1-s2.0-S2215dffgdf098615000105-main

ble at ScienceDirect

Engineering Science and Technology, an International Journal 18 (2015) 270e277

Contents lists availa

Engineering Science and Technology,an International Journal

journal homepage: http: / /www.elsevier .com/locate/ jestch

Full length article

Role of welding parameters on interfacial bonding in dissimilarsteel/aluminum friction stir welds

Z. Shen a, Y. Chen b, M. Haghshenas b, *, A.P. Gerlich b

a Tianjin Key Laboratory of Advanced Joining Technology, School of Material Science and Engineering, Tianjin University, Tianjin 300072, Chinab Mechanical and Mechatronics Engineering, University of Waterloo, Waterloo, Canada

a r t i c l e i n f o

Article history:Received 23 October 2014Received in revised form8 December 2014Accepted 16 December 2014Available online 4 February 2015

Keywords:DissimilarInterfacial bondingFriction stir weldingIntermetallic compound

* Corresponding author. Tel.: 1 519 777 8978; faxE-mail address: [email protected] (M. HaPeer review under responsibility of Karabuk Univ

http://dx.doi.org/10.1016/j.jestch.2014.12.0082215-0986/Copyright 2015, The Authors. Productiolicense (http://creativecommons.org/licenses/by-nc-n

a b s t r a c t

In this study, lap welds between Al5754 to DP600 steel (aluminum plate top, and steel plate bottom)were manufactured by friction stir welding (FSW). The effects of welding parameters (i.e. travel speedsand penetration depth into lower steel sheet) on the interfacial bonding, tensile strength, and failuremechanism were investigated. The results show that intermetallic compound of Fe4Al13 was detected atthe Al/Fe interface. The weld strength increases significantly by increasing the penetration depth into thelower steel substrate at all travel speeds. The failure mode under overlap shear loadings is prematurefailure through the aluminum substrate when the penetration depth is more than 0.17 mm, and shearfracture when the penetration depth is less than 0.17 mm.Copyright 2015, The Authors. Production and hosting by Elsevier B.V. on behalf of Karabuk University.This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-

nc-nd/4.0/).

1. Introduction

The use of friction stir welding (FSW) for joining of dissimilarmetals combinations in the automotive and manufacturing in-dustries has been widely studied thanks to the fact that FSW offersa number of advantages for dissimilar materials, including:enhanced mechanical properties (i.e. tensile and fatigue), improvedprocess quality, avoiding consumables, lower health and environ-mental issues, and reduced operating costs [1e3]. In the automo-tive industry, the focus on the application of FSW has mainlyinvolved: the joining of extruded parts to form larger extrusions,sheet joining for tailor welded blanks, and joining of light-weightmaterials. FSW offers numerous advantages and potential for costreductions in each of these cases. However, cost-effective andreliable joints between light-weight materials will demand signif-icant development and further consideration. A compellingexample of dissimilar FSW can be found in the 2013 Honda Accord,where this technique has been applied for joining the castaluminum and stamped steel parts of the engine cradle [4e6]. Inthis case, a notable innovation is the use of a C-frame linear FSWsystemwhich exerts all the axial loading on the tool, thus avoiding

: 1 519 515 0020.ghshenas).ersity.

n and hosting by Elsevier B.V. on bd/4.0/).

the need for an extremely stiff and high load capacity robot andfixture to apply the tool force.

The main advantage common to nearly all the techniques is thatsolid state processing limits the temperature rise within the weldregion. This limits the formation or growth of undesirable andbrittle intermetallic compounds (IMCs) within the weld whichdeteriorate strength. Lower peak temperatures also minimizethermal distortion and residual stresses, which can often lead to thefracture of the joint immediately upon cooling of the weld in thecase when intermetallic compounds are present and cracks areformed in the joint. Chen and Kovacevic [7] pointed out that themaximum temperature in dissimilar FSW Al/steel is 631 C on thesteel side, which is drastically lower than that in fusion welding.Nevertheless, local melting of aluminumwas observed in the weld,which can promote diffusion rate between the steel and aluminumsubstrates, thus IMCs tend to be formed in the Al/Fe system [7]. Ithas been reported that Fe-rich IMCs (i.e. FeAl) are not as detri-mental to the mechanical performance of the joint as other Al-richIMCs (i.e. Fe4Al13), since it has been argued that FeAl is more ductile[8]. Also, an IMCs layer will not drastically deteriorate weldstrength when the thickness of which is less than 2 mm [9]. Hence,the mechanical properties of weld can be improved by altering thetypes, distribution and thickness of IMCs, through selecting weld-ing parameters such as travel speed, and penetration depth.

During lap welding of dissimilar alloys, the key parameters to beconsidered include the tool geometry, rotation speed, and travel

ehalf of Karabuk University. This is an open access article under the CC BY-NC-ND

Z. Shen et al. / Engineering Science and Technology, an International Journal 18 (2015) 270e277 271

speed, as with all other FSW procedures. However, lap welding ofdissimilar alloys also requires careful control of the tool pin length,and its penetration depth into the lower sheet material. Forexample, when aluminum or magnesium alloys are joined to steel,the pin penetration into the steel will rapidly wear away steel-based tools, and to avoid this one may maintain the pin abovethe sheet in order to promote diffusion bonding between the sheets[10]. That is, bonding could be promoted by an indirect diffusionjoining mechanism while maintaining the tool pin around0.05e0.1 mm above the surface of the lower steel sheet duringAl5754/DP600 friction stir lap welding. This maintains the flatinterface profile between the sheets, and results in fewer inter-metallic compounds at the interface. This approach, however,precludes the contribution of mechanical interlocking between thesheets by deformation of the lower sheet into the upper sheet. Itcan also be difficult to maintain this small distance between thetool pin and the lower sheet steel surface. However, when a WC-based tool is used, the tool may penetrate into the steel sheetduring joining without encountering severe wear. In prior work byChen and Nakata [11], the influence of tool penetration wasconsidered in Mg/Steel FSW joining, and it was shown that a thininterfacial reaction zone could be promoted when new layer ofsteel is exposed by the tool. Deformation of the steel sheet duringtool penetration will also promote mechanical interlocking, whichwill contribute to joint strength [8]. However, this will also promoteformation of intermetallic compounds when aluminum and steelsalloys are joined, which may contain pre-existing cracks, have highhardness, and thus limit joint strength [10,12,13]. It should be notedthat in comparison, other works involving diffusion bonding andadhesive bonding have always found that strengths are maximizedwhen the thickness of reaction layer or intermetallic compoundregions is minimized. For example, in the case of friction stir spotwelding of aluminum to steel joining, it has been shown that bondstrength deteriorates drastically once the reaction layer thicknessexceeds 1.5 mm [14]. Considering this fine scale, it would appearthat the FSW technique presents great potential in achieving themaximum theoretical strength between dissimilar joints, since thelow temperatures and rapid speed of the process can be mosteffective in suppressing the growth of intermetallic compounds.

It is obvious that controlling the structure and phases at theinterfacial region of dissimilar joints produced by FSW is verycomplex due to transient thermal cycles and short diffusion time.Since the influence of welding parameters on the structure andstrength of the interfacial region remains unclear, the present workaims to determine the contributions of metallurgical bonding (viadiffusion of aluminum and iron in the stir zone) and mechanicalinterlocking due to deformation of the lower steel sheet duringFSW lap joining of AA5754 aluminum and DP600 steel sheets. Thecontributions of each will be assessed using a combination of mi-croscopy, mechanical testing, and fractography.

2. Experimental procedure

The base materials examined consisted of 2.2 mm thick AA5754aluminum and 2.5 mm thick DP600 dual phase steel, with thecompositions shown in Table 1. A displacement controlled manualmilling machine was utilized to fabricate FSW dissimilar joints,

Table 1Nominal chemical composition of 5754 aluminum and DP600 steel (wt%).

AA5754 Mg Si Cr Cu Zn Fe Al3.13 0.05

Fig. 2. SEM micrograph of bonded region in weld produced using 16 mm/min, (a) atthe edge of the stir zone, and (b) Al/Fe interface.

Fig. 1. Optical micrographs of AA 5754/DP600 dissimilar FSW joints produced using (a)45 mm/min and (b) 16 mm/min, (c) microstructure of steel directly under the tip ofpin.

Z. Shen et al. / Engineering Science and Technology, an International Journal 18 (2015) 270e277272

extensive IMCs is detected in both locations on the Al side which isestimated to be Al-rich IMCs. The layer is formed due tomechanicalmixing of Fe and Al during friction stir welding. As shown in Fig. 2a,some cracks and voids are also observed at the corner of the hookbecause of inadequate material flow. Here, the DP600 material isdisplaced upwards into the AA5754 alloy to a distance of 920 mm.The average thickness of the intermetallic near the centerline at theAl/Fe interface shown in Fig. 2b is measured to be < 10 mm. The EDXquantification at zone A reveals a composition of 34.8% Fe, 63.6% Al,and 1.7% Mg, which is consistent with Fe4Al13 with a small amountof Mg in solution.

An overview of the FSW dissimilar joint produced using 45 mm/min is shown in Fig. 3a, where a defect-free weld wasmadewith novoids. However, a large amount of IMCs are observed within the stirzone, which can facilitate crack propagation along the Al/steelinterface. The composition measured in zone B was 33.3%Fe, 64.9%Al, and 1.8%Mg, which is nearly the same as that measured in zoneA and consistent with Fe4Al13. Meanwhile, the average thickness ofintermetallic in the center of Al/Fe is over 170 mm, which exhibitslayered structure made of steel, aluminum and intermetallic com-pounds (see Fig. 3c). As shown in Fig. 3b and c, many steel frag-ments with different sizes scattered into the aluminum, with thelargest fragments near the center. This is due to the fact that thesteel at the interface was stirred into the aluminum by the tip oftool pin, and the stirring intensity at the edge is greater than that atthe center of the tip of pin due to the higher tangential velocity at

the edge. It also should be noted that the boundary betweenaluminum substrate and the long steel flash can be clearly identi-fied. However, the aluminum and steel were mixed sufficientlywithin the hook region (see Fig. 3b) because the stirring process ismore severe close to the hook region. As indicated in Figs. 2 and 3,lower travel speed does not produce more IMCs, however in thepresent work it may be likely that both of the travel speeds appliedwere comparatively low, and hence did not produce a significantdifference in this regard. It can be suggested thatmost of the energyinput was consumed by stirring of steel for higher penetrationdepths.

In order to determine the distribution of Al and Fe, the inter-facial region of the joint in Fig. 3a was further analyzed by EPMA, asshown in Fig. 4. The EPMA map shows that the majority of thematerials produced around the interface are consistent with theFe4Al13 phase with a similar composition across the bonded region.Here, many steel particles fractured and interspersed within thisintermetallic. The EPMAmap for Al indicates that the steel particlesalso have a boundary layer (appearing in yellow color), all with asimilar small fraction of Al and large Fe content, suggesting thesesteel particles may be outlined by an Fe-rich intermetallic (otherthan Fe4Al13).

Following approximately 10 welding trials with 140 mm lengtheach, and various plunge depths, the tool was examined with amacro microscope, as shown in Fig. 5. The observations of thesurface indicate that negligible wear has been imposed on the tool

Fig. 4. EPMA maps for Fe and Al of joint produced using 45 mm/min.

Fig. 5. Image of the tool pin after FSW tests.

Fig. 3. SEM micrograph of joint produced using 45 mm/min, (a) overview, (b) edge ofbonded region, and (c) center of bonded region layered structure.

Z. Shen et al. / Engineering Science and Technology, an International Journal 18 (2015) 270e277 273

pin following dissimilar welding. This suggests that the high tem-peratures imposed at the interface were sufficient to soften theDP600 steel, and suppress the wear of the WC based pin.

In order to determine the thermal history during the process,temperature measurements were conducted using K-type ther-mocouples positioned at the interface of the sheets and peripheryof the pin. Several thermocouples were positioned, however, most

of which were damaged by the deformation induced by the toolpin. Hence, the temperature measured in the present investigationis the temperature at the outer periphery of the weld. Themaximum temperature profile successfully detected using a travelspeed of 45 mm/min as shown in Fig. 6 (as seen a peak of 424.8 Cwas measured). This is consistent with the steel microstructuresobserved in Fig. 1c, which suggest that no phase transformations inthe steel occurred above the Ac1 temperature. The temperatureincreases to approximately 150 C at a slow heating rate due tooriginal preheating of the sheets, then immediately to the peaktemperature at a fast heating rate, the duration for the temperaturehigher than 400 C is approximate 12 s. The heating and coolingrates near the peak are on the order of 6.47 to 2.12 C/sec, respec-tively. It should be noted that the maximum stable temperature ofthe Al/Fe and Fe4Al13 is much higher than 424.8 C, according toFeeAl binary phase diagram (see Fig. 7).

3.2. Mechanical responses

The distribution of hardness along the centerline of the weld inthe vertical direction is indicated in Fig. 8. As seen the hardnessdecreases gradually to the minimum (66.4 HV) from the top surfaceof aluminum sheet to the Al/Fe interface, then increases dramati-cally to the maximum (349 HV) at the layered structure in the Al/Feinterface, and then drops to another minimum (182.3 HV) in the

Fig. 6. Temperature history at the boundary of weld. Fig. 8. Hardness distribution along the centerline of the weld in the vertical direction.

Fig. 9. The locations where hardness tests were performed in the intermetalliccompound.

Z. Shen et al. / Engineering Science and Technology, an International Journal 18 (2015) 270e277274

heat affected zone (HAZ) of the steel, and then increases to the steelhardness of up to 200 HV.

Being a non-heat-treatable (or work-hardened) aluminum alloy,the mechanical properties of Al5754 are greatly influenced bydislocation contribution (i.e. density) and grain size refinementrather than precipitates in the structure. Therefore, softening in theAl sheet can be attributed to the fact that the recovery occurs andthe grain sizes near the Al/Fe interface was coarser than that on theupper surface of the Al [16]. Meanwhile, the variation of hardness insteel also can be attributed to the variation of grain size as a whole(see Fig. 1c), and softening in the HAZ of the steel sheet can beattributed to the tempering of the martensite islands in that basematerial.

The maximum hardness was measured at the Al/Fe interfacedue to the formation of intermetallic compound there. In order toinvestigate the IMCs at the Al/Fe interface in more detail, EDXanalysis was conducted at the indents (see Fig. 9 and Table 2,respectively). The hardness at the layered structure in location B is399 HV, and the composition in location B is 32.5%Fe, 63.62%Al, and3.88%Mg, which is also consistent with Fe4Al13. The hardness valuehere is consistent with 470 HV measured for the intermetallic inprior FSW Al/steel joints in prior work by Kundu et al. [17] In

Fig. 7. FeeAl equilibrium phase diagram [21].

comparison, the average hardness in the steel close to the Al/Feinterface is 290e293 HV, where the composition is mainly Fe(98.31%), with a small amount of Al, due to that aluminum haslimited solubility in iron (see Fig. 7). These hardness values in thesteel are consistent with the temperatures measured suggestingthe stir zone region remained below the steel transformationtemperature.

In order to investigate the effects of tool pin penetration into thesteel on the strength properties, the overlap shear tests were per-formed for 30 and 20 mm wide joints produced using differentpenetrations. As indicated in Fig. 10, the results suggest that thetensile strength decreases and then increases with the increasing ofpenetration depth for both travel speeds. The maximum failurestrength of 236.4 N/mm was obtained at the welding condition ofthe travel speed of 45mm/min and penetration depth of 0.389mm.

Table 2EDX quantification results (wt%) indicated in Fig. 9.

Spectrum In stats. Mg Al Fe Total

A Yes 1.69 98.31 100.00B Yes 3.88 63.62 32.50 100.00

Fig. 10. Correlation between fracture load and penetration depth of the pin into thesteel.

Fig. 12. Failed tensile specimens, a) shear fracture occurred through interface, and b)premature fracture occurred through aluminum substrate.

Z. Shen et al. / Engineering Science and Technology, an International Journal 18 (2015) 270e277 275

A weld with comparable strength can be obtained by main-taining the tip of pin approximately 0.1 mm above the Al/Feinterface, since this promotes an interfacial layer with fewer cracksat the Al/Fe interface through an indirect diffusion joining mecha-nism [10]. As shown in Fig. 11, the aluminum surface is imposedonto the lower steel sheet by the tool at the Al/Fe interface, and thewidth of the bonded region is consistent with the diameter of pin.However, the surface of steel is rather flat, preventing the formationof mechanical interlocking at the Al/Fe interface (by the displace-ment of the lower steel sheet into the upper aluminum sheet). Theinterfacial layer is similar to what has been found by Gendo et al.[18], in which diffusion bonding was formed by diffusion of thecoating layer at the steel surface into the aluminum sheet.

As indicated in Fig. 11, chaotic mixed structures with a mass ofdefects such as cracking and voids, were produced by penetratingthe pin a small distance into the steel substrate (less than0.078 mm), which counteracts the contribution of the mechanicalinterlocking effect and is responsible for the decreasing of weldstrength. However, when the penetration depth reaches 0.092 mm,intermixing of the two sheet was enhanced and metallurgicalbonding occurs with fewer defects at the Al/Fe interface within stirzone, thus improving the strength of the weld. It is worth notingthat a more IMCs were formed in the interface with intermediatepenetration depths (0.092e0.17 mm) than that at higher or lowerpenetration depth (see Figs. 2 and 3), which tends to deteriorate theweld strength when the crack propagates along the interface.Beyond this penetration depth, much more DP600 material is

Fig. 11. Interfacial bonding obtained

displaced upwards into upper Al5754 sheet, the elongated steelflash promoted a mechanical interlocking effect at the weld edges[19]. The displaced DP600 material in the aluminum sheet appearsto have provided more surface area to disperse the intermetalliccompounds, thus resulting in a slightly lower overall thickness.Hence, it can be concluded that the penetration depth plays acrucial role on determining the strength of the weld.

Two types of failure modes were observed during overlap shearloading: either shear fracture occurred through interface, or pre-mature fracture through the aluminum substrate. As shown inFig. 12a, failure occurred at the Al/Fe interface, when the pinpenetrated less than 0.17 mm into the lower steel substrate, whichcan be attributed to the inferior bonding at the interface. In addi-tion, the steel flashes at the weld edge are not strong enough topreclude the crack from propagating into the interface (see Fig. 11).Furthermore, the formation of brittle IMCs is also a critical factordeteriorating the mechanical properties of dissimilar Al/steelwelds. When the penetration depth is higher than 0.17 mm, a weldwith significant steel flash or hook-shaped features at the weldedge is formed (see Fig. 1). Hence, when the weld is subjected toexternal load, the crack initiates at the tip of the long steel flash andpropagates along a short distance and finally into aluminum sub-strate. Therefore, the failure occurred through aluminum substrate(see Fig. 12b). Under such circumstances, the IMCs at the interfacescarcely influence the weld strength since the crack does notpropagate along the center interface below the pin.

To further investigate the failure mechanism and identify theintermetallic compounds formed at the Al/Fe interface, SEM and

at different welding parameters.

Fig. 13. SEM micrographs obtained from the fracture surface of the a) interface and b)Al substrate at the long steel flash indicated in Fig. 12b.

Fig. 14. XRD patterns obtained from fracture surfaces of the a) shear fracture (16 mm/min), and b) premature failure through the Al sheet (45 mm/min).

Z. Shen et al. / Engineering Science and Technology, an International Journal 18 (2015) 270e277276

XRD analysis were performed on the failed fracture surfaces. Fig. 13reveals SEMmicrographs of fracture surface from the Al/Fe interfaceand the Al substrate near the elongated steel flash (see Fig. 12b).Fig. 14 displays the XRD spectrums obtained from the fracture sur-face of shear fracture and Al substrate. As indicated in Fig. 13a, thefracture surface at theAl/Fe interface,whose location corresponds tothe interface in Fig. 2b, is rather brittle. IMC corresponding to Fe4Al13was detected at this surface (see Fig. 14a), which partially contrib-utes to the brittle fracture surface. In addition, the presence of a peakcorresponding to AlFe was observed at the fracture surface (seeFig. 14b), which is consistent with the Al/Fe interfaces appearing asan intermediate chemistry in Fig. 4, however this is onlysuggested tobe AlFe since only one peak was detected. As a Fe-rich IMC, AlFe ismuch ductile than Fe4Al13 [20]. As indicated in Fig. 13b, the fracturesurface of Al substrate at the long steelflash is comparatively ductile,as suggested by the boundary between Al substrate and long steelflash clearly identified in Figs. 2a and 3b.

4. Conclusions

The role of welding parameters (penetration depth into lowersteel sheet and travel speed) on the interfacial bonding and me-chanical performance of friction stir lap welded AA5754 and DP600were investigated. The following conclusions can be drawn:

1. Weld of Al 5054 plate and DP600 steel plate (Al plate top, steelplate bottom) with excellent mechanical properties was suc-cessfully manufactured by friction stir welding.

2. Higher penetration depth resulted in less intermetallic com-pounds at the Al/Fe interface.

3. Penetration depth into the steel substrate plays a decisive role indetermining the weld strength.

4. The micro-hardness distribution across the joint indicates thatthemicro-hardness in the joint interface is greater than the basematerials.

5. There is a correlation between the penetration depth into thelower steel sheet and the failure mode. In other words, pre-mature failure through the Al sheet occurs when the penetra-tion depth is not lower than 0.17 mm into the lower steelsubstrate. Shear fracture occurs when the penetration depth islower than 0.17 mm.

6. Intermetallic compound of Fe4Al13 was detected at the fracturesurface, which are responsible for the deteriorated weldstrength at lower penetration depth.

Acknowledgments

The authors acknowledge the financial support provided toZhikang Shen by the China Scholarship Council (CSC) during the

Z. Shen et al. / Engineering Science and Technology, an International Journal 18 (2015) 270e277 277

present investigation. Further financial support from the NaturalScience and Engineering Research Council (NSERC) and the Cana-dian Foundation for Innovation (CFI) are greatly appreciated. Sup-port from Linamar (CAMTAC Manufacturing) in Guelph, Ontario forproviding theWC tools is also appreciated. Also, the authors wish tothank Dr. Yuquan Ding for his valuable help on the SEM analyses.

References

[1] P.L. Threadgill, A.J. Leonard, H.R. Shercliff, P.J. Withers, Friction stir welding ofaluminum alloys, Int. Mater. Rev. 54 (2009) 49e93.

[2] C.J. Dawes, W.M. Thomas, Friction stir process welds aluminum alloys, Weld. J.75 (1996) 41e45.

[3] R.S. Mishra, Z.Y. Ma, Friction stir welding and processing, Mater. Sci. Eng.: R 50(2005) 1e78.

[4] Y. Kusuda, Ind. Robot: An Int. J. 40 (2013) 208e212.[5] B.T. Gibson, D.H. Lammlein, T.J. Prater, W.R. Longhurst, C.D. Cox, M.C. Ballun,

K.J. Dharmaraj, G.E. Cook, A.M. Strauss, Friction stir welding: process, auto-mation, and control, J. Manuf. Processes 16 (2014) 56e73.

[6] M. Haghshenas, S. Sahraeinejad, A. Abdel-Gwad, B. Gocke, A.P. Gerlich, Dis-similar joining of automotive sheet materials by friction stir welding, in:CanWeld Conference, 2013. Niagara Falls, Ontario.

[7] C.M. Chen, R. Kovacevic, Joining of Al 6061 alloy to AISI 1018 steel by com-bined effects of fusion and solid state welding, Int. J. Mach. Tools Manuf. 44(2004) 1205e1214.

[8] T. Liyanage, J. Kilbourne, A.P. Gerlich, T.H. North, Joint formation in dissimilarAl alloy/steel and Mg alloy/steel friction stir spot welds, Sci. Technol. Weld.Joining 14 (2009) 500e508.

[9] C.Y. Lee, D.H. Choi, Y.M. Yeon, S.B. Jung, Dissimilar friction stir spot welding oflow carbon steel and AleMg alloy by formation of IMCs, Sci. Technol. Weld.Joining 14 (2009) 216e220.

[10] M. Haghshenas, A. Abdel-Gwad, A.M. Omran, B. Goke, S. Sahraeinejad,A.P. Gerlich, Friction stir weld assisted diffusion bonding of 5754 aluminumalloy to coated high strength steels, Materi. Des. 55 (2014) 442e449.

[11] Y.C. Chen, K. Nakata, Effect of tool geometry on microstructure and me-chanical properties of friction stir lap welded magnesium alloy and steel,Mater. Des. 30 (2009) 3913e3919.

[12] M. Movahedi, A.H. Kokabi, S.M. Seyed Rehani, N. Najafi, Effect of tool traveland rotation speeds on weld zone defects and joint strength of aluminiumsteel lap joints made by friction stir welding, Sci. Technol. Weld. Joining 17(2012) 162e167.

[13] Y.C. Chen, K. Nakata, Effect of the surface state of steel on the micro-structure and mechanical properties of dissimilar metal lap joints ofaluminum and steel by friction stir welding, Metall. Mater. Trans. A 39(2008) 1985e1992.

[14] R. Qiu, C. Iwamoto, S. Satonaka, The influence of reaction layer on the strengthof aluminum/steel joint welded by resistance spot welding, Materials Charact.60 (2009) 156e159.

[15] H.H. Cho, S.H. Kang, S.H. Kim, K.H. Oh, H.J. Kim, W.S. Chang, H.N. Han,Microstructural evolution in friction stir welding of high-strength linepipesteel, Mater. Des. 34 (2012) 258e267.

[16] A. Elrefaey, M. Gouda, M. Takahashi, K. Ikeuchi, Characterization of aluminum/steel lap joint by friction stir welding, J. Mater. Eng. Perform. 14.1 (2005)10e17.

[17] S. Kundu, D. Roy, R. Bhola, D. Bhattacharjee, B. Mishra, S. Chatterjee,Microstructure and tensile strength of friction stir welded joints betweeninterstitial free steel and commercially pure aluminium, Mater. Des. 50(2013) 370e375.

[18] T. Gendo, K. Nishiguchi, M. Asakawa, S. Tanioka, in: Proceedings of the SAEWorld Congress, 2007. Detroit (USA), paper # 2007-01-1703.

[19] Y. Wei, J. Li, J. Xiong, F. Huang, F. Zhang, Microstructure and mechanicalproperties of magnesium alloy and stainless steel weld-joint made by frictionstir lap welding, Mater. Des. 33 (2012) 111e114.

[20] M. Potesser, T. Schoeberl, H. Antrekowitsch, J. Bruckner, in: EPD Congress,TMS (The Minerals, Metals & Materials Society), San Antonio Texas, 2006, pp.167e176.

[21] B. Sundman, I. Ohnuma, N. Dupin, U.R. Kattner, S.G. Fries, An assessment of theentire AleFe system including DO3 ordering, Acta Mater. 57 (2009)2896e2908.