The relation between ductility at high temperature and solid solution in Mg alloys

D.F.Shi, C.M.Cepeda-Jiménez, M.T.Pérez-Prado

a IMDEA Materials Institute, C/ Eric Kandel, 2, 28906 Getafe, Madrid, Spain

b Department of Materials Science, Polytechnic University of Madrid/Universidad Politécnica de Madrid, E.T.S.de Ingenieros de Caminos, 28040, Madrid,Spain

c Department of Physical Metallurgy, CENIM, CSIC, Av.Gregorio del Amo 8, 28040 Madrid, Spain

Abstract This work investigates the effect of solid solution on ductility and on the activation of individual deformation mechanisms at moderate temperatures and at quasi-static strain rates in Mg-Zn and Mg-Al alloys.With that aim, four solid solution Mg-Zn and Mg-Al binary alloy ingots containing 1 and 2 wt.% solute atoms were subjected to hot rolling and subsequent annealing to generate polycrystals with similar average grain size and basal-type texture for each composition.The activity of the different slip systems after tensile testing at 150°C and at 250°C was evaluated in pure Mg and in the alloys by EBSD-assisted slip trace analysis.In addition, segregation of Zn and Al atoms at grain boundaries during the thermo-mechanical processing was characterized by HAADF-STEM and EDX.It was found that while the addition of Al and Zn atoms to pure Mg does not lead to major changes in the mechanical strength at the investigated temperatures, it does enhance ductility significantly, especially at 250°C.Our results show that this increase in ductility cannot be attributed to a higher activation of non-basal systems in the alloys, as reported earlier, as the incidence of non-basal systems is indeed considerably higher in pure Mg.This work suggests, on the contrary, that the ductility increase may be attributed to the presence of a more homogenous basal activity in the alloys due to a lower degree of orientation clustering, to grain boundary solute segregation, and to a higher slip diffusivity at grain interiors.

Keywords: Magnesium; Solute segregation; Slip trace analysis; Ductility; Deformation mechanisms.

1.Introduction

Magnesium and its alloys are attractive structural materials owing to their low density, good castability and recyclability.However, the widespread industrial application of magnesium is hindered because of its inherent limitations of strength and formability, which are related to the prevalence of basal slip and to the strong basal-type texture that develops during processing [1,2].Improving formability requires strategies to increase the activity of alternative deformation mechanisms,including pyramidal slip and twinning, as well as to weaken the basal texture [3,4].

Solid solution alloying has proven somewhat effective to enhance the strength and the ductility of Mg alloys and,thus, to improve their formability at moderate temperature.The classical solid-solution hardening models are based on the interaction between immobile solute atoms and dislocations gliding in the slip plane under stress [5,6].The addition of alloying elements reportedly leads to a reduction of the anisotropy in the critical resolved shear stress (CRSS) values of basal and non-basal deformation modes, and thus to a larger activity of non-basal systems.Solute addition also alters the strain rate and temperature dependence as well as the work hardening of the active deformation mechanisms[7].Moreover, recent research by D.F.Shiet al[8]found that Aland Zn solutes additionally modify the topology of the grain boundary network in such a way that intergranular slip localization (i.e., the formation of shear bands traversing several grains) and thus early failure, which are commonly observed in pure magnesium, are prevented [9].In agreement with this work, Sandl?beset al[1]also reported a reduction of shear band formation and, thus, improved ductility in Mg-Y solid solutions with respect to pure Mg.

Grain boundary (GB) segregation of alloying elements can also be effectively employed to improve formability as well as other materials′properties such as damage resistance, creep behavior, corrosion resistance, etc [10].These beneficial effects have been attributed to a decrease in the GB energy by solute segregation, which leads to increased GB cohesion and thus hinders grain growth at moderate temperatures.GB segregation has been reported to occur in magnesium alloys belonging to the Mg-Zn [4], Mg-Y [2], Mg-Zn-Ca [11], and Mg-Zn-Y systems [12].

Finally, it is well known that increasing the processing temperature also leads to improved formability in magnesium alloys,owing to a decrease of the CRSS of non-basal systems,which are difficult to activate at room temperature [13-15].In turn, high temperature deformation may result in grain coarsening and in an increase of the processing cost [11].To date,the influence of solid solution additions on the activity of individual deformation mechanisms and on the ductility of Mg alloys at moderate temperatures remains poorly understood.

The purpose of the current study is to investigate the influence of Zn and Al solutes(in concentrations ranging from 1 to 2 wt.%) on the tensile deformation modes and on their contribution to ductility during deformation of binary Mg alloys at 150 and 250°C and at quasi-static strain rates.These elements have been chosen because they are the most common alloying elements in commercial Mg alloys.Electron backscattered diffraction-assisted slip trace analysis [16]has been utilized to quantitatively assess the influence of the mentioned solutes on the relative activities of various slip systems.In addition,grain boundary segregation and local strain heterogeneities have been characterized and compared with those corresponding to pure magnesium polycrystals with similar grain size and texture.The findings of this study provide guidelines for the development of new high performance magnesium alloys.

2.Experimental procedure

Four solid solution Mg-Zn and Mg-Al binary alloy ingots with solute contents of 1 and 2 wt.% were produced by induction melting under Ar atmosphere from 99.95% pure Mg, Al, and Zn.The as-cast alloys were subjected to a multi-step thermomechanical processing, including solution treatment (450°C,12h), hot rolling and subsequent annealing to generate polycrystals with similar average grain size and basal-type texture for each composition.The annealing conditions ranged between 300°C for 5-10 min for the Mg-Zn alloys and between 350°C for 5-20 min for the Mg-Al alloys.The grain size and texture of the binary Mg alloys were designed to be similar to those corresponding to pure Mg analyzed in a previous study (d～13 μm) [17], in order to compare and assess the influence of solutes on ductility and deformation mechanisms at moderate temperatures.A more detailed description of the processing method can be found in [8].

The microstructure and the microtexture of all materials were examined by EBSD using a field emission gun (FEG)SEM (Helios NanoLab 600i, FEI) equipped with an HKL EBSD system, a CCD camera, as well as both the Aztec and the Channel 5.0 data acquisition and analysis software packages.EBSD measurements were conducted using a step size of 0.8 μm at an accelerating voltage of 15 kV, and with a beam current of 2.7 nA.The average grain size values were calculated by the linear intercept method from inverse pole figure (IPF) maps in the normal direction (ND)to the rolling plane, using only GBs with misorientation angles greater than 15°.Sample preparation for EBSD included mechanical mirror-polishing using diamond pastes of increasingly finer particles, and a final finishing using a colloidal silica slurry.

The macrotexture of the processed samples was analyzed by the Schulz reflection method using an Empyrean Panalytical X-ray diffractometer with a Cu-Kαradiation source and a parallel beam operated at 45 kV and 40 mA.The 2θangles ranged from 25 to 65° with a step of 0.262° and an acquisition time of 2 s.The surface area examined was ～1 cm2.The (0001), (10-10), (10-11), (10-12), (10-13), and (11-20) pole figures were measured.X-ray diffraction data were corrected for background and defocusing using the software X’Pert HighScore Plus.From the incomplete measured pole figures, the orientation distribution function (ODF) and the complete calculated pole figures were constructed using the MTEX MATLAB code [18].

Dog-bone tensile samples with a gage length of 10 mm and a transversal section of 3 × 2.5 mm2were electrodischargemachined out of the rolled and annealed sheets, with the tensile axis parallel to the rolling direction(RD).Uniaxial tensile tests were then carried out on an Instron 3384 universal testing machine at 150 and 250°C and at initial strain rate of 10?3s?1.Two tests per sample were carried out to failure in order to evaluate the full mechanical response (yield stress(σ0.2),ultimate tensile stress(σUTS),and ductility),and others were stopped at a plastic strain of ～10% in order to analyze the active deformation mechanisms by slip trace analysis, as described below.

The activity of the different slip systems was evaluated by EBSD-assisted slip trace analysis (Fig.1).This method makes use of SEM micrographs at different magnifications (Figs.1a and 1b) and EBSD orientation data input into a MATLAB code, which allows to identify the crystallographic plane corresponding to selected traces appearing at the surface of deformed single or polycrystals [19](highlighted using a red line in Figs.1a and 1b in two SE micrographs at different magnifications).In particular, first, selected large areas of the microstructure at the center of the gage length were mapped by EBSD before each tensile test (Figs.1c).After a tensile strain of ～10%, the orientation of grains in which slip tracesbecome apparent during straining was determined by postmortem EBSD examination (Fig.1d).In general, only one set of parallel traces were detected in each grain for all the studied materials and each set of traces was counted as one single trace.Subsequently, the Euler angles of each of the selected grains were input into the MATLAB code, which provided as output a visual representation of all the possible plane traces corresponding to that particular orientation under analysis (Fig.1e).Finally, comparison of each trace under study with those simulated by the code allows assigning the operative slip system by choosing the trace providing the best match (grain A in Fig.1).In addition, the MATLAB code also gives the Schmid Factor of the corresponding slip system under the assumption of uniaxial stress along RD (Fig.1f).This process of slip trace analysis is repeated for as many traces as possible.In general more than one hundred traces are evaluated in each sample, as the analysis accuracy relies on having statistically relevant information [16,20,21].The activity of the different slip systems is finally related to the frequency of the corresponding observed traces.

Fig.1.Procedure for slip trace analysis.a) and b) SEM micrographs at two magnifications showing the appearance of slip traces (highlighted using a red line in grain A).c) and d) EBSD IPF maps in the ND corresponding to the SEM micrograph in b) before and after deformation, respectively, to identify the orientations of grains in which traces were observed.e) Calculation of the 12 possible traces and determination of the active slip system, and f) Schmid factors for each grain in which a trace was detected.

The density of geometrically necessary dislocations(GND)was calculated from EBSD misorientation data with the help of the ATEX software [22], which calculates the misorientation between a reference pattern and each point within a selected region.The EBSD-GND maps were then related to the local dislocation content.

Fig.2.EBSD IPF maps in the ND and XRD {0001} pole figures corresponding to the investigated polycrystalline materials.The average grain size values are included as insets in each map.a) Pure Mg; b) Mg-1 wt.% Zn; c) Mg-2 wt.% Zn; d) Mg-1 wt.% Al; e) Mg-2 wt.% Al.

Finally, the microstructure of the binary Mg alloys as well as the presence of GB solute segregation were also analyzed by high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) and by energy dispersive X-ray spectroscopy (EDX) in a Talos F200X FEI TEM operating at 200 kV.Specimens for TEM observation were prepared by ion milling using a dual-beam SEM-FIB Helios NanoLab 600i, FEI.A trenching-and-lift-out technique, described in great detail in [23], was adopted to extract lamellae, which were thinned down to approximately 50 nm for electron transparency.

3.Results

3.1.Microstructure of pure Mg and of the Mg-Zn and Mg-Al binary alloys

The microstructures of the pure Mg and of the four binary Mg-Zn and Mg-Al alloys (with 1 and 2 wt.% alloying content) fabricated for the present study are illustrated in Fig.2 by means of several EBSD inverse pole figure (IPF) maps in the normal direction (ND) to the rolling plane.High angle boundaries (θ＞15°) are colored in black.The average grain size,d, measured by the linear intercept method on the EBSD maps, is similar in all materials (10 μm＜d＜14 μm).The IPF EBSD maps reveal that the microstructures of the pure Mg and of the binary alloys are equiaxed and have a homogenous grain size distribution.They consist mainly of high angle grain boundaries, which suggests the occurrence of recrystallization during hot processing and annealing.The XRD macrotextures corresponding to pure Mg and to all four binary alloys are also shown as insets in Fig.2 by means of the (0001) direct pole figures.Both the pure Mg and the alloys present similarly strong basal textures, with＜0001＞poles tilted a few degrees away from ND towards RD, as has been commonly observed in rolled magnesium alloys [24].The average Schmid factors for basal slip (SFbasal), assuming a tensile stress along RD, are also included in Fig.2 asinsets for each investigated composition.It can be seen that SFbasalvalues are very similar for all the investigated materials.In particular, they fall within the range (SFbasal～0.222-0.231) for the binary alloys and are slightly lower for pure Mg (SFbasal～0.204).

Fig.3.a) Correlated and b) uncorrelated misorientation distribution histograms corresponding to pure Mg, Mg-Zn, and Mg-Al binary alloys.

The correlated and uncorrelated distributions of misorientation angles corresponding to pure Mg and to each alloy composition are shown in Fig.3.Correlated measurements consider the misorientation between each scanned point and its nearest neighbors while in uncorrelated measurements the misorientation is calculated with respect to a reference grain.As is characteristic of hot rolled Mg and Mg alloys, all histograms present a maximum frequency of boundaries with misorientation angles close to 30° [9].The fraction of grain boundaries withθ＜30° (fθ＜30°) in the correlated distribution,which is indicative of the degree of clustering of grains with similar orientations [8], is 61% in pure Mg, 49% in the Mg-1%Zn alloy, 45% in the Mg-2%Zn alloy, 44% in the Mg-1%Al alloy, and 45% in the Mg-2%Al alloy.Thus, the addition of solutes in agreement with earlier studies [8], reduces significantly the degree of orientation clustering.

The microstructure of the Mg-1wt.% Zn (Figs.4a,b) and Mg-1wt.%Al (Figs.4c,d) solid solutions was analyzed at higher magnification by HAADF STEM and by EDX.Two or three GBs were analyzed per sample, as a function of the lamella size.Altogether, these imaging techniques evidence the occurrence of segregation of Zn and Al atoms at grain boundaries (highlighted with white arrows in Figs.4a and 4c)during hot rolling, which is in agreement with other previous reported studies [25-27].It has been reported [11]that solute segregation is an energetically favorable process that leads to a decrease in the energy of grain boundaries, thus effectively reducing their mobility and causing uniform grain growth during recrystallization.However, this experimental observation is different from that made in previous studies [2,4], which reported that Zn atoms do not segregate or have a low grain boundary segregation tendency, as reported in the latter.It is our contention that probably the amount of segregated atoms depends on the GB type, on the alloy composition and the specific alloying species, as well as on temperature [4].

3.2.Effect of solutes on the mechanical response at moderate temperatures

Figure 5 depicts the tensile true stress-true plastic strain curves of the pure Mg (Fig.5a) and of the binary Mg-Zn(Fig.5b) and Mg-Al alloys (Fig.5c) deformed until failure at 150°C and 250°C at an initial strain rate of 10?3s?1.It is shown for each material composition that,as expected,raising the testing temperature gives rise to a decrease in the yield(σ0.2) and maximum (σUTS) strengths and in general to an increase in ductility.The addition of solutes does not lead to major changes in the mechanical strength at each temperature.Theσ0.2values for pure Mg at 150°C and 250°C are 60 and 50 MPa, respectively, while the averageσ0.2values for the alloys at the same temperatures are 65 MPa and 37 MPa.However, solute addition does have a positive influence in the ductility.While the solid solution alloys exhibited no signs of mechanical instability up to very high strains, the flow curves corresponding to pure Mg show much higher work hardening ability, especially at 150°C, with clear evidences of plastic instability, which is known to be detrimental for ductility.The origin of the instabilities observed in pure magnesium at high temperature could be associated to clusters of grains well oriented for basal slip that form during the rolling process,as will be described below, which lead to pronounced slip localization and to the formation of shear bands.The largest plastic strains (εplastic) were measured in the Mg-1 wt.% Zn and Mg-2 wt.% Zn alloys at 250°C (εplastic～75% and 70%,respectively).These values are significantly higher than those measured in pure Mg, which amounted to 12% at 150°C and to 20% at 250°C.

3.3.Effect of solutes on high-temperature slip activity

The activity of the different slip systems after a tensile strain of ～10 % at 150°C and at 250°C has been evalu-ated quantitatively in pure Mg and in the alloys by EBSDassisted slip trace analysis with the aim of determining the effect of the addition of solutes and of temperature.More than 100 slip traces per composition were analyzed.Traces in the grain mantle regions, which usually reflect the local activation of multiple slip systems to maintain strain compatibility at grain boundaries, and which often possessed an ill defined or wavy nature, were not taken into account for this analysis.However, although this methodology does not allow to capture absolutely all the slip activity, it is our contention that this approach when sufficient statistic is considered does provide a qualitative estimation of the main operative deformation mechanisms.In particular, the formation of well defined slip traces on the grain surface indicates that straining is mostly accommodated by the deformation system leading to the slip line.

Fig.4.Grain boundary segregation in (a,b) Mg-Zn and (c,d) Mg-Al alloys.For each alloy, HAADF micrographs (a,c) and EDX maps of the corresponding solute element (b,d) are shown.

Fig.5.Tensile true stress-true plastic strain curves corresponding to a) pure Mg, b) Mg-Zn alloys (1 and 2 wt.% Zn) and c) Mg-Al alloys (1 and 2 wt.% Al)deformed at 150 and 250°C and at initial strain rate of 10?3 s?1.

The frequencies of the basal and non-basal traces corresponding to each alloy and to pure Mg [17]are plotted inFig.6.It can be seen that, in pure Mg, deformed at 150°C,approximately 74% of traces correspond to basal slip, while the contribution of non-basal systems (prismatic plus＜c+a＞pyramidal slip) was observed to be around 26%.When the testing temperature is increased to 250°C the frequency of basal traces decreases to 58% and the frequency of nonbasal traces increases to 42% [17].This is consistent with the well known decrease of the CRSS of non-basal systems with increasing temperature.Second, in all the binary Mg alloys tested at 150°C, an average of 98% of traces correspond to basal slip and non-basal activity was observed to be negligible (2%).When the testing temperature is raised to 250°C, basal slip continues to be the main deformation mechanism in the four binary alloys under investigation, with an incidence close to 95%, and the contribution of non-basal systems remains below 5%.An exception to this behavior is the Mg-2wt.%Zn alloy, where the incidence of basal slip decreases to 85% and that of non-basal systems increases up to 15% (Fig.6a).Nevertheless, in general, our results show that the addition of solutes leads to a comparatively higher activity of basal slip at the moderate temperatures investigated.

Fig.6.Frequency of basal and non-basal slip traces in pure Mg and in the binary alloys after tensile deformation along RD to a strain of 10 % at 150 and 250°C and at an initial strain rate of 10?3 s?1.a) Pure Mg and Mg-Zn alloys (1 and 2 wt.% Zn) and b) Mg-Al alloys (1 and 2 wt.% Al).

Fig.7.Schmid factor distribution corresponding to grains in which basal (light blue), prismatic (yellow), and pyramidal (green) traces were detected after a tensile strain of 10% along RD at the two temperatures investigated.(a,d) Pure Mg, (b,e) Mg-1 wt.% Zn, (c,f) Mg -2 wt.% Zn).The {0001} pole figures included as insets indicate the orientation of the grains in which basal (blue) and non-basal (red) traces were found.

Figures 7 and 8 show the SF distribution histograms with respect to the global external stress corresponding to grains in which basal, prismatic, and pyramidal＜c+a＞, traces were detected in pure Mg and in Mg-Zn (Fig.7) and Mg-Al alloys(Fig.8) after a tensile strain of 10% at 150 and at 250°C.In addition,the orientations of the grains in which basal and nonbasal slip traces were detected are depicted in {0001} pole figures by means of blue and red dots, respectively.These results show that in general, and irrespective of composition,most basal slip traces were detected in grains with SFbasalhigher that 0.3, indicating that basal slip is predominantly activated in response to the applied stress.However,non-basal traces were observed in grains with a wide range of SFs,suggesting that prismatic and pyramidal slip systems act both in response to the applied stress as well as accommodation mechanisms in response to local stresses developed during straining [28,29].

Fig.8.Schmid factor distribution corresponding to grains in which basal (light blue), prismatic (yellow), and pyramidal (green) traces were detected after a tensile strain of 10% along RD at the two temperatures investigated.(a,c) Mg-1 wt.% Al, (b,d) Mg -2 wt.% Al).The {0001} pole figures included as insets indicate the orientation of the grains in which basal (blue) and non-basal (red) traces were found.

Figures 9, 10 and 11 show selected areas of the deformed gage length and their corresponding EBSD-derived GND maps for pure Mg (Fig.9) and for the Mg-Zn (Fig.10)and Mg-Al (Fig.11) alloys after a tensile strain of ～10 % at 150°C and 250°C.It must be noted that,in these post-mortem measurements, only the presence of the GNDs that remain after deformation is captured [2].The GND density gives a measure of the local stresses.In pure Mg, irrespective of temperature (Fig.9), GNDs are seen to accumulate along transgranular bands as a consequence of high stress localization,which ultimately appears to lead to cracking and early fracture [6], as evidenced by the presence of non-indexed points(in black) in the vicinity of high GND density regions.In contrast, the deformed Mg-Zn (Fig.10) and Mg-Al (Fig.11)microstructures are characterized by a much more homogeneous GND distribution, in the absence of localized transgranular shear bands, at all the testing temperatures investigated.The GND density decreases significantly with increasing temperature, reflecting a higher prevalence of recovery processes.In summary, at the moderate temperatures investigated, the addition of solutes leads to a more homogeneous local stress distribution and prevents premature crack formation along transgranular bands.

The origin of the transgranular bands of high GND density can be appreciated in Fig.12.This figure illustrates the EBSD GND density maps at higher magnification corresponding to selected areas of the gage length of pure Mg (Fig.12a)and of the Mg-1wt.%Zn alloy (Fig.12b) after tensile testing at 150°C and a strain of ～10%.It can be seen that, in pure Mg, the observed basal slip traces lie on clearly defined slip bands that are indicated by dotted lines.These bands are formed by “soft grains”, that are both well oriented for basal slip (high SFbasal) and relatively large in size [20], and which therefore contain a low GND density.In contrast, in the grains that are adjacent to the slip band, which are poorly oriented for basal slip, a higher GND density is present (red regions) as the need to preserve boundary continuity leads to a higher degree of lattice curvature [30,31].The presence of such high stress regions in the vicinity of the basal slip bands acts as a precursor to crack initiation.Indeed, slip localization is well known to be related to early fracture [9,32].It is our contention that the presence of the mentioned high stress regions may be the origin of the relatively high incidence of hard non-basal slip in pure Mg (Fig.6).On the other hand, in accordance with figures 10 and 11, Fig.12b provides a highermagnification view of the more homogeneous distribution of strain in the solid solution alloys.The presence of solutes appears to prevent the formation of transgranular basal slip bands, and thus to avoid the accumulation of high stresses in highly localized regions.It seems reasonable, then, to assume that the observed low incidence of hard non-basal slip in the solid solution alloys(Fig.6)may be related to the more homogeneous distribution of basal slip.

Fig.9.SEM micrographs and EBSD-GND density maps corresponding to pure Mg after tensile testing at (a) 150°C and (b) 250°C up to a strain of ～10%.The black areas correspond to unindexed regions or to microcracks.

Fig.10.SEM micrographs and the corresponding EBSD-GND density maps for (a,b) Mg-1 wt.% Zn and (c,d) Mg-2 wt.% Zn after tensile testing at (a,c)150°C and (b,d) 250°C up to a strain of ～10%.

Fig.11.SEM micrographs and the corresponding EBSD-GND density maps for (a,b) Mg-1 wt.% Al and (c,d) Mg-2 wt.% Al after tensile testing at (a,c)150°C and (b,d) 250°C up to a strain of ～10%.

Fig.12.SEM micrographs and GND density maps overlapped on the same areas corresponding to (a) pure Mg (b) Mg-1wt.%Zn after tensile testing at 150°C up to a strain of ～10%.

Figure 13 illustrates the EBSD IPF maps in the ND corresponding to all the studied materials after tensile testing at 150°C at an initial strain rate of 10?3s?1and up to a strain close to fracture, which amounts to 13% in pure Mg and to 50-60% in the alloys (see insets in Fig.13).It can be seen that, while grains in pure Mg retain a relatively equiaxed shape at fracture, significant grain elongation in the direction of the tensile axis is observed inthe binary alloys due to the much larger strains to fracture recorded.The presence of a small fraction of fine, equiaxed grains at the grain boundaries at strains close to fracture in the binary alloys (Figs.13c and e) reflects an incipient,but still limited, occurrence of recrystallization [33].As expected, tensile twins are basically absent at these moderate temperatures, in agreement with the fact that twinning becomes less frequent with increasing temperature, and also due to the strong basal texture present in the current alloys.Fig.14 shows the corresponding inverse pole figures(IPFs) in the tensile direction after a strain of ～10% and up to fracture.The gradual strengthening of the＜11-20＞pole with strain in the alloys is consistent with the operation of basal slip [34], and confirms the limited degree to which recrystallization takes place at 150°C, even at relatively large strains [35,36].

Fig.13.EBSD IPF maps in the ND after tensile testing at 150°C at an initial strain rate of 10?3 s?1 and up to a strain close to fracture.a) Pure Mg; b)Mg-1 wt.% Zn; c) Mg-2 wt.% Zn; d) Mg-1 wt.% Al; e) Mg-2 wt.% Al.

Figure 15 illustrates the EBSD IPF maps in the ND corresponding to all the studied materials after tensile testing at 250°C at an initial strain rate of 10?3s?1and up to a strain close to fracture, which amounts to 24% in pure Mg,to 70-78% in the Mg-Zn alloys, and to 54-63% in the Mg-Al alloys.In pure Mg, the presence of wavy boundaries is indicative of the early stages of recrystallization.In the binary alloys, the absence of elongated grains at failure is consistent with the occurrence of dynamic recrystallization (DRX) and grain growth, as is commonly observed in Mg alloys at this temperature[37].Grain growth is more pronounced in Mg-Zn than in Mg-Al alloys, likely due to the larger strains to failure attained [38].Figure 16 depicts the IPFs showing the orientation of the tensile axis before and after testing of all the materials investigated at 250°C up to 10% and up to fracture.In all cases, a clear intensity maximum appears around the＜11-20＞pole after a tensile strain of 10%, consistent with the operation of basal slip, and then this maximum gradually fades as strain progresses due to the simultaneous occurrence of DRX.

Fig.14.Inverse pole figures in the tensile direction corresponding to the (a) pure Mg and binary (b,c) Mg-Zn and (d,e) Mg-Al alloys before and after tensile testing at 150°C up to a strain of ～10% and up to fracture (εF).

4.Discussion

This work reports the effect of solid solution on ductility and deformation mechanisms at moderate temperatures in Mg-Zn and Mg-Al alloys containing 1 and 2wt.% of solute atoms.As shown in Fig.5, the addition of Al and Zn atoms to pure Mg enhances ductility at high temperature.

Earlier studies have attributed the ductility increase associated to the addition of solutes to a higher incidence of nonbasal systems [4,5,39-46].For example, some studies [5,44-46]report a more intense activity of non-basal dislocation slip due to the fact that solutes improve the cross-slip probability of basal dislocations onto prismatic planes.The good ductility of Mg-Zn solid solutions has also been associated to the activation of＜c+a＞slip due to the minimization in the difference in binding energy and thus in CRSS anisotropy between slip systems [4].However, our results do not support the thesis that a higher incidence of non-basal slip contributes to enhanced ductility.On the contrary, we have shown that the ductility is significantly higher in the alloys than in pure Mg (Fig.5), despite the fact that the incidence of non-basal systems at the two temperatures investigated is considerably higher in the latter (Fig.6).

It might also be speculated that the high ductility observed in the Mg alloys is due to an enhanced activity of grain boundary sliding (GBS).However, earlier research has shown that in pure Mg and in Mg alloy polycrystals with similar microstructure as that of the materials investigated here, GBS activity is negligible.In particular, former studies based on strain-rate-change tests [17]have demonstrated that, even at 250°C, the stress exponent of pure Mg polycrystals with the same grain size and texture as those of the present study, at a strain rate of 10?3s?1, is higher than 5.According to the well-stablished understanding of creep [47,48]such a stress exponent value indicates that deformation is controlled by dislocation slip, and that the contribution of GBS is minimal.Moreover, alloying additions are known to increase the stress exponent [49]and, thus, to reduce the likelihood of GBS activation.Finally, it is also known that GBS resistance increases with solute segregation [50]due to the associated decrease in grain boundary energy, as indicated by the Gibbs adsorption equation [51].In summary, the observed ductil-ity increase at moderate temperatures in the Mg alloy polycrystals can also not be attributed to an enhanced GBS activity.

Fig.15.EBSD IPF maps in the ND after tensile testing at 250°C at an initial strain rate of 10?3 s?1 and up to a strain close to fracture.a) Pure Mg; b)Mg-1 wt.% Zn; c) Mg-2 wt.% Zn; d) Mg-1 wt.% Al; e) Mg-2 wt.% Al.

Our results prove, however, that the higher ductility observed in the Mg alloys can be attributed to a more homogeneous strain distribution (Figs.10 and 11) and to the absence of slip localization along transgranular basal slip bands,which triggers cracking and early fracture in pure Mg (Fig.9).In agreement with earlier studies at room temperature [8], the absence of localized basal slip bands in the alloys at the moderate temperature investigated may be at least partially attributed to the lower degree of clustering of orientations(Fig.3), which would hinder basal slip transfer.The solid solution alloys can thus accommodate a larger amount of strain,which at a high temperature of 250°C induces DRX, a mechanism that also contributes to enhanced ductility (Fig.15).Therefore, although five slip systems are required for achieving homogeneous deformation, fewer systems can also accommodate the imposed strain if neighboring grains deforms in a compatible way, as observed in the Mg-Zn and Mg-Al alloys.

Another factor that might contribute to impede basal slip transfer in the alloys is the occurrence of GB segregation(Fig.3) [52].It has been reported that the main driving force for segregation of solute atoms along GBs is the minimization of the elastic strains of the dislocations in GBs induced by size misfit between solute atoms and Mg [53].This segregation results in considerable reduction of the GB energy.It is known that low energy grain boundaries constitute strong obstacles to slip transfer and dislocation nucleation [54].Finally, slip transfer might also be more limited in the alloys because defects such as dislocations and vacancies interact attractively with solute atoms, promoting diffuse slip within the grains and therefore decreasing the local stress state on GBs and the constraints in the surrounding grains.

Fig.16.Inverse pole figures in the tensile direction corresponding to the (a) pure Mg and binary (b,c) Mg-Zn and (d,e) Mg-Al alloys before and after tensile testing at 250°C up to a strain of ～10% and up to fracture (εF).

5.Conclusions

This research aims to investigate the effect of solid solution alloying on the tensile ductility and on the active deformation mechanisms at moderate temperatures (150°C and 250°C) and quasi-static strain rates (10?3s?1) in Mg-Zn and Mg-Al alloys containing 1 and 2 wt.% of solute atoms.With that purpose, four polycrystalline Mg-Zn and Mg-Al alloys were rolled and annealed at the appropriate conditions to design microstructures with similar grain sizes, textures and misorientation distributions, and their behavior was compared to that of similar pure Mg polycrystalline samples.The incidence of basal, prismatic and pyramidal＜c+a＞slip was estimated by EBSD-aided slip trace analysis for all the testing conditions investigated.The following conclusions may be drawn from this study:

1 The addition of solutes does not lead to major changes in the mechanical strength at the investigated temperatures with respect to pure Mg.However, the ductility is significantly higher in the alloys than in pure Mg, especially at 250°C.

2 The addition of solutes reduces significantly the degree of orientation clustering.The fraction of grain boundaries withθ＜30° (fθ＜30°) in the correlated distribution, which is indicative of the degree of clustering of grains with similar orientations,was ～61%in pure Mg and ～45%in the solid solution alloys.

3 Basal slip is the dominant deformation mechanism in all the solid solution Mg-Zn and Mg-Al alloys investigated,irrespective of composition and testing temperature.The incidence of non-basal systems (prismatic and pyramidal)was considerably lower in the Mg alloys than in pure Mg.Thus,contrary to what has been reported earlier,the higher ductility observed in the alloys cannot be attributed to an enhancement of non-basal activity due to solute addition.

4 The higher ductility observed in the alloys can be attributed to the following factors: i) a more homogeneous strain distribution associated to a lower degree of orientation clustering, which hinders basal slip localization and transferalong transgranular slip bands; ii) the occurrence of segregation of Zn and Al atoms at grain boundaries, which decreases the grain boundary energy and also contributes to impede basal slip transfer; iii) an enhancement of diffuse slip due to the interaction of dislocations and vacancies with solute atoms at grain interiors, which, in turn, might decrease the local stress state on GBs and the constraints in the surrounding grains.

5 All the above-mentioned factors facilitate the accommodation of a larger amount of deformation, thus allowing the alloys at 250°C to reach the critical strain at which dynamic recrystallization occurs.The onset of DRX further contributes to increase ductility at this temperature.

Acknowledgments

The research leading to these results has received funding from the Madrid region under programme S2018/NMT-4381-MAT4.0-CM project.Funding from projects PID2019-111285RB-I00 and PID2020-118626RB-I00 awarded by the Spanish Ministry of Science, Innovation and Universities, is also acknowledged.DFS acknowledges the financial support from the China Scholarship Council (Grant no 201706050154).Dr Fernando Carre?o (CENIM-CSIC,Madrid) is sincerely thanked for allowing the use of the rolling equipment for this research work.