M A T E R I A L S C H A R A C T E R I Z A T I O N 9 4 ( 2 0 1 4 ) 6 9 – 7 9

j

Ava i l ab l e on l i ne a t www.sc i enced i r ec t . com

ScienceDirectwww.e l sev i e r . com/ loca te /matcha r

Microstructure stability of ultra-fine grained

magnesium alloy AZ31 processed by extrusion andequal-channel angular pressing (EX–ECAP)

Jitka Stráskáa,⁎, Miloš Janečeka, Jakub Čížekb, Josef Stráskýa, Branislav Hadzimac,d

aDepartment of Physics of Materials, Faculty of Mathematics and Physics, Charles University in Prague, Ke Karlovu 5, 121 16 Praha 2, CzechRepublicbDepartment of Low Temperature Physics, Faculty of Mathematics and Physics, Charles University in Prague, V Holešovičkách 2, 180 00 Praha8, Czech RepubliccUniversity of Žilina, Research Centre, Univerzitná 1, 010 26 Žilina, Slovak RepublicdDepartment of Materials Engineering, Faculty of Mechanical Engineering, University of Žilina, Univerzitná 1, 010 26 Žilina, Slovak Republic

A R T I C L E D A T A

⁎ Corresponding author.E-mail addresses: straska.jitka@gmail.com

osef.strasky@gmail.com (J. Stráský), branislav

http://dx.doi.org/10.1016/j.matchar.2014.05.01044-5803/© 2014 Elsevier Inc. All rights rese

A B S T R A C T

Article history:Received 19 November 2013Received in revised form28 March 2014Accepted 22 May 2014

Thermal stability of the ultra-fine grained (UFG) microstructure of magnesium AZ31 alloywas investigated. UFG microstructure was achieved by a combined two-step severe plasticdeformation process: the extrusion (EX) and subsequent equal-channel angular pressing(ECAP). This combined process leads to refined microstructure and enhanced microhard-ness. Specimens with UFG microstructure were annealed isochronally at temperatures150–500 °C for 1 h. The evolution of microstructure, mechanical properties and dislocationdensity was studied by electron backscatter diffraction (EBSD), microhardness measurementsand positron annihilation spectroscopy (PAS). The coarsening of the fine-grained structure athigher temperatureswas accompanied by a gradual decrease of themicrohardness anddecreaseof dislocation density. Mechanism of grain growth was studied by general equation for graingrowth and Arrhenius equation. Activation energies for grain growth were calculated to be 115,33 and 164 kJ/mol in temperature ranges of 170–210 °C, 210–400 °C and 400–500 °C (443–483 K,483–673 K and 673–773 K), respectively.

© 2014 Elsevier Inc. All rights reserved.

Keywords:Microstructure stabilitySevere plastic deformation (SPD)Ultrafine-grained (UFG) materialsEqual-channel angular pressing(ECAP)AZ31 magnesium alloy

1. Introduction

Magnesium alloys belong to materials with potential to replacesome conventional structural materials in automotive, aircraft

(J. Stráská), janecek@met.hadzima@fstroj.uniza.sk

13rved.

and other industry branches. Magnesium is a very light metalwith relatively good mechanical properties which result inits expanding use in weight-critical applications. Interest inmagnesium-basedmaterials has recently been revived primarily

.mff.cuni.cz (M. Janeček), jcizek@mbox.troja.mff.cuni.cz (J. Čížek),(B. Hadzima).

70 M A T E R I A L S C H A R A C T E R I Z A T I O N 9 4 ( 2 0 1 4 ) 6 9 – 7 9

due to its gradually decreasing cost and the determination ofscientists, researchers and engineers to cut down energy con-sumption and greenhouse gas emissions [1].

Ultra-fine grained (UFG) materials with submicrometeror even nano-scale grain sizes can be produced by severe plasticdeformation (SPD) techniques. These methods are very efficientin achieving significant grain refinement in polycrystallinemetals. These UFG materials have usually superior mechanicalproperties including high strength and, if the UFGmicrostructureis sufficiently stable, a superplastic capability at elevated tem-peratures [2,3]. Nowadays, the most attractive procedures forprocessing by SPD are equal-channel angular pressing (ECAP) [4]or combined process of extrusion followed by ECAP (EX–ECAP) [5],high-pressure torsion (HPT) [6] and accumulative roll-bonding(ARB) [7]. In practice, ECAP or EX–ECAP processes are especiallyattractive because of their simplicity in laboratory operation.Moreover, these processes can produce relatively large billets forindustrial applications [8]. There are several reports to dateof the successful processing of magnesium AZ31 alloy usingECAP at elevated temperatures and employing different pro-cessing procedures [9–13].

Extremely small grain sizes of polycrystalline materialsusually lead to enhanced mechanical properties. However, grainsize refinement achieved by ECAP or EX–ECAP may not play amajor role in the yield stress of thematerial. H.K. Kim [14] claimsthat the AZ31 alloy after 1 and 2 ECAP passes with smaller grainsizes shows a lower yield stress compared to the extruded only(no ECAP passes) material with large grains due to texturemodification during the ECAP process. According to hispredictions of flow stress, the strain hardening is the largestcontributor to flow stress whereas grain size refinementplays a relativelyminor role in ECAPed specimens.Masoudpanahet al. [15] also reported lower yield stress of the AZ31 alloyafter 4 ECAP passes compared to the extruded state. Jin et al.[16] studied in detail the microstructure evolution duringECAP, proving the rearranging of dislocations induced in theinitial stage of ECAP and their formation into sub-grainboundaries. The newly created dislocations in sub-grains areabsorbed by the sub-boundaries with the increasing straininduced by additional ECAPpasses. Themisorientations betweensub-boundaries gradually increase and evolve to low and high-angle grain boundaries. Nevertheless, functional properties likebiocompatibility, special magnetic properties or corrosionresistance (investigated in our previous paper [17]) should bealso considered.

Thepractical applicationsof theUFGmaterials are limiteddueto low microstructure stability at elevated temperatures thatcomplicates the processing of final products. Thermal stabilitydepends on many variables, such as stacking fault energy ofthe material, processing or properties of grain boundaries [18].Microstructure stability can be improved by various alloyingelements or composite reinforcements. Microstructure stabilityof the AZ31 alloy after ECAP was studied by Kim [19] or by RadiandMahmudi [20], who investigated the AZ31 alloy reinforced byalumina nano-particles. Both papers present calculations ofactivation energies for grain growth which identified two orthree temperature regimes with significantly different values ofactivation energy.

The main objective of this work is the investigation ofmicrostructure stability during annealing of the UFG AZ31

magnesium alloy prepared by extrusion and 4 passes ofECAP.

2. Material and methods

As-cast commercial AZ31 alloy (nominal composition of Mg–3%Al–1%Zn) was first extruded at 350 °C with an extrusionratio of 22. Subsequently, it was processed by equal channelangular pressing. ECAP pressing was performed at 180 °Cfollowing route BC, i.e. rotating the sample 90° between theindividual passes, with the velocity of 50 mm/min. The angleΦ between two intersecting channels and the corner angle Ψwere 90° and 0°, respectively. Both channels have a squarecross section of 10 mm × 10 mm. The ECAP die was equippedwith an ejector that allows pushing the sample out of the dieimmediately after pressing from the feed-in channel to the exitchannel. Molybdenum disulfide grease was used as a lubricant.All the specimens were pressed four-times through the ECAPdie.

Investigated flat specimens with approx. 2 mm thicknesswere cut from the middle part of the billets perpendicular tothe pressing direction. Microstructural observations andmicrohardness measurements on planes parallel to thepressing direction are very similar to the perpendicular one[21] and they are not studied in this work.

Series of specimens for thermal-stability investigationwere prepared by isochronal annealing at the temperatures150–500 °C for 1 h followed by water-quench.

Specimens were mechanically grinded on watered abra-sive papers, and then polished with polishing diamondsuspension of grades 3, 1 and 1/4 μm and alumina sus-pension of grade 0.05 μm. Using this procedure, flat samplesfor Vickers microhardness measurements (load 100 g, 10 s)with minimum surface scratches were obtained. Finally,the specimens' surface was polished by argon ions (GatanPIPS™) which enables EBSD measurements. Specimens forpositron annihilation spectroscopy measurements wereadditionally etched for 30 s in the mixture of nitric acid,water and ethylene glycol.

Microhardness was measured by LECO M-400-A micro-hardness meter. A FEI Quanta 200 FX scanning electronmicroscope equipped with EDAX EBSD camera and OIMsoftware was utilized for EBSD observations.

A 22Na2CO3 positron source with the activity of 1.5 MBqwas used in positron lifetime measurements. The sourcespot with diameter of 1 mm was deposited on a 2 μm thickMylar foil and sealed between two identical specimens of thestudied material. The source contribution consists of twocomponents with lifetimes of 368 ps (intensity 8%) and1.5 ns (intensity 1%) which come from positrons annihilatedin the 22Na2CO3 source spot and in the covering Mylar foil,respectively.

A fast–fast spectrometer [22] with the time resolution of150 ps (FWHM 22Na) was employed for positron lifetimemeasurements. At least 107 positron annihilation eventswere accumulated in each positron lifetime spectrum whichwas subsequently decomposed into individual exponentialcomponents by a maximum likelihood procedure [23].

Fig. 2 – Microstructure of the extruded AZ31 alloy (lightmicroscopy, polarized light).

71M A T E R I A L S C H A R A C T E R I Z A T I O N 9 4 ( 2 0 1 4 ) 6 9 – 7 9

3. Results

3.1. Mechanical properties

For clarity and further comparisons, we present the resultsof Vickers microhardness HV0.1 measurement (at least 15indents were made in each sample) that were presentedalready in our previous work [24]. The first data point in Fig. 1(HV0.1 = 86) corresponds to the initial severely deformedspecimen. Microhardness values after annealing at 150 and170 °C do not differ significantly. However, the microhardnessdeclines abruptly in the range of 170–230 °C and thencontinuously to 500 °C.

3.2. Microstructure

The microstructure of the extruded AZ31 alloy was observedusing light microscopy. The microstructure is bimodal con-taining large grains elongated in the extrusion direction (tensof microns) and smaller grains (few microns). Typical micro-structure of the extruded sample cut perpendicular to theextrusion direction is shown in Fig. 2.

The microstructure was investigated using EBSD. Themicrostructure and grain size distribution of the specimen inthe initial non-annealed condition (after extrusion and 4passes of ECAP) are shown in Fig. 3(a) and (b), respectively.The microstructure is homogeneous comprising very finegrains of the average size of 0.9 μm. The microstructure andgrain sizes of the samples after 1 h of isochronal annealing at150 and 170 °C (not shown here) are very similar to the non-annealed specimen.

Inhomogeneous grain growth is observed at higher an-nealing temperatures (Figs. 4–10). Some grains start to grow attemperatures of 190 and 210 °C (the microstructure of thesample after annealing at 210 °C is very similar to that of190 °C, and is not shown here). The fraction of coarse grainsincreases with increasing annealing temperature. At a

Fig. 1 – Microhardness of the AZ31 specimens processed byextrusion and 4 passes of ECAP after 1 h of isochronalannealing.

Fig. 3 – Microstructure of the AZ31 sample after extrusionand 4 passes of ECAP, (a) inverse pole figure and (b) grain sizedistribution.

Fig. 4 – Microstructure of the AZ31 EX–ECAP specimen after1 h of isochronal annealing at 190 °C, (a) inverse pole figureand (b) grain size distribution.

Fig. 5 – Microstructure of the AZ31 EX–ECAP specimen after1 h of isochronal annealing at 250 °C, (a) inverse pole figureand (b) grain size distribution.

72 M A T E R I A L S C H A R A C T E R I Z A T I O N 9 4 ( 2 0 1 4 ) 6 9 – 7 9

temperature of 250 °C, some areas with original fine grainsare still observed. However, the small grains are continuouslydisappearing at higher annealing temperatures and nearly nosmall grains are observed after annealing at 400 °C (see Fig. 8).Please note that the magnification of Figs. 7(a) and 8(a) is twotimes smaller than the magnification of the previous inversepole figures; simultaneously, four times bigger area of thesample (100 × 100 μm) is depicted in Figs. 7 and 8. Microstruc-ture of the specimens annealed at 450 and 500 °C was observedby light microscope— see Figs. 9 and 10, respectively.

Grain size distribution and average grain size of thesamples studied by EBSD were calculated from data mea-sured in the area 100 × 100 μm for all annealing tempera-tures to achieve better statistics. Statistical values from thesamples studied by light microscopy were calculated fromtwo or more images in order to get more than 1000 grains foreach sample. Grains intersecting the edge of EBSD maps andlight microscopy images were not included in the analysis.The dependence of average grain size (number average) onannealing temperature is plotted in Fig. 11. In the case of the

samples annealed at 250 and 300 °C, the average values arecalculated from the bimodal grain size distribution.

Annealing twins observed after annealing at 250–400 °C(see Figs. 5(a)–8(a)) were excluded from grain size calcula-tions to achieve true grain size values (twin boundaries wereignored in grain size calculations, i.e. the twin is consideredto be a part of the grain). All these twins were determined asthe tensile twins with misorientation angle of 86° [25].

3.3. Dislocation density

The plastic shear deformation by EX–ECAP causes the accumu-lation of large plastic strain and the increase of density ofstructural defects. These defects are stable at room tempera-ture, but they annihilate during annealing.

Positron lifetime spectra of all samples studied can be wellfitted by two exponential components. The shorter com-ponent with the lifetime τ1, which is lower than the bulkpositron lifetime in Mg τΒ = 225 ps [26], represents a contri-bution of free positrons not trapped at defects. The longer

Fig. 6 – Microstructure of the AZ31 EX–ECAP specimen after1 h of isochronal annealing at 300 °C, (a) inverse pole figureand (b) grain size distribution.

Fig. 7 – Microstructure of the AZ31 EX–ECAP specimen after1 h of isochronal annealing at 350 °C, (a) inverse pole figureand (b) grain size distribution.

73M A T E R I A L S C H A R A C T E R I Z A T I O N 9 4 ( 2 0 1 4 ) 6 9 – 7 9

component with lifetime τ2 ≈ 260 ps arises from positronstrapped at dislocations [27].

The development of positron lifetimes τ1 and τ2 withincreasing annealing temperature is plotted in Fig. 12(a), whileFig. 12(b) shows the dependence of the intensity I2 of positrontrapped at dislocations on annealing temperature. The lifetimeτ2 of positrons trapped at dislocations remains approximate-ly constant during annealing confirming that the nature ofpositron traps does not change. The intensity I2 of positronstrapped at dislocations decreases with increasing annealingtemperature and becomes immeasurable at the temperatureof 300 °C.

Dislocation density ρD can be calculated from positronlifetime data using the two-state simple trapping model [28].The details of this procedure are described elsewhere [29]. Thedependence of the mean dislocation density ρD (calculatedfrom Eq. (1) in [29]) for the samples subjected to annealingtreatment at various temperatures is plotted in Fig. 13. Disloca-tion density decreases with increasing annealing temperatureand falls below the detection limit of positron annihilationspectroscopy at temperatures T ≥ 300 °C.

4. Discussion

4.1. Correlation between mechanical properties, microstructureand lattice defect evolution

It follows frommicrohardnessmeasurements (Fig. 1) that UFGmicrostructure of AZ31 alloy is stable up to 170 °C. Afterannealing at temperatures higher than 190 °C, a sharp drop ofmicrohardness occurred. A detailed inspection of the micro-hardness–temperature plot indicates a two-step character ofthe microhardness decline. In the lower annealing tempera-ture range (170–210 °C) the decline is much sharper while forhigher annealing temperatures (T > 210 °C) the slope of thecurve is significantly lower.

This two-step character of the curve suggests a change ofthe mechanism controlling the degradation of mechanicalproperties.

The strength and hardness of severely deformed ultra-finegrained material are affected mainly by the dislocation density[30] and the grain size according to Hall–Petch relation [31,32].

Fig. 8 – Microstructure of the AZ31 EX–ECAP specimen after1 h of isochronal annealing at 400 °C, (a) inverse pole figureand (b) grain size distribution.

Fig. 9 – Microstructure of the AZ31 EX–ECAP specimen after1 h of isochronal annealing at 450 °C, (a) light microscopyimage and (b) grain size distribution.

74 M A T E R I A L S C H A R A C T E R I Z A T I O N 9 4 ( 2 0 1 4 ) 6 9 – 7 9

Therefore the grain coarsening and the annihilation of disloca-tions during annealing are expected to control the materialstrength and hardness.

Fig. 11 showing the grain size evolution and Fig. 13 showingthe dislocation density evolution with annealing temperaturemust be simultaneously inspected. In the low temperatureregion of the microhardness drop (T ≈ 170–210 °C) the graingrowth is negligible, whereas the dislocation density graduallydeclines indicating a recovery of dislocation structure. Mostprobably rearrangement and mutual annihilation of disloca-tions with opposite signs take place during annealing in thistemperature range. As seen in Fig. 5 the fine grain structurebecomes unstable and significant grain growth is observed attemperatures T > 210 °C. In this temperature range, the dislo-cation density is very low, falling below the detection limit ofPAS (ρD ≈ 1012 m−2) at T ≈ 300 °C.

Frommicrostructure observation (EBSD) and lattice defectdensity estimation (PAS) one can conclude that in the lowerannealing temperature region (T ≈ 180–210 °C) it is mostlythe annihilation of dislocations which causes the drop of

microhardness. At higher temperatures (T > 210 °C), proba-bly the grain growth influences significantly the hardness ofinvestigated material.

4.2. Grain growth analysis

The determination of grain size in ultra-fine grain materialallows us to analyze the mechanisms of grain growth duringisothermal annealing. Two microstructural aspects may bedetermined:

a) The activation energy of grain growthThe grain growth mechanism during static annealing canbe determined from calculated activation energy for graingrowth. For this analysis we can use the general equationfor the grain growth

dn−dn0 ¼ kt; ð1Þ

Fig. 10 – Microstructure of the AZ31 EX–ECAP specimen after1 h of isochronal annealing at 500 °C, (a) light microscopyimage and (b) grain size distribution.

Fig. 11 – Dependence of the average grain size (numberaverage, excluding twins) of the EX–ECAP AZ31 alloy onannealing temperature after 1 h of isochronal annealingprocess.

Fig. 12 – Results of positron lifetime measurements for AZ31EX–ECAP specimens subjected to 1 h of isochronal annealingat various temperatures. (a) Lifetimes τ1 (empty squares) andτ2 (full circles) of the free positron and the dislocationcomponent. (b) Intensity I2 of the components arising frompositrons trapped at dislocations.

Fig. 13 – Dislocation density of the specimens after 1 h ofisochronal annealing.

75M A T E R I A L S C H A R A C T E R I Z A T I O N 9 4 ( 2 0 1 4 ) 6 9 – 7 9

76 M A T E R I A L S C H A R A C T E R I Z A T I O N 9 4 ( 2 0 1 4 ) 6 9 – 7 9

where d is the grain size at given annealing time, d0 is theinitial grain size, n is the grain growth exponent, t is theannealing time and k is a temperature-dependent constantwhich can be described by Arrhenius equation

k ¼ k0 exp −QRT

� �; ð2Þ

where k0 is a constant, Q is the activation energy forgrain growth, R is the gas constant and T is the absolutetemperature.Thevalueof the stress exponentn is of primary importance. Inthe ideal case (infinite crystal with no defects), the graingrowth exponent n should be equal to 2. However, very oftenhigher values of n are found and attributed to various factorsaffecting grain growth kinetics, such as the free surface effect,impurity-drag, texture, dislocation substructure and micro-structure heterogeneities [33]. Previous studies [34–36] report-ed a value of n in a range from 2 to 8 for various magnesiumalloys and magnesium-based composites. Higher values of n(n ≥ 5)were observedmainly in ultra-fine grainedmagnesiummaterials produced by mechanical alloying [35,36]. The valueof grain growth exponent n observed in ultrafine-grainedmagnesium alloy AZ31 produced by various techniques ofsevere plastic deformation ranges between 2 and 4 [34,37,38].The AZ31 alloy processed by the most similar conditions(ECAP without previous hot extrusion, where the averagegrain size after 4 passes was equal to 2.5 μm) was studied byKim and Kim [19,34]. The grain growth exponent n used intheir calculations was equal to 2. We use the same value of n,whichwill allowus tomake comparisonwithKim's and otherauthors' works.Substituting Eq. (2) into Eq. (1) one can determine theactivation energy Q as the slope of the dependence ofln(d2 − d02) on T−1which is shown in Fig. 14. Three temperatureranges with different Q values can be distinguished. Thecalculated values of activation energy for grain growthare 115,33 and 164 kJ/mol in the temperature ranges 170–210 °C,210–400 °C and 400–500 °C (443–483 K, 483–573 K and 573–673 K), respectively. These three temperature ranges with

Fig. 14 – Plot of ln(d2 − d02) vs. T−1 for the estimation of the

activation energy for grain growth of the EX–ECAPmagnesium alloy AZ31.

different Q values were observed in other fine-grained AZ31alloys in various conditions and the respective temperatureranges are very similar with our temperature ranges [20,34].In the low temperature range (T<210 °C), the activationenergy is relatively high— higher than activation energy forgrain boundary diffusion in pure magnesium (92 kJ/mol[39]). This result is consistent with a well-known fact thatactivation energy of alloys should be higher than activationenergy of pure metals. In this temperature range, thedislocation density within the grains decreases with in-creasing temperature but it remains relatively high.In the high temperature range (T>400 °C), the activationenergy Q is equal to 164 kJ/mol, which is higher than latticeself-diffusion in pure magnesium (135 kJ/mol [40]). Thelattice self-diffusion is activated and grain growth isenhanced leading to fully-recrystallized structure.In the intermediate temperature range, the value of Q isabnormally low. Similarly low value of Q was reported byWang et al. [41] in the ECAPed Al–Mg alloy annealed atthe temperatures T ≤ 275 °C. The authors attribute theabnormally low value of Q to the non-recrystallized micro-structure with a certain fraction of non-equilibrium grainboundaries. This conclusion is consistent with the conceptof reduced activation energy for grain boundary diffusion inultra-fine grained materials produced by severe plasticdeformation caused by the ability of the non-equilibriumgrain boundaries to provide enhanced atomic mobility[42,43]. The AZ31 alloy after extrusion and 1 pass of ECAPcontains a significant number of non-equilibrium grainboundaries. However, the fraction of non-equilibrium grainboundaries decreases with increasing number of ECAPpasses so that nearly no such grain boundaries are observedin more deformed AZ31 alloy (see our previous paper [44]).Better explanation of this abnormally low activation energyin the intermediate temperature range is provided by Kimand Kim [34] — they claim that the value of Q in thistemperature range ismeaningless. This can happenwhen k0in Eq. (2) varies with temperature in the intermediatetemperature range.Note that in the intermediate temperature range, thedislocation density decreases very fast with increasingannealing temperature and the dislocation density getsbelow the limit of PAS resolution in the high temperaturerange. High dislocation density is known to increasediffusivity of material due to the so-called pipe diffusion. Itis therefore argued in line with Kim and Kim [34] that k0declines with increasing temperature due to decreasingdislocation density causing decrease in diffusivity. Assump-tion of constant k0 is invalid in this case and leads to asignificant underestimation of activation energy.

b) Hall–Petch relationEBSD analysis allows us to determine the validity of Hall–Petch (HP) relation for isochronally annealed UFG speci-mens in the temperature range up to 400 °C. For thisanalysis the HP relation yields

HV ¼ H0 þ KHd−12; ð3Þ

where H0 and KH are the material constants and HV is themeasured value of microhardness.

Table 1 –Microhardness values and average grain sizes at different annealing temperatures.

Annealing temperature [°C] – 170 190 210 250 300 350 400 450 500Microhardness HV0.1 85.8 84.1 78.0 71.6 67.6 65.4 63.2 59.3 57.7 51.6Average grain size d [µm] 0.94 0.99 1.05 1.48 1.83 2.06 3.04 3.79 10.09 24.53

77M A T E R I A L S C H A R A C T E R I Z A T I O N 9 4 ( 2 0 1 4 ) 6 9 – 7 9

The results of HV and d obtained from measured valuesshown in Figs. 1 and 11, respectively, are summarized inTable 1 and plotted in Fig. 15.

The constants H0 and KH may be calculated from theparameters of a straight line depicted in Fig. 15. The bestlinear fit was applied only to data corresponding to higherannealing temperatures (from 250 to 500 °C). In this temper-ature range, only the grain size affects material hardness asthe dislocation density is low. At low temperatures, both grainsize and dislocation density contribute to strengtheningand the linear fit of microhardness data fails. Data for lowannealing temperatures (i.e. high dislocation density condi-tions) lie clearly above the Hall–Petch fit (marked by the redarrow).

The calculated material constants from the high temper-ature fit of HV vs. d-1/2 are: H0 = 47 ± 2 and KH = 27 ± 3 μm1/2.These values are partly comparable to those reported onECAPed Al alloys with H0 = 35–47 and KH = 35–50 μm1/2 [45]but different from those reported on the only ECAPed AZ31alloy by Kim and Kim [34] (H0 = 38, KH = 42). But theirconstants were calculated from the linear fit of the wholetemperature range because the changes of dislocation densi-ties were not taken into consideration. It causes that theircalculated constants H0 and KH are underestimated andoverestimated, respectively, in comparison with our calculat-ed constants. Our value of the constant H0 is closer to themicrohardness value of the AZ31 in annealed condition(HV0.1 = 58 ± 3, see our previous paper [46]) in comparisonwith the value of H0 calculated by Kim and Kim [34].

Fig. 15 – Hall–Petch relationship for the isochronallyannealed EX–ECAP AZ31 alloy based on HV0.1microhardness data.

5. Conclusions

Extrusion and 4 passes of ECAP pressing (EX–ECAP) resultedin strong microstructure refinement. The average grain sizewas determined by EBSD to be approximately 900 nm. Thismaterial was isochronally annealed at temperatures 150–500 °C for 1 h. Several experimental techniques (micro-hardness measurements, EBSD and PAS) were employed toidentify the softening processes operating in this temperaturerange. The following conclusions can be drawn from thisinvestigation:

1) At lower annealing temperatures (≈190 °C) softeningmanifested by microhardness drop is mainly controlledby dislocation annihilation (proved by PAS). On the otherhand, at higher annealing temperatures (≈300 °C andhigher), grain growth becomes the principal softeningmechanism (proved by EBSD).

2) The kinetics of grain growth was described by the kineticequation d2 − d02 = kt and the activation energies for graingrowth were determined to be 115, 33 and 164 kJ/molin temperature ranges of 170–210 °C, 210–400 °C and400–500 °C (443–483 K, 483–673 K and 673–773 K), respec-tively. The value of activation energy in the intermediatetemperature range is underestimated due to the decreas-ing dislocation density causing decrease in diffusivity.

3) Vickers microhardness evolution in higher temperaturerange (250 °C–500 °C) can be successfully explained byHall–Petch relation. In lower temperature range, thedislocation density plays significant role and the micro-hardness is significantly higher than extrapolation of Hall–Petch fit. The effect of dislocations must be taken intoaccount to get reliable estimates of parameters of Hall–Petch relation.

Acknowledgments

This study was financially supported by the GACR under theGrant P107/13/13616S. J. Stráská acknowledges the financialsupport by student grants GAUK 530712/2012 and SVV2014-267303. B. Hadzima acknowledges financial support bythe European Regional Development Fund and Slovak StateBudget under the project “Research Centre of the University ofŽilina” with ITMS code 26220220183.

R E F E R E N C E S

[1] Gupta M, Sharon NML. Magnesium, Magnesium Alloys, andMagnesium Composites. New Jersey: John Wiley & Sons, Inc.;2011. p. 1–3.

78 M A T E R I A L S C H A R A C T E R I Z A T I O N 9 4 ( 2 0 1 4 ) 6 9 – 7 9

[2] Valiev RZ, Islamgaliev RK, Alexandrov IV. Bulk nanostructuredmaterials from severe plastic deformation. Prog Mater Sci2000;45:103–89.

[3] Valiev RZ, Estrin Y, Horita Z, Langdon TG, Zehetbauer MJ, ZhuYT. Producing bulk ultrafine-grained materials by severeplastic deformation. JOM 2006;58:33–9.

[4] Valiev RZ, Langdon TG. Principles of equal-channel angularpressing as a processing tool for grain refinement. Prog MaterSci 2006;51:881–981.

[5] Matsubara K, Miyahara Y, Horita Z, Langdon TG.Developing superplasticity in a magnesium alloy through acombination of extrusion and ECAP. Acta Mater2003;51:3073–84.

[6] Zhilyaev AP, Langdon TG. Using high-pressure torsion formetal processing: fundamentals and applications. Prog MaterSci 2008;53:893–979.

[7] Saito Y, Utsunomiya H, Tsuji N, Sakai T. Novel ultra-highstraining process for bulk materials— development of theaccumulative roll-bonding (ARB) process. Acta Mater1999;47:579–83.

[8] Horita Z, Fujinami T, Langdon TG. The potential forscaling ECAP: effect of sample size on grain refinementand mechanical properties. Mater Sci Eng A2001;318:34–41.

[9] Agnew SR, Horton JA, Lillo TM, Brown DW. Enhancedductility in strongly textured magnesium produced byequal channel angular processing. Scr Mater2004;50:377–81.

[10] Janeček M, Popov M, Krieger MG, Hellmig RJ, Estrin Y.Mechanical properties and microstructure of a Mg alloy AZ31prepared by equal-channel angular pressing. Mater Sci Eng A2007;462:116–20.

[11] Figueiredo RB, Langdon TG. Developing superplasticity in amagnesium AZ31 alloy by ECAP. J Mater Sci2008;43:7366–71.

[12] Figueiredo RB, Langdon TG. Grain refinement andmechanical behavior of a magnesium alloy processed byECAP. J Mater Sci 2010;45:4827–36.

[13] Xu J, Shirooyeh M, Wongsa-Ngam J, Shan D, Guo B, LangdonTG. Hardness homogeneity and micro-tensile behavior in amagnesium AZ31 alloy processed by equal-channel angularpressing. Mater Sci Eng A 2013;586:108–14.

[14] Kim HK. The grain size dependence of flow stress in anECAPed AZ31 Mg alloy with a constant texture. Mater Sci EngA 2009;515:66–70.

[15] Masoudpanah SM, Mahmudi R. Effects of rare-earth elementsand Ca additions on the microstructure and mechanicalproperties of AZ31 magnesium alloy processed by ECAP.Mater Sci Eng A 2009;526:22–30.

[16] Jin L, Lin D, Mao D, Zeng X, Chen B, Ding W. Microstructureevolution of AZ31 Mg alloy during equal channel angularextrusion. Mater Sci Eng A 2006;423:247–52.

[17] Vrátná J, Hadzima B, Bukovina M, Janeček M. Roomtemperature corrosion properties of AZ31 magnesium alloyprocessed by extrusion and equal channel angular pressing. JMater Sci 2013;48:4510–6.

[18] Gubicza J, Chinh NQ, Dobatkin SV, Khosravi E, Langdon TG.Stability of ultrafine-grained microstructure in fcc metalsprocessed by severe plastic deformation. Key Eng Mater2011;465:195–8.

[19] Kim HK. Activation energies for the grain growth of an AZ31Mg alloy after equal channel angular pressing. J Mater Sci Lett2004;39:7107–9.

[20] Radi Y, Mahmudi R. Effect of Al2O3 nano-particles onthe microstructural stability of AZ31 Mg alloy afterequal channel angular pressing. Mater Sci Eng A2010;527:2764–71.

[21] Vrátná J, Janeček M, Čížek J, Lee DJ, Yoon EY, Kim HS.Mechanical properties and microstructure evolution in

ultrafine-grained AZ31 alloy processed by severe plasticdeformation. J Mater Sci 2013;48:4705–12.

[22] Bečvář F, Čížek J, Lešták L, Novotný I, Procházka I, Šebesta F. Ahigh-resolution BaF2 positron-lifetime spectrometer andexperience with its long-term exploitation. Nucl Inst MethodsPhys Res A 2000;443:557–77.

[23] Procházka I, Novotný I, Bečvář F. Application ofmaximum-likelihood method to decomposition ofpositron-lifetime spectra to finite number of components.Mater Sci Forum 1997;255–257:772–4.

[24] Vrátná J, Janeček M. Experimental study of ultrafine-grained commercial magnesium alloy AZ31 prepared bysevere plastic deformation. METAL 2013: 22nd

International Conference on Metallurgy and Materials;2013. p. 1603–7.

[25] Aghababaei R, Joshi SP. Micromechanics of tensile twinningin magnesium gleaned from molecular dynamicssimulations. Acta Mater 2014;69:326–42.

[26] Čížek J, Procházka I, Smola B, Stulíková I, Kužel R, Matěj Z,Cherkaska V. Thermal development of microstructure andprecipitation effects in Mg–10 wt%Gd alloy. Phys Status Solidi(a) 2006;203:466–77.

[27] Čížek J, Procházka I, Smola B, Stulíková I, Očenášek V.Influence of deformation on precipitation process inMg–15 wt.%Gd alloy. J Alloys Compd 2007;430:92–6.

[28] West R. In: Hautojärvi P, editor. Positrons in Solids. BerlinHeidelberg: Springer-Verlag; 1979. p. 89.

[29] Janeček M, Čížek J, Gubicza J, Vrátná J. Microstructureand dislocation density evolutions in MgAlZn alloyprocessed by severe plastic deformation. J Mater Sci2012;47:7860–9.

[30] Estrin Y, Tóth LS, Molinari A, Bréchet Y. A dislocation-basedmodel for all hardening stages in large strain deformation.Acta Mater 1998;46:5509–22.

[31] Hall EO. The deformation and ageing of mild steel: IIIdiscussion of results. Proc R Soc 1951;B64:747–53.

[32] Petch NJ. The cleavage strength of polycrystals. J Iron SteelInst 1953;174:25–7.

[33] Higgins GT. Grain-boundary migration and grain growth.Metal Sci 1974;8:143–50.

[34] Kim HK, Kim WJ. Microstructural instability and strength ofan AZ31 Mg alloy after severe plastic deformation. Mater SciEng A 2004;385:300–8.

[35] Cao P, Lu L, Lai MO. Grain growth and kinetics fornanocrystalline magnesium alloy produced by mechanicalalloying. Mater Res Bull 2001;36:981–8.

[36] Thein MA, Lu L, Lai MO. Kinetics of grain growth innanocrystalline magnesium-based metal–metal compositesynthesized by mechanical alloying. Compos Sci Technol2006;66:531–7.

[37] Miao Q, Hu L, Wang X, Wang E. Grain growth kinetics of afine-grained AZ31 magnesium alloy produced by hot rolling. JAlloys Compd 2010;493:87–90.

[38] Chao HY, Sun HF, Chen WZ, Wang ED. Static recrystallizationkinetics of a heavily cold drawn AZ31magnesium alloy underannealing treatment. Mater Charact 2011;62:312–20.

[39] Frost HJ, Ashby MF. Deformation-Mechanism Maps. Oxford:Pergamon Press; 1982 43–5.

[40] Shewman PG. Trans AIME 1956;206:918.[41] Wang J, Iwahashi Y, Horita Z, Furukawa M, Nemoto M, Valiev

RZ, Langdon TG. An investigation of microstructural stabilityin an AlMg alloy with submicrometer grain size. Acta Mater1996;44:2973–82.

[42] Valiev RZ, Kozlov EV, Ivanov YF, Lian J, Nazarov AA, BaudeletB. Deformation behaviour of ultra-fine-grained copper. ActaMetall Mater 1994;42:2467–75.

[43] Lian J, Valiev RZ, Baudelet B. On the enhanced grain growthin ultrafine grained metals. Acta Metall Mater1995;43:4165–70.

79M A T E R I A L S C H A R A C T E R I Z A T I O N 9 4 ( 2 0 1 4 ) 6 9 – 7 9

[44] Janeček M, Yi S, Král R, Vrátná J, Kainer KU. Texture andmicrostructure evolution in ultrafine-grained AZ31 processedby EX–ECAP. J Mater Sci 2010;45:4665–71.

[45] Furukawa M, Horita Z, Nemoto M, Valiev RZ. Microhardnessmeasurements and the Hall–Petch relationship in an Al + Mgalloy with submicrometer grain size. Acta Mater1996;51:4619–29.

[46] Vrátná J, Janeček M, Stráský J, Kim HS, Yoon EY.Investigation of microhardness and microstructure of AZ31alloy after high pressure torsion. Magnesium Technology;2011. p. 589–94.