Seediscussions,stats,andauthorprofilesforthispublicationat:http://www.researchgate.net/publication/266148591

Micro-frictionstirweldingoftitanzincsheets

ARTICLEinJOURNALOFMATERIALSPROCESSINGTECHNOLOGY·FEBRUARY2015

ImpactFactor:2.24·DOI:10.1016/j.jmatprotec.2014.08.029

READS

113

3AUTHORS:

SpyrosPapaefthymiou

NationalTechnicalUniversityofAthens

15PUBLICATIONS94CITATIONS

SEEPROFILE

ConstantinosGoulas

DelftUniversityofTechnology

3PUBLICATIONS0CITATIONS

SEEPROFILE

EvangelosGavalas

RWTHAachenUniversity

1PUBLICATION0CITATIONS

SEEPROFILE

Availablefrom:ConstantinosGoulas

Retrievedon:27October2015

M

Sa

Zb

c

d

a

ARRAA

KFZT

1

i(sat

wTmtfaaWd

m

(

h0

Journal of Materials Processing Technology 216 (2015) 133–139

Contents lists available at ScienceDirect

Journal of Materials Processing Technology

jo ur nal ho me page: www.elsev ier .com/ locate / jmatprotec

icro-friction stir welding of titan zinc sheets

. Papaefthymioua,∗, C. Goulasb,c, E. Gavalasd

National Technical University of Athens, School of Mining & Metallurgical Engineering, Laboratory of Physical Metallurgy, 9, Her. Polytechniou str.,ografos, GR-15780 Athens, GreeceMaterials Innovation Institute (M2i), Mekelweg 2, 2628 CD Delft, The NetherlandsDelft University of Technology, Department of Materials Science and Engineering, Mekelweg 2, 2628 CD Delft, The NetherlandsELKEME S.A., Piraeus str. 252, GR-17778 Tavros, Athens, Greece

r t i c l e i n f o

rticle history:eceived 11 March 2014eceived in revised form 4 August 2014ccepted 29 August 2014vailable online 6 September 2014

eywords:

a b s t r a c t

Aim of this research is first to evaluate the applicability of micro-friction stir welding (�FSW) to wroughtzinc alloy sheets and then to improve the structural integrity of such joints. �FSW tool design was basedon an algorithm that considers material and process limitations. Joining trials were performed at differentfeed rates. It is proven that joining by �FSW thin ZnTiCu sheets is possible and it offers extremely finemicrostructures and �-phase distribution due to the mechanical fragmentation which is the outcome ofthe stirring. The �-phase particles were homogenized and precipitated inside deformed zinc grains and

riction stir weldingnTiCuiZn15

not at the grain boundaries, where they used to be in fusion welds. Electron microscopy showed thatits size was limited to 150 nm, which is in average 13 times smaller than the size of the 2 �m that theyget when sheets are TIG welded. Macroscopically, the �FSW joint mechanical properties are comparablewith industrially fusion-welded material. The relative low elongation achieved, similar to fusion-weldedsheets, is explained by the occurrence of three main defects: root opening, thinning and kissing bond.

. Introduction

Friction stir welding (FSW) is, a solid state joining technique,nvented by The Welding Institute (TWI), United Kingdom in 1992Thomas et al., 1992). Although initially developed for Al-alloys,oon it was applied successfully to many other metals and materi-ls, especially to metals hard to weld using the given fusion weldingechniques.

In FSW frictional heat is generated between the wear-resistantelding tool shoulder and pin, and the material of the work-piece.

he frictional heat and surrounding temperature cause the stirredaterials to be softened and mixed avoiding their melting. Thus,

he bonding is a solid state process. However, the grains are trans-ormed and relocated. Material flow patterns under the shoulderre similar to the forging process, while these around the tool pinre like an extrusion process as stated by (Mishra and Ma, 2005).

eld quality is strongly affected by the tool geometry as shown in

etail by Sued et al. (2014) and Groche et al. (2014).Groche et al. (2014) in their review paper described in detail

ethods of joining by forming, among others FSW, which seems

∗ Corresponding author. Tel.: +30 210 772 4710.E-mail addresses: spapaef@metal.ntua.gr, spyrospapaef@gmail.com

S. Papaefthymiou), K.Goulas@tudelft.nl (C. Goulas).

ttp://dx.doi.org/10.1016/j.jmatprotec.2014.08.029924-0136/© 2014 Elsevier B.V. All rights reserved.

© 2014 Elsevier B.V. All rights reserved.

to be of paramount importance. In general, regarding the process,during FSW the rotational movement of a tool is overlapped bythe feed speed. The tools are primarily machined from tool steelssuch as H13 and subsequently hardened. In research settings, thesetools would typically exhibit little wear and would be consideredpractically non-consumable. The side aligning the directions ofboth speeds is called the advancing side, whereas the other sideis called the retreating side. Plastic flow and frictional heating cre-ate a characteristic stir zone. Friction stir welded microstructuresconsist of four zones: the dynamically recrystallized zone (DXZ) ornugget, the thermo-mechanical affected zone (TMAZ) and the heataffected zone (HAZ). The DXZ or nugget is located where the pin isin contact with the base material. This zone is characterized by suf-ficient values of thermal and strain energy for recrystallization tooccur leading to fine grains of equal size. The thermo-mechanicallyaffected zone (TMAZ) is located at the side of the nugget, is affectedby high strains and heat, but the energy provided is insufficient fora complete recrystallization. The heat affected zone (HAZ) consistsof undeformed base metal and is thermally affected by the heatproduced by the stirring of the tool.

As Groche et al. explicitly point out during FSW the materialnear the top of the work piece is stirred under by the threads

and is deposited in the weld nugget, if fusion welding, meltingand homogenization would take place. However, larger concentra-tion differences can be found in the welding zone, with diffusioncouples at the edge. The joining mechanism can be described as

1 als Pro

am

GFasltnmwitmCej

FtFcer

ist

ihivbafpbsjfdimiwsoewh

tmttG(adtLLpm

34 S. Papaefthymiou et al. / Journal of Materi

combination of cold pressure welding, diffusion welding andechanical alloying at the same time (Groche et al., 2014).Defects in FSW might be of any orientation, size, or shape.

roche et al. describe and categorize very well the defects inSW. Like arc welding, the process moves in a linear fashion, usu-lly at a constant rate along the joint line, and therefore has aimilar tendency to produce defects, which propagate for someength and have their major dimensions parallel to the travel direc-ion. Defects are divided into inclusions, volumetric defects, andon-volumetric (laminar) defects. Flash is produced by displace-ent of material from the face (tool-side surface) of friction stirelded components. Root flaws are surface-breaking discontinu-

ties that are present on the material surface which is opposite ofhe tool. The mechanical effect of a root defect is dependent pri-

arily on its through-wall height and degree of bonding or width.hanges in root defect orientation also change mechanical prop-rties. Especially in thin sheets root defect is detrimental for theoint’s structural integrity.

According to Kim et al. (2006) the defects which occur duringSW can be caused by an overheating or insufficient heating of par-ial areas. Overheating can lead to the formation of an excess flash.urthermore, the grain structure can be affected by exceeding aertain temperature value. Cavities can be formed due to an inad-quate material flow, which can be caused by a disadvantageousotational speed to travel speed of the tool.

Thus, from all the above it is clear that the FSW process can benfluenced by the tool design, the travel speed and the rotationalpeed, the vertical pressure of the tool, the tilt angle of the tool andhe material characteristics.

Mishra and Ma (2005) underline the difficulty of creating highntegrity welds of aerospace aluminium alloys able to withstandigh-strength and fatigue loads, not to crack due to the poor solid-

fication microstructure and porosity in the fusion zone and theery significant loss in mechanical properties as compared to thease material. These aluminium alloys (e.g. highly alloyed 2XXXnd 7XXX series) are generally classified as non-weldable as theseactors make the joining of such alloys by conventional weldingrocesses unattractive. It is known that some aluminium alloys cane resistance welded, but the surface preparation is expensive, withurface oxide being a major problem. Thus, FSW is an alternativeoining process to fusion welding. It is a suitable joining techniqueor sheet metals to be welded by butt, lap or fillet joints. The rapidevelopment of FSW in aluminium and its successful commercial

mplementation has motivated its application to other non-ferrousaterials (Mg, Cu, Ti, as well as their composites), steel, dissim-

lar alloys/metals, and even thermoplastics. As many metals areelded using given fusion techniques, it is not only important to

how the feasibility of FSW, but also to delineate its advantagesver existing fusion techniques. Additionally, FSW gains potentialngineering importance due to the absence of problems associatedith conventional welding such as distortion, grain growth in theeat affected zone and porosity.

As previously stated a defining characteristic of FSW is thathe joint is created by a non-consumed cylindrical rotating tool,

echanically traversed through the materials. The pin thoughhat might be threaded or polished, can be consumed. Especiallyhreaded pins are usually altered (polished) during processing asibson et al. (2014) describe in their review paper. Lorrain et al.

2010) state that material flows during FSW are very complexnd not fully understood. Given the fact that most of the studiesescribe the use of threaded pins and that the initially threadedools may become unthreaded due to tool wear (Gibson et al., 2014;

orrain et al., 2010), the material flow becomes even more complex.orrain et al. have shown that the material flow with unthreadedin undergoes the same features with classical threaded pins;aterial is deposited in the advancing side (AS) in the upper part

cessing Technology 216 (2015) 133–139

of the weld and in the retreating side (RS) in the lower part of theweld; a rotating layer appears around the tool.

Nishihara and Nagasaka (2004) successfully adapted the FSWprocess to materials with thicknesses of 1000 �m or less. Theauthors named this modified technique micro-friction stir welding(�FSW). �FSW could be applied as well in thin walled structures,electrical, electronic and micro-mechanical assemblies.

Significant challenges related to tool design and critical pro-cedure parameters are to be overcome when �FSW is applied.Different materials with various tool geometries and processparameters are to be tested before excellent joints with opti-mized conditions can exist. Teh et al. (2012) achieved homogeneousmicrostructure distribution after �FSW of thin aluminium sheets;even 180◦ bend was carried out without cracking. This techniquecould be appropriate for special welding uses requiring propertiessuch as high flexural strength, ductility and durability, combinedwith smooth joint appearance.

Béré et al. (2000) indicated the importance of the c/a ratio inrelation to formability. Zinc with its hexagonal close packed crys-tal structure and with lattice parameter c/a ratio 1.856, which ishigher than the ideal 1.633, has a very limited formability, animportant factor that needs to be taken into consideration whendesigning new techniques that involve high degree of deforma-tion. FSW up to now has mostly been applied in easily deformablemetals like aluminium and steel. But when materials with limitedformability (such as metals with hexagonal crystal structure e.g.magnesium, zinc) are to be joined, literature and industrial experi-ence are limited. Unfortunately, especially for welding of zinc alloysthe available scientific literature, even with current fusion weldingtechniques, is very limited. Furthermore, FSW of zinc alloys hasnot been industrially applied yet and experimental results are notpublished (if available).

Copper (Cu) and titanium (Ti) are the main alloying elements inthe ZnTiCu alloy, from which rolled zinc obtains its desired mechan-ical properties. Their chemical composition is defined by thestandard EN 988:1997. ZnTiCu alloy contains usually 0.06–0.20%Ti. Titanium has very little solubility in Zn (0.02%), and it also reactswith Zn and forms a hard intermetallic phase with chemical com-position TiZn15 called �-phase. Copper is in solid solution, forming�-Zn solution.

The presence of the intermetallic �-phase is of high importance,as it refines the cast Zn grains and prevents grain growth during hotrolling. The �-phase precipitates very close to zinc’s melting point(m.p. = 421 ◦C; �-phase forms 2–3 ◦C below the m.p.), and formsstringers along the rolling direction in the �-Zn matrix. Their sizeand distribution are crucial for the alloy’s mechanical propertiesas they can be responsible for precipitation hardening based ontheir size, but also they can act as stress raisers and crack initiationpoints, especially when they are enlarged and placed at the grainboundaries.

Zinc is easily joined with fusion techniques such as TIG/MIG(Squillace et al., 2004) and/or ultrasonic welding (Patel et al., 2014)but the structural integrity of such joints is easily deteriorated.Especially, when very thin zinc sheets are to be welded for con-struction purposes, such as gutters, industrial roofing systems andother architectural applications, both suppliers and applicants facesevere brittle fracture occurrences even at low loading conditionsleading to cracking within the weld zone.

After fusion welding, �-phase redistributes and is totally con-centrated at the grain boundaries. The precipitation hardeningeffect is thus cancelled and the overall structural integrity of thewelded part can be compromised.

Pantazopoulos and Sampani (2007), in a case study of a zinctube failure, found that zinc wrought alloys can suffer from failurecaused by crack formation initiated at such hard intermetallic �-phase precipitates.

als Pro

f(tca

aiiTbepl

sfpdztifscBt(taso

iw

Fao

uctstastf

otffitr

tm1pttei

S. Papaefthymiou et al. / Journal of Materi

Zinc alloys industrially are welded using highly productiveusion welding techniques. In a previous work of the authorsGoulas et al., 2013) it was indicated that these welds are suscep-ible to cracking due to their inhomogeneous microstructure withoarse columnar grains and segregated intermetallic precipitatest the grain boundaries.

The application of FSW is of great interest for such applications,s it will prevent the �-phase enlargement that is caused by highnput energy of the fusion technique and will contribute to mechan-cal fragmentation of the precipitates caused by the stirring tool.he production of welds by plastic deformation, at temperatureselow the melting temperature of the base materials, can be anffective way for reducing the formation of brittle intermetallichases and consequently to create joints with higher integrity and

ack of cracking in the weld zone.In our case where brittle, elongated �-phase particles can cause

tress concentration at grain boundaries and can lead to earlyailure of the zinc sheets, it is important to define the properrocessing conditions to avoid overloading and grain degradationue to overheating-melting and/or mechanical stirring. Similar toinc’s hard �-phase, Simoncini and Forcellese (2012) pointed outhat the formation of brittle inter-metallic compounds near thenterface between Al and Mg alloys, when these two metals areusion welded deteriorate the integrity of the joints. Presence ofuch inter-metallic phases in the fused zone is the main reasonausing brittleness and low strength of the joints, as observed byen-Artzy et al. on joints in Mg-Al alloys obtained by means of gasungsten arc and electron beam welding techniques, and by Liu et al.2013) on laser weld bonded joints in Mg–Al alloys. Such defect,hat strongly reduces the mechanical properties of the assembly,lso occurs in unconventional welding, such as high energy den-ity laser welding. An emerging and very attractive technology tovercome such drawback is the friction stir welding (FSW).

The problem of the second phase brittleness is even morentense and has severe impact in the structural integrity of the joints

hen thin and ultra-thin sheets are welded.Lately spot FSW is applied to successfully joint very thin sheets.

or micro-electronic, medical, micro devices, soldering and otherpplications micro-FSW and spot-FSW are effective joining meth-ds as dissimilar metals can be joined together (Wang et al., 2010).

Wang et al. (2010) characterized numerous spot-FSW of 300 �mltra-thin aluminium 1050 sheets investigating also the failureauses that were found to be related to the processing parame-ers as failure initiated near the possible original notch tip in thetir zone and the failure propagated along the circumference ofhe nugget to cause final fracture. The interface between the heatffected zone and the thermal-mechanical affected zone was theoftest region, where the cracks of friction stir spot microwelds inhe lap-shear specimens under the loadings initiated and lead toracture of the specimens.

Yin et al. (2010) correlated FSW processing parameters to failuref AZ31 Magnesium alloy after spot welds. Tool shoulder pene-ration and dwell time are crucial for material mixing and theormation of hook regions. Different grain sizes contributed to thenal material-joint integrity. Buffa et al. (2009) showed how par-ially bonded regions with its curved profiles, referred to as hookegions, create defects (hooking defects).

Many researchers applied FSW in very thin sheets, most ofhem joined aluminium, copper and/or dissimilar alloys and

etals with each other. Galvão et al. (2013) successfully joined mm-thick dissimilar copper sheet to 6 mm thick aluminiumlates (aluminium–copper/different Al alloys – copper) proving

he importance of the processing parameters in combination tohe alloy (heat-treatable or non-heat-treatable alloy). Barcellonat al. (2006) discussed the critical microstructural phenomena thatnfluence the mechanical properties of different aluminium alloys

cessing Technology 216 (2015) 133–139 135

when FSW is applied and even post-weld heat treatments areapplied.

Understanding the efforts of researchers and the technologydrive to join thin sheets for applications of a broad scope, it is atremendous challenge of FSW technology to be coupled togetherwith material characteristics that will enable a successful widen-ing of the process application spectrum. In this framework, the aimof this work is to create joints of thin titanium zinc sheets withfiner structure and improved ductility in order to overcome thebrittleness of current fusion welded sheets for structural and archi-tectural applications. Additionally, as zinc has a similar lattice withmagnesium and thus shows similar limitations in forming, FSWwas one more challenge to overcome including process parameterdetermination and successful joining.

For this purpose �FSW was applied. Using �FSW, thin sheetscould be joined avoiding the columnar microstructure and thestringer like morphology of intermetallic phases while creatingjoints with good integrity and smooth appearance, comparable toindustrially welded sheets.

2. Experimental procedure

The usual FSW tool designs are mostly empirical and difficultto optimize, as time consuming series of experiments are required.For this reason, Gratecap et al. (2008) developed a model to definethe tool geometry and dimensions, based on a selected workingtemperature (0.7–0.8Tm). In their study, this model was appliedto design the tool to join 5 mm thick aluminium plates. A majorchallenge in our work was to apply the same algorithm aiming toproduce sound joints of thin (700 �m thick) ZnTiCu sheets. Apartfrom the significant thickness difference in the study of Gratecapet al., our material, ZnTiCu, has remarkably different physical prop-erties than aluminium.

According to the model of Gratecap et al., the temperature devel-oped during the joining process is a function of tool geometryand dimensions, physical properties of the material and the join-ing parameters, such as tool rotation speed and feed rate. In ourcase, the working temperature was selected to be 320 ◦C, corre-sponding to the 0.75Tm. The used experimental setup faced thefollowing limitations (that were considered in the tool design):maximum rotational speed and feed rate were 1000 rpm and318 mm/s respectively. Additionally, the tool geometry was tunedby the model with aim to minimize material loss and at the sametime to have sufficient material displacement around the pin andunder the shoulder.

The final geometry and dimensions of the used tool are shown inFig. 1. The FSW tool was made of X40CrMoV5-1 tool steel temperedat 500 ◦C, with the following dimensions: shoulder diameter 5 mm,shoulder height 5 mm, pin diameter 0.5 mm and pin height 0.7 mm.The tool was applied to the sheets under a 3◦ tilt angle.

For a range of different feed rates, two of the above sheets werebutt welded longitudinally to rolling direction. The joining lengthwas 200 mm on 250 mm total sample length. The tool penetrationdepth was manually adjusted to achieve maximum stirring com-bined with the lowest possible weld flash under the given processconditions. Trials were executed in a Mill with an adapted �FSWtool. Additionally, according to the practice of Teh et al. (2012) spe-cial hold-downs in combination with a low thermal conductivityback plate were used as shown in Fig. 2. The presence of the backplate assisted the process by reducing heat extraction from the weldzone (Teh et al., 2012).

Metallographic, tensile and three point bending sampleswere cut from �FSW welded ZnTiCu pieces with original size0.7 mm × 250 mm × 50 mm. The chemical composition of the usedZnTiCu alloy is shown in Table 1.

136 S. Papaefthymiou et al. / Journal of Materials Processing Technology 216 (2015) 133–139

Fig. 1. Geometry and dimensions of the used �FSW tool.

Table 1Chemical composition of ZnTiCu sheets used for FSW experiments (wt%).

Zn Pb Al Cu Sn Fe Ti

99.7665 0.0039 0.0093 0.1498 0.0001 0.0009 0.0690

Fig. 3. FSW with 127 mm/min (at 1000 rpm). As-polished micrograph showing joint area

is observed. The characteristics of �FSW (stir zone, onion rings, weld flash) are indicated

Fig. 2. The FSW set-up used for the current experiments.

All welds produced were examined visually for their integrity.The macroscopically inspected and approved welds were thencross-sectioned for detailed examination by means of light opti-cal microscopy (LOM) using an inverted microscope and scanningelectron microscopy (SEM) equipped with secondary electron andback scattered electron sensors. Energy dispersive spectroscopywas carried out for local elemental chemical composition analy-sis and for the identification of the �-phase intermetallic particles.Finally, from each weld, specimens for tensile and 3-point-bendingtests were also prepared in sets of three (3).

The cross sections were prepared according to the metallo-graphical standards; chemical etching with Nital 2% (2% HNO3 98%ethanol) was applied to reveal �-phase distribution. For the ten-sile and the 3-point bending tests an electromechanical testingmachine with 30 kN load cell was used.

3. Results

Metallographic evaluation proved successful joining at different

feed rates (travel speed). Optical and electron micrographs showingthe microstructure of ZnTiCu FSW welds are illustrated in Figs. 3–7.

LOM examination discerned the characteristic FSW “wine-glass” and “bowl” morphologies as described by Khaled (2005).

Fig. 4. FSW with 318 mm/min (at 1000 rpm). As-polished micrograph showing jointarea in optical microscope under polarized light. Characteristic “bowl morphology”is observed. Intense weld flash is evident affecting the overall performance of theweld.

in optical microscope under polarized light. Characteristic “wineglass” morphology.

S. Papaefthymiou et al. / Journal of Materials Processing Technology 216 (2015) 133–139 137

Fig. 5. Electron micrograph (SEM) illustrating the distribution of the �-phase (sec-ondary electron detector). �FSW was performed at 127 mm/min. (a) Stir zone andbw�

Tmae

ase metal. (b) Details of stir zone showing the sub-microscopic dispersoids thatere formed due to the extended mechanical fragmentation. The difference in the-phase particles in comparison to (a) is clear.

he characteristic FSW regions i.e. Nugget, stir zone, thermo-echanically affected zone and onion ring patterns are identified in

ll joints (Figs. 3 and 4). Weld flash was observed as well. The exist-nce of heat affected zones (HAZ) was not able to be distinguished

Fig. 6. Electron micrograph showing �-phase stringers in the base metal.

Fig. 7. Electron micrograph. �FSW performed at 318 mm/min. Details of the sub-microscopic dispersoids of the �-phase in the stir zone. The �-phase is no more instringers, but in finely dispersed round particles.

as no microstructural changes were observed near the stir zone;instead base metal structure was present, neighbouring with thethermo-mechanically affected zone (Fig. 3).

Base metal microstructure consisted of highly elongated grainstowards rolling direction and �-phase stringers, with average grainsize of 1000 nm.

The microstructure of the stir zone was very fine. The �-phaseparticles were found in the SEM to be significantly smaller in thestir zone than in the base metal (Fig. 5). They exhibited a size vari-ation from 50 nm to 250 nm, with an average size at about 170 nm.For comparison, note that the base metal’s �-phase particle sizewas approximately 2000 nm. The �-phase distribution in the weld’sregions was not in stringers as it was in the as-rolled material(Fig. 6), but in spherical sub-microscopic particles, evenly dispersedin the �-Zn matrix (Fig. 7). The average grain size was also reducedat about 400 nm in average within the stir zone (Fig. 7).

Mechanical properties obtained for FSW samples are shownbelow. The mechanical properties of industrial TIG welded zincsheets, approved by quality inspection, are used for refer-ence/comparison reasons (Figs. 8 and 9). Specimens were testedparallel to the rolling direction.

Macroscopically, the mechanical properties of �FSW and fusionwelded sheets are comparable. TIG welded samples achieved

slightly greater UTS, but with less elongation. From the 3-pointbending test of �FSW samples, fusion-welded and as-received sam-ples, stress-flexure extension curves were obtained. Specimens

Fig. 8. Tensile test results of �FSW-welded ZnTiCu sheets; industrial TIG sample isused as reference.

138 S. Papaefthymiou et al. / Journal of Materials Pro

Ft

wh1b

4

dictpsowo

raerahppii

1

2

ig. 9. (a) Bending test results of �FSW ZnTiCu sheets. (b) Statistical evaluation ofhe results and scatter plots.

ere tested transverse to rolling direction. Fusion welds exhibitedigher flexural strength, but they reached lower extension (up to0 mm). On the other hand, �FSW joints showed more consistentehaviour even at extensions of about 20 mm.

. Discussion

Grain boundary embrittlement of fusion welded TiZnCu sheetsue to hard �-phase precipitates compromises the structural

ntegrity of very thin welded sheets. As �-phase particles pre-ipitate at temperature just below the solidus, there is no heatreatment to bring those into solution and then to allow their re-recipitation in a fine, homogeneous dispersed manner. Frictiontir welding, through the high deformation in the joining region,ffers a fine microstructure with evenly distributed intermetallics,hich could give an effective solution to the embrittlement issue

f very thin TiZnCu sheet joints.In order to achieve joints of high integrity, all possible flaws

elated to the FSW process parameters must be first determinednd then the occurrence of flaws must be limited if not directlyliminated. Despite the successful �FSW joints, three main defectselated to process parameters were evident: root opening, thinningnd kissing bond. Occurrence of kissing bond (Fig. 3) indicates loweat input, usually barely enough to create a joint. The kissing bondhenomenon was observed and analysed by Sato et al. (2005). Poorerformance in the 3-point bending test of the 318 mm/min sample

n comparison to the 127 mm/min can be attributed to the followingssues:

. Thinning by 15% further weakened the material’s resistance, asit reduced significantly the area to which the tension is applied.

. The “kissing bond” weakened the material as it acted like a pre-crack in the 45◦ direction to the loading.

cessing Technology 216 (2015) 133–139

Both these facts explain the reason for the �FSW 3-point bend-ing samples not being able to reach the strain level of the referenceTIG samples. Flexure extension though is superior. This is attributedto the very fine microstructure in the joint area (Rodrigues et al.,2009).

At the retreating side (Fig. 3), the microstructure inconsistencybetween stir zone and base metal is clear. An increase of the toolrotation speed or shoulder diameter would increase the heat gener-ation. This would lead to smoother and more homogeneous joints.

Heat caused by friction is dependent on the vertical load. Due tothe fact that load varies because of:

1. inhomogeneities in the base metal;2. variation of the sheet’s cross section;3. disturbances in the base metal or tool drives;4. dynamic loads generated by acceleration/deceleration of moving

components;5. vibration transmitted by the environment;6. self-excited vibration generated by the process or friction, in

order to achieve uniform welds throughout their length, automa-tion of the process with force control system is essential. In thisway, over or under loading would be avoided and, thus, bettercontrol of the process and higher precision could be achieved aswell as better repeatability.

Workpiece alignment could also benefit significantly by processautomation. Manual workpiece alignment can lead to asymmetricheat transfer, material flow and properties of both sides of the weldas Nandan et al. (2007) well indicate. Of course in terms of materialsthat are not easily formable and/or have low heat transfer coeffi-cient, as in the case investigated, such characteristics can play asignificant role in successful joining and will be determinative tothe process conditions.

Process parameters, indeed, were found to be directly affectedby the weld morphology as presented by Cavaliere et al. (2006).Specifically, tool penetration depth as well as travel speed affectedthe joint profile. A shallower penetration combined with a lowertravel speed produced a wineglass-like joint, while a deeper pene-tration with higher travel speed produced a bowl-like joint. Weldflash was evident leading to thinning of the joint area by average15% affecting the overall quality of the weld. The integrity of theweld was significantly affected by the occasional root opening. Thiswas a common defect that seems to be not travel speed dependentas it appeared to both low and high speeds. It can be attributed tothe variant distance between the bottom of the sheet and the pinand the manually controlled process parameters.

As Groche et al. (2014) explicitly point out lack of penetra-tion (LOP), lack of consolidation (LOC), and kissing bond typeroot defects are caused by excessive penetration ligament (dis-tance between bottom of probe and root-side surface of material),inadequate tool-joint alignment (missing the joint), or inadequatedisruption (poor stirring) of the abutted parent material surfacesnear the root respectively. Both LOP and kissing bonds imply thatthere is some portion of the abutted joint surfaces which remainunbonded or inadequately disrupted. The kissing bond is generallyless severe and can involve significant distortion of the jointline.According to the literature, depending on location and extent,LOP defects can have significant effects on mechanical proper-ties.

A distinct difference in the microstructure between stir zoneand base metal was revealed in the SEM. Due to extended stirring,the �-phase particles were mechanically fragmented, thus becom-

ing significantly smaller in the stir zone than in the base metal(Fig. 5). The intensity of the stirring affected also the average grainsize, which is between 170 nm (feed rate 318 mm/min) and 400 nm(128 mm/min) in the stir zone (Fig. 7) compared to 2000 nm in the

als Pro

brtttt

easfit

c9tf

tu

poLlph

5

•

•

•

•

•

•

A

fctMa

A

i2

S. Papaefthymiou et al. / Journal of Materi

ase metal. This is of high importance as the �-phase is mostlyelated to embrittlement issues found in fusion welds. Fragmenta-ion of the �-phase particles is caused mechanically, as a result ofhe high deformation applied in the stir zone. This contributes alsoo weld strength increase as proven through tensile and bendingests.

The �FSW samples exhibited 15% lower strength than the refer-nce TIG samples. This is mainly caused due to the 10–15% thinningnd the occasional existence of root opening. Root opening repre-ents a notch perpendicular to the loading direction. However, thener �-phase distribution in the stir zone of the �FSW contributedo increase of total elongation.

The solid-state joining prevented intermetallic phases fromoarsening. FSW samples managed to achieve bending angles above0◦ without breaking (in contrast with the fusion welded sampleshat broke at much smaller angles). �FSW joints showed greaterormability potential than fusion welded pieces.

The metallographic analysis revealed onion ring patterns withinhe Nugget zone. The onion ring pattern is a banded microstructuresually caused by threaded pins.

Most probably the Nugget zone’s ring pattern is an effect oferiodic differences in the crystallographic orientation of grains,r varying relative orientation in adjacent grains as described byarsson et al. (2000). The used pin though was not threaded, so it isikely that the pin had poor surface quality, wear, or possible ZnTiCuickup. This could be avoided by finer tool finishing as well as byigher rotational speeds.

. Conclusions

ZnTiCu sheets 700 �m thick were successfully joined by applyingmicro-friction stir welding (�FSW) with a tool specially designedfor this purpose.The �FSW parameters of 1000 rpm and 318 mm/min resulted inthe best joint integrity.The �FSW joints were comparable to the sheets welded by pro-cesses of present state-of-the-art industrial practice TIG welds.The �-phase (TiZn15) particles are fragmented through the �FSWprocess, finely dispersed and their size is reduced in the stir zone.The grain structure of the �FSW welds consists of small, equiaxedgrains, which differ significantly from the large columnar grainsof the fusion welds.The �FSW resulted in better formability.

cknowledgments

The authors express their gratitude to Prof. Dr. D. Manolakosor providing his lab and to Mr. N. Melissas and Mr. G. Michas forarrying out numerous �FSW trials. Special thanks to Mrs. V. Pan-eleakou and Mr. M. Mariou, vital members of the team. Thanks to

r. A. Vazdirvanidis and Mr. A. Toulfatzis for their support in LOMnd mechanical testing respectively.

ppendix A. Supplementary data

Supplementary data associated with this article can be found,n the online version, at http://dx.doi.org/10.1016/j.jmatprotec.014.08.029.

cessing Technology 216 (2015) 133–139 139

References

Barcellona, A., Buffa, G., Fratini, L., Palmeri, D., 2006. On microstructural phenomenaoccurring in friction stir welding of aluminium alloys. J. Mater. Process. Technol.177, 340–343.

Béré, A., Chen, J., Hairie, A., Nouet, G., Paumier, E., 2000. Determination of the high c/aratio of hexagonal metals with a semi-empirical tight-binding method. Comput.Mater. Sci. 17, 249–254.

Buffa, G., Campanile, G., Fratini, L., Prisco, A., 2009. Friction stir welding of lap joints:influence of process parameters on the metallurgical and mechanical properties.Mater. Sci. Eng. A 519, 19–26.

Cavaliere, P., Campanile, G., Panella, F., Squillace, A., 2006. Effect of welding param-eters on mechanical and microstructural properties of AA6056 joints producedby friction stir welding. J. Mater. Process. Technol. 180, 263–270.

Galvão, I., Leal, R.M., Rodrigues, D.M., Loureiro, A., 2013. Influence of tool shouldergeometry on properties of friction stir welds in thin copper sheets. J. Mater.Process. Technol. 213, 129–135.

Gibson, B.T., Lammlein, D.H., Prater, T.J., Longhurst, W.R., Cox, C.D., Ballun, M.C.,Dharmaraj, K.J., Cook, G.E., Strauss, A.M., 2014. Friction stir welding: process,automation, and control. J. Manuf. Process. 16, 56–73.

Goulas, C., Papaefthymiou, S.A., Pantazopoulos, G., Panagopoulos, C., 2013. Weldingof thin titanium zinc sheets. In: 5th Hellenic Conference on Metallic Materials,Volos, Greece.

Gratecap, F., Racineux, G., Marya, S., 2008. A simple methodology to define conicaltool geometry and welding parameters in friction stir welding. Int. J. Mater.Form. 1, 143–158.

Groche, P., Wohletz, S., Brenneis, M., Pabst, C., Resch, F., 2014. Joining by forming—areview on joint mechanisms, applications and future trends. J. Mater. Process.Technol. 214, 1972–1994.

Khaled, T., 2005. An Outsider Look at Friction Stir Welding. Federal Aviation Admin-istration.

Kim, Y.G., Fujii, H., Tsumura, T., Komazaki, T., Nakata, K., 2006. Three defect typesin friction stir welding of aluminum die casting alloy. Mater. Sci. Eng. A 415,250–254.

Larsson, H., Karlsson, L., Svensson, L.E., 2000. Friction stir welding of aluminium alloy5083 and aluminium alloy 6082. In: Svetsaren, p. 54.

Liu, H.J., Hou, J.C., Guo, H., 2013. Effect of welding speed on microstructure andmechanical properties of self-reacting friction stir welded 6061-T6 aluminumalloy. Mater. Des. 50, 872–878.

Lorrain, O., Favier, V., Zahrouni, H., Lawrjaniec, D., 2010. Understanding the mate-rial flow path of friction stir welding process using unthreaded tools. J. Mater.Process. Technol. 210, 603–609.

Mishra, R.S., Ma, Z.Y., 2005. Friction stir welding and processing. Mater. Sci. Eng. R:Rep. 50, 1–78.

Nandan, R., Roy, G.G., Lienert, T.J., Debroy, T., 2007. Three-dimensional heat andmaterial flow during friction stir welding of mild steel. Acta Mater. 55, 883–895.

Nishihara, T., Nagasaka, Y., 2004. Development of micro-FSW. In: 5th InternationalFriction Stir Selding Symposium, Metz.

Pantazopoulos, G., Sampani, A., 2007. Analysis of a weld failure of a rolled Zn-alloystrip – a case study. Eng. Fail. Anal. 14, 642–651.

Patel, V.K., Bhole, S.D., Chen, D.L., 2014. Characterization of ultrasonic spot weldedjoints of Mg-to-galvanized and ungalvanized steel with a tin interlayer. J. Mater.Process. Technol. 214, 811–817.

Rodrigues, D.M., Loureiro, A., Leitao, C., Leal, R.M., Chaparro, B.M., Vilac a, P.,2009. Influence of friction stir welding parameters on the microstructural andmechanical properties of AA 6016-T4 thin welds. Mater. Des. 30, 1913–1921.

Sato, Y.S., Takauchi, H., Park, S.H.C., Kokawa, H., 2005. Characteristics of the kissing-bond in friction stir welded Al alloy 1050. Mater. Sci. Eng. A 405, 333–338.

Simoncini, M., Forcellese, A., 2012. Effect of the welding parameters and tool con-figuration on micro- and macro-mechanical properties of similar and dissimilarFSWed joints in AA5754 and AZ31 thin sheets. Mater. Des. 41, 50–60.

Squillace, A., De Fenzo, A., Giorleo, G., Bellucci, F., 2004. A comparison betweenFSW and TIG welding techniques: modifications of microstructure and pittingcorrosion resistance in AA 2024-T3 butt joints. J. Mater. Process. Technol. 152,97–105.

Sued, M.K., Pons, D., Lavroff, J., Wong, E.H., 2014. Design features for bobbin frictionstir welding tools: development of a conceptual model linking the underlyingphysics to the production process. Mater. Des. 54, 632–643.

Teh, N.J., Goddin, H., Whitaker, A., 2012. In: Institute, T.W. (Ed.), Developments inmicro applications of friction stir welding.

Thomas, W.M., Nicholas, E.D., Needham, J.D., Murch, M.G., Templesmith, P., Dawes,

C.S., 1992. Friction Stir Butt Welding.

Wang, D.-A., Chao, C.-W., Lin, P.-C., Uan, J.-Y., 2010. Mechanical characterization offriction stir spot microwelds. J. Mater. Process. Technol. 210, 1942–1948.

Yin, Y.H., Sun, N., North, T.H., Hu, S.S., 2010. Hook formation and mechanical proper-ties in AZ31 friction stir spot welds. J. Mater. Process. Technol. 210, 2062–2070.