Materials and Design 34 (2012) 258–267

Contents lists available at SciVerse ScienceDirect

Materials and Design

journal homepage: www.elsevier .com/locate /matdes

Microstructural evolution in friction stir welding of high-strength linepipe steel

Hoon-Hwe Cho a, Suk Hoon Kang b, Sung-Hwan Kim a, Kyu Hwan Oh a, Heung Ju Kim c,Woong-Seong Chang c, Heung Nam Han a,⇑a Department of Materials Science & Engineering and Center for Iron & Steel Research, RIAM, Seoul National University, San 56-1, Kwanak-gu, Seoul 151-744, Republic of Koreab Nuclear Materials Research Division, Korea Atomic Energy Research Institute, 1045 Daedeok-daero, Yuseong-gu, Daejon 305-353, Republic of Koreac New Materials Research Department, Research Institute of Industrial Science & Technology, San 32, Hyoja-dong, Nam-gu, Pohang 790-600, Republic of Korea

a r t i c l e i n f o a b s t r a c t

Article history:Received 6 July 2011Accepted 5 August 2011Available online 12 August 2011

Keywords:WeldingFerrous metals and alloysMicrostructure

0261-3069/$ - see front matter Crown Copyright � 2doi:10.1016/j.matdes.2011.08.010

⇑ Corresponding author. Tel.: +82 2 880 9240.E-mail address: hnhan@snu.ac.kr (H.N. Han).

Microstructural evolution during friction stir welding (FSW) of a high-strength linepipe steel was studied.The various grain structures developed through a complex process including the rearrangement of low-angle boundaries, continuous dynamic recrystallization and phase transformation. In most parts of thestir zone (SZ), acicular-shaped bainitic ferrites were formed by the phase transformation during theFSW process. A fine-grained microstructure developed mainly in the thermomechanically affected zone(TMAZ), where continuous dynamic recrystallization occurs. The shear texture in the SZ became consid-erably weak due to the phase transformation during the FSW process. The hardness of the SZ was signif-icantly higher than that of the other FSWed regions due to the bainitic ferrites.

Crown Copyright � 2011 Published by Elsevier Ltd. All rights reserved.

1. Introduction

In recent years, there has been increasing demand for high-performance linepipe steels with high strength, excellent low-temperature toughness, good weldability and superior corrosionresistance because pipelines are normally installed in severe envi-ronments such as permafrost or seismic regions. Thus, recently, APIX100 grade linepipe steel with those properties has been noted inthe field of pipeline manufacturing.

The electric resistance welding (ERW) process is normally usedto join linepipe steel in the manufacturing process due to efficiencyand little pollution. However, the heat generated during the ERWprocess of linepipe steel leads to grain coarsening in the weldedregion because it solidifies directly from the liquid to ferrite phasewithout any intermediate phase transformation [1]. Although thissteel has useful properties in the wrought condition, the conven-tional ERW process reduces the toughness, ductility and corrosionresistance of this steel because of grain coarsening during thewelding process.

Therefore, this steel should be joined with a low heat input andhigh welding speed, and friction stir welding (FSW) has attractedattention as an alternative welding process in the manufacture ofa pipeline. FSW is a solid state joining process patented by TheWelding Institute (TWI, Cambridge, UK) in 1991 [2]. The FSW pro-cess is performed using a rotating and traversing cylindrical toolwith a shoulder to generate frictional heat and an integral protrud-ing smaller thread pin or probe that penetrates into and plasticizes

011 Published by Elsevier Ltd. All r

the material in the vicinity of the joint line. The tool is plunged intoand then traverses along the joint line between two abutting work-pieces. The work-pieces are ultimately joined by the stirring actionof the softened material. The advantages of this process includelow residual stress, low energy input and a fine homogeneousmicrostructure compared to the conventional liquid–solid weldingprocess.

Initially, the FSW process was used only for non-ferrous alloyswith a low melting temperature, such as aluminum alloys[3–10]. Sato and Kokawa [3] examined dominant microstructuralfactors governing the global tensile properties of a FSWed jointof 6063 Al by estimating distribution of local tensile properties cor-responding to local microstructure and hardness. Kang et al. [7]also investigated the material flow and crystallographic orientationin aluminum alloy sheets joined by FSW. They found that themicrotexture of the material near the tool is very close to the sim-ple shear texture. Most recently, Zhang et al. [10] studied changesof microstructure and mechanical properties as a function of rota-tion speed in under water FSWed aluminum alloy joints. In con-trast, application of FSW to ferrous alloys including linepipe steelwith high melting temperatures has been limited due to high tem-peratures and severe wear conditions induced by the welding toolduring the process. However, continued research into this processhas brought some success in the joining of ferrous alloys [11–13],and this practical success requires a clearer understanding of theFSW joints of ferrous alloys. Sato et al. [11] reported that FSW suc-cessfully produced the defect-free weld in ultrahigh carbon steelhaving the (ferrite + cementite) duplex structure using a polycrys-talline cubic boron nitride (PCBN) tool, and FSW changed the (fer-rite + cementite) duplex structure into the martensitic structure in

ights reserved.

Table 1Chemical composition and carbon equivalent (Ceq) of the examined API grade X100 linepipe steel.

Chemical composition (wt.%)

C Si Mn P S Nb V Mo Ceq

0.05–0.07 0.25 62.0 60.01 60.001 60.05 60.05 60.3 0.46–0.48

H.-H. Cho et al. / Materials and Design 34 (2012) 258–267 259

the weld center. Cho et al. [13] also examined the microstructureand hardness of FSW joint of 409 ferritic stainless steel. They foundthat the increase in plunging depth increased the amounts of bothheat input and plastic deformation to the welded material duringFSW, which led to an increase in the fraction of low-angle bound-ary (LAB), a decrease in the grain size, and an increase in the hard-ness in the stir zone (SZ). In particular, it is very important tounderstand the microstructural evolution of ferrous alloys duringFSW because this understanding would enable a prediction ofthe weld properties resulting from the microstructures.

Sato et al. [14] examined the recrystallization phenomenon dur-ing friction stirring of 304L austenitic stainless steel by experimen-tal approaches and found that friction stir processed 304Laustenitic stainless steel underwent partial static recrystallizationfollowing dynamic recrystallization. Mironov et al. [15] examinedthe microstructural evolution of pure iron during friction stir-pro-cessing. They reported that the development of grain structure wasdriven mainly by grain subdivision, geometrical effect of strain andlocal grain-boundary migration.

The above-mentioned studies focused on face-centered cubic(FCC) and body-centered cubic (BCC) metals, in which a phasetransformation does not occur during FSW. On the other hand, asolid-state phase transformation should be considered in themicrostructural evolution of ferrous alloys including linepipe steelduring FSW. This phenomenon is considered as a main factoraffecting the properties of ferrous alloys, and studies concerningthis phenomenon have been performed by many researchers

Fig. 1. Characteristic microstructural features of the BM: (a) band contrast (BC) map showangle distribution (random distribution is shown by a dotted line), (c) {1 1 0} pole figur

[16–21]. Sato et al. [11] examined the microstructural evolutionof ultrahigh carbon steel during FSW. However, this research fo-cused mainly on the relationship between the microstructuresand mechanical properties instead of the microstructural evolutionof ultrahigh carbon steel. Furthermore, other studies in the field ofFSW for ferrous alloys focused mainly on the development of ade-quate tool materials due to the above-mentioned problems [22,23].

It should be noted that the development of grain structure offerrous alloy is comprehensively investigated. Recently, linepipesteel has attracted noticeable attention in structural materials forits high strength, excellent low-temperature toughness and goodweldability, and the application of FSW for this steel has been be-gun to use the advantages of a solid-state welding process. There-fore, in the field of FSW, studies of the microstructural evolution inhigh-strength linepipe steel (API grade X100) are greatly challeng-ing. The present study focused on the fundamental issue of grainstructure development during FSW of high-strength linepipe steel.For this purpose, an electron back-scatter diffraction (EBSD)equipped with field emission scanning electron microscope (FE-SEM) and a transmission electron microscope (TEM) were usedto provide a comprehensive understanding of microstructural evo-lution during FSW of high-strength linepipe steel.

2. Experimental procedure

API X100 grade linepipe steel, 10 mm in thickness, was used inthe FSW studies; Table 1 lists the chemical composition and carbon

ing the typical microstructure of API grade X100 linepipe steel, (b) misorientation-e.

260 H.-H. Cho et al. / Materials and Design 34 (2012) 258–267

equivalent provided by a steel supplier, POSCO. The test plate wassectioned in half along the rolling direction and prepared for a buttjoint. Oxide scale was removed by sand grinding followed bydegreasing with methanol. An Ar atmosphere was maintainedusing a gas cup located around the tool to minimize surface oxida-tion during the weld cycle. Friction stir welding was performedwith a PCBN tool (MegaStir Technologies, UT, USA) in an inertgas environment. The tool rotation and welding speeds were450 rpm and 127 mm/min, respectively. The PCBN tool had ashoulder diameter of 16 mm and a pin diameter of 4 mm. The

3 mm

RS

Weld cen

SZTMAZHAZ*BM

Fig. 2. Cross-sectional macrograph of the welded sample. See text for detai

100 µm

Re

Region 3

T

ND

TDWD

Fig. 3. Composite EBSD map of HAZ, TMAZ and SZ on the RS of the weld. The individual gcolor code triangle is shown in the bottom right corner. See text for details.

principal directions of FSW geometry are denoted as the weldingdirection (WD), transverse direction (TD) and normal direction(ND).

Following FSW, the plate was prepared perpendicular to theWD and examined by optical microscope (OM), EBSD equippedwith FE-SEM and TEM. The surface for OM was prepared bymechanical polishing with a 1 lm diamond paste followed bychemical etching in 3 ml nitric acid + 97 ml methanol solution.For the EBSD observations, the sample was mechanically groundand electrolytically polished in a 10 ml perchloric acid + 90 ml

ND

TDWD

AS

ter

ls. RS and AS indicate the retreating and advancing sides, respectively.

gion 1

Region 2

HAZ

MAZ

SZ

111

001 101

rains are colored according to their crystallographic direction relative to the WD; the

H.-H. Cho et al. / Materials and Design 34 (2012) 258–267 261

ethanol solution using a Struers Lectropol-5 electrolytic polisher.For TEM analysis, a thin slice was cut using a low-speed diamondsaw and then mechanically thinned to nearly 100 lm thickness.Three millimeter disks were punched out and electro-polished ina Struers Tenupol-3 twin-jet polisher using a solution containing10% perchloric acid in methanol.

OM analysis was carried out using an Olympus GX51 opticalmicroscope. High-resolution EBSD studies were performed usinga Jeol JSM6500F FE-SEM equipped with a HKL Channel 5 EBSDsystem. The accelerating voltage was 20 kV, the probe currentwas 4 nA and the working distance was 15 mm with the samplestage tilted by 70�. The camera resolution was 978 � 733 pixelsin the operation of 8 � 8 binning. The mapping grid was a regu-lar square in 0.3 lm steps. The limits of the low-angle bound-aries (LABs) and high-angle boundaries (HABs) were set to 2�and 15�, respectively. Based on this criterion, the grain sizeswere measured using the linear intercept method. TEM analysiswas carried out using a TECNAI F20 operating at a voltage of200 keV.

The Vickers hardness test was conducted in a 4 � 11 array at2 mm intervals on a cross section perpendicular to the WD usinga commercial AIS 3000R (Frontics Inc., Seoul, Korea) with depthand load resolutions of 0.1 lm and 14.7 mN, respectively. Themaximum depth, dwell time and testing speed were 70 lm, 0.5 sand 5 lm/s, respectively.

All microstructural analyses were made on a transversal(TD–ND) plane, which is identical to the cross-section perpendicu-lar to the WD.

50 μm

(a)

50 μm

(b)

Fig. 4. The orientation maps (a–c) of some selected areas (regions 1–3 in Fig. 3). The indiWD; the color code triangle is shown in the bottom center. LABs and HABs are depicted

3. Results and discussion

3.1. Base material

Fig. 1 shows the microstructural features for the base material(BM). The microstructure consists mainly of the dual phases of fer-rite and bainite with an average grain size of approximately2.8 lm, as shown in Fig. 1a. The ferrite phase is almost equiaxed,whereas the bainite phase is subdivided into very fine laths. Themisorientation-angle distribution is characterized by a sharplow-angle peak and some clustering of the misorientation axesnear the h1 1 1i (Fig. 1b) because the BM was produced by the roll-ing process. Therefore, the BM has a very weak rolling texture, asshown in Fig. 1c.

3.2. Low-magnification overview

Fig. 2 shows a cross-sectional macrograph of the FSWed zone. Inthe cross section, the left- and right-hand sides of the weld centercorrespond to the retreating side (RS) and advancing side (AS) ofthe rotation tool, respectively. The FSW joint consists of a SZ, BM,broad transition region, which is commonly called the thermome-chanically affected zone (TMAZ), and a heat affected zone (HAZ)between the TMAZ and the BM, even though it is difficult to clearlydistinguish each zone. The SZ shows a basin-like shape that widensconsiderably towards the upper surface. The border region be-tween the SZ and the BM is the TMAZ and HAZ.

50 µm

(c)

5 µm

111

001 101

NDTD

WD

vidual grains are colored according to their crystallographic direction relative to theas white and black lines, respectively. See text for details.

(a) TD

WD

max 2.34min 0.35

1

(b) TD

WD

max 1.79min 0.41

1

(c) TD

WD

max 4.31min 0.14

123

Fig. 5. The {1 1 0} pole figures showing textures in (a) HAZ, (b) TMAZ and (c) SZ.

Fig. 6. Profile of LABs fraction measured from the region indicated by the redrectangle in Fig. 2.

262 H.-H. Cho et al. / Materials and Design 34 (2012) 258–267

3.3. Microstructural evolution

3.3.1. Approach methodAn examination how the original grain structure transforms to

the SZ microstructure is needed to understand grain structuredevelopment during FSW [15,24,25]. Therefore, in the presentstudy, the region marked by a red1 rectangle in Fig. 2 was carefullyanalyzed. For the sake of simplicity, only microstructural evolutionin the RS is described.

Fig. 3 shows the EBSD maps taken from the region indicated bythe red rectangle in Fig. 2. Some selected areas (regions 1–3) inFig. 3 are shown at higher magnification in Fig. 4a–c, respectively.In the EBSD maps in Fig. 4a–c, the LABs and the HABs are depictedas white and black lines, respectively. The large EBSD map in Fig. 3is divided into three regions (HAZ, TMAZ and SZ) to quantify themicrostructural evolution in the FSWed region, and the micro-structural data measured from these regions are presented in Figs.5–8.

3.3.2. HAZFig. 3 shows an orientation map including HAZ, TMAZ and SZ.

The parent grains in the HAZ tend to grow due to the significantamount of heat generated during FSW. In the HAZ, the averagegrain size is about 3.3 lm, which is larger than that (2.8 lm) ofthe BM shown in Fig. 1. On the center in the HAZ (Fig. 4a), the ori-ginal grains are homogeneously distributed, and the microstruc-

1 For interpretation of color in Figs. 1–8, 11–13, the reader is referred to the webversion of this article.

ture of this region is similar to that of the BM except for thegrain size.

The crystallographic orientation of the parent grains does notchange practically, as shown in Fig. 5a. The fraction of LABs inthe HAZ (Fig. 6) is 39.6%, which is slightly smaller than 43.8% inthe BM. This indicates that the removal or rearrangement of LABsoccurred in the HAZ during the FSW process due to the increasingtemperature of the material. Fig. 7a shows the misorientation-angle distribution in the HAZ. The characteristic feature is a clus-tering of the misorientation axis near h1 1 0i. The LABs rearrangearound h1 1 0i rotation axes by the significant heat generated dur-ing FSW, and consequently, a strong cluster in the misorientation-angle distribution develops.

3.3.3. TMAZIn the TMAZ (Figs. 3 and 4b), considerably fine homogeneous

grains similar to those observed in the SZ of other steel alloysdeveloped [11–15,25–27]. The fine-grained microstructure in theTMAZ was attributed to the continuous dynamic recrystallizationinduced by the substantial shear deformation and significantamount of heat generated during FSW. The average grain size isapproximately 2.2 lm, which is the smallest value among theFSWed regions.

The continuous dynamic recrystallization, which is an interest-ing microstructural characteristic of the TMAZ, would be associ-ated with the progressive transformation of sub-grains into newgrains within the deformed grains as well described in the litera-ture [28]. In other words, LABs accumulate progressively in theLABs within the deformed grains during FSW, leading to an in-crease in their misorientation angle and the formation of HABs,when a critical value of the misorientation angle is reached. Then,migration of HABs eventually causes fine homogeneous grains inthe TMAZ. This result is in agreement with the literature in whichthe continuous dynamic recrystallization well occurs in ferritewith a relatively larger stacking fault energy [29,30].

The shear texture weakly developed in the TMAZ, as shown inFig. 5b. The fraction of LABs in the TMAZ (Fig. 6) is approximately30%, which is the smallest value among the FSWed regions. This isdue to an increase in the HABs induced by the continuous dynamicrecrystallization. The misorientation-angle distribution is featuredby the increase in HABs in the overall range and the dispersed clus-tering of the misorientation axes near h1 1 0i, as shown in Fig. 7b.

3.3.4. SZIn the SZ (Figs. 3 and 4c), the structure morphology, texture and

misorientation-angle distribution are completely different fromthe other FSWed regions. The structure morphology observed inthis region is quite distinguishing and can be elucidated

Fig. 7. Misorientation-angle distributions in (a) HAZ, (b) TMAZ and (c) SZ.Misorientation-axis distributions are shown in the top right corners of the graphs,respectively.

Fig. 8. Profiles of grain size and acicular-shaped phase fraction with aspect ratioover 3 measured from the region indicated by the red rectangle in Fig. 2.

0 200 400 600 8000

200

400

600

800

1000

Tem

pera

ture

, o C

Time, sec

Fig. 9. Temperature–time profile measured at the root of the tool shoulder.

H.-H. Cho et al. / Materials and Design 34 (2012) 258–267 263

adequately in terms of the phase transformation that occurs duringFSW. Fig. 9 shows the temperature–time profile measured at thetool shoulder during FSW. According to thermodynamic calcula-tion using Thermo-Calc software [31], the A1 and A3 temperaturesof the API grade X100 linepipe steel are 683 and 792 �C, respec-tively. During the experiment, the peak temperature measured atthe root of the tool shoulder is 983 �C (Fig. 9), which is significantlyhigher than the A1 and A3 temperatures. Since the temperature of

the material during FSW might be considerably higher than that ofthe root of the tool shoulder, it is likely that the temperature of thematerial in the SZ during FSW becomes higher than the A1 and A3

temperatures. Therefore, the austenite-to-ferrite transformationoccurs in the SZ during FSW, and this phenomenon would resultin a specific SZ completely different from the other FSWed regions.

Region 3 shows a typical SZ with acicular-shaped bainitic fer-rites formed by a phase transformation during FSW (Fig. 4c). In re-gion 3, the original ferrite structure transforms fully to a single-phase austenite structure during the heating cycle of FSW. Thissteel then undergoes a solid-state transformation from austeniteto acicular-shaped bainitic ferrite during the cooling cycle due torapid cooling and a high strain induced by severe plastic deforma-tion caused by frictional stirring [32,33]. The high-magnificationinset in the bottom right corner of Fig. 4c shows typical acicular-shaped bainitic ferrite with many LABs, and this is in accordancewith the literature in which the dislocation density of bainitic fer-rite in high-strength steel is quite high [34].

Fig. 10a shows typical acicular-shaped bainitic ferrites in the SZ.As mentioned above, high density dislocations are distributed inthese bainitic ferrites. In addition, shaves of acicular-shaped bain-itic ferrites tend to grow as a series of parallel platelets emanatingfrom the austenite grain boundary. According to Borrato et al. [35],the recrystallization stop temperature (RST) in austenite, which isthe minimum temperature that recrystallization can occur duringhot rolling, is expressed as follows:

RSTð�CÞ ¼ 887þ 464Cþ ð6645Nb� 664ffiffiffiffiffiffiffi

Nbp

Þ þ ð732V

� 230ffiffiffiffi

VpÞ þ 890Tiþ 363Al� 357Si ð1Þ

500 nm 500 nm

(a) (b)Fig. 10. TEM micrographs showing the morphological characteristics of (a) SZ and (b) BM.

264 H.-H. Cho et al. / Materials and Design 34 (2012) 258–267

where the chemical compositions are in wt.%.The RST value of API grade X100 linepipe steel is 999 �C, which

is slightly higher than the peak temperature (983 �C) measured atthe root of the tool shoulder. Since the temperature of the materialduring FSW may be considerably higher than that of the root of thetool shoulder as mentioned above, it is likely that dynamic recrys-tallization occurs in the austenite phase in the SZ during FSW pro-cess. Therefore, very fine austenite grains would develop during

Fig. 11. Characteristic microstructural features of the RS of the SZ: (a) band contrast (Bdotted line), (c) {1 1 0} pole figure.

FSW, which would lead to the formation of acicular-shaped bainit-ic ferrties that nucleate mainly at the austenite grain boundary. Asa result, Fig. 10a shows that the morphological characteristics ofthe SZ are completely different from those of the BM composedof ferrite and bainite, as shown in Fig. 10b.

Fig. 8 shows the profiles of the grain size and acicular-shapedphase fraction with an aspect ratio over 3 measured from theregion indicated by the red rectangle in Fig. 2. The criterion to

C) map, (b) misorientation-angle distribution (random distribution is shown by a

H.-H. Cho et al. / Materials and Design 34 (2012) 258–267 265

classify the FSWed regions in the API grade X100 linepipe steelwould not be the grain size but rather the acicular-shaped phasefraction. In the distribution of the acicular-shaped phase fraction,boundaries among the FSWed regions are quite distinct, whereasthe boundary between the TMAZ and the SZ is ambiguous in thatof the grain size. In other words, the acicular-shaped phase fractiondecreases drastically at the boundary between the HAZ and theTMAZ due to the newly formed homogeneous grains in the TMAZ,and it considerably increases at that between the TMAZ and SZbecause of the acicular-shaped bainitic ferrites developed by aphase transformation that occurs in the SZ.

3 mm

RS

Wel

TMAZHAZ*BM

(a)

(b)

(c)

(d)

Fig. 12. Cross-sectional macrograph of the w

Fig. 13. Profiles of hardness on each line

The texture sharpens significantly close to h1 1 0i (achieving�4.3 times random) as shown in Fig. 5c. The fraction of LABsbroadly increases as it approaches the center of the SZ (Fig. 6) be-cause typical acicular-shaped bainitic ferrites have many LABs asmentioned above. The misorientation-angle distribution is charac-terized by the increase in HABs in the angular range of 55–60�. Thenewly developed acicular-shaped bainitic ferrites are typically sur-rounded by HABs, thus increasing the fraction of HABs.

Fig. 11 summarizes the microstructural features for the RS inthe SZ taken from the region indicated by the black dashed rectan-gle in Fig. 8. The microstructural evolution in this region can be

ND

TDWD

AS

d center

SZ

elded sample with indentation marks.

(a–d in Fig. 12) for welded sample.

266 H.-H. Cho et al. / Materials and Design 34 (2012) 258–267

explained by its cooling from the (ferrite + austenite) duplex struc-ture. According to the reported calculation result [36], the peaktemperature of the RS in the SZ is lower than that of the centerin the SZ, which may allow the presence of a (ferrite + austenite)duplex structure below the A3 temperature. Since the austenitephase transforms to bainitic ferrite during the cooling cycle, themicrostructure of the RS in the SZ includes both acicular-shapedbainitic ferrite developed by a phase transformation and finehomogeneous ferrite induced by dynamic recrystallization, asshown in Fig. 11a. The misorientation-angle distribution is dis-tinctly different from the random distribution, but is similar tothe center in the SZ (Fig. 11b). Although the predominant deforma-tion mode during FSW is expected to be simple shear [37–40], thetexture of the RS in the SZ is broadly random, and this may be asso-ciated with a weakness of shear texture caused by a phase trans-formation, which occurs during FSW. The microstructuralfeatures of the AS in the SZ are very similar to those of the RS ofthe SZ, and are omitted in this paper.

3.4. Hardness distribution

Fig. 12 shows a cross-sectional macrograph of the FSWed regionwith indentation marks. Fig. 13 presents the corresponding hard-ness profile variations on the transversal lines (a–d) of indentationmarks shown in Fig. 12. From these figures, it can be verified thatthe hardness of the SZ is significantly higher than that of the otherFSWed regions. This may be due to the development of bainitic fer-rite caused by a phase transformation that occurs in the SZ duringFSW [41]. On the other hand, the hardness of the HAZ is lowestamong the FSWed regions because the grains in the HAZ tend togrow due to the significant heat generated during FSW.

4. Conclusion

The microstructural evolution of API X100 grade linepipe steelduring FSW was examined using EBSD equipped with FE-SEM,OM and TEM. Grain structure evolution was found to be a complexprocess involving the rearrangement of LABs, continuous dynamicrecrystallization and phase transformation. In the HAZ, the LABsrearrange around the h1 1 0i rotation axes by the significant heatgenerated during FSW. Fine homogeneous grains were mainlydeveloped in the TMAZ, where continuous dynamic recrystalliza-tion occurs. Acicular-shaped bainitic ferrites were observed inmost parts of the SZ, and the fraction of these could be a quite dis-tinct criterion to distinguish the FSWed regions. Although FSWcauses shear texture in the SZ as known, the texture in the SZwas quite random due to the weakness of shear texture inducedby a phase transformation that occurs in the SZ during FSW. Thehardness of the SZ is significantly higher than that of the otherFSWed regions because of the bainitic ferrites that developed asa result of a phase transformation during FSW.

Acknowledgements

This work was supported by National Research Foundation ofKorea grant funded by the Ministry of Education, Science and Tech-nology (2010-0018936) and was sponsored by POSCO.

References

[1] Folkhard E. Welding metallurgy of stainless steels. 1st ed. Wien NewYork: Spring-Verlag; 1988.

[2] Thomas WM, Nicholas ED, Needham JC, Murch MG, Templesmith P, Dawes CJ.International patent application no. PCT/GB92/02203; GB patent applicationno. 9125978.8; 1991.

[3] Sato YS, Kokawa H. Distribution of tensile property and microstructure infriction stir weld of 6063 aluminum. Metall Mater Trans A 2001;32A:3023–31.

[4] Scialpi A, De Filippis LAC, Cavaliere P. Influence of shoulder geometry onmicrostructure and mechanical properties of friction stir welded 6082aluminum alloy. Mater Des 2007;28:1124–9.

[5] Steuwer A, Dumont M, Peel M, Preuss M, Withers PJ. The variation of theunstrained lattice parameter in an AA7010 friction stir weld. Acta Mater2007;55:4111–20.

[6] Kang SH, Chung H-S, Han HN, Oh KH, Lee CG, Kim SJ. Relationship betweenformability and microstructure of Al alloy sheet locally modified by friction stirprocessing. Scripta Mater 2007;57:17–20.

[7] Kang SH, Han HN, Oh KH, Cho J-H, Lee CG, Kim S-J. Investigation of the materialflow and texture evolution in friction-stir welded aluminum alloy. Met MaterInt 2009;15:1027–31.

[8] Kim D, Badarinarayan H, Ryu I, Kim JH, Kim C, Okamoto K, et al. Numericalsimulation of friction stir spot welding process for aluminum alloys. Met MaterInt 2010;16:323–32.

[9] Li B, Shen Y. The investigation of abnormal particle-coarsening phenomena infriction stir repair weld of 2219-T6 aluminum alloy. Mater Des2011;32:3796–802.

[10] Zhang HJ, Liu HJ, Yu L. Microstructure and mechanical properties as a functionof rotation speed in underwater friction stir welded aluminum alloy joints.Mater Des 2011;32:4402–7.

[11] Sato YS, Yamanoi H, Kokawa H, Furuhara T. Microstructural evolution ofultrahigh carbon steel during friction stir welding. Scripta Mater2007;57:557–60.

[12] Bhadeshia HKDH, Debroy T. Critical assessment: friction stir welding of steels.Sci Technol Weld Join 2009;14:193–6.

[13] Cho H-H, Han HN, Hong S-T, Park J-H, Kwon Y-J, Kim S-H, et al. Microstructuralanalysis of friction stir welded ferritic stainless steel. Mater Sci Eng A2011;528:2889–94.

[14] Sato YS, Nelson TW, Sterling CJ. Recrystallization in type 304L stainless steelduring friction stirring. Acta Mater 2005;53:637–45.

[15] Mironov S, Sato YS, Kokawa H. Microstructural evolution during friction stir-processing of pure iron. Acta Mater 2008;56:2602–14.

[16] Han HN, Lee JK. A constitutive model for transformation superplasticity underexternal stress during phase transformation of steels. ISIJ Int 2002;42:200–5.

[17] Han HN, Lee CG, Oh C-S, Lee T-H, Kim S-J. A model for deformation behaviorand mechanically-induced martensitic transformation of metastableaustenitic steel. Acta Mater 2004;52:5203–14.

[18] Han HN, Lee JK, Suh D-W, Kim S-J. Diffusion controlled transformationplasticity of steel under externally applied stress. Philos Mag 2007;87:159–76.

[19] Han HN, Lee CG, Suh D-W, Kim S-J. A microstructure-based analysis fortransformation induced plasticity and mechanically induced martensitictransformation. Mater Sci Eng A 2008;485:224–33.

[20] Han HN, Oh C-S, Kim G, Kwon O. Design method for TRIP-aided multiphasesteel based on a microstructure-based modelling for transformation-inducedplasticity and mechanically induced martensitic transformation. Mater Sci EngA 2009;499:462–8.

[21] Lee M-G, Kim S-J, Han HN. Crystal plasticity finite element modeling ofmechanically induced martensitic transformation (MIMT) in metastableaustenite. Int J Plasticity 2010;26:688–710.

[22] Mahoney M, Nelson TW, Sorenson C, Packer S. Friction stir welding of ferrousalloys: current status. Mater Sci Forum 2010;638:41–6.

[23] Mishra RS, Mahoney MW. Friction stir welding and processing. 1sted. USA: ASM International; 2007.

[24] Mironov S, Sato YS, Kokawa H. Development of grain structure during frictionstir welding of pure titanium. Acta Mater 2009;57:4519–28.

[25] Suhuddin UFHR, Mironov S, Sato YS, Kokawa H, Lee CW. Grain structureevolution during friction-stir welding of AZ31 magnesium alloy. Acta Mater2009;57:5406–18.

[26] Reynolds AP, Tang W, Gnaupel-Herold T, Prask H. Structure, properties, andresidual stress of 304L stainless steel friction stir welds. Scripta Mater2003;48:1289–94.

[27] Sato YS, Nelson TW, Sterling CJ, Steel RJ, Pettersson C-O. Microstructure andmechanical properties of friction stir welded SAF 2507 super duplex stainlesssteel. Mater Sci Eng A 2005;397:376–84.

[28] Gourdet S, Montheillet F. A model of continuous dynamic recrystallization.Acta Mater 2003;51:2685–99.

[29] Schramm RE, Reed RP. Stacking fault energies of seven commercial austeniticstainless steels. Metall Mater Trans A 1975;6A:1345–51.

[30] Vecchio KS, Hertzberg RW. The influence of stacking fault energy on ductilefracture micromorphology. J Mater Sci 1988;23:2220–4.

[31] Thermo-Calc, Version S, Foundation of computational thermodynamics.Stockholm, Sweden: Royal Institute of Technology; 2008.

[32] Strangwood M, Bhadeshia HKDH. The mechanism of acicular ferrite formationin steel weld deposits. Adv Weld Technol Sci 1987;1:209–14.

[33] Yang JR, Bhadeshia HKDH. Thermodynamic of the acicular ferritetransformation in alloy-steel weld deposits. Adv Weld Technol Sci1987;1:187–91.

[34] Farrar RA, Harrison PL. Acicular ferrite in carbon–manganese weld metals: anoverview. J Mater Sci 1987;22:3812–20.

[35] Borrato F, Barbosa R, Yue S, Jonas JJ. Effect of chemical composition on thecritical temperatures of microalloyed steels. Thermec-88 1988;1:383–90.

[36] Cho J-H, Kang S-H, Han HN, Oh KH. Modeling friction stir welding process ofaluminum alloys. Met Mater Int 2008;14:247–58.

H.-H. Cho et al. / Materials and Design 34 (2012) 258–267 267

[37] Prangnell PB, Heason CP. Grain structure formation during friction stir weldingobserved by the ‘stop action technique’. Acta Mater 2005;53:3179–92.

[38] Fonda RW, Bingert JF, Colligan KJ. Development of grain structure duringfriction stir welding. Scripta Mater 2004;51:243–8.

[39] Reynolds AP, Hood E, Tang W. Texture in friction stir welds of Timetal 21S.Scripta Mater 2005;52:491–4.

[40] Park SHC, Sato YS, Kokawa H. Basal plane texture and flow pattern in frictionstir weld of a magnesium alloy. Metall Mater Trans A 2003;34A:987–94.

[41] Sung HK, Shin SY, Hwang BC, Lee CG, Kim NJ, Lee SH. Effects of alloyingelements and the cooling condition on the microstructure, tensile properties,and charpy impact properties of high-strength bainitic steels. Kor J Met Mater2010;48:798–806.