Dynamic recrystallization of Mg and Mg–Y alloysCrystallographic textu

Materials Science and Engineering A485 (2008) 375–382Dynamic recrystallization of Mg and Mg–Y alloys:Crystallographic texture developmentR.Cottam a,∗,J.Robson a,G.Lorimer a,B.Davis ba School of Materials,Materials Science Centre,The University of Manchester,Grosvenor Street,Manchester M17HS,United Kingdomb Magnesium Elektron Ltd.,PO Box23,Rake Lane,Swinton,Manchester M270D,United KingdomReceived21June2007;received in revised form30July2007;accepted11August2007AbstractPure magnesium and three binary Mg–Y alloys(0.23,0.84and2.71wt%Y)have been deformed in plane strain compression under conditions where dynamic recrystallization(DRX)is observed.The development of crystallographic texture during deformation has been determined for both the parent and DRX grains.In all but the highest Y alloy,the texture of the DRX grains was found to follow that of the parent grains closely in all alloys,implying that the DRX texture is dominated by the deformation conditions,rather than preferred nucleation or growth.In the highest Y alloy,the DRX texture is randomized,which also suggests that preferred nucleation or growth is not responsible for texture formation in this alloy.A transition in macrotexture development was observed in going from pure Mg deformed at250◦C,to Mg–2.71wt%Y deformed at450◦C.This can be attributed to activation of additional slip modes,and a concomitant decrease in the contribution of twinning to deformation.© 2007 Elsevier B.V. All rights reserved.Keywords:Dynamic recrystallization;Magnesium;Texture;EBSD1.IntroductionImproving the poor low temperature formability of magne-sium alloys has been the focus of renewed research efforts in recent years,driven by a need for light weighting in the trans-portation sector.Low temperature formability of magnesium is limited by the plastic anisotropy of individual grains of mag-nesium[1–3],such that deformation of a polycrystal results in generation of incompatibility stresses between grains that cause premature failure[3,4].However,it has also been shown that by using non-conventional processing methods,such as equal channel angular extrusion,and producing a combination of a favourable texture andfine grain size,greatly improved tensile ductilities can be obtained[5,6].Current research is focussed on obtaining the desired microstructure and texture utilizing conventional methods com-bined with alloy and process optimization.To do this,it is essential to have a fundamental understanding of the processes that influence thefinal texture and grain structure in deformed magnesium alloys.Dynamic recrystallization(DRX)is one such ∗Corresponding author.Tel.:+441613063588;fax:+441613063586.E-mail address:ryan.cottam@(R.Cottam).process that is commonly observed during elevated temperature deformation of magnesium and its alloys.DRX in magnesium has been the focus of several studies[7–11]aimed at under-standing the fundamental DRX mechanisms.Little work has yet been done to investigate the effect of DRX on the overall texture evolution or the effect of changes in solute content on DRX.The focus of this paper is on the global(average)texture that results from DRX,and the effect of yttrium additions on this texture development.Yttrium has been used in this study since it is known to alter the relative activity of the deformation modes (slip and twinning)in magnesium alloys[12].One goal of this work was to investigate whether this effect has an influence on the texture of DRX grains that is distinct from that of the original (parent)grains and might thus be exploited for texture control.Studies of texture development during deformation in mag-nesium alloys under conditions of DRX are limited.Ion et al.[10]and Xiong and Davies[11]have addressed texture evo-lution of Mg–Al and AZ31alloys,respectively,in both cases finding that the DRX grains adopted the basal texture of the par-ent grains.Xiong and Davies also evaluated the effect of strain path changes,but again found the basal texture dominant in the DRX grains.Current understanding of texture evolution during DRX has been developed largely for cubic metals[13,14]undergoing0921-5093/$–see front matter© 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.msea.2007.08.016376R.Cottam et al./Materials Science and Engineering A 485 (2008) 375–382“conventional”DRX.In such cases,strain induced bound-ary migration(SIBM)acts as a precursor to formation of DRX grains.However,in magnesium and its alloys,it is accepted that a number of other DRX mechanisms operate, such as continuous rotational DRX[10],which are quite dif-ferent from the conventional mechanism observed in cubic crystals.By changing deformation conditions,it is possible to change the dominant DRX mechanism(e.g.from rotation recrystallization at lower temperatures to SIBM at higher tem-perature[7]).The textures that result from the combinations of DRX mechanisms that operate in magnesium are not yet well characterized.By varying deformation temperature and solute yttrium,this study aims to determine the texture evolution during DRX of magnesium alloys across a range of conditions where the DRX mechanism and relative deformation system activity are expected to change.2.ExperimentalThree binary Mg–Y alloys and commercially pure Mg were prepared by Magnesium Elektron at the Manchester site with compositions showing in Table1.Impurity elements Zn,Si, Cu,Mn and Fe were all below0.01wt%.The alloys were pro-duced as billets7.5cm diameter,solution treated at525◦C for 8h and extruded into3cm diameter bar at390◦C.The initial textures of the extruded billets were measured by EBSD using a Philips XL30FEG SEM with a working distance of20mm and an accelerating voltage of20kV.Samples were prepared for EBSD by grinding and mechanically polishing to colloidal silica and then electro-polishing in a solution of30%nitric acid and ethanol solution,chilled to−30◦C,using a12V potential. The extruded grain size was determined using the linear intercept method.Channel die samples10mm×11mm×12mm were cut from the bar with the compression direction parallel to the extru-sion direction.The samples were coated in powdered graphite which acted as a lubricant during the deformation.An ini-tial strain rate of1×10−3s−1was employed to deform the samples.To obtain statistically useful texture data,it was necessary to choose deformation conditions that led to a significant frac-tion of DRX.Adding yttrium suppresses DRX,requiring higher deformation temperatures to achieve comparable DRX frac-tions.Therefore,it was not possible to vary the yttrium content and deformation temperature independently across the full com-position range studied,since when yttrium was added at high Table1Test temperatures used during channel die compression of pure magnesium and the three Mg–Y alloysAlloy Temperature(◦C)Mg200250300 Mg–0.23wt%Y250300 Mg–0.84wt%Y400 Mg–2.71wt%Y450levels,an increased in deformation temperature was needed to obtain DRX.The test temperatures used are shown in Table1.For each condition,samples were compressed to logarithmic strains of0.1,0.3,0.6and0.9.Channel5EBSD post-processing software was used to create subsets of texture data,based on grain size.A threshold size was determined by examining the grain size histogram after deformation and determining the grain size that divided the DRX grains from the parent grains.The transition size was chosen byfirst selecting a size just above the peak in the grain size histogram due to the DRX grains.This size was verified by identifying automatically all grains below this size and checking whether all of the DRX grains,but none of the parent grains,had been captured.It was apparent from the Euler maps that DRX grains do not undergo further DRX.3.Results3.1.Extruded billet characterizationAfter extrusion,Mg had a mean grain size of57␮m, Mg–0.23wt%Y51␮m,Mg–0.84wt%Y24␮m and Mg–2.71wt%Y21␮m.In pure magnesium and all three alloys the grain structure was fully dynamically recrystallized during extrusion.All the materials show a typical extruded magnesium texture [15]where the basal poles are radially distributed perpendicular to the extrusion axis(Fig.1).Increasing the yttrium concentra-tion from0.23to2.71wt%Y not only refines the grain structure of the material during dynamic recrystallization(DRX)that occurred during extrusion but also weakened the extruded tex-ture.3.2.Mechanical testingThe true stress–logarithmic strain response during channel die deformation is shown in Fig.2.All curves initially show hardening followed by a region of steady state deformation and then some secondary hardening up to the highest true strain used in the test(0.9).As expected,the addition of yttrium strengthens the material,even at the higher test temperatures used.It is notable that in both the pure magnesium and alloy specimens there is little evidence of the strain softeningthat Fig.1.Inverse polefigures for extruded magnesium alloys where the scale is times random:(a)Mg,(b)Mg–0.23wt%Y,(c)Mg–0.84wt%Y and(d) Mg–2.71wt%Y.R.Cottam et al./Materials Science and Engineering A 485 (2008) 375–382377Fig.2.True stress–logarithmic strain response for the three Mg–Y alloys and pure magnesium.often accompanies dynamic recrystallization[16](Fig.2). However,this softening is also in competition with harden-ing associated with dislocation accumulation,twin formation (Hall–Petch effect)and texture changes.For the initial tex-ture and deformation conditions employed here,this offsets the softening due to DRX,and gives theflat response observed in the stress–strain curve.Once DRX has refined the grain structure and twinning has reoriented the mate-rial(into a“hard”orientation;c-axis compression),work and texture hardening begin again to dominate the stress–strain response and the strain hardening exponent increases(secondary hardening).3.3.Optical micrographs after deformationOptical microscopy was performed to investigate the occur-rence of deformation twins in the microstructure after the initial deformation(Fig.3).Since the compression axis was along the extrusion direction,the majority of grains have their c-axes near perpendicular to the compression direction at the start of testing,a favourable orientation for formation of the{10¯12} c-axis tension twin.The presence of deformation twins can be found in the pure Mg at all deformation temperatures and in the Mg–0.23wt%Y alloy(Fig.3).In the Mg–0.84wt%Y alloy the microstructure shows small deformation twins,but these only occur in less than half of the grains.This alloy and deformation temperature therefore appears to produce a mixed (transition)mode of deformation,involving some twinning but also enhanced slip(most likely due to activation non-basal c+a slip).The Mg–2.71wt%Y alloy shows no evidence of twinning.3.4.Euler maps of initial DRXLocal microtexture EBSD maps,showing regions containing a mixture of parent and DRX grains are shown in Fig.4.The superimposed unit cells illustrate the orientation of each grain. From these examples,it is clear that the orientation of the DRX grains are generally similar of that of the original grains.Itcan Fig.3.Optical micrographs after channel die compression:(a)pure magnesium compressed to a logarithmic strain of0.1at250◦C;(b)Mg–0.23wt%Y compressed to a logarithmic strain of0.1at250◦C;(c)Mg–0.84wt%Y compressed to a logarithmic strain of0.1at400◦C;(d)Mg–2.71wt%Y compressed to a logarithmic strain of0.3at450◦C.378R.Cottam et al./Materials Science and Engineering A485 (2008) 375–382Fig.4.Euler maps after PSC to a true strain of 0.3:(a)pure magnesium deformed at 250◦C;(b)Mg–0.23wt%Y deformed at 250◦C;(c)Mg–0.84wt%Y deformed at 400◦C;(d)Mg2.71wt%Y deformed at 450◦C.Grains labelled ‘P’are the parent grains.also be seen that the orientation of neighbouring DRX grains can be quite different,showing different angles and axes of rotation from the parent grain.This will be discussed in detail in a future paper.Pole figures {0002}for the pure magnesium deformed at different temperatures are shown in Fig.5.Rather than plot-ting intensity contours,the position of individual measurement points is shown.This is done to identify whether any new orien-tations are created during DRX,to characterize the orientation spread of poles and to identify regions for which orientations are absent.As discussed already,pole figures for DRX and parent grains are plotted separately,with the two-grain classes distinguished on the basis of their size.Fig.6shows the evolution of the {0002}pole figures for the yttrium containing alloys with increasing strain.Again,data for the DRX and parent grains are plotted separately.Fig.7shows {0002}pole figures for the final texture (after compression to a logrithmic strain of 0.9).In this case,intensity contours are shown to give an indication of the sharpness of the final texture.4.Discussion of results4.1.The development of texture with strain and temperature for pure MgIn the pure magnesium,distinct differences can be observed in the texture evolution at the three test temperatures.Thesedifferences can be largely attributed to changes in the relative contribution of twinning and slip to deformation.Formation ofthe common {10¯12}c -axis tension twin will lead to rapid reori-entation of the {0002}pole towards the ND,whereas basal slip will lead to a progressive rotation of the {0002}poles from the TD–RD plane to the ND.The high concentration of {0002}poles oriented close to the ND after deformation to a strain of 0.3at 200◦C is indica-tive of the high level of twinning that has already occurred under these conditions.As the deformation temperature is increased (at a strain of 0.3),an increasingly pronounced ring type tex-ture forms,with an absence of poles aligned in the ND.The ring texture most probably arises from partial rotation of the basal poles towards the ND by basal slip,with the markedly reduced intensity of poles around the ND reflecting the sharp reduction in twin formation with increasing deformation temper-ature [17].For the highest deformation temperature and strain (300◦C,0.6),rotation of the basal poles towards the ND by slip is,as expected,most advanced,and a (reoriented)basal texture is obtained.Of most interest in this work is the comparison of the ori-entations of the parent and DRX grains.It is clear from Fig.5that in pure magnesium,the orientation distribution of the DRX grains is similar to that of the parent grains.On a local level it has been shown by Galiyev et al.[8]that under the condi-tions used in this study,DRX in pure magnesium occurs largely by a continuous rotation mechanism.What the present result shows is that although any given DRX grain is rotating awayR.Cottam et al./Materials Science and Engineering A 485 (2008) 375–382379Fig.5.{0002}polefigures generated form EBSD Euler maps for pure magnesium where the horizontal axis is the rolling direction and the vertical is the transverse direction.from its parent grain,the development of its orientation is dom-inated by the macroscopic plastic strain imposed,and thus no special new orientations are created by DRX.In other words, although each DRX grain may rotate away from its neighbour during deformation,taken globally these rotations follow those of the parent grains,This confirms on a microtexture scale the conclusions of Kawalla et al.[18]and Al-Samman and Gottstein [19]who performed bulk texture analysis of uniaxial compres-sion at increasing strain intervals and found similar behaviour. This conclusion is supported by recent texture modelling work applied to equal channel angular extrusion at elevated tempera-ture[20].This shows that even when DRX is ignored,prediction offinal textures is good.The results presented here suggest that the average response of the DRX grains to deformation follows that of the parent grains.4.2.The effect of yttrium on texture developmentAs with an increase in temperature for pure magnesium,the combined increase in yttrium and deformation temperature used in this study leads to a transition in deformation behaviour and texture development.Unfortunately,despite taking care to sam-ple a large number of grains(over6000),in some cases(as shown in Fig.6d,f and h in particular)it is clear that orientations that should be present from a consideration of symmetry have not been detected.Despite this limitation,the data are adequate to draw qualitative conclusions about the influence of yttrium on the texture evolution of the parent and DRX grains.Fig.7shows that thefinal texture after deformation to a strain of0.9is similar for the pure magnesium and Mg–0.23wt%Y deformed at250and300◦C.The most notable difference in tex-tures is after deformation at250◦C,where the Mg–0.23wt%Y shows a basal,rather than split basal texture,as observed in the pure Mg and the Mg–0.23wt%Y deformed at300◦C.The rea-son for this difference is not yet clear,but may be due to the observed change in the fraction and scale of twins with yttrium (Fig.3).The0.23and0.84wt%Y alloys show similar behaviour with respect to the relationship between DRX and parent grain ori-entations.As in the case of pure Mg,in these alloys,the DRX380R.Cottam et al./Materials Science and Engineering A485 (2008) 375–382Fig.6.{0002}pole figures generated form EBSD Euler maps for the three Mg–Y alloys compressed in plane strain compression at elevated temperature.The horizontal axis is the rolling direction and the vertical is the transverse direction.and parent grains show similar orientation distributions,with evidence that the spread of orientations in the DRX grains is slightly greater than in the parent grains.This is in contrast to pure Mg,where a similar spread in orientation is observed.DRX in these alloys randomizes the texture.The final texture observed in the 0.84wt%Y alloy (Fig.7)is relatively weak,and represents a transition in behaviour from the macroscopic textures observed in the both the pure Mg and Mg–0.23wt%Y to that observed in the Mg–2.71wt%Y alloy.Increasing the addition of Y to 2.71wt%Y and the defor-mation temperature to 400◦C results in a change in the texture formation of both the parent and DRX grains.This change in behaviour can most likely be attributed to a reduction in the relative critical resolved shear stress (CRSS)for c +a slip com-pared to basal slip and twinning deformation modes.The initial texture of the extruded material was relatively weak in this alloy,and this makes interpretation of the results for low strain defor-mation difficult.However,once a logarithmic strain of 0.6is reached,it can be seen that a strikingly different texture has developed,and this is reflected in the final texture intensity.In addition,the DRX grains show a different,and far more random distribution of orientations than the parent grains.The randomization of texture has been observed after the static recrystallization of Mg–Y containing alloy [21].In this work,it was argued that shear banding formed random nuclei,which during recrystallization led to the random texture.How-ever,this does not explain the results seen here,where DRX grains were found to form at grain boundaries,not shear bands,and are also subject to extensive deformation after forming.The most likely explanation comes form the effect of adding yttrium on the mechanism of DRX,which will be discussed in a future paper.It should also be noted that while the initial DRX texture is random,the macroscopic texture after deformation to 0.9log-arithmic strain is quite strong.This suggests that when the DRXR.Cottam et al./Materials Science and Engineering A 485 (2008) 375–382381Fig.7.Polefigures after channel die deformation to a logarithmic strain of0.9for the alloys and pure Mg at a range of temperatures.grains initially form in this alloy they do not have a strong pre-ferred orientation,but subsequent deformation causes them to rotate towards the TD,hence producing the strongfinal texture.A relatively random texture has also been observed after deformation of AZ61at elevated temperatures[22],under condi-tions where extensive DRX occurred.The origins of this texture were not given but,as in this work,it was noted that the random-ization process occurred only with higher solute levels,which again points to the influence of solute in randomizing the DRX texture.5.ConclusionsIn this study,the evolution of texture of DRX and parent grains was studied for pure magnesium at a range of temper-atures,and binary magnesium–yttrium alloys at temperatures where extensive DRX occurs.As yttrium suppresses dynamic recrystallization,it was necessary to deform the high yttrium containing alloys at higher temperatures to obtain comparable DRX fractions.Increasing deformation temperature and yttrium content had the effect of reducing the importance of twinning as a deforma-tion mode and increasing the activity of non-basal slip modes. This was reflected in the texture evolution.In pure magnesium and Mg–0.23wt%Y,basal or split basal textures were observed. In Mg–2.71wt%Y,a completely different texture was seen,with basal poles aligned in the TD and RD.In the Mg–0.84wt%Y alloy a transition texture between these extremes was observed.It has been shown that in pure magnesium,the texture of the DRX grains followed closely that of the parent grains,i.e. it is the deformation conditions that determine the texture of the DRX grains,rather than special conditions of nucleation or growth.The addition of yttrium at low levels(0.23wt%)appears to slightly randomize the texture of the DRX grains compared to that in pure magnesium deformed at the same temperature. This effect is also observed in Mg–0.84wt%Y alloy deformed at higher temperature.Addition of higher levels of yttrium (2.71wt%)combined with a higher deformation temperature lead to a change in behaviour,with a near random distribution of DRX grains being in contrast to strongly textured parent grains. AcknowledgementsThis work has been carried out as part as part of the‘Light Alloys Portfolio Partnership’EPSRC grant EP/C002210/1.The authors would like to thank Magnesium Elektron for theirfinan-cial and technical support,Mr.Michael Faulkner for his technical support.References[1]B.C.Wonsiewicz,W.A.Backofen,TMS-AIME(1967)1422–1431.[2]S.R.Agnew,O.Duygulu,Int.J.Plasticity21(2005)1161–1193.[3]S.R.Agnew,C.N.Tome,D.W.Brown,T.M.Holden,S.C.V ogel,ScriptaMater.48(2003)1003–1008.[4]D.W.Brown,S.R.Agnew,M.A.M.Bourke,T.M.Holden,S.C.V ogel,C.N.Tome,Mater.Sci.Eng.A399(2005)1–12.[5]J.Koike,T.Kobayashi,T.Mukai,H.Watanabe,M.Suzuki,K.Maruyama,K.Higashi,Acta Mater.51(2003)2055–2065.[6]S.R.Agnew,J.A.Horton,T.M.Lillo,D.W.Brown,Scripta Mater.50(2004)377–381.[7]A.Galiyev,R.Kaibyshev,G.Gottstein,Acta Mater.49(2001)1199–1207.382R.Cottam et al./Materials Science and Engineering A 485 (2008) 375–382[8]A.Galiyev,R.Kaibyshev,T.Sakai,Mater.Sci.Forum419–422(2003)509–514.[9]X.Yang,H.Miura,T.Sakai,Mater.Sci.Forum467–470(2004)531–536.[10]S.E.Ion,F.J.Humphreys,S.H.White,Acta Metall.30(1982)1909–1919.[11]F.Xiong,C.H.J.Davies,Magnesium Technology2005,TMS,San Fran-cisco,2005.[12]S.R.Agnew,M.H.Yoo,C.N.Tome,Acta Mater.49(2001)4277–4289.[13]A.M.Wusatowska-Sarnek,H.Miura,T.Sakai,Mater.Sci.Eng.A323(2002)177–186.[14]O.V.Mishin,V.Y.Gertsman,G.Gottstein,Mater.Character.38(1997)39–48.[15]M.R.Barnett,Z.Keshavarz,A.G.Beer,D.Atwell,Acta Mater.52(2004)5093–5103.[16]M.R.Barnett,D.Atwell,A.G.Beer,Mater.Sci.Forum467–470(2004)435–440.[17]A.Jain,S.R.Agnew,Mater.Sci.Eng.A462(2007)29–36.[18]R.Kawalla, C.Schmidt,H.Reidel, A.Prakash,DGM,Dresden,2006.[19]T.Al-Samman,G.Gottstein,DGM,Wiley-VCH,Dresden,2006.[20]S.R.Agnew,P.Mehrotra,T.M.Lillo,G.M.Stoica,P.K.Liaw,Acta Mater.53(2005)3135–3146.[21]J.W.Senn,S.R.Agnew,Magnesium Technology in the Global Age,Mon-treal,2006.[22]L.Helis,K.Okayasu,H.Fukutomi,DGM,Wiley-VCH,Dresden,2006.。