You are on page 1of 8

Materials and Design 59 (2014) 344351

Contents lists available at ScienceDirect

Materials and Design

journal homepage:

Material ow and thermo-mechanical conditions during Friction

Stir Welding of polymers: Literature review, experimental results
and empirical analysis
F. Simes a,b, D.M. Rodrigues b,
Instituto Politcnico de Coimbra, ISEC, DEM, Portugal
CEMUC Department of Mechanical Engineering, University of Coimbra, Portugal

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

Article history: Meanwhile the thermo-mechanical conditions during Friction Stir Welding (FSW) of metals have already
Received 27 October 2013 been subject of extensive analysis and thoroughly discussed in literature, in which concerns the FSW of
Accepted 16 December 2013 polymers, the information regarding this subject is still very scarce. In this work, an analysis of the mate-
Available online 18 February 2014
rial ow and thermo-mechanical phenomena taking place during FSW of polymers is performed. The
analysis is based on a literature review and on the examination of friction stir welds, produced under var-
Keywords: ied FSW conditions, on polymethyl methacrylate (PMMA). Due to the high transparency of this polymer,
Friction Stir Welding
it was possible to analyse easily the morphological changes induced by the welding process on it. Results
of the weld morphologic analysis, of the residual stress elds in the different weld zones and of temper-
Morphology ature measurements during welding are shown, and its relation with welding conditions is discussed.
Residual stresses From the study it was possible to conclude that, due to the polymers rheological and physical properties,
Temperature the thermo-mechanical conditions during FSW are very different from that registered during welding of
metals, leading to completely different material ow mechanisms and weld defect morphologies.
2013 Elsevier Ltd. All rights reserved.

1. Introduction process involves only the plunge and retraction of the FSW tool,
being exclusively used in discontinuous lap joining. The works al-
Although Friction Stir Welding (FSW) process was initially ready performed on FSSW addressed the joining of high density
developed for the joining of Al-alloys [1], a large number of studies polyethylene (HDPE) [1921], polypropylene (PP) [2224] and
addressing similar and dissimilar FSW of other metallic materials polymethyl methacrylate (PMMA) [25]. The dissimilar joining of
were also already developed [2,3]. However, in which concerns PMMA to Acrylonitrilebutadienestyrene (ABS) was also investi-
the joining of polymers by FSW, the amount of work performed gated [26].However, despite the growing interest of the FSSW tech-
is still very small and most of the studies published were devel- nique, the following review will exclusively address the works
oped considering a very narrow range of polymers, more precisely, performed using the FSW technology, which is the one explored
polyethylene (PE) [410], high density polyethylene (HDPE) in the experimental program conducted in present work.
[4,11,12], polypropylene (PP) [1315], acrylonitrilebutadiene Bozkut [4] employed a conventional FSW tool, with a 18 mm
styrene (ABS) [16], polyethylene-terephthalate-glycol (PETG) [17] diameter shoulder and a 6 mm diameter pin, in the welding of high
and the polyamide Nylon 6 [18]. Those works can be divided in density polyethylene (HDPE). The tool rotation speed (x) and the
two categories, according to the type of FSW tool in use, i.e., works traverse speed (t) ranged between 1500 and 3000 rpm and 45
in which are used conventional FSW tools [47,1012,15,18], sim- 115 mm/min, respectively. This author measured temperature
ilar to that used in FSW of metals, and works in which stationary variations between 120 and 165 C during FSW of this polymer
shoulder tools are used [8,9,13,16,17]. which has a melting temperature of 132 C. The author observed
In literature it is also possible to nd a relatively large number voids and root cracks in the welds, which were responsible for
of works addressing the joining of polymers using Friction Stir Spot their poor tensile properties.
Welding (FSSW), which is a FSW related technique. The FSSW An investigation on the effects of critical process parameters on
FSW of polyethylene was also performed by Saeedy and Givi [10],
Corresponding author. Tel.: +351 239790700; fax: +351 239790701.
who tested rotational speeds and welding speeds ranging from
E-mail address: (D.M. Rodrigues).
1000 to 1800 rpm and 1220 mm/min, respectively, and tool tilt
0261-3069/ 2013 Elsevier Ltd. All rights reserved.
F. Simes, D.M. Rodrigues / Materials and Design 59 (2014) 344351 345

angles of 1 and 2. An optimum weld strength value of 75% of that to the low thermal conductivity of the polymer, the heat generated
of the base material was achieved in this study, but for only one during welding was concentrated in the weld nugget, resulting in
optimised set of welding parameters. These results showed the the formation of a pool of semi-molten polymer around the pin.
important inuence of process parameters on welding results. The size of the molten pool was increased by increasing the rota-
The same was concluded by Rezgui et al. [12] who used ANOVA tion speed, resulting in improved tool stirring action and weld
in optimising FSW process parameters for the joining of HDPE. quality. However, when the shoulder temperature was set to
Arici and Selale [5] and Arici and Sinmaz [6] also performed 80 C and 110 C, weld material degradation was observed. Incom-
FSW of PE using conventional FSW tools. In order to avoid root plete tool penetration and lack of connection between the weld
crack defects, such as that reported by Bozkurt [4], these authors nugget and the adjacent base material were the main defects re-
performed double pass butt welding and used rotation speeds up ported. Defect incidence was observed to be reduced when the
to 1000 rpm and traverse speeds up to 60 mm/min. They reported heat input was sufcient to melt the whole nugget material.
insufcient heat generation for rotation speeds of 600 and 800 rpm Rezgui et al. [9] also used a stationary shoulder tool to weld
and suitable welding conditions at 1000 rpm. Those authors also HDPE, but with a shoulder made of wood, and three cylindrical
observed that, depending on the amount of heat input, some mate- threaded pins (M10, M12 and M14). The Taguchi method was used
rial was expelled from the joining area during welding. In which by the authors to optimise the FSW parameters. Rotation speeds
concerns the double pass FSW, Arici and Selale [5] also studied from 900 to 1700 rpm and traverse speeds from 16 to 44 mm/
the effect of tool tilt angle on FSW of PE observing that the thick- min were selected. They reported that welding took place by melt-
ness of the welds, as well as the tensile strength, decreased with ing the polymer and that the low thermal conductivity of this
increasing tool tilt angle. material was responsible for the formation of a small size HAZ.
Aydin [7] used a tool and process parameters similar to that The tensile behaviour of the welds was analysed, being referred
used by Arici and Selale [5], but performed single pass butt welding that the presence of discontinuities in the weld caused specimens
with pre-heating of the base material. They report that pre-heating failure for very small strain values.
insures adequate heating even at low rotation speeds. Welds sur- Kiss and Czigny [13,17] performed FSW using a non-heated
face morphology and global properties were also appointed to be PTFE stationary shoulder and different pin diameters and FSW
improved by pre-heating. The use of pre-heating in FSW of poly- parameters. In order to analyse the internal stresses, using stress
ethylene, with conventional rotating tools, was also studied by optics, as well as to visualize the weld nugget and HAZ, those
Squeo and Quadrini [11], who performed pre-heating of the tool, authors changed the base material from PP (semicrystalline), to
as well as pre-heating of the plates to be joined. Two different PETG, which displays a high level of transparency and low crystal-
pin diameters were tested and several combinations of tool tra- lization tendency. For the PP welds, no signicant differences were
verse and rotational speeds. As the previous authors, Squeo and reported by the authors between the advancing and retreating
Quadrini also conclude that pre-heating of the tool enabled sides of the welds. Spherulitic structures, similar to that of the base
improving welds quality. However, according to the authors, material, were observed in the nugget, which they attributed to the
achieving optimum FSW conditions still needed further optimisa- slow cooling rate of the weld. Concerning to the PETG welds, it was
tion of process parameters. reported the presence of a sharp discontinuity between the weld
Panneerselvam and Lenin [15] used conventional FSW tools to and the base material. The colour fringes, registered in the stress
analyse the inuence of pin geometry on FSW results, but using optics photographs, were attributed by the authors to the molecu-
PP as base material. They tested square, triangular, threaded and lar orientation induced by tool rotation and to the residual stresses
tapered pin geometries, in a range of rotation speeds varying from resulting from heating.
1500 to 2250 rpm and traverse speeds varying from 30 to 60 mm/ Finally, Bagheri et al. [16], used a stationary shoulder tool with
min. According to the pin geometry and welding parameters, sev- an heated shoe (T = 50; 80; 100 C) and a non-conventional
eral defects were reported, such as lack of consolidation, porosities, threaded pin, 10 mm in diameter and 7 mm in pitch, to weld
cavities and inclusions. Poor joining was also reported to occur ABS. The rotational speeds used were 800, 1250 and 1600 rpm
mainly at the retreating side of the welds. The authors claim that and the traverse speeds were 20, 40 and 80 mm/min. It was con-
the threaded tool pin prole generated the best welding results. cluded that increasing the rotational speed, conducted to an
Finally, Panneerselvam and Lenin [18] conducted FSW tests, important increase in temperature that lead to burning of the work
using a conventional rotating tool, as in previous works, but testing piece and squeezing of melted material from the weld. On the
a different base material, the Nylon 6 polyamide. These authors other hand, for low rotation speeds, inadequate material mixing
used a specially designed left hand threaded tool pin prole, oper- and lack of joining between the weld and the base material were
ated at a rotational speed of 1000 rpm and at a welding speed of observed. In the same way, at the highest travel speeds, the weld
10 mm/min, to analyse the inuence of tool rotation direction on line was observed to be very discontinuous. Those results were im-
weld quality. They found that the FSW joints fabricated with the proved by heating the shoe. The maximum tensile strength welds
tool operating with counter clockwise rotation were free of defects were produced with the shoe heated to 100 C. However, despite
and had improved strength, relative to the welds performed with the optimised welding conditions, the tensile samples were ob-
clockwise tool rotation. They attributed the poor quality of the served to fail at the retreating side of the welds, were a discontinu-
welds obtained with the clockwise rotating tool to the expulsion ity was present. The maximum tensile strength was reported for
of the stirred material from the weld seam. the welds performed with the maximum heat input, i.e., maximum
In order to avoid some of the problems registered in FSW with tool rotation speed and shoe temperature and the lowest traverse
conventional tools, such as the squeezing of the melted polymer speed.
from the weld nugget, some researchers developed modied tools, From the literature review on FSW it becomes apparent that in
with a stationary shoulder, sometimes called shoe. Mostafapour order to obtain good quality welds, stationary shoulder tools
and Azarsa [8], for example, used a stationary aluminium shoulder, should be used [8,9,13,16,17]. Using this type of tools, the rotating
heated by electric resistance, to weld high-density polyethylene pin generates the friction heat necessary to melt and weld the
(HDPE). Shoulder temperatures were adjusted to 80, 110 and abutting surfaces by solidication. The stationary shoulder also
140 C. The pin was threaded, with 10 mm diameter. The rotation avoids the plastic material, swirling around the pin, of being ex-
speeds tested ranged between 1000 and 1600 rpm and the traverse pelled from the weld seam [13]. High shoulder temperature and
speeds between 10 and 40 mm/min. According to the authors, due high tool rotational speeds, conducting to base materials melting
346 F. Simes, D.M. Rodrigues / Materials and Design 59 (2014) 344351

Table 1
Thermal properties of PMMA cast sheet [45].

Glass temperature (C) Maximum service temperature (C) Minimum service temperature (C) Thermal conductivity (W/m C)
96104 80105 75 to 65 0.170.25

5 mm cylindrical pin. Tool 2 is a 35 mm diameter shoulder tool,

Tool 1 Tool 2 with a at base on the outside and a conical geometry in the centre,
and a 6 mm cylindrical pin. The inner conical cavity in tool 2 shoul-
der was used in order to avoid the material of being expelled from
the joint during welding. The pin length, of 6 mm, was the same for
both tools. By using a pin length lower than the PMMA plates
Fig. 1. Schematic representation of FSW tool geometries.
thickness it was possible to have a full picture of the heat and
material transfer under the pin during welding.
The same advancing speed, of 25 mm/min, and rotation speed,
and slow cooling of the welds, are indicated as important factors in
of 700 rpm, were used for both tools. During welding, the temper-
achieving successful welding [8,16]. However, since melting is in-
ature evolution at the surface of the weld beads was registered
volved in materials joining, the FSW of polymers, unlike the FSW
using a infrared camera from Flir, model ThermaCAM E300, that
of metals, cannot be considered a solid state welding process. An-
produces fully radiometric images, allowing to measure the tem-
other important issue is that in spite of the well-established, and
perature of objects. Transverse and longitudinal samples were
patented, tools and procedures for the FSW of polymers, no sys-
cut, as schematized in Fig. 2, from several welds in order to enable
tematic knowledge on material ow mechanisms and thermo-
a deep analysis of the welds morphology. Due to the high transpar-
mechanical conditions during FSW, which implies analysing the
ency of the PMMA, a qualitative photoelastic analysis of the resid-
temperature distribution and stressstrain evolution during weld-
ual stress elds in the welds was accomplished by using a
ing, as well as its relation with processing parameters and defect
polarised light device.
formation, can be found in literature. In current work, the material
ow and thermo-mechanical conditions during FSW of polymers
are analysed by comparison with the metalworking ow-parti-
tioned model proposed by Arbegast [27]. 3. Weld morphology analysis
Despite the absence of publications on material ow during
FSW of polymers, several works can be found in literature analys- In order to characterise the main features of the polymeric
ing material ow during similar [2837] and dissimilar [3844] welds, in Fig. 3 is show an image of the surface of a weld, produced
FSW of metals. Those works describe the different methodologies with tool 1. In this gure it is possible to clearly distinguish the pin
used to analyse the ow mechanisms (experimental and numerical inuence zone, in the centre, surrounded by a well-dened shoul-
simulation procedures) and address a varied range of welding con- der inuence zone. Meanwhile the pin inuence zone displays the
ditions, i.e., different base materials, tool geometries and process characteristic FSW arc shaped striations, resulting from the peri-
parameters. The Arbegast ow partitioned model [27], cleverly odic deposition of material, by the pin, at the trailing side of the
synthetizes the main ow mechanisms identied in that literature, tool, the shoulder inuence zone displays a very smooth surface
and for that reason, was selected in current work for performing a with no evidence of material stirring, into the pin inuence zone,
comparative analysis on material ow in FSW of polymers and during welding. Tool 2 surface displayed similar surface
metals. morphology.
In Figs. 4 and 5 are shown two (Figs. 4a and 5a) and three
(Figs. 4b and 5b) dimensional images of tool 1 and tool 2 transverse
2. Experimental procedure weld samples, respectively. In both gures are discernible impor-
tant differences in morphology between the advancing and retreat-
Bead on plate friction stir welds on 10 mm thick PMMA plates ing sides of both welds. Meanwhile the advancing side material
were made in a CNC machine. In Table 1 are shown the thermal displays the same full transparency features of the original base
properties of the base material [45] which are considered relevant material, with only very small and dispersed discontinuities, indi-
in present analysis. It is interesting to stress that the heating and cating an almost perfect merging between the weld and the
cooling of the metallic and polymeric materials are substantially adjacent materials, the retreating side material is fully non-trans-
different, due to the large difference in thermal conductivity. For parent, displaying a straight interface between the pin inuence
instance, the PMMA thermal conductivity presented in Table 1 zone and the adjacent material. In Fig. 6, where are shown images
(0.170.25 W/m C) is only 0.13% of aluminium alloys thermal con- of tool 1 longitudinal samples, which illustrate individually the
ductivity (around 135 W/m C). retreating (Fig. 6a) and advancing (Fig. 6b) sides of the weld, it is
By performing bead on plate welding it was possible to simplify possible to conrm that those characteristics are repeated consis-
the experimental work, excluding simultaneously the inuence of tently along the weld. Optical microscopy images of the non-trans-
plates clamping and positioning on welding results. As shown in parent retreating side material in Fig. 6a, are shown in Fig. 7
Fig. 1, two different left-hand thread tools were used. Tool 1 is a showing that this feature results from the presence of a large num-
conventional FSW tool, with a 15 mm conical shoulder and a ber of macro and micro voids in its structure.

Transverse cut Longitutinal cut

Fig. 2. Weld sampling scheme.

F. Simes, D.M. Rodrigues / Materials and Design 59 (2014) 344351 347

zone, are perceptible in both gures, as usually reported for the

friction stir welds in metals. So, the straight interface at the retreat-
ing side of both welds indicates that the thermo-mechanical af-
fected zone (TMAZ), for the polymeric materials, is restricted to
the pin inuence zone. The backlight images also display, for both
welds, the presence of a heat affected zone (HAZ) surrounding the
TMAZ. For both welds, the HAZs displays a perfectly regular and
well dened geometry, extending from the under shoulder area
to the bottom of the pin. Unlike the TMAZ, the HAZ displays per-
fectly symmetrical features for both welds, reecting the symme-
try in heat production and dissipation under the shoulder.
Finally, in Fig. 8c and d are shown polarised light images of the
welds in Fig. 8a and b, revealing the residual stress elds induced
by the welding process. Like the TMAZ morphology, the residual
stress eld in its neighbourhood, is also not symmetric for both
welds, being possible to notice a stress concentration at the
advancing side of the TMAZ for both welds. In spite of this, for
3 mm the tool 2 weld, the residual stress elds discernible in the HAZ lo-
cated at the outer side of the shoulder inuence zone have sym-
metrical features.

Fig. 3. Picture of the surface of a weld produced with tool 1.

4. Analysis of the thermo-mechanical conditions during FSW of

In Fig. 8a and b are now shown backlight images of the same According to Arbegast studies on FSW of metals [27], and as is
welds, which enable to distinguish not only the same features al- schematized in Fig. 9a, material processing during welding pro-
ready addressed when analysing Figs. 46, i.e., a straight interface gresses sequentially from a pre-heating phase, at the leading side
between the pin inuence zone at the retreating side of the weld, of the tool, followed by the initial deformation, extrusion and forg-
with a portion of opaque material next to it, but also a less pro- ing of the stirred material under the tool, and nally, by its cool-
nounced interface at the advancing side of the weld, which was down at the trailing side of the tool. Due to the very low thermal
not perceptible in the previous gures. No signs of material being conductivity of the polymers, the pre-heating and initial deforma-
stirred by the shoulder, across that interface, into pin inuence tion areas in front of the tool are shrunk in a thin layer, as is shown

3 mm

(a) (b)
Fig. 4. Two (a) and three (b) dimensional images of tool 1 weld.

3 mm

(a) (b)
Fig. 5. Two (a) and three (b) dimensional images of tool 2 weld.
348 F. Simes, D.M. Rodrigues / Materials and Design 59 (2014) 344351

Fig. 6. Longitudinal tool 1 samples displaying the retreating (a) and the advancing (b) side.

II) material deformation zones form inside the pin inuence zone.
Due to tool rotation, Zone II material converges to Zone I, by ow-
ing around the pin and downward it, through Zone IV. Zone III
material, in the shoulder inuence zone, is also dragged across
the top toward the advancing side Zone I. The size and shape of
each material ow zone represented in Fig. 11a is related to the
temperature eld, the constitutive properties of the materials
being joined, the strain rate and the tool geometry. Arbegast re-
lated inadequacies in this ow partitioning to specic FSW features
and defect types.
Coupling the weld morphological analysis performed in the pre-
vious item, for the PMMA welds, with the Arbegast ow-parti-
tioned deformation zone model for welds in metals, it becomes
possible to access the thermo-mechanical conditions during FSW
500 m
of polymers, as well as to understand the mechanisms of defect
formation for these materials. Actually, comparing the cross sec-
tion of the polymer weld in Fig. 11b, with Arbegast model in
Fig. 7. Optical microscopy image of the retreating side material, showing macro
and micro-voids developed during FSW. Fig. 11a, it becomes apparent that Zone III is absent from the poly-
meric welds. Effectively, the interface of the pin inuence zone re-
mains straight, and parallel to the pin prole, from the bottom to
in Fig. 9b, and the nal cool down of the weld is very slow as the top of the welds, which means that no material ow could have
shown in Fig. 10a and b, where the temperature elds registered took place across it. This assumption is also in accordance with
at the surface of the plate, with an infrared camera, during FSW, Fig. 3 where it is possible to distinguish, clearly delimited, the
are represented. According to Fig. 10b, where are shown tempera- shoulder and pin inuence zones at the top of the weld. In the
tures proles at the trailing side of the tool, the temperature there same way, the straight line at the bottom of the TMAZ, discernible
remain high even after a signicant tool displacement. in Fig. 8a and b, indicates that Zone IV is also absent from the fric-
Arbegast [27] also stated that, for metals, the material ow at tion stir welds in PMMA.
the trailing side of the tool is partitioned in four distinct deforma- Analysing again the temperature distribution at the trailing side
tion zones, distributed around the pin and beneath the shoulder, as of the tool, represented in Fig. 10, and taking into account Table 1,
it is shown in Fig. 11a. The advancing (Zone I) and retreating (Zone it is possible to conclude that, during FSW, the temperatures under

5 mm 5 mm

(a) (b)

retreating advancing retreating side advancing side

side side

(c) (d)
Fig. 8. Backlight and polarised light images of tool 1 (a and c) and tool 2 (b and d) welds.
F. Simes, D.M. Rodrigues / Materials and Design 59 (2014) 344351 349

Fig. 9. Schematic representation of Arbegast model [27] for FSW of metals (a) and longitudinal sections of tool 1 sample (b).

137.6 C


Temperature [C]

2 1

Infrared width (IW)

20.7 C

(a) (b)
Fig. 10. Tool and material image captured with an infrared camera after FSW (a) and temperature elds surface and temperature prole at different weld positions (b).

Advancing Side Retreating Side

Zone I Zone II
(a) (b)
Fig. 11. Representation of deformation zones and ow material for FSW of metals according Arbegast model (a) and transverse sections of tool 1 sample (b).

the shoulder were higher than 135 C, which means that polymer [16], the burning and ejection of the polymer from the weld bead
melting should have took place in contact with the shoulder. Most can be avoided by an appropriate control of the rotation speed.
of this melted material is squeezed from the under shoulder area Analysing again Fig. 11b it is possible to see that the non-trans-
under the centrifugal action of the rotating tool, instead of entering parent material at the retreating side of the welds (Zone II) displays
into Zone I, which explains the absence of Zone III from the PMMA features which indicate that the material ows from it to the
weld proles. Actually, it is possible to see some lack of material at advancing side of the tool (Zone I), which is in accordance with
the top of all welds. As observed in current work, Panneerselvam Arbegast model. The magnication view of the non-transparent
and Lenin [18], Arici and Selale [5] and Bagheri et al. [16] also re- material of Zone II, shown in Fig. 7, also shows that this material
ported material ejection from the weld seam during FSW. Accord- has a deeply non-homogeneous structure, with large voids inside
ing to Panneerselvam and Lenin [18], in FSW with conventional of it. Assuming that material melting occurs not only in the under
shoulder tools, this phenomenon may be avoided by welding with shoulder area, but also in the pin inuence zone, it is reasonable to
a left hand threaded pin operating in counter clockwise direction. accept that melted material is also squeezed out of the weld from
In FSW with stationary shoulder tools, according to Bagheri et al. this zone. This will occur more easily at the retreating side of the
350 F. Simes, D.M. Rodrigues / Materials and Design 59 (2014) 344351

Fig. 12. Optical microscopy image of the sample after welding (a) and image of the weld one month after welding (b) showing additional defects.

tool, were the tangential ow of the material is opposite to the tra- Acknowledgements
verse speed of the tool. The squeeze of melted material to the un-
der shoulder zone III, by one side, and to the advancing side Zone I, This research is sponsored by FEDER funds through the program
by the other side, originate the severe discontinuities reported in COMPETE Programa Operacional Factores de Competitividade
Fig. 7. At the advancing side of the tool, where the partially melted and by national funds through FCT Fundao para a Cincia e a
material coming from the retreating side is extruded against the Tecnologia under the Project PEst-C/EME/UI0285/2013.
base material side wall, the pressure generated by the ow of
material is sufcient to consolidate it, giving rise to the almost con-
tinuous joint interface shown in Figs. 46. It is important to stress
that in FSW of metals, the advancing side of the weld is the prefer- [1] Threadgill PL, Leonard AJ, Shercliff HR, Withers PJ. Friction stir welding of
ential location for defect formation [4650], which is opposite to aluminium alloys. Int Mater Rev 2009;52(2):4993.
that registered in current work for the polymer welds. The forma- [2] DebRoy T, Bhadeshia HKDH. Sci Technol Weld Joi 2010;15(4):26670.
[3] am G. Friction stir welded structural materials: beyond Al-alloys. Int Mater
tion of important discontinuities at the retreating side of the poly- Rev 2011;56:148.
mer friction stir welds can also be observed in most of the works [4] Bozkurt Y. The optimization of friction stir welding process parameters to
published by other authors, but is only explicitly mentioned in ref- achieve maximum tensile strength in polyethylene sheets. Mater Des
erences [15,16]. [5] Arici A, Selale S. Effects of tool tilt angle on tensile strength and fracture
An important question is whether the retreating side disconti- locations of friction stir welding of polyethylene. Sci Technol Weld Joi
nuities can be avoided by performing the welds at lower tempera- 2007;12:5369.
[6] Arici A, Sinmaz T. Effects of double passes of the tool on friction stir welding of
tures, avoiding in that way material squeezing due to melting. In polyethylene. J Mater Sci 2005;40:33136.
Fig. 12a is shown an image of a cross-section of the initial part of [7] Aydin M. Effects of welding parameters and pre-heating on the friction stir
a tool 2 weld. Since no dwell time was used, this part of the weld, welding of UHMW-polyethylene. Polym Plast Technol Eng 2010;49:595601.
[8] Mostafapour A, Azarsa E. A study on the role of processing parameters in
performed with the tool still cold, can be considered a low tem-
joining polyethylene sheets via heat assisted friction stir welding:
perature weld. As it is possible to see from the picture, no impor- investigating microstructure, tensile and exural properties. Int J Phys Sci
tant discontinuities can be reported at both the advancing and 2012;7(4):64754.
[9] Rezgui A, Ayadi M, Cherouat A, Hamrouni K, Zghal A, Bejaoui S. Application of
retreating sides of this part of the weld, which supports the
Taguchi approach to optimize friction stir welding parameters of polyethylene.
hypothesis that the defects results from melting and squeezing of EPJ Web Conf 2010;6:18.
material from the weld bead. However, in Fig. 12b is shown an- [10] Saeedy S, Givi M. Investigation of the effects of critical process parameters of
other image of the same sample, but acquired one month after cut- friction stir welding of polyethylene. Proc Inst Mech Eng, Part B: J Eng Manuf
ting the sample. In this image is now possible to see, two thin [11] Squeo E, Quadrini F. Friction stir welding of polyethylene sheets. The Annals of
cracks, located at the advancing and retreating sides of the weld. DUNA REA DE JOS University of Galati Fascicle V. Technologies in Machine
Those cracks, which result from the residual stress eld in the Building; 2009. ISSN 1221-4566.
[12] Rezgui M, Trabelsi A, Ayadi A, Hamrouni K. Optimization of friction stir
weld, revealed by the polarised light image of Fig. 8d, show that welding process of high density polyethylene. Int J Prod Qual Eng
residual stresses can be very important in FSW of polymers. 2011;2(1):5561.
[13] Kiss Z, Czigny T. Microscopic analysis of the morphology of seams in friction
stir welded polypropylene. Express Polym Lett 2012;6:5462.
5. Conclusions [14] Kiss Z, Czigny T. Applicability of friction stir welding in polymeric materials.
Periodica Polytech, Mech Eng 2007;51:158.
[15] Panneerselvam K, Lenin K. Effects and defects of the polypropylene plate for
From current analysis important differences in material ow different parameters in friction stir welding process. Int J Res Eng Technol
and temperature distribution between metallic and polymeric 2013;2(2):14352.
materials, during FSW, were identied and critically analysed [16] Bagheri A, Azdast T, Doniavi A. An experimental study on mechanical
properties of friction stir welded ABS sheets. Mater Des 2013;43:4029.
based on Arbegast ow-partitioned deformation zone model. The [17] Kiss Z, Czigny T. Effect of welding parameters on the heat affected zone and
differences in thermo-mechanical conditions are explained based the mechanical properties of friction stir welded poly(ethylene-terephthalate-
on the strong dissimilarities in weld morphology and defect loca- glycol). J Appl Polym Sci 2012;125:22318.
[18] Panneerselvam K, Lenin K. Joining of Nylon 6 plate by friction stir welding
tion. Meanwhile for metals the welds result from the conuence process using threaded pin prole. Mater Des 2014;53:3027.
of the pin and shoulder driven ow volumes in the TMAZ, for poly- [19] Bilici M, Ykler A. Inuence of tool geometry and process parameters on
mers, only the pin driven ow volume enters the TMAZ, since the macrostructure and static strength in friction stir spot welded polyethylene
sheets. Mater Des 2012;33:14552.
shoulder driven material, melted during welding, is squeezed from
[20] Bilici M, Ykler A, Kurtulmus M. The optimization of welding parameters for
the joint. The formation of important discontinuities, at the friction stir spot welding of high density polyethylene sheets. Mater Des
retreating side of the welds, indicated in literature as one of the 2011;32:40749.
main weldability problems for polymers, is also explained by the [21] Bilici M, Ykler A. Effects of welding parameters on friction stir spot welding of
high density polyethylene sheets. Mater Des 2012;33:54550.
squeezing of melted material from both the shoulder and pin dri- [22] Bilici M. Effect of tool geometry on friction stir spot welding of polypropylene
ven ow volumes. sheets. Express Polym Lett 2012;6(10):80513.
F. Simes, D.M. Rodrigues / Materials and Design 59 (2014) 344351 351

[23] Mert S, Arici A. Design of optimal joining for friction stir spot welding of [38] Li Y, Murr L, McClure J. Solid-state ow visualization in the friction-stir
polypropylene sheets. Sci Technol Weld Joi 2011;16(6):5227. welding of 2024 Al to 6061 Al. Scr Mater 1999;40:10416.
[24] Arici A. Friction stir spot welding of polypropylene. J Reinf Plast Compos [39] Li Y, Murr L, McClure J. Flow visualization and residual microstructures
2008;27(18):20014. associated with the friction-stir welding of 2024 aluminum to 6061
[25] Oliveira P, Amancio-Filho S, Santos J, Hage E. Preliminary study on the aluminum. Mater Sci Eng, A 1999;271:21323.
feasibility of friction spot welding in PMMA. Mater Lett 2010;64:2098101. [40] Guerra M, Schmidt C, McClure J, Murr L, Nunes A. Flow patterns during friction
[26] Dashatan S, Azdast T, Ahmadi S, Bagheri A. Friction stir spot welding of stir welding. Mater Charact 2003;49:95101.
dissimilar polymethyl methacrylate and acrylonitrile butadiene styrene [41] Leal R, Leito C, Loureiro A, Rodrigues D, Vilaa P. Material Flow in
sheets. Mater Des 2013;45:13541. heterogeneous friction stir welding of thin aluminium sheets: effect of
[27] Arbegast William J. A ow-partitioned deformation zone model for defect shoulder geometry. Mater Sci Eng, A 2008;498(12):38491.
formation during friction stir welding. Scr Mater 2008;58:3726. [42] Galvo I, Leal R, Loureiro A, Rodrigues D. Material ow in heterogeneous
[28] Colligan K. Material ow behavior during friction stir welding of aluminum. friction stir welding of aluminium and copper thin sheets. Sci Technol Weld Joi
Weld J 1999;78(7):229s37s. 2010;15(8):65460.
[29] Seidel T, Reynolds A. Visualization of the material ow in AA2195 friction-stir [43] Esmaeili A, Besharati Givi M, Zareie Rajani H. Experimental investigation of
welds using a marker insert technique. Metall Mater Trans A material ow and welding defects in friction stir welding of aluminumbrass.
2001;32A:287984. Mater Manuf Processes 2012;27(12):14028.
[30] Guerra M, Schmidt C, McClure JC, Murr L, Nunes A. Flow patterns during [44] Pourahmad P, Abbasi M. Materials ow and phase transformation in friction
friction stir welding. Mater Charact 2003;49:95101. stir welding of Al 6013/Mg. Trans Nonferr Metal Soc China
[31] Fratini L, Buffa G, Palmeri D, Hua J, Shivpuri R. Material ow in FSW of AA7075- 2013;23(5):125361.
T6 butt joints: numerical simulations and experimental verications. Sci [45] Ashby M. CES Edupack. Granta Design; 2009.
Technol Weld Joi 2006;11:41221. [46] Leito C, Louro R, Rodrigues DM. Analysis of high temperature plastic
[32] Schmidt H, Dickerson T, Hattel J. Material ow in butt friction stir welds in behaviour and its relation with weldability in friction stir welding for
AA2024-T3. Acta Mater 2006;54:1199209. aluminium alloys AA5083-H111 and AA6082-T6. Mater Des 2012;37:4029.
[33] Zhao Y, Lin S, Qu F, Wu L. Inuence of pin geometry on material ow in friction [47] Kim Y, Fujii H, Tsumura T, Komazaki T, Nakata K. Three defect types in friction
stir welding process. Mater Sci Technol 2006;22:4550. stir welding of aluminum die casting alloy. Mater Sci Eng, A 2006;415:2504.
[34] Leito C, Leal R, Rodrigues D, Vilaa P, Loureiro A. Material ow in friction stir [48] Chen H, Yan K, Lin T, Chen S, Jiang C, Zhao Y. The investigation of typical
welding. Microsc Microanal 2008;14:8790. welding defects for 5456 aluminum alloy friction stir welds. Mater Sci Eng, A
[35] Chen Z, Pasang T, Qi Y. Shear ow and formation of nugget zone during friction 2006;433:649.
stir welding of aluminium alloy 5083-O. Mater Sci Eng, A 2008;474:3126. [49] Iordachescu M, Iordachescu D, Ocana J, Vilaa P, Scitelnicu E. Characteristic
[36] Lorrain O, Favier V, Zahrouni H, Lawrjaniec D. Understanding the material ow aws in aluminium alloys joints. Metal Int 2009;12:1358.
path of friction stir welding process using unthreaded tools. J Mater Process [50] Kumar K, Kailas S. The role of friction stir welding tool on material ow and
Technol 2010;210(4):6039. weld formation. Mater Sci Eng, A 2008;485:36774.
[37] Fonda R, Reynolds A, Feng C, Knipling K, Rowenhorst D. Material ow in
friction stir welds. Metall Mater Trans A 2013;44(1):33744.