dispersed in fine distribution due to fractures during FSP,and the overall grains were refined.After corrosion-induced grain separation,the corrosion cracks could not easily become deeper because of the small area involved.As mentioned above,the corrosion form of the alloy in solution changed from primarily severe local pitting corrosion to largely uniform corrosion.Considering a consistent total amount of corrosion,the corrosion proceeded evenly across the surface,rather than causing deeper penetrations;essentially,as more areas were affected,the depth of corrosion was significantly reduced.

These results indicated that FSP not only decreased the corrosion current density and increased the impedance of the Mg-5Zn alloy in 3.5 wt.% NaCl aqueous solution,but it also greatly changed the corrosion mode,which significantly reduced the corrosion depth.

Assuming that the corrosion mode after FSP was completely uniform corrosion,the corrosion ratevdepth(mm/y)can be estimated based on Faraday’s law.In electrochemical corrosion,anodic dissolution leads to metal corrosion,i.e.,M→Mx++xe-.According to Faraday’s law,for every mole of metal dissolved at the anode,xFFaraday’s power was passed(wherexis the number of gained and lost electrons in the electrode reaction equation,andFis the Faraday constant(96,500 C/mol)).If the current intensity isI0and the energizing time ist,the amount of electricity passed isI0t.If the atomic weight of the metal isM0,the mass of metal dissolved at the anode,Δm,is described by Eq.(4):

For uniform corrosion,the entire metal surface areaS0can be considered as the anode area,so the corrosion current density of the entire metal surface area isi0=I0/S0.Therefore,the relationship between the corrosion ratevdepthand the corrosion current densityi0can be derived from Eq.(5),which indicates that the corrosion ratevdepthis proportional to the corrosion current densityi0:

Therefore,the corrosion current densityi0can be used to express the electrochemical corrosion rate of the metal.If the unit ofi0is μA/cm2and the unit ofρis g/cm3,then Eq.(6) can be obtained:

Substituting the data for Mg-5Zn into Eq.(6) gives the value shown in Eq.(7):

Therefore,assumed completely uniform corrosion corresponds to a corrosion rate of 0.1 mm/y (if it is local pitting,the corrosion rate is much higher than 0.1 mm/y at the location where pitting occurs).This indicated that the FSPed Mg-5Zn alloy would exhibit excellent corrosion restistance in engineering applications,with broad application prospects.

4.Conclusion

(1) After FSP,the as-cast Mg-5Zn sample grains were refined from 133.8 μm to 1.3 μm.A YS of 102 MPa,a UTS of 214 MPa,and an elongation of 11% along the PD were obtained for the FSPed Mg-5Zn alloy;these values were 10.9%,21.6%,and 61.5% greater than those of the as-cast sample,respectively.In addition,along the TD,the mechanical properties exhibit significant anisotropy compared to the FSPed sample along the PD,with a YS of 152 MPa,an UTS of 193 MPa,and an elongation of only 5.75%.

(2) After FSP,the precipitates in the Mg-5Zn alloy were uniformly refined and dissolved.According to the PDP test in 3.5 wt.% NaCl aqueous solution,theIcorrdecreased from 3.1 × 10-5to 4.1 × 10-6A/cm2,and the impedance increased from 894 to 1534Ω/cm2.

(3) The FSPed Mg-5Zn exhibited largely uniform corrosion characteristics,rather than severe localized corrosion.The corrosion depth was significantly reduced due to an increase in the total corrosion area.

(4) The decrease in corrosion current density and corrosion depth caused by the change in corrosion mode (from severe localized to largely uniform corrosion) together indicated that FSP improved the corrosion resistance of the Mg-5Zn alloy.

Declaration of competing interest

The authors declare no conflict of interest related to this work.

CRediT authorship contribution statement

Fei Long:Investigation,Formal analysis,Visualization,Writing– review &editing.Gaoqiang Chen:Investigation,Writing– review &editing.Mengran Zhou:Methodology,Writing– review &editing.Qingyu Shi:Methodology,Conceptualization,Resources,Supervision.Qu Liu:Methodology,Supervision,Formal analysis,Writing– review &editing.

Acknowledgements

This study was financially supported by the National Natural Science Foundation of China (grant no.51705280 and 52035005).