Materials and Design 32 (2011) 4825–4831

Contents lists available at ScienceDirect

Materials and Design

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

Improvement of weld temperature distribution and mechanical properties of7050 aluminum alloy butt joints by submerged friction stir welding

Rui-dong Fu a,b,⇑, Zeng-qiang Sun a, Rui-cheng Sun a, Ying Li a, Hui-jie Liu c, Lei Liu a

a State Key Laboratory of Metastable Materials Science and Technology, Qinhuangdao 066004, Hebei Province, PR Chinab College of Materials Science and Engineering, Yanshan University, Qinhuangdao 066004, Hebei Province, PR Chinac State Key Laboratory of Advanced Welding Production Technology, Harbin 150001, Heihongjiang Province, PR China

a r t i c l e i n f o

Article history:Received 29 March 2011Accepted 13 June 2011Available online 21 June 2011

Keywords:A. Non-ferrous metals and alloysD. WeldingE. Mechanical

0261-3069/$ - see front matter � 2011 Elsevier Ltd. Adoi:10.1016/j.matdes.2011.06.021

⇑ Corresponding author at: State Key Laboratory ofand Technology, Qinhuangdao 066004, Hebei Provinc

E-mail address: rdfu@ysu.edu.cn (R.D. Fu).

a b s t r a c t

Submerged friction stir welding (FSW) in cold and hot water, as well as in air, was carried out for 7050aluminum alloys. The weld thermal cycles and transverse distributions of the microhardness of the weldjoints were measured, and their tensile properties were tested. The fracture surfaces of the tensile spec-imens were observed, and the microstructures at the fracture region were investigated. The results showthat the peak temperature during welding in air was up to 380 �C, while the peak temperatures duringwelding in cold and hot water were about 220 and 300 �C, respectively. The temperature at the retreatedside of the joint was higher than that at the advanced side for all weld joints. The distributions of microh-ardness exhibited a typical ‘‘W’’ shape. The width of the low hardness zone varied with the weld ambientconditions. The minimum hardness zone was located at the heat affected zone (HAZ) of the weld joints.Better tensile properties were achieved for joint welded in hot water, and the strength ratio of the weldjoint to the base metal was up to 92%. The tensile fracture position was located at the low hardness zoneof the weld joints. The fracture surfaces exhibited a mixture of dimples and quasi-cleavage planes for thejoints welded in cold and hot water, and only dimples for the joint welded in air.

� 2011 Elsevier Ltd. All rights reserved.

1. Introduction

FSW was invented by The Welding Institute in 1991, and hasbeen successfully applied in welding of aluminum alloys, particu-larly 2XXX or 7XXX series aluminum alloys, which are difficult toweld using traditional melting welding methods [1–3]. However,degradation of the mechanical properties of weld joints of heattreatable strengthening aluminum alloys in the HAZ due to the ef-fects of the welding thermal cycle remains a key issue. Many tech-nical methods have been developed to limit the degradation ofjoint performance in the HAZ. Submerged welding is consideredan effective method of welding, and is employed in various weld-ing processes.

In simple terms, the principle of submerged welding processis that the welds are placed in a liquid medium, and weld pro-cessing takes place under a specific ambient temperature. Thisprocess is suitable for alloys that are sensitive to overheatingduring the welding process. Tokisue et al. were the first to usesubmersion in a rotary friction weld for 6061 aluminum alloys[4]. The results of their study showed that it was possible to

ll rights reserved.

Metastable Materials Sciencee, PR China.

generate enough friction for welding even though the sampleswere submerged. In recent years, some remarkable results wereobtained through the use of submerged FSW. Thomas adoptedsubmerged FSW to improve the strength of the FSW joint of6061 aluminum alloy [5]. Upadhyay and Reynolds investigatedthermal boundary conditions and their effects on the mechanicalproperties of AA7050-T7 FSW joints welded in sub ambientwater and a �25 �C liquid medium. The ultimate tensile strengththroughout the range of parameters tested showed improve-ments [6]. Nelson et al. demonstrated that 7075-T7351 alumi-num alloys could be considered ‘quench sensitive’, where thecooling rate from thermal exposure has an important influenceon the mechanical properties of the friction stir welds, specifi-cally in the case of natural aging. The addition of cooled watermist behind the FSW tool or a water-cooled anvil resulted in10% and 8% increase in strength, respectively [7]. Liu et al. foundthat submerged FSW improved the tensile strength of FSW jointsof 2219 aluminum alloy [8]. However, the range of temperatureused in the above research was limited below room temperature.There are few reports on FSW under the welding ambient tem-perature, which is above room temperature.

In the present study, the submerged FSW in cold and hot waterfor 7050 aluminum alloys is conducted. For comparison, normalFSW in air is also performed. Variations in the weld temperature

4826 R.D. Fu et al. / Materials and Design 32 (2011) 4825–4831

fields and mechanical properties of the weld joints with weldambient conditions are investigated and discussed.

2. Experiment procedures

As-received hot rolled and then naturally aged plates of 7050high strength aluminum alloy with a gauge thickness of 5.5 mmwere selected as experimental materials. The chemical composi-tion of this alloy is 6.00% Zn, 2.2% Mg, 2.24% Cu, and 0.05% Ni, withthe balance made up of Al. The ultimate tensile strength of theplates was 396 MPa. Friction stir welds were produced under cold(about 8 �C) and hot (about 90 �C) water. For comparison, frictionstir welds exposed to air were also produced. A H13 tool steelFSW tool consisting of a 12 mm diameter shoulder and a 6.2 mmdiameter pin was employed in submerged FSW. The weld travelspeed was 100 mm/min and the rotational speed of the FSW toolwas 800 rpm. The welding direction was parallel to the plate roll-ing direction, and the tool rotation axis was normal to the plane ofthe plate. During the welding process, an eight-channel thermo-detector was used for the measurement of the transverse distribu-tion of the weld temperature. The positions and the marks of thethermocouple are illustrated in Fig. 1.

The samples for metallographic observation were ground, pol-ished, and etched using Keller’s reagent. DSC test samples werecut from the welded zone and the HAZ of the weld joints using awire electric discharge machine. The sample dimension was5.5 mm in diameter and 2.5 mm in thickness. Thermal analysiswas conducted using a NETZSCH differential scanning calorimeter(DSC). Samples were heated in an inert atmosphere (Ar2) at a con-stant heating rate of 20 K/min from room temperature to 500 �C.

The Vickers microhardness distribution was measured under aload of 1.96 N for a dwell time of 10 s along the centerlines ofthe cross-section with an interval of 0.5 mm. Foil specimens wereprepared by the twin-projecting method under low temperatureand observed on a JEOL-2010 transmission electron microscope(TEM) at 200 keV. Tensile samples were cut along the transversedirection of the weld joints. The tensile tests were conducted atroom temperature and a cross-head speed of 3 mm/min. The frac-ture surfaces were observed after the tensile tests using a HITACHIS-4800 scanning electron microscope (SEM).

3. Results

3.1. Distribution of the temperature and microhardness

The transverse distribution of the temperature and microhard-ness under different weld ambient conditions are shown in

Welding direction Retreating side

Advancing side

R4

R3

R2

R1

A1

A2

A3

A4

180mm

160m

m

12m

m

Fig. 1. Illustration of friction stir welds and the positions of the thermocouples.

Fig. 2a and b, respectively. In Fig. 2a, the weld temperature distri-bution under different weld ambient conditions exhibits the samevariation with the distance away from the central line of the weldseam. Furthermore, the temperature at the retreated side is higherthan that at the advanced side under three weld ambientconditions.

The differences in the three weld temperature distributioncurves are the peak temperature and the width of the critical tem-perature above which the dissolution or precipitation of some sec-ondary phase particles in the alloys occurs. It is evident that thepeak temperature for the joint welded in air is up to 380 �C, thehighest amongst the three cases. This indicates that the weld heatinput in this case is higher than that in cold and hot water. Mean-while, due to the diffusion of weld heat mainly relying on the basemetal, its HAZ is the widest, while the peak temperatures at thesame position are only about 220 and 300 �C, respectively, forjoints welded in cold and hot water. This also results in the reduc-tion of the HAZ width.

The microhardness distribution of the three weld joints areshown in Fig. 2b. It is known that the hardness variation of a weldjoint is related to the microstructure variation, which is a result ofthe weld heat affecting. For heat treatable strengthening aluminumalloy, grain growth, dissolution or coarsening of the secondaryphase particles often occur in the HAZ. These variations in themicrostructure, which strongly rely on the weld temperature dis-tribution, result in the decrease of the hardness of the weld joints.Therefore, the widths of the low hardness zone of the weld joint in-crease with rising heat input and falling cooling rate. For example,when welded in air, the width of the region above the temperatureof 200 �C was estimated to be about 40 mm (Fig. 2a), correspond-ing to the width of the low microhardness zone in Fig. 2b. Simi-larly, it is easy to understand the microhardness distribution ofthe joints welded in cold and hot water. The minimum hardnessvalues for the weld joints welded in air, cold water, and hot waterare 105, 114, and 115 HV, respectively.

3.2. Metallographic observations

The macrocross-sectional morphologies of weld joints underdifferent welding ambient conditions are shown in Fig. 3. Due tothe difference in the weld conditions, the macrofeatures of thethree joints exhibit different characters.

For the joint welded in air, the welded zone can be divided intotwo zones, i.e., a shoulder active zone (upper section in the cross-section) and a stir pin active zone (bottom section in the cross-sec-tion), where the ‘‘onion’’ feature is clearly seen (Fig. 3a). Fig. 2bshows the macrocross-sectional features of the joint welded in coldwater. Due to the stronger cooling rate during welding, the bound-ary line between the HAZ and parent metal is easily seen (the blackarrow in Fig. 3b). Moreover, the separated zone feature as that inFig. 2a has disappeared. The normal feature of a ‘‘V’’ shape is verydistinct. The macrofeature of the joint welded in hot water (Fig. 3c)is similar to that in cold water. The difference is that the boundaryline of the HAZ cannot be seen in the figure due to the weaker cool-ing rate in hot water. In addition, a feature called the ‘‘S’’ line(white arrow shown in Fig. 3b and c) is clearly observed in the weldzone of the joints welded in cold and hot water. Some researchersconsider this S a weld defect because it is a weaker interface andthe tensile fracture occurs at this position.

3.3. Tensile properties

The tensile test results of the weld joints welded under differentambient temperatures are shown in Fig. 4. From Fig. 4a, it can beseen that the joint welded in hot water has the best tensile prop-erty, i.e., it has higher strength and better elongation. The weld

80

120

160

200

240

280

320

360

400

A4

A3

A2A1

R4R3

R2

Distance from the central line (mm)-14 -9 -6 -3 0 3 6 9 14

Advancing sideRetreating side

Tem

pera

ture

/

(a) R1

-30 -20 -10 0 10 20 30

100

110

120

130

140

150

160

170 in air in cold water in hot water

shoulder

retreated sideadvanced side

Har

dnes

s (H

V)

Distance from the central line (mm)

(b) in air in cold water in hot water

Fig. 2. Distribution of the weld temperature field and microhardness under different welding ambient conditions.

retreated side advanced side 1mm

retreated side advanced side 1mm

retreated side advanced side 1mm

(a)

(b)

(c)

Fig. 3. Macromorphology of weld joints under different welding ambient condi-tions (a) in air, (b) in cold water, and (c) in hot water.

R.D. Fu et al. / Materials and Design 32 (2011) 4825–4831 4827

joint efficiency (defined as the ratio of the strength of the weldjoint to that of the base metal) achieved 92% of the base metal.The gauge elongation is 9.2%, which is higher than the 6.5% ofthe base metal. In comparison, the strength of the joint weldedin cold water was slightly higher than that welded in air, althoughthe gauge elongation decreased.

In comparing the variations of the tensile strength and the elon-gation of weld joints with the welding ambient conditions (Fig. 4b),it can be observed that controlling weld temperature distributionimproves the ultimate tensile strength of the weld joint. However,overcooling does not improve its elongation. Due to the lower for-mation temperature of the weld zone, the hardening of the weldjoint resulted in decreased elongation. Based on the present weldparameters, variation of the ambient conditions of welding in hotwater is an alternative approach to improve the mechanical prop-erties of weld joints.

3.4. Fracture surface observations

The fracture positions of tensile specimens are shown in Fig. 5.For the joint welded in air (Fig. 5a), the tensile fracture position islocated at the HAZ of the advanced side. The distance of the frac-ture region to the central line of the weld seam corresponds to

the position with the lowest hardness (Fig. 2b). However, the fail-ures of the joints welded in cold and hot water occur at the re-treated side (Fig. 5b and c). Similarly, the fracture positions arealso located at the regions with the lowest hardness. The trans-verse tensile specimens consist of several different microstructuralregions, which allow deformation to localize at the weakest region.This phenomenon is known as strain localization and has beendemonstrated to occur within the HAZ (strength can be as low as60% of the base metal strength) of friction stir welds produced fromprecipitation strengthened aluminum alloys [9–14]. Severe strainlocalization can reduce elongation to very low values, much lowerthan would be expected for aluminum alloys. Strain localization intransverse tensile tests influences the measured strength since thespecimen will always neck and fail within the weakest location(corresponding to the low hardness location). For weld jointswelded in air and hot water, evident necking can be seen and oc-curs only at the HAZ before fracture (Fig. 5a and c), whereas forthe joint welded in cold water, strain localization occurs at boththe welded zone and the HAZ.

In Fig. 5, the cracks propagate along the direction of an angle of45� parallel to the plate plane. The differences in the failure fea-tures of the three joints include the distance and manner of crackpropagation. For the joint welded in cold water, the fracture occursin simple shearing mode, which implies poor elongation. For thejoint welded in air and hot water, the cracks propagate in a ‘‘Z’’shaped manner, which indicates increased resistance to crackpropagation even though the fracture also occurs in shearingmode. Consequently, the elongation of the weld joint may be in-creased. This is proven by the results of the tensile test presentedin Fig. 4b. The variations of the strength do not exhibit the samerelation to the fracture feature.

Further observations of the fracture surfaces of the tensile spec-imens are presented in Fig. 6. The macrofeatures of the three frac-ture surfaces show that the fracture mode is transformed frommulti-shearing for the joint welded in air (Fig. 6a) to solo-shearingfor the joint welded in cold water (Fig. 6c). By increasing theambient temperature, the fracture mode is transformed back tomulti-shearing (Fig. 6e). Considering the fracture positions andweld temperature field distribution, the evolution of the fracturemode is not difficult to understand. However, the microfeaturesof the central region in the fracture surfaces are different for thethree weld joints. Although the fracture surfaces of the three weldjoints exhibit ductile characteristics with respect to having largequantities of dimples, the size and depth of the dimples in the frac-ture surface of the joint welded in air (Fig. 6b) are different from

300

310

320

330

340

350

360

370

380

390

400

5.0

5.5

6.0

6.5

7.0

7.5

8.0

8.5

9.0

9.5

10.0

Ulti

mat

e Te

nsile

Stre

ngth

/ M

Pa

Strength

in hot water

Elo

ngat

ion

(%)

Elongation

(b)

in air in cold water0 1 2 3 4 5 6 7 8 9 100

50

100

150

200

250

300

350

400

450

Stre

ss /

MPa

Strain

Base metal in air in cold water in hot water

(a)

Fig. 4. Tensile properties of weld joints welded in different welding ambient conditions.

Retreated sideAdvanced side

Advanced side

Advanced side Retreated side

(a)

(b)

(c)

Fig. 5. The tensile fracture position of weld joints welded in different ambientconditions (a) in air, (b) in cold water, and (c) in hot water.

4828 R.D. Fu et al. / Materials and Design 32 (2011) 4825–4831

those of the joints welded in water. The dimples in the fracture sur-face of the joint welded in water (Fig. 6d and e) look small andeven. In addition, quasi-cleavage planes and secondary cracks canalso been found in some positions. These indicate high ultimatetensile strength and are consistent with the results presented inFig. 4.

4. Discussion

It has been reported in the literature that temperature historyplays a significant role in determining properties within a frictionstir weld [15,16]. Typically, if the peak temperature is greater thanthe solution heat treatment temperature for the FSW joints ofaging strengthening aluminum alloys, a characteristic ‘‘W’’ shapedhardness distribution is observed. This arises due to solution heattreatment of the nugget and overaging of the heat affected zone(HAZ) [17–19]. If the weld is performed at relatively low power(with a stir zone peak temperature less than approximately300 �C) the characteristic W shape in the hardness profile willnot be observed. It can be proved by the present results of microh-ardness distribution under different weld conditions. However, theaverage value of microhardness in the nugget zone for the jointwelded in hot water is higher than that in cold water, althoughthe minimum value in the HAZ is lower. It indicates that reprecip-

itating process of the second phase particles in the nugget zone isaccelerated by hot water, whereas, the coarsening process of theparticles in the HAZ is restricted.

Compared with the joint welded in air, the increments ofstrength for the weld joints welded in cold and hot water were5% and 14%, respectively. This agrees to the results reported inthe literature [7,20,21]. However, the increment of the elongationfor the joint welded in cold water was negative, i.e., submergedFSW in cold water decreases the elongation of the weld joint, whilethe considerable increase in elongation was up to 9% for the jointwelded in hot water. The results confirm that submerged FSW inhot water is the best way to achieve optimal mechanical propertieswith higher strength and elongation to reduce the integrated ther-mal exposure of a given location.

Another finding from the tensile tests is that all failure positionsof the weld joints are located in the HAZ. The variation of the vol-ume fractions and the size of the precipitates in the HAZ play animportant role in the final mechanical properties of the weld joint.The grain growth resulting from overheat input in the HAZ is an-other factor that affects its mechanical property. Regardless ofthe weld parameters, the improvement of the tensile propertiesfor the joint welded in hot water is related to the restriction effectsof the hot water cooling on the coarsening process of the secondphase particles in the HAZ.

Numerous studies have been devoted to understand the rela-tionships between properties and welding parameters for 7XXXseries alloys [15–19,22,23]. The extent of the modification inmicrostructure during FSW process will be primarily governedby the temperature history which in turn is dependent on weld-ing parameters used and the thermal boundary conditions dur-ing the weld [6]. TEM observations of the fracture regions ofthe as-weld joints welded in different ambient conditions areshown in Fig. 7. This illustrates that not only has the grain sizein the HAZ of the joint welded in air (Fig. 7a) grown, the popu-lation of the precipitates has also significantly increased. Mean-while, some of the particles in this joint became coarsecompared with the other two joints welded in cold and hotwater (Fig. 7b and c).

In order to further investigate the features of the particles at thefracture regions of the weld joint, DSC analysis of the HAZ and theweld thermal cycles at the fracture regions was performed. Fig. 8shows the DSC curves for the HAZ of the weld joints welded in dif-ferent ambient conditions. In general, the endothermic peaks of theDSC curve indicate the occurrence of dissolution reaction; the exo-thermic peaks indicate the occurrence of precipitation reaction.

1mm

20μm

20μm

20μm

1mm

1mm

(a) (b)

(c) (d)

(e) (f)

Fig. 6. The tensile fracture surfaces of weld joints welded in different ambient conditions (a and b) in air, (c and d) in cold water, and (e and f) in hot water.

(c)(b)(a)

Fig. 7. The second phase particles in the fracture region of the joints welded in different ambient temperatures (a) in air, (b) in cold water, and (c) in hot water.

R.D. Fu et al. / Materials and Design 32 (2011) 4825–4831 4829

The area under a given DSC peak is related to the precipitatevolume fraction, and the peak temperature is related to the averageprecipitate size [24,25].

In Al–Zn–Mg series aluminum alloys, the supersaturated solid-solution decomposes in the following sequence: Supersaturatedsolid-solution ? GP zone ? g0 (MgZn2) ? g (MgZn2) [26,27]. Thetemperature of a given reaction is dependent on the alloy compo-sition, sample quench rate, and DSC heating rate [24]. The reactionpeak (labeled I) below 160 �C in the DSC curve is attributed to thedissolution of unstable GP zones or g0 phases. The reaction peak(labeled II) at a temperature range of 160–300 �C is attributed to

the precipitation of stable g phases. The reaction peak (labeledIII) above 300 �C is attributed to the dissolution of the remainingphases in the alloy.

From the areas under the peak I in Fig. 8, we can deduce the vol-ume fraction of the GP zone or g0 phases in the HAZ of the as-welded joints. The larger the area under the peak, the larger thevolume fraction of the precipitates in the HAZ. The area underthe peak II cannot be used to deduce the factual volume fractionsof the precipitates in the HAZ due to the effects of the dissolutionof the GP zone or g0 phases, which increase the variation of the ma-trix composition before the formation of the peak II. However, the

0 5 10 15 20 25 30 35 40 45 5020

40

60

80

100

120

140

160

180

200

220

240

260

280

300

Tem

pera

ture

/ ΟC

Time / s

in air in cold water in hot water

A4

R3R1

Fig. 9. The weld thermal cycles of the fracture position in the HAZ.

4830 R.D. Fu et al. / Materials and Design 32 (2011) 4825–4831

variations of the amplitudes of the peak II are evidently based onthe peak I for the three DSC curves in Fig. 8. This implies that theincrements of the g phase rely on the decrements of the GP zoneor g0 phase in the DSC specimens. Furthermore, if the areas underthe peaks II, and I, for a DSC curve are larger than those of the oth-ers, the volume fractions of the precipitates in the correspondingas-welded joints should be smaller, and vice versa. Thus, it canbe deduced that the volume fractions of the precipitated particlesobserved in Fig. 7c are larger than those in Fig. 7a and b.

The above evolutions of the precipitated phases can be ex-plained by the features of weld thermal cycles in the HAZ. The weldthermal cycles at the fracture positions of the joints welded in dif-ferent weld ambient conditions are shown in Fig. 9. The labels A4,R1, and R3 in Fig. 9 represent the thermal couples embedded inwelds corresponding to the fracture positions. For the joint weldedin air, the peak temperature of the weld thermal cycle at the A4 po-sition is up to 275 �C, exceeding the peak temperature of the pre-cipitating g phases (refer to the peak II in Fig. 8). For jointswelded in cold and hot water, the peak temperature of the weldthermal cycle at the fracture regions is about 220 �C. At this tem-perature, the GP zone or g0 phases have dissolved and the g phaseshave begun to form for the joint welded in hot water. Meanwhile,for the joint welded in cold water, there only occurred the dissolu-tion of the GP zone or g0 phases (see peak I in Fig. 8). The variationsof the microstructures in the HAZ of the weld joints welded in hotand cold water are not only related to the peak temperature of theweld thermal cycles but also to the residence time above the crit-ical temperature of the precipitates. Although the peak tempera-tures at the R1 and R3 positions are almost same, the residencetime above the critical temperature at the R1 position is twice thatat the R3 position. The difference in the weld thermal cycle resultsin variations in the microstructures.

Based on the above discussion, the volume fraction of the unsta-ble GP zone or g0 phases in the HAZ for the joint welded in hotwater is the largest amongst the three as-welded joints. Theseunstable phases are the more important strengthening phases forheat treatable aluminum alloys and determine their final mechan-ical properties. Thus, the tensile property of the joint welded in hotwater is the best among the three weld joints. In the case of thejoint welded in cold water, due to the lower weld peak tempera-ture, the volume fraction of the unstable GP zone or g0 phases inthe HAZ is small and similar to that of the base metal. The factorthat affects the failure of the joint welded in cold water is notthe variation of the precipitates but strain hardening, which occurs

100 150 200 250 300 350 400 450-0.5

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

Hea

t flo

w/ (

mw

/mg)

Temperature/

in air in cold waterin hot water

Exothermic

Fig. 8. DSC analysis in the HAZ of weld joints welded at different ambienttemperatures.

at the fracture position in the as-welded joints. In conclusion,welding in either cold or hot water, rather than air, improves theweld temperature distribution and mechanical properties of joints.However, the optimal weld parameters for submerged FSW shouldbe further investigated.

5. Conclusion

The present study on submerged FSW for 7050 aluminum alloysaimed to determine an optimized process to improve the strengthand toughness of weld joints. Several general trends were ob-served, and some useful correlations were found:

(1) The transverse distribution of weld temperature fields forjoints welded in three ambient conditions showed that thetemperature at the retreated side is higher than that at theadvanced side. The peak temperature of the weld tempera-ture field when welded in air was the highest at 380 �C. Inthe case of joints welded in cold and hot water, the peaktemperatures of the weld temperature fields were about220 and 300 �C, respectively.

(2) The microhardness distribution of the as-welded jointsunder three ambient conditions exhibited typical ‘‘W’’ shapefeatures. The width of the minimum hardness zone variedwith the ambient conditions and corresponded accordinglywith the range of the HAZ.

(3) The mechanical properties of weld joints welded in hotwater are the best amongst all weld joints tested. The ratioof the ultimate tensile strength and elongation of the jointwelded to the base metal in hot water reached 92% and150%, respectively.

(4) The failure positions of weld joints were all located in HAZzones, where the microhardness value was the lowest inthe entire weld joint. However, due to differences in the vol-ume fractions of the unstable precipitates and the grain sizeof the fracture region, the features and lengths of crack prop-agation differ for all the weld joints.

Acknowledgments

The authors thank the Modern Welding Production TechnologyState Key Laboratory and the National Science Foundation for Dis-tinguished Young Scholars (No. 50925522) for their financialsupport.

R.D. Fu et al. / Materials and Design 32 (2011) 4825–4831 4831

References

[1] Thomas WM, Nicholas ED, Needham JC, Murch MG, Temple Smith P, Dawes CJ.Friction stir butt welding. International patent application No. PCT/GB92/02203; December 1991.

[2] Nandan R, DebRoy T, Bhadeshia HKDH. Recent advances in friction-stirwelding-process, weldment structure and properties. Prog Mater Sci2008;53:980.

[3] Uday MB, Ahmad Fauzi MN, Zuhailawati H, Ismail AB. Advances in frictionwelding process: a review. Sci Technol Weld Join 2010;15(7):534.

[4] Sakurada D, Katoh K, Tokisue H. Underwater friction welding of 6061aluminum alloy. J Jpn Inst Light Met 2002;52(2):2.

[5] Thomas Bloodworth. On the immersed friction stir welding of AA6061-T6: ametallurgic and mechanical comparison to friction stir welding. Thesis for thedegree of master of science in mechanical engineering, Nashville, Tennessee;May 2009.

[6] Upadhyay P, Reynolds AP. Effects of thermal boundary conditions in frictionstir welded AA7050-T7 sheets. Mater Sci Eng 2010;A527:1537.

[7] Nelson TW, Steel RJ, Arbegast WJ. In situ thermal studies and post-weldmechanical properties of friction stir welds in age hardenable aluminiumalloys. Sci Technol Weld Join 2003;8(4):283.

[8] Liu HJ, Zhang HJ, Yu L. Effect of welding speed on microstructures andmechanical properties of underwater friction stir welded 2219 aluminumalloy. Mater Des 2011;32:1548.

[9] Mahoney MW, Rhodes CG, Flintoff JG, Spurling RA, Bingel WH. Properties offriction-welded 7075-T651 aluminum. Metall Mater Trans A 1998;29A:1955.

[10] von Strombeck A., dos Santos JF, Torster F, Laureano P, Kocak M. Fracturetoughness behaviour of FSW joints in aluminium alloys. In: Proceedings of the1st international friction stir welding conference, Thousand Oaks, CA, USA.Cambridge (UK): TWI; 1999.

[11] Mishra RS, Sharma SR, Mara NA, Mahoney MW, ASM International joining ofadvanced and specialty materials III. Singh M. et al., editors; 2001. p. 157.

[12] Reynolds AP, Lockwood WD, Seidel TU. Processing-property correlation infriction stir welds. Mater Sci Forum 2000;331:1719.

[13] Sato YS, Park SHC, Kokawa H. Distribution of tensile property andmicrostructure in friction stir weld of 6063 aluminum. Metall Mater Trans A2001;32A:3023.

[14] Heinz B, Skrotzki B. Metall characterization of a friction-stir welded aluminumalloy 6013. Mater Trans B 2002;33B:489.

[15] Sato YS, Kokawa H, Enomoto M, Jogan S, Hashimoto T. Microstructureevolution of 6063 aluminum during friction-stir welding. Metall Mater TransA 1999;30A:2429.

[16] Reynolds AP, Tang W, Khandkar Z, Khan JA, Linder K. Relationships amongweld parameters, hardness distributions, and temperature histories in alloy7050 friction stir welds. Sci Technol Weld Join 2005;10(2):190.

[17] Jata KV, Sankaran KK, Ruschau JJ. Friction-stir welding effects onmicrostructure and fatigue of aluminum alloy 7050-T7451. Metall MeterTrans A 2000;31A:2181.

[18] Sato YS, Urata M, Kokawa H. Metall parameter controlling microstructure andhardness during friction-stir welding of precipitation-hardenable aluminumalloy 6063. Mater Trans A 2002;33(3):625.

[19] Hassan KAA, Prangnell PB, Norman AF, Price DA, Williams SW. The effect ofwelding parameters on the nugget zone microstructure and properties in highstrength Al-alloy friction stir welds. Sci Technol Weld Join 2003;8(4):257.

[20] Benavides S, Li Y, Murr LE, Brown D, McClure JC. Low-temperature friction-stirwelding of 2024 aluminum. Scripta Mater 1999;41(8):809.

[21] Sakurada D, Kathoh K, Tokisue H. Underwater friction welding of 6061aluminum alloy. J Jpn Inst Light Meter 2002;52(2):2.

[22] Liu G, Murr LE, Niou CS, McClure JC, Vega FR. Microstructural aspects of thefriction stir welding of 6061-T6 aluminum. Scripta Mater 1997;37:355.

[23] Su J-Q, Nelson TW, Mishra R, Mahoney M. Microstructural investigation offriction stir welding 7075-T651 aluminum. Acta Mater 2003;51:713.

[24] Jiang XJ, Noble B, Holme B, Waterloo G, Tafto J. Differential scanningcalorimetry and electron diffraction investigation on low-temperature agingin Al–Zn–Mg alloys. Metall Mater Trans A 2000;31A:339.

[25] Shenoy RN, Howe JM. A differential scanning calorimetric study of aweldalite™ type alloy. Scripta Metall Mater 1995;33(4):651.

[26] Fuller CB, Mahoney MW, Calabrese M, Micona L. Evolution of microstructureand mechanical properties in naturally aged 7050 and 7075 Al friction stirwelds. Mater Sci Eng 2010;A527:2233.

[27] Starink MJ, Dion A. DSC study of precipitation in an Al–Mg–Mn alloymicroalloyed with Cu. Thermo-chim Acta 2004;417:5.