Download - The influence of ECAP on microstructure evolution of aluminium alloys during in-situ heating in TEM

Michaela Poková, Mariia Zimina, Miroslav CieslarCharles University in Prague, Faculty of Mathematics and Physics, Department of Physics of Materials, Prague, Czech Republic

The influence of ECAP on microstructureevolution of aluminium alloys duringin-situ heating in TEMPaper presented at \XV International Conference on Electron Microscopy",15–18 September 2014, Cracow, Poland

Twin-roll casting of aluminium alloys produces sheets withhigh solid-solution supersaturation. During high tempe-rature processing the solid solution decomposes and cubica-Al(Mn,Fe)Si precipitates form. Processing by equalchannel angular pressing introduces new grain boundariesinto the material and shifts the precipitation to lower tem-peratures. In-situ annealing in a transmission electron mi-croscope enables dynamic observation of recovery, precipi-tation, particle dissolution and grain growth. However,most of the processes observed by in-situ transmission elec-tron microscopy occur at temperatures that are lower thantemperatures estimated from specimens annealed conven-tionally in a furnace.

Keywords: Transmission electron microscopy; In-situ an-nealing; Precipitation; Equal channel angular pressing;Electrical resistivity

1. Introduction

Twin-roll casting (TRC) of aluminium alloys is a moderncasting method, which produces sheets with thickness ofseveral mm and is less energy and material demanding thanconventional casting (CC). Owing to high solidificationrate, the initial microstructure is in a thermodynamicallyunstable state and exhibits high solid-solution supersatura-tion [1]. In the course of annealing at elevated temperaturessolid solution decomposes via second-phase particles preci-pitation [2–4]. Nevertheless, the different initial micro-structure of TRC materials does not allow direct use of re-sults obtained on CC materials and therefore processesoccurring during thermo-mechanical treatment of TRCsheets have been intensively studied [5–7].AA3003 aluminium alloys are commonly used in the

automotive industry. Type, size and distribution of second-phase particles in the aluminium matrix affect mechanicalproperties of the final product [8]. To achieve the requiredproperties the microstructure changes occurring during thedown-stream process must be explored in detail.Besides primary particles, which form during the casting

process, new particles precipitate in the course of annealing

at elevated temperatures [9]. In AA3003 alloys thetemperature of intensive precipitation has been observedbetween 300 8C and 500 8C [10, 11]. The composition andmorphology of precipitates depend on the content ofalloying elements. Two phases frequently observed inAl–Mn–Fe–Si alloys are orthorombic Al6(Mn,Fe) andcubic a-Al15(Mn,Fe)3Si2, but others such as hexagonala-Al8Fe2Si have also been reported in the literature [12,13]. The type of predominant phase depends mainly on thesilicon content in the alloy [14].The production of material with finer grain size is of

great importance as the reduction in grain size results gen-erally in a strength increase at lower temperatures and in aformability enhancement at elevated temperatures. One ofthe techniques refining aluminium alloys is severe plasticdeformation (SPD), which can produce grains smaller than1 lm. The most common SPD technique is equal channelangular pressing (ECAP) [15]. It is a process where a billetis pressed through a special die consisting of two channelsof the same cross-section. They intersect at an angle U(908 £ U < 1808). The shape of the billet remains nearly un-changed after the pressing [16]. The ECAP procedure canbe thus repeated several times and the stored deformationenergy can be multiplied. Consequently, ultra-fine grainedmaterial with a high fraction of high-angle grain boundaries(HAGB) can be produced [17].The microstructure after ECAP processing is influenced

by many parameters; the most important is the number ofECAP passes. Higher number of passes induces largerstrain in the material and simultaneously more uniform mi-crostructure and higher fraction of HAGB (e.g. [18, 19]).In the present study the influences of ECAP processing

and heat pre-treatment on precipitation processes occurringduring in-situ annealing in a transmission electron micro-scope (TEM) were studied.

2. Experimental procedure

Twin-roll cast AA3003 series aluminium alloy with a nom-inal composition 1.0–1.5 wt.% Mn, £0.7 wt.% Fe,£0.6 wt.% Si and 0.05–0.2 wt.% Cu with a small addition(0.16 wt.%) of zirconium was studied. The alloy was an-nealed in an air furnace with a heating rate 0.5 K · min–1

M. Poková et al.: The influence of ECAP on microstructure evolution of aluminium alloys during in-situ TEM heating

676 Int. J. Mater. Res. (formerly Z. Metallkd.) 106 (2015) 7

Inte

rnat

iona

l Jou

rnal

of

Mat

eria

ls R

esea

rch

dow

nloa

ded

from

ww

w.h

anse

r-el

ibra

ry.c

om b

y C

arl H

anse

r V

erla

g on

Jul

y 21

, 201

5Fo

r pe

rson

al u

se o

nly.

up to 450 8C, held for 8 h at 450 8C and subsequently waterquenched. The alloy without heating will be referred to inthe text as \D" and the annealed one as \DZ". The initial an-nealing at 450 8C results in the formation of coherent Al3Zrprecipitates in the material DZ. This treatment also leadsto the precipitation of a high number of a-Al(Mn,Fe)Si par-ticles.Both materials D and DZ were subjected to severe plastic

deformation by ECAP at room temperature. Used ECAPchannels had square cross-section 10 · 10 mm2 with the in-tersection angle of 908. Route BC was employed (the speci-men is rotated after each pass by 908 around its longitudinalaxis [20]).Both as-cast and deformed materials were step-by-step

isochronally annealed in an air furnace with a heatingscheme of 50 K/50 min. Resistivity annealing spectra,which are sensitive to solid solution decomposition and sol-ute redistribution [21, 22], were measured in liquid nitrogenin order to evaluate temperature ranges of precipitation andparticle dissolution. This measurement was completed byobservations in a conventional TEM JEOL JEM 2000FXand also by in-situ heating in a TEM. Electron back-scatterdiffraction (EBSD) in a QUANTA FEG 200 scanning elec-tron microscope (SEM) was used for the crystallographicidentification of particles.

3. Results

3.1. Electrical resistivity

Evolution of electrical resistivity during isochronal anneal-ing was measured in all studied materials in order to moni-tor the redistribution of solute atoms. A representation inthe form of the electrical resistivity spectrum (a negativederivative of the resistivity evolution) was chosen to high-light temperature ranges connected with the most signifi-cant redistribution of solutes (Fig. 1). A peak in positivevalues generally represents precipitation, while negativevalues are linked to particle dissolution and reversion of so-lutes into solid solution. A significant difference betweenmaterials annealed and non-annealed prior to ECAP wasobserved. In non-annealed materials pronounced peaks ap-pear between 250 8C and 480 8C. With increasing number

of ECAP passes the maxima of peaks are shifted to lowertemperatures by nearly 100 8C when compared to the as-cast material. The same maximum is notably suppressed inthe annealed specimen. Above 480 8C the resistivity in-creases in all materials and local minima in resistivity an-nealing spectra are detected.

3.2. Electron back-scatter diffraction

EBSD was applied for identification of crystallographicphases. Two phases were identified – Al and cubica-Al(Mn,Fe)Si. More details are given in Table 1. All pri-mary phases, which are present in materials after casting,are of cubic a-Al(Mn,Fe)Si phase. The crystallographicstructure of particles does not change after annealing at450 8C for 8 h (see Fig. 2). Precipitates formed during heattreatment are also of cubic a-phase.

3.3. In-situ heating

All alloys were subjected to in-situ annealing in a TEM.This type of experiment allows observation of recovery ofthe dislocation substructure, precipitation and particle coar-sening and dissolution in one selected area of the specimenduring the whole heating cycle. The heating Scheme 50 K/50 min was applied – the sample was heated to the requiredtemperature, held for 50 min at this temperature and after-wards the temperature was increased again by 50 K. Theimages from in-situ heatings can be found in Figs. 3–5. Inthe as-cast material first precipitates of a-Al(Mn,Fe)Siphase form at subgrain boundaries at 300 8C. After longerannealing times they also appear in the grain interior. Theirvolume fraction (number and also the size) increases at350 8C. Above this temperature their number density de-creases, finer particles dissolve back to the solid solution,while the coarser ones further grow. Nevertheless, thecoarse particles are also unstable at higher annealing tem-peratures and almost no particles are present in the matrixat 550 8C. No grain boundary motion was observed duringthe whole annealing cycle.The ECAP treatment considerably modifies grains in the

material. They are fragmented into numerous subgrainscontaining a high density of dislocations. The intensity offragmentation increases with increasing number of ECAPpasses. While subgrains are still elongated and subgrainboundaries are not well defined after one pass, smallerequiaxed subgrains with an average size lower than 1 lmand sharp boundaries are observed after 4 ECAP passes. Inthe material D the dislocation density is much higher thanin the material DZ, where nearly dislocation-free subgrainsare often observed. However, the main difference betweenthese two materials is the presence of secondary particles


Int. J. Mater. Res. (formerly Z. Metallkd.) 106 (2015) 7 677

Fig. 1. Resistivity annealing spectra during isochronal annealing.D = as-cast material, D 1P = material after one ECAP pass, D 4P = ma-terial after four ECAP passes and DZ 4P = material annealed at 450 8Cand processed by four ECAP passes.

Table 1. Phases identified by EBSD and their respective colors inFig. 2.

Phase Structure Lattice constant Color

Al fcc a = 0.404 nm

a-Al(Mn,Fe)Si bc a = 1.2643 nm

Inte

rnat

iona

l Jou

rnal

of

Mat

eria

ls R

esea

rch

dow

nloa

ded

from

ww

w.h

anse

r-el

ibra

ry.c

om b

y C

arl H

anse

r V

erla

g on

Jul

y 21

, 201

5Fo

r pe

rson

al u

se o

nly.

of a-Al(Mn,Fe)Si phase in the material DZ due to the heat-treatment before ECAP (Fig. 6). In-situ heating experi-ments reveal substantial microstructural changes in thecourse of annealing. Full recovery of the entire dislocationsubstructure occurs in the material D below 200 8C. Newparticles of a-Al(Mn,Fe)Si phase are observed above250 8C preferentially on subgrain boundaries. Their volumefraction and density gradually grow and reach maximum at350 8C. At higher annealing temperatures their ripeningand partial reversion to the solid solution occurs. At 500 8Conly a small number of coarser particles remain undis-solved in the matrix. Significant subgrain growth takesplace at 450 8C resulting in the formation of a bimodalstructure – a small fraction of grains preserve their submi-cron size while the majority of the matrix contains largergrains with an average size of several micrometers.In the D material after one ECAP pass the dislocation

density is high and sub-grains are elongated. Recovery of

dislocation substructure takes place during annealing to300 8C and HAGB form. These boundaries start to migrateat 450 8C. a-Al(Mn,Fe)Si particles nucleate from 300 8C,their partial dissolution begins at 400 8C. They emerge bothat the grain boundaries and inside the grains.The a-Al(Mn,Fe)Si particles were already present in the

pre-annealed material DZ before ECAP and therefore nonew precipitates form during the in-situ heating. Neverthe-less, these particles also grow at annealing temperatures be-low 350 8C, while above this temperature their ripening anddissolution occur. Only a small number of coarse particlesis observed in the matrix at 500 8C. Similarly as in the mate-rial D the recovery of dislocation occurs at temperatures be-low 200 8C. At 450 8C a coalescence of subgrains and rapidgrain growth begin and at 500 8C the majority of grainsreaches a size of about 10 lm.



Fig. 2. EBSD map of phases of alloy an-nealed at 450 8C for 8 h (left). Details of iden-tified phases are given in Table 1. Darkgrey = Al matrix, light grey = a-Al(Mn,Fe)Si.Corresponding image in back-scattered elec-trons (right).

Fig. 3. Evolution of microstructure of the as-cast material D during heating 50 K/50 min. Post-mortem observation (top) and in-situ heating inTEM (bottom).

Inte

rnat

iona

l Jou

rnal

of

Mat

eria

ls R

esea

rch

dow

nloa

ded

from

ww

w.h

anse

r-el

ibra

ry.c

om b

y C

arl H

anse

r V

erla

g on

Jul

y 21

, 201

5Fo

r pe

rson

al u

se o

nly.

3.4. Post-mortem observations

Post-mortem TEM observations were performed on materi-als isochronally annealed in an air furnace with the samestep as the one used during in-situ experiments since in-situexperiments might be affected by a confined specimenthickness (Figs. 3–5). During in-situ observations particlesform at lower temperatures and their average diameter dur-

ing annealing is higher while their number density is lower.At 500 8C nearly all particles are already dissolved in the in-situ processing. On the other hand the precipitate density at500 8C is still relatively high in materials annealed in the airfurnace. In addition, recovery of the dislocation substruc-ture occurs more rapidly and the grain growth is limitedand postponed to higher temperatures during in-situ experi-ments. In materials processed by ECAP the final grain size



Fig. 4. Evolution of microstructure of material D after one ECAP pass during heating 50 K/50 min. Post-mortem observation (top) and in-situ heat-ing in TEM (bottom).

Fig. 5. Evolution of microstructure of material D after four ECAP passes during heating 50 K/50 min. Post-mortem observation (top) and in-situheating in TEM (bottom).

Inte

rnat

iona

l Jou

rnal

of

Mat

eria

ls R

esea

rch

dow

nloa

ded

from

ww

w.h

anse

r-el

ibra

ry.c

om b

y C

arl H

anse

r V

erla

g on

Jul

y 21

, 201

5Fo

r pe

rson

al u

se o

nly.

after in-situ annealing is several microns while the grainsize is much larger than 10 lm in conventionally annealedones.

4. Discussion

The ECAP processing significantly modifies the micro-structure of the TRC material. The main feature is the frag-mentation of original grains and formation of numerousnew subgrain and grain boundaries. They serve as preferen-tial sites for heterogeneous precipitation of a-Al(Mn,Fe)Siphase. Therefore, the higher the strain induced by ECAP,the higher the fracture of boundaries in the material andalso the probability of precipitate formation at lower tem-peratures. Thus, the temperature of the most intensive pre-cipitation shifts to lower temperatures with increasingstrain. In the material which was annealed before ECAPthe precipitates are already present in the matrix and there-fore no new ones form during subsequent annealing.According to in-situ TEM observations, recrystallization

is accelerated by the dissolution of secondary particles. Atlower annealing temperatures precipitates are formedmainly on subgrain boundaries and consequently the mo-tion of boundaries is hindered by particles. Nevertheless,at 450 8C the majority of precipitates dissolves or they aretoo coarse to effectively pin subgrain boundaries. As a re-sult intensive subgrain coalescence is observed followedby full recrystallization and grain growth.TEM observations disclosed discrepancies between in-

situ and post-mortem results. For example in material Dafter 4 ECAP passes precipitates are already observed dur-ing in-situ heating at 250 8C while the number of precipi-tates observed by conventional TEM is very low even at300 8C. The highest volume fraction of precipitates was de-tected at 350 8C for in-situ and at 500 8C for post-mortemobservations. Also, the number density and average size ofprecipitates are different. During the in-situ heating experi-ment precipitates are coarser and their number density islower. The recovery of the dislocation substructure is fasterduring in-situ annealing, but in contrast the grain growth issuppressed and at 550 8C the grain size is much lower thenin specimens annealed conventionally. The origin of thesedifferences arises from the two-dimensional nature of in-situ observations. During the in-situ annealing the recoveryis accelerated because dislocations can more easily annihi-late at the free surface, which forms a high fraction of the

TEM foil. As precipitates nucleate at lower temperaturethe nucleation frequency is reduced, the density of createdparticles is lower and due to the diffusivity of solute ele-ments enhanced by surface diffusion, solutes are preferen-tially consumed by coarsening of existing precipitatesrather than on nucleation of new ones. On the other handHAGB are more stable because their mobility is retardedby the free surface and therefore higher annealing tempera-ture is necessary for their motion. Moreover, as dislocationdensity is reduced more quickly during in-situ annealing,the driving force for recrystallization and grain growth islower as compared to conventionally annealed materials.Thus, recrystallization is postponed to higher temperatures.

5. Conclusions

Intensive precipitation of cubic a-Al(Mn,Fe)Si phase oc-curs in the TRC material at 400 8C, first on (sub)grainboundaries but after longer annealing times and higher tem-peratures also in the grain interior. In the material subjectedto ECAP, which has a much higher fraction of grain andsubgrain boundaries serving as nucleation sites for precipi-tation, new precipitates already form at 300 8C. On the otherhand, the pre-annealing before ECAP causes decomposi-tion of the supersaturated solid solution and consequentlyno new precipitates form during the following post-ECAPannealing.In-situ experiments enable observation of dynamic pro-

cesses occurring during annealing. However, due to thetwo-dimensional character of TEM foil the observed pro-cesses are not in accordance with post-mortem observa-tions. During in-situ experiments recovery, precipitationand particle dissolution start at lower temperatures, whilegrain growth is shifted to higher temperatures.

The financial supports of grants GAUK 1428213 and GACR P107-12-0921 are gratefully acknowledged.

References

[1] M. Slámová, M. Karlík, M. Cieslar, B. Chalupa, P. Merle: KovoveMater. 40 (2002) 389.

[2] M. Karlík, T. Mánik, H. Lauschman: J. Alloys Compd. 515 (2012)108. DOI:10.1016/j.jallcom.2011.11.101

[3] M. Cieslar, M. Slámová, J. Uhlír, C. Coupeau, J. Bonneville:Kovove Mater. 45 (2007) 91.

[4] M. Poková, M. Cieslar, J. Lacaze: Acta Phys. Pol. A 122 (2012)625.



Fig. 6. Microstructure of materials D (left)and DZ (right) after four ECAP passes. Someparticles of a-Al(Mn,Fe)Si phase in DZ aremarked by circles.

Inte

rnat

iona

l Jou

rnal

of

Mat

eria

ls R

esea

rch

dow

nloa

ded

from

ww

w.h

anse

r-el

ibra

ry.c

om b

y C

arl H

anse

r V

erla

g on

Jul

y 21

, 201

5Fo

r pe

rson

al u

se o

nly.

[5] M. Slámová, A. Kubosová, J. Politello, M. Cieslar: Int. J. Mater.Res. 100 (2009) 428. DOI:10.3139/146.110044

[6] M. Poková, M. Cieslar: Mater. Sci. Eng.: IOP Conference Series63 (2014) 012086. DOI:10.1088/1757-899X/63/1/012086

[7] C. Gras, M. Meredith, J.D. Hunt: J. Mater. Process. Technol. 169(2005) 156. DOI:10.1016/j.jmatprotec.2005.03.034

[8] T. Carlberg, M. Jaradech, H.K. Kamga: JOM-J. Min. Met. Mater.Soc. 58 (2006) 56. DOI:10.1007/s11837-006-0025-9

[9] M. Poková, M. Cieslar, M. Slámová: Int. J. Mater. Res. 100(2009) 391. DOI:10.3139/146.110029

[10] M. Cieslar, M. Slámová, M. Hájek, J. Vesely: Mater. Sci. Forum567–568 (2008) 325.

[11] M. Dehmas, P. Archambault, M. Serriere, E. Gautier, C.A. Gan-din: Aluminium 78 (2002) 864.

[12] N.A. Belov, D.G. Eskin, A.A. Askenov: Multicomponent PhaseDiagrams: Application for Commercial Aluminum Alloys, Else-vier (2005).

[13] Y. Birol: J. Alloys Compd. 471 (2009) 122.DOI:10.1016/j.jallcom.2008.04.005

[14] T. Pettersen, Y. Li, T. Furu, K. Marthinsen: Mater. Sci. Forum558–559 (2007) 301.DOI:10.4028/www.scientific.net/MSF.558-559.301

[15] I. Nikulin, A. Kipelova, S. Malopheyev, R. Kaibyshev: Acta Ma-ter. 60 (2012) 487. DOI:10.1016/j.actamat.2011.10.023

[16] Y. Iwahashi, Z. Horita, M. Nemoto, T.G. Langdon: Acta Mater.45 (1997) 4733. DOI:10.1016/S1359-6454(97)00100-6

[17] M. Cabibbo: Mater. Sci. Eng. A-Struct. 560 (2013) 413.DOI:10.1016/j.msea.2012.09.086

[18] P. Málek, M. Poková, M. Cieslar: Metal 2014 Conference Pro-ceedings (2014) 2839.

[19] A.B. Naizabekov, V.A. Andreyachshenko, R. Kocich: Micron 44(2013) 210. DOI:10.1016/j.micron.2012.06.011

[20] K. Nakashima, Z. Horita, M. Nemoto, T.G. Langdon: Mater. Sci.Eng. A-Struct. 281 (2000) 82.DOI:10.1016/S0921-5093(99)00744-3

[21] M. Hájek, J. Vesely, M. Cieslar: Mater. Sci. Eng. A-Struct. 462(2007) 339. DOI:10.1016/j.msea.2006.01.175

[22] M. Vlach, I. Stulíková, B. Smola, T. Kekule, H. Kudrnová, S. Da-nis, R. Gemma, V. Ocenásek, J. Málek, D. Tanprayoon, V. Neu-bert: Mater. Charact. 86 (2013) 59.DOI:10.1016/j.matchar.2013.09.010

(Received December 4, 2014; accepted March 9, 2015; on-line since April 9, 2015)

Correspondence address

RNDr. Michaela Poková, Ph.D.Ke Karlovu 5121 16, Prague 2Czech RepublicTel.: +420221911386Fax: +420221911490E-mail: [email protected]

BibliographyDOI 10.3139/146.111243Int. J. Mater. Res. (formerly Z. Metallkd.)106 (2015) 7; page 676–681# Carl Hanser Verlag GmbH & Co. KGISSN 1862-5282



Inte

rnat

iona

l Jou

rnal

of

Mat

eria

ls R

esea

rch

dow

nloa

ded

from

ww

w.h

anse

r-el

ibra

ry.c

om b

y C

arl H

anse

r V

erla

g on

Jul

y 21

, 201

5Fo

r pe

rson

al u

se o

nly.

Download - The influence of ECAP on microstructure evolution of aluminium alloys during in-situ heating in TEM

Top Related