JMRTEC-470; No.

of Pages 11
ARTICLE IN PRESS
j m a t e r r e s t e c h n o l . 2 0 1 8;x x x(x x):xxx–xxx

Available online at www.sciencedirect.com

www.jmrt.com.br

Original Article

Effect of welding speed on friction stir welds of GL

E36 shipbuilding steel

Pedro Henrique Costa Pereira da Cunha a , Guilherme Vieira Braga Lemos b,∗ ,
Luciano Bergmann c , Afonso Reguly b , Jorge Fernandez dos Santos c ,
Ricardo Reppold Marinho d , Marcelo Torres Piza Paes d
a Pós-Graduação em Engenharia Mecânica - PPMec. Escola de Engenharia, Universidade Federal do Rio Grande (FURG), Rio Grande, RS,
Brazil
b Laboratório de Metalurgia Física (LAMEF), Programa de Pós-Graduação em Engenharia de Minas, Metalúrgica e de Materiais

(PPGE3M), Universidade Federal do Rio Grande do Sul (UFRGS), Porto Alegre, RS, Brazil
c Helmholtz-Zentrum Geesthacht (HZG), Institute of Materials Research, Materials Mechanics, Solid-State Joining Processes (WMP),

Geesthacht, Germany
d Centro Pesquisa da Petrobras – CENPES – Petróleo Brasileiro S.A., Rio de Janeiro, RJ, Brazil

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

Article history: The aim of this study was to characterize mechanical and microstructural characteristics of
Received 14 November 2017 friction-stir-welded GL E36 shipbuilding steel. The tool rotational speed was kept constant at
Accepted 12 July 2018 500 rpm and different welding speeds were used (1, 2 and 3 mm/s) to achieve different heat
Available online xxx inputs. Thermal cycles were monitored by thermocouples placed near the weld face. The
welded joints showed a very good esthetics and homogeneous surface quality that indicate
Keywords: a stability of the considered process parameters. The welded joints properties were ana-
Friction stir welding lyzed by metallography and mechanical tests such as microhardness, tensile and bending.
Shipbuilding steel Macrostructural observations were done at the beginning, middle and ending of the welded
Heat input length. In addition, radiographic inspection was carried out. The pcBN tool exhibited good
Microstructure wear behavior even after welding around 8.5 m where no apparent loss in dimensions and
Mechanical properties geometrical features of the probe and shoulder were found. The macrographs displayed dif-
ferent microstructural features and material flow pattern among the heat inputs achieved. A
large microstructure gradient was observed, especially within the stirred zone. All the tensile
samples broke at the base material showing that the joints achieved higher strength. Micro-
hardness peaks of about 400 HV were also encountered in all the joints. Finally, for welding
speeds of 2–3 mm/s the thermocouples presented the most uniform thermal profiles.
© 2018 Published by Elsevier Editora Ltda. on behalf of Brazilian Metallurgical, Materials
and Mining Association. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction

Friction stir welding (FSW) has expanded rapidly since its

development in 1991 and has found applications in a wide
∗
Corresponding author. variety of industries, including aerospace, automotive, rail-
E-mail: guilherme.lemos@ufrgs.br (G.V. Lemos). way, and maritime [1]. The process had a great success when
https://doi.org/10.1016/j.jmrt.2018.07.014
2238-7854/© 2018 Published by Elsevier Editora Ltda. on behalf of Brazilian Metallurgical, Materials and Mining Association. This is an
open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Please cite this article in press as: Cunha PH, et al. Effect of welding speed on friction stir welds of GL E36 shipbuilding steel. J Mater Res Technol.
2018. https://doi.org/10.1016/j.jmrt.2018.07.014
JMRTEC-470; No. of Pages 11
ARTICLE IN PRESS
2 j m a t e r r e s t e c h n o l . 2 0 1 8;x x x(x x):xxx–xxx

applied to aluminum alloys, and this fact was a driving force

Table 1 – Chemical composition of the shipbuilding steel
to stimulated exploration of its applicability to other materials used.
such as steel. Steels represent by far the greatest opportu-
Grade
nity for any new process given their undisputed prominence
in structural applications due to the fact that the material is GL E36 C Mn Si P Cu Ni Cr Mo Al N Nb
strong, versatile, cost effective and reliable [2]. Review publi-
0.170 1.400 0.390 0.013 0.030 0.020 0.060 0.006 0.027 0.007 0.025
cations for FSW are available, Gibson et al. [1], Mishra et al. [3],
Nandan et al. [4], Çam [5], Rai et al. [6] and Padhy et al. [7] made
comprehensive surveys for FSW technique. Moreover, Pradeep tensile, charpy and microhardness tests as well as microstruc-
[8], Venkatesh Kannan et al. [9] and Liu et al. [10] published ture characterization [25].
specific reviews concerning FSW applied on steels. According to this scenario, the present study intends to ful-
Cui et al. welded the following carbon steels: IF steel, S12C, fill the lack of technical and scientific knowledge concerning
S20C, S35C, S50C under various welding conditions, and the friction stir welding application to GL E36 shipbuilding steel
mechanical properties and the microstructure were evaluated and provide, therefore, mechanical and microstructure fea-
[11]. Ghosh et al. reported results for plain carbon steel (0.44% tures of the welded joints as a function of welding speed.
carbon content) under variable rotational and welding speeds
[12]. Matsushita et al. performed FSW with single-sided one-
pass butt welding using 12-mm thick structural steel grade 2. Methods
plates with 400 N mm−2 of tensile strength. The microstruc-
tural and mechanical properties such as hardness, tensile The base material used for this study was the shipbuilding
strength, and Charpy V-notch toughness were investigated steel grade GL E36 supplied with a normalized heat treat-
[13]. In addition, a special issue of Science and Technology of ment condition. Plates of 1200 mmL × 200 mmW × 6 mmT were
Welding and Joining (Vol. 14, No. 3) is available only with papers processed according to the rolling direction using butt joint
regarding FSW applied on carbon and stainless steels. configuration. The chemical composition given by the supplier
The potential for friction stir welding of steel is high, but is presented in Table 1.
up to now only for niche applications since for most applica- Fig. 1 shows the microstructure of this material; it consists
tions, traditional arc welding processes are well established of very refined ferrite and pearlite with banded structure in the
and cost effective. The FSW process is being actively devel- rolling direction with 180 vickers hardness. The mechanical
oped for various shipbuilding activities (most of these are properties informed by the supplier inspection certificate are:
military), and for other critical applications such as welding 436 MPa for yield strength, 564 MPa for tensile strength, elon-
of high strength pipelines, where conventional arc weld- gation of 24% and toughness of 87 J at −40 ◦ C. The welds were
ing consumables are not adequately developed. Furthermore, processed at Helmholtz-Zentrum Geesthacht using a friction
the potential for lower distortion has also attracted interest stir welding portal equipment using a force control process. A
from shipyards was reported by Konkol et al. [14]. In addi- threaded polycrystalline cubic boron nitride (pcBN) tool from
tion to these advantages, the problem of hydrogen cracking MegaStirTM with W-Re binder was used, commercially sup-
is not expected to be encountered in FSW due to the elimi- plied with the specification Q70. The tool angle was tilted to
nation of the normal source of hydrogen in fusion welding [3] 1.5◦ .
which translates into very low diffusible hydrogen levels as Before the process the plates were cleaned with sandpaper
those measured even for FSW welds performed underwater to remove the oxide layer from the surface and additionally
[15]. cleaned with ethanol. Argon was used as shielding gas to pro-
Research focused on the application of FSW on shipbuild- tect both the tool and the weld area from surface oxidation.
ing steel is scarce; nevertheless, studies concerning HSLA The dimensions of the tool consisted of shoulder diameter of
grades are more often available [16–21]. Reynolds et al. pro-
duced joints of approximately 470 mm in length on DH36 grade
with 6.4 thickness. They reported single pass, full penetration
at welding speeds as high as 7.6 mm/s substantially over-
matched. Bainite and martensite were found in the nugget
region [22]. Cater et al. made a comparison between friction
stir and submerged arc welding to join DH36 and E36 grades.
An assessment of the distortion for the two welding tech-
niques was reported and an initial comparative data on weld
tensile strength, toughness and fatigue was provided. Based
on their results, friction stir welding was shown to outperform
submerged arc welding [23]. Azevedo et al. published fatigue
results for GL-A36 steel; 4 mm plates thick were joined using
different process parameters and tools. Their experimental
results showed that the joints exhibited a fatigue behavior
similar to the base material [24]. Toumpis et al. proposed a
process envelope for the grade DH36 varying rotational and Fig. 1 – Base material microstructure. Etched with Nital 3%,
welding speeds for 6 mm plates. They evaluated the joints by scale bar 100 ␮m.

Please cite this article in press as: Cunha PH, et al. Effect of welding speed on friction stir welds of GL E36 shipbuilding steel. J Mater Res Technol.
2018. https://doi.org/10.1016/j.jmrt.2018.07.014
JMRTEC-470; No. of Pages 11
ARTICLE IN PRESS
j m a t e r r e s t e c h n o l . 2 0 1 8;x x x(x x):xxx–xxx 3

36.8 mm and probe length of 6 mm. The welding parameters

employed were a constant rotational speed of 500 rpm with
1 mm/s, 2 mm/s and 3 mm/s welding speeds to achieve three
different heat input values. The process parameters have been
selected based on the available literature [1,3–7] and on the
know-how from Helmholtz Zentrum Geesthacht.
The heat input (HI) index was calculated according to Eq.
(1):

ω. 
HI = (1)
v

where ω represents the rotational speed (radians per second),

 represents the torque supplied from the machine (Nm) and v
the welding speed (m/s). During the process, the tool tempera-
ture was measured with an infrared pyrometer pointed at the
face of the locking collar (Fig. 2). For the thermal cycles exper- Fig. 2 – Thermal acquisition by the pyrometer pointed at
imented for the plates during the welding, 18 thermocouples the locking collar of the tool.
(K type) were placed on the plates as shown in Fig. 3.
The location of the samples removed for bending (BD1, BD2
and BD3), tensile (TT1, TT2, TT3, TT4 and TT5) and microstruc- microhardness samples, the numbers 1, 2 and 3 stand for the
ture and microhardness (M1, M2 and M3) investigations are beginning, the middle and the ending of the joint, respectively.
shown in Fig. 4. In case of the bending, microstructure and X-ray inspection was done for the whole length of the joint.

1200
5

06 12 5
05 11
04 10
18
17
16

5
5
Welding Rotation Direction
15
400

5
5
13
14
01 07 15
5
02 08
03 09 5
50
500

Fig. 3 – Thermocouple positions used for thermal data acquisition.

1200
947
497
4 47
4 4 4 4

BDI TT1 TT2 TT3 TT4 TT5 BD2 BD3

215.50

107.75

M1 M2 M3
100

200
200

100

200

100

400
25

25

25

24 24 24

Fig. 4 – Location of samples used for characterization of the joints, values in mm.

Please cite this article in press as: Cunha PH, et al. Effect of welding speed on friction stir welds of GL E36 shipbuilding steel. J Mater Res Technol.
2018. https://doi.org/10.1016/j.jmrt.2018.07.014
JMRTEC-470; No. of Pages 11
ARTICLE IN PRESS
4 j m a t e r r e s t e c h n o l . 2 0 1 8;x x x(x x):xxx–xxx

Table 2 – Etchants used to reveal the microstructure.

Etchant Composition

Nital Ethanol, nitric acid

Beraha I 1 g potassic sulfite, 100 ml parent dilution
Beraha I (1000 ml distilled water, 200 ml
hydrochloric acid 32%, 24 g ammonium
hydrogen difluoride
Nital + Na2 S2 O5 Pre-etching with Nital, wet etching with
Na2 S2 O5 10% in aqueous dilution
LePera 50 ml Na2 S2 O5 1% in aqueous dilution,
50 ml picric acid 4% in ethanol

Fig. 5 – Surface appearance of the joint processed with

The tensile tests were performed using a Zwick Roell uni- 1 mm/s, region close to the exit hole.
versal testing machine with a 100 kN load cell performed at
1 mm/min at room temperature according to the standard
DIN EN 895:1995-08 for the joints specimens and for the base
well. Very high pressures lead to overheating and thinning of
material DIN 50125:2009-07. The bending tests have been car-
the joint while very low pressures lead to insufficient heating
ried out according to the standard DIN EN ISO 5173:2010-08,
and voids [4].
using HIDROALFA press machine model 20. The microhard-
The weld face quality along the whole length is a very good
ness profiles were measured by a Zwick Roell indentec ZHV
indicative that the forces involved during the process were sta-
2 system using HV0.5 and 0.4 mm distance between inden-
ble and the absence of flash after the process proved that the
tations. Additionally, microhardness mapping measurements
parameters used were suitable to maintain the stirred mate-
were performed using a hardness scanner model UT 100
rial within the joint, with an exception of the joint processed
supplied by BAQ. The macrograph and micrograph samples
with 3 mm/s of welding speed because in this case the bending
were prepared using different grades of sandpaper with the
test showed a small defect at the root of the weld.
final polishing done using 1 ␮m diamond suspension. The
The results from the X-ray inspection did not show any
specimens were etched for micrograph evaluation with the
macro defects along the whole length of the three different
reagents listed in Table 2. Nevertheless, for the macrograph
weld conditions processed. In this sense, Fig. 6 shows the
inspection only Nital was used.
images obtained at the middle of the joints that exemplify
these tests. Usually, radiographic test can detect broken tool
3. Results and discussion pieces and defects such as voids and wormholes. The tech-
nique easily identifies volumetric defects larger than 0.5 mm
3.1. Process parameters relationship and joint quality in diameter; however, this test has some difficulty with non-
assessment

Table 3 shows the welding parameters used to process the

joints as well as the torque, the axial force and the heat input
calculated for each process condition. The torque generated
among the three different conditions increased only slightly
with the increase in welding speed and this was expected to
happen according to the literature [4]. The weld face of the
joints displayed very good esthetic and homogeneous surface
quality without virtually any flash after the welding procedure,
as illustrated in Fig. 5. Flash is produced by material displace-
ment from the face of friction stir welded components. It is
often used as a visual sign that the proper tool depth has been
used. For example, if the insertion depth is too deep, excessive
flash is created. Excessive flash may also result from improper
tooling or parameter settings. Shoulder scrolling and reduced
rotation rate are examples of flash mitigation techniques [1]. Fig. 6 – X-ray images of the joints processed with (a)
Axial force has an important role on the quality of the weld as 1 mm/s, (b) 2 mm/s and (c) 3 mm/s as welding speed.

Table 3 – Process parameters utilized to produce the joints.

Condition Rotational speed (rpm) Welding speed (mm/s) Torque (Nm) Axial force (kN) Heat input (kJ/mm)

1 500 1 59 40 3.089
2 500 2 66 40 1.728
3 500 3 75 40 1.309

Please cite this article in press as: Cunha PH, et al. Effect of welding speed on friction stir welds of GL E36 shipbuilding steel. J Mater Res Technol.
2018. https://doi.org/10.1016/j.jmrt.2018.07.014
JMRTEC-470; No. of Pages 11
ARTICLE IN PRESS
j m a t e r r e s t e c h n o l . 2 0 1 8;x x x(x x):xxx–xxx 5

volumetric defects (particularly those oriented in the plane tion was the same as the forward velocity vector [30]. This
parallel to the surface) [1]. observation has been confirmed for almost all the thermocou-
ples measurements at the present study by the values from
3.2. Thermal profiles and heat input approach Fig. 8. The large temperature difference for the 304L alloy is
higher than the values typically observed in the FSW of other
Wei and Nelson have studied four different process heat materials under most welding conditions. However, the asym-
indexes and reported that the heat input provides the best metry in temperature can be explained due to the very low
correlation with post-friction stir weld microstructures [17]. thermal conductivity of stainless steel compared to most other
Nevertheless, it is important to keep in mind that this equa- alloys [4]. The heat input index approach should consider the
tion only gives an estimate of the heat input into the joint, thickness of the plates being welded because for different
since it does not take into account tool heat losses through plate thickness the peak temperatures and the cooling rates
radiation or by conduction [18]. Furthermore, it was reported are expected to be different because there will be different
that the maximum temperature reached for FSW on steels is bulk material to dissipate the thermal cycle supplied by the
less than 1200 ◦ C and the time t8–5 taken to cool over the friction heat of the process. Therefore, the same heat input
range 800–500 ◦ C is approximately 11 s. This is comparable to, value adopted for different plate thickness will result in differ-
for example, t8–5 for a manual metal arc weld with a heat ent macrograph features and different joint microstructures
input of about 1.3 kJ/mm. Therefore, the metallurgical trans- characteristics resulting in distinct joint performance.
formations expected on the basis of cooling rates alone are
not expected to be remarkably different from ordinary welds 3.3. Mechanical and microstructural features
[4,26]. The insufficient heat input accompanied by the lack in
material flow may be caused in a stagnant zone formation as Considering that X-ray inspection has a limitation to detect
proved by modeling studies [27]. The material in this region discontinuities smaller than 0.5 mm in diameter, simple bend-
will be vulnerable to crack formation under the normal plunge ing tests are a useful and fast way to evaluate more precisely
force [28]. And lower heat inputs are associated with higher the existence of smaller defects not detected by the radio-
welding speeds if kept the rotational speed. graph technique. Fig. 9 presents the bent samples and their
The temperatures measured with the pyrometer pointed to respective angles. The sample conditions processed with 1 and
the side face of the locking collar showed peak temperatures 2 mm/s have been bent completely proving absence of defects
of 850 ◦ C, 800 ◦ C and 760 ◦ C during the process respectively within the joint. However, during the bend test, the condi-
for the condition processed with 1, 2 and 3 mm/s of welding tion processed with 3 mm/s of welding speed, the sample BD2
speed. These values confirmed that keeping the same rota- (middle of the joint) showed a lack of penetration (LOP); Fig. 10
tional speed with the increase of the welding speed the tool shows this defect at the joint root at the bent specimen and
is going to reach a lower peak temperature because it stays the defect detail at the metallographic specimen.
less time in contact at the same heated position on the plate A root bend test, for instance, is a good predictor of fatigue
during the process, as illustrates in Fig. 7. Additionally, the test performance related to root flaws. The failure in root bend-
temperature measurements acquired on the plates with the ing is accompanied by a measurable and generally significant
thermocouples, whose were located near to the face of the reduction of fatigue life [1]. Poor bonding of the joint can often
joints, proved the similar phenomenon, i.e., the increase of the be attributed to inadequate pressure on the workpiece mate-
welding speed decreases the peak temperature undergone by rial due to a lack of axial force placed onto the surface of the
the workpiece as it is shown in Fig. 8. workpieces owning inadequate axial load [14]. Stevenson et al.
Cho et al. have reported based on the experimental and [31] also reported that as weld root defects resulted from the
modeling results at 304L stainless steel that temperatures lack in plunge depth of the FSW tool. Therefore, a recommen-
were higher on the advancing side than on the retreating side dation to solve this issue and to be able to produce sound
around 100 K (173 ◦ C) [29]. Arbegast and Hartley have reported welds using higher welding speeds should be to increase the
that there was a slightly higher temperature on the advancing axial load to make sure that the material stirring will be bond
side of the joint because the tangential velocity vector direc- properly, i.e., without defects that could compromise the per-
formance of the joint. Rameshi et al. reported the tendency to
form macroscopic defects with the increase in welding speed
Tool Temperature and that root flaw and groove defect were observed at higher
900
welding speeds [32]. Moussawi et al. reported results from
800
700
DH36 and EH46 steel grades FSW joints correlating defects
Temperature (ºC)

600
3 mm/s 2 mm/s 1 mm/s with high tool welding speeds such voids, weld root defect
500 and kissing bonds. It was found that the lack in material flow
1 mm/s
400 as a result of stagnant zone formation was the main reason of
2 mm/s
300
3 mm/s these defects [28].
200 Fig. 11 exhibited tensile results for the joints and the base
100 material mechanical properties were taken from data given by
0
0 50000 100000 150000 200000 the supplier. In general, friction stir welds of steels displayed
Time (s) higher mechanical properties than the respective base mate-
rial [5]. This is explained because of the thermo mechanical
Fig. 7 – Pyrometer tool measurements. effect imposed by the welding process on the base material

Please cite this article in press as: Cunha PH, et al. Effect of welding speed on friction stir welds of GL E36 shipbuilding steel. J Mater Res Technol.
2018. https://doi.org/10.1016/j.jmrt.2018.07.014
JMRTEC-470; No. of Pages 11
ARTICLE IN PRESS
6 j m a t e r r e s t e c h n o l . 2 0 1 8;x x x(x x):xxx–xxx

1 mm/s - RS 1 mm/s - AS
800 800
750 750
T4
700 700
650 T13
T4 650
600 T7 T1
600

Temperature (ºC)
T10 T16 T1
Temperature (ºC)

550 T5 550 T2
500 500
T6 T8 T3
450 450
T11 T17 T2
400 T10 400 T7
350 T5
T11 350 T15 T8
300 T12 T18 T3
T6 300
250 T12 T9 T9
250
200 T16 200 T13
150 150
100 T17 T14
100
50 T18 50 T15
0 0
0 500 1000 1500 2000 0 500 1000 1500 2000
Time (s) Time (s)

2 mm/s - RS 2 mm/s - AS
800 800
750 750
700 700 T13
650 650
T4 T1
600 600
Temperature (ºC)

Temperature (ºC)
550 T5 550 T2
500 500 T1
T4 T10 T16 T6 T7 T3
450 450
400 T10 400 T7
T2
350 T11 350 T8
T5 T11 T17 T8
300 300
250 T12 250 T9 T15 T9
T12 T3
T6 T18
200 T16 200 T13
150 150
T17 T14
100 100
50 T18 50 T15
0 0
0 200 400 600 800 1000 0 200 400 600 800 1000

Time (s) Time (s)

3 mm/s - RS 3 mm/s - AS
800 800
750 750
700 700
650 650
T4 T1
600
Temperature (ºC)

600
Temperature (ºC)

550 T5 550 T2
500 500 T1
T6 T3
450 450
T16 T10 T7
400 T10 400 T7
T4
350 T11 350
300 T11 T2 T8
T17 300 T8
250 T5 T12 T9
T12 T18 250 T3 T15
200 T16 200 T9
T6 T13
150 150
100 T17 T14
100
50 T18 50 T15
0 0
0 200 400 600 800 1000 0 200 400 600 800 1000

Time (s) Time (s)

Fig. 8 – Thermal profiles measured by the thermocouples, retreating side (RS) and advancing side (AS).

1 mm/s,3.089 kJ/mm 2 mm/s,1.728 kJ/mm 3 mm/s,1.309 kJ/mm

152º
151º 151º

152º 99º
152º

151º 152º
151º

20mm 20mm 20mm

Fig. 9 – Bent samples, heat input and the measured angle for each welding speed condition.

Please cite this article in press as: Cunha PH, et al. Effect of welding speed on friction stir welds of GL E36 shipbuilding steel. J Mater Res Technol.
2018. https://doi.org/10.1016/j.jmrt.2018.07.014
JMRTEC-470; No. of Pages 11
ARTICLE IN PRESS
j m a t e r r e s t e c h n o l . 2 0 1 8;x x x(x x):xxx–xxx 7

a b

12mm
50µm

Fig. 10 – (a) Top view from the root of the specimen BD2 joined with 3 mm/s showing the LOP (b) cross section image taken
at the bottom of the specimen near to the bending sample; the morphology is similar to a forging lap; scale bar is 50 ␮m.

Tensile Tests the increase of the welding speed, i.e., decrease of the heat
input would result in a smaller heat affected zone and narrow
564 565 562 556
stirred zone. The macrographs results are useful for quality
600
inspection and an initial guidance for microstructure evalu-
500 453 437 438
413 ation, as well as for better understanding of the mechanical
400 performance of the joints.
MPa

YS
300 UTS
Microhardness profiles were made at the beginning, the
200
middle and ending of the joints, but since that these results
were quite similar, and the macrographs, the profile at the
100
beginning of the joint are representative (Fig. 13). These results
0
Base Material 1mm/s 2mm/s 3mm/s exhibit that the microhardness profiles increase from the base
material until the stirred zone due the gradual microstruc-
Fig. 11 – Tensile testing results. ture transformation of the original microstructure toward the
direction of the stirred zone. The perlite and ferrite from the
base material were gradually modified by the FSW process first
that produces a refined and harder microstructure within the to a degenerated perlite, then to a spheroidized microstruc-
stirred zone. Lee et al. found that the FSW improved the yield ture (heat affected zone) and finally to a complex mixed
strength due to both grain boundary strengthening by grain microstructure consisting mainly of ferrite, martensite and
refinement and dislocation hardening by the introduction of bainite within the stirred zone. The values measured within
dislocations with the increasing welding speed. They inves- the stirred zone ranged by 300–400 HV at the middle thickness
tigated the influence of this parameter on the metallurgical of the joint, with slightly higher values at the retreating side.
and mechanical behavior of twinning-induced plasticity steels The color maps exhibited in Fig. 14 offers a better way
using three welding speeds [33]. to evaluate the microhardness than the single lines profiles
The macrographs images (Fig. 12), at the first approach, because they demonstrate the hardness overview behavior.
prove the stability of the processed joints because they pre- However, these results should be interpreted carefully, since
sented the same flow pattern and the same microstructure they are qualitative and should be used as a reference because
features among the same parameter. These results also do not the machine takes a calibration measure from the base mate-
demonstrate presence of voids. Also, they demonstrated that rial of the sample. According to this calibration, the machine

1 mm/s - beginning (sample M1) 2 mm/s - beginning (sample M1) 3 mm/s - beginning (sample M1)

1 mm/s - middle (sample M2) 2 mm/s - middle (sample M2) 3 mm/s - middle (sample M2)

1 mm/s - ending (sample M3) 2 mm/s - ending (sample M3) 3 mm/s - ending (sample M3)

Fig. 12 – Macrographs of the welded joints.

Please cite this article in press as: Cunha PH, et al. Effect of welding speed on friction stir welds of GL E36 shipbuilding steel. J Mater Res Technol.
2018. https://doi.org/10.1016/j.jmrt.2018.07.014
JMRTEC-470; No. of Pages 11
ARTICLE IN PRESS
8 j m a t e r r e s t e c h n o l . 2 0 1 8;x x x(x x):xxx–xxx

450 processed with 2 mm/s of welding speed. The microhard-

RS AS
ness color maps demonstrated a suitable correlation with the
400 1 mm/s
2 mm/s macrographs specimens, therefore being appropriate to real-
Microhardness (HV0.5)

350
3 mm/s
ize the location of the hardness gradients.
The etching trials made, using the solutions listed in Table 2
300 for the microstructure characterization, proved that Nital is
the suitable choice since it revealed the microstructure more
250
clearly and easily. Beraha I and Nital + Na2 S2 O5 did not show
200 results better than Nital, and LePera presented heteroge-
neous, unclear and not reliable results. The mixed, highly
150
deformed and refined microstructure presented at the stirred
100 zone of the samples probably contributed for the unsuccessful
0 5 10 15 20
other etching trials, since these reagents should reveal such
Position (mm) microstructures properly [34].
Fig. 15 indicates the microstructures of the center at the
Fig. 13 – Microhardness profiles at the mid-thick cross
middle thickness of the stirred zones of the joints. They con-
section, retreating side (RS) and advancing side (AS).
sisted of a complex mixed refined microstructure of ferrite
(highlighted as ¨F¨), martensite (highlighted as ¨M¨) and bainite
(highlighted as ¨B¨). As higher peak temperatures are achieved
from lower welding speeds, they will result in larger austen-
makes all the individual microhardness measurements and
ite grains. This grain size then influences the hardenability
builds the overview map. For a precise measurement a stan-
of the material and thus the phase type and morphology
dard test machine should be used. These results revealed
that will form during the cooling [18]. Since a way to eval-
lower values at the bottom and higher values at the retreat-
uate the weldability for steels is the carbon equivalent, this
ing side. These facts were explained because close to the root
value calculated according to the International Institute for
of the joint a higher amount of ferrite was found and at the
Welding (IIW) for the base material joined in the current work
retreating side a higher amount of martensite and bainite was
was 0.48, by itself this index predicts that the grade has high
found. The first observation leads to the assumption that cool-
tendency to form martensite. A strong correlation for hard-
ing rate within this region was lower than at the top of the
ness as a function of the IIW carbon-equivalent, according
joint. The second remark could be explained since the retreat-
to the Eq. (2) has been reported, and the highest hardness
ing side achieved a lower peak temperature which leads to a
values reported correspond to microstructures which contain
higher cooling rate.
substantial quantities of martensite and bainite in carbon-
Cui et al. reported that as the peak temperature decreases,
containing steels [4].
the cooling rate increases (with the increasing welding speed)
and these values are independent of the types of steels [11].
Mn + Si Ni + Cu Cr + Mo + V
Besides, the argon used as a shielding gas for the tool might CE = C + + + (2)
6 15 5
play a role as a forced cooling source leading to the forma-
tion of harder microstructures at the top of the joint. The The condition welded with the lower heat input of this
maps also demonstrated that with the increase of the weld- study, i.e., processed with the higher welding speed of 3 mm/s
ing speed the microhardness behavior showed less gradient reached the lowest peak temperatures and hence had a higher
between the advancing and the retreating sides and more cooling rate leading to a higher amount of displacive transfor-
homogenous hardness distribution, especially the condition mation such as bainite, Widmanstätten ferrite and acicular

HV 0.5
163
1 mm/s 207
250
294
338
382
425
469
513
556
600
2 mm/s

3 mm/s

Fig. 14 – Color microhardness maps, retreating side (RS) at the left side and advancing side (AS) at the right side.

Please cite this article in press as: Cunha PH, et al. Effect of welding speed on friction stir welds of GL E36 shipbuilding steel. J Mater Res Technol.
2018. https://doi.org/10.1016/j.jmrt.2018.07.014
JMRTEC-470; No. of Pages 11
ARTICLE IN PRESS
j m a t e r r e s t e c h n o l . 2 0 1 8;x x x(x x):xxx–xxx 9

10um 10um 10um

F F M M B B

Fig. 15 – Complex mixed microstructures of the mid thickness from the center of the stirred zone, ferrite highlighted as ¨F¨,
martensite as ¨M¨ and bainite as ¨B¨.

Fig. 16 – Overview of the new tool (left) and the same tool (right) after 8.5 m of welds.

ferrite instead of diffusional transformations such as allotri- Notwithstanding, it is very important to keep in mind that
omorphic ferrite and pearlite. The microstructures identified the prior base material microstructure features, i.e., grain size
at the stirred zone of the welds produced on the present study and heat treatment supply condition, are going to play an
showed consistency with the evaluation of Reynolds et al. [22] important role concerning all the metallurgical transforma-
when they applied the FSW process on the grade DH36 with tions induced by the thermo and mechanical process.
6.4 mm in thickness, as well as the results reported by Toumpis
et al. [25] also for the DH36 alloy with 6 mm thick and with the
work published by Barnes et al. for the grade HSLA-65 with 6.35
3.4. Tool performance
thickness [18]. Rameshi et al. found the microstructure of the
stirred zone consisted of upper bainite and fine ferrite phases
Fig. 16 demonstrates the condition of the tool before any welds
and that the amount of upper bainite reduced with welding
and after approximately 8.5 m of welds. The tool visual eval-
speed due to decrease in peak temperature and faster cooling
uation showed good wear behavior without apparent loss in
[32].
dimensions and geometrical features of the probe and shoul-
Imam et al. reported that the volume fraction of marten-
der. This is an indicative that the process parameters selected
site and average hardness in the stirred zone decrease
were sufficiently appropriate to reach the necessary heat and
with the increasing welding speed and that the peak tem-
plasticize the material right in front of the tool during its lin-
peratures decreased and cooling rates increase with the
ear motion. Consequently, ensuring that the forces and the
increasing welding speed [35]. Toumpis and his cowork-
torque experienced by the tool were suitable to minimize its
ers reported that slow welding speed welds, 100 mm/min
wear.
and 200 rpm, presented a ferrite predominant, homoge-
The literature claims that tool wear was found to increase
neous microstructure with significant grain refinement; a
with rotational speed and decrease at lower traverse speed,
heterogeneous microstructure was exhibited by the weld at
which suggests that process parameters can be adjusted to
the intermediate welding speeds, 250 mm/min and 300 rpm,
increase tool life [6]. Furthermore, it was reported that due
consisted of acicular shaped bainitic ferrite rich regions
to the low fracture toughness of the pcBN tool, this mate-
and ferrite predominant regions of either acicular shape
rial is susceptible to thermal cycling cracking as a result of
or of random geometry and the fast welding speed welds,
differences of heating and cooling rates between the pin and
500 mm/min and 700 rpm, features to be a heterogeneous but
shoulder [18]. Despite of this, pcBN tool is still preferred FSW
predominantly acicular shaped bainitic ferrite microstructure
of steels and other high melting temperature alloys due to its
with small regions of acicular ferrite [36].
high strength and hardness at elevated temperatures along

Please cite this article in press as: Cunha PH, et al. Effect of welding speed on friction stir welds of GL E36 shipbuilding steel. J Mater Res Technol.
2018. https://doi.org/10.1016/j.jmrt.2018.07.014
JMRTEC-470; No. of Pages 11
ARTICLE IN PRESS
10 j m a t e r r e s t e c h n o l . 2 0 1 8;x x x(x x):xxx–xxx

with high temperature stability. Furthermore, the low coeffi- [4] Nandan R, DebRoy T, Bhadeshia HKDH. Recent advances in
cient of friction for pcBN results in smooth weld surface [6]. friction-stir welding – process, weldment structure and
properties. Prog Mater Sci 2008,
http://dx.doi.org/10.1016/j.pmatsci.2008.05.001.
4. Conclusions [5] Çam G. Friction stir welded structural materials: beyond
Al-alloys. Int Mater Rev 2013,
The effect of welding speed on friction stir welds of GL E36 http://dx.doi.org/10.1179/095066010X12777205875750.
[6] Rai R, De A, Bhadeshia HKDH, DebRoy T. Review: friction stir
shipbuilding steel was evaluated in the present study. The
welding tools. Sci Technol Weld Join 2011,
findings of the current study can be summarized as follow: http://dx.doi.org/10.1179/1362171811Y.0000000023.
[7] Padhy GK, Wu CS, Gao S. Friction stir based welding and
• The process parameters were appropriately chosen since processing technologies – processes, parameters,
all the joints showed very good esthetics and homogeneous microstructures and applications: a review. J Mater Sci
surface quality. Nevertheless, when used 3 mm/s as weld- Technol 2018;34:1–38,
ing speed it was found lack of penetration, but this could http://dx.doi.org/10.1016/j.jmst.2017.11.029.
[8] Pradeep A. A review on friction stir welding of steel. Int J Eng
be bypass increasing the axial force. Concerning the tool, it
Res Dev 2012,
did not exhibit considerable wear even after 8.5 m of weld http://dx.doi.org/10.4028/www.scientific.net/AMR.922.553.
length. [9] Venkatesh KM, Arivarsu M, Manikandan M, Arivazhagan N.
• The macrographs at the beginning, middle and ending Review on friction stir welding of steels. Mater Today Proc
of the welded joints had similar microstructure features 2018;5:13227–35,
proving the process stability, especially between the joints http://dx.doi.org/10.1016/j.matpr.2018.02.313.
[10] Liu FC, Hovanski Y, Miles MP, Sorensen CD, Nelson TW. A
produced using 2 and 3 mm/s as welding speed.
review of friction stir welding of steels: tool, material flow,
• The X-ray inspection has limitation to detect smaller
microstructure, and properties. J Mater Sci Technol
defects, so it was mandatory to additionally perform bend- 2018;34:39–57, http://dx.doi.org/10.1016/j.jmst.2017.10.024.
ing tests to evaluate the joint integrity. [11] Cui L, Fujii H, Tsuji N, Nakata K, Nogi K, Ikeda R, et al.
• For the welds processed with welding speeds of 2–3 mm/s, Transformation in stir zone of friction stir welded carbon
the temperature measures showed most uniform thermal steels with different carbon contents. ISIJ Int 2007,
profiles. Hence, the microstructure and mechanical proper- http://dx.doi.org/10.2355/isijinternational.47.299.
[12] Ghosh M, Hussain M, Gupta RK. Effect of welding parameters
ties obtained with these two conditions are more uniform
on microstructure and mechanical properties of friction stir
along the whole joint length.
welded plain carbon steel. ISIJ Int 2012;52:477–82,
• A high microstructure gradient was found within the stirred http://dx.doi.org/10.2355/isijinternational.52.477.
zone. The microstructure found consisted mainly of ferrite, [13] Matsushita M, Kitani Y, Ikeda R, Endo S, Fujii H.
martensite and bainite with different levels of refinement Microstructure and toughness of friction stir weld of thick
and morphologies showing that joints were obtained with structural steel. ISIJ Int 2012,
a balance between strength and toughness. http://dx.doi.org/10.2355/isijinternational.52.1335.
[14] Lohwasser D, Chen Z. Friction stir welding: from basics to
applications; 2009, http://dx.doi.org/10.1533/9781845697716.
Conflicts of interest [15] Hoyos JJ, Pereira VF, Giorjao RR, McNelley TR, Ramírez AJ.
Effect of friction stir welding on hydrogen content of ISO
3183 X80M steel. J Manuf Process 2016,
The authors declare no conflicts of interest. http://dx.doi.org/10.1016/j.jmapro.2016.01.012.
[16] Konkol PJ, Mathers JA, Johnson R, Pickens JR. Friction stir
welding of HSLA-65 steel for shipbuilding. J Ship Prod
Acknowledgments
2003;19:159–64.
[17] Wei L, Nelson TW. Correlation of microstructures and
The authors of this work would like to thank CNPq process variables in FSW HSLA-65 steel. Weld J 2011.
(National Council for Scientific and Technological Develop- [18] Barnes SJ, Bhatti AR, Steuwer A, Johnson R, Altenkirch J,
ment) and CAPES (Brazil) for their support (grant number Withers PJ. Friction stir welding in HSLA-65 steel: Part I.
236657/2012-7). In addition, the authors thank Helmholtz- Influence of weld speed and tool material on microstructural
Zentrum Geesthacht and Petrobras. development. Metall Mater Trans A Phys Metall Mater Sci
2012, http://dx.doi.org/10.1007/s11661-012-1110-z.
[19] Steuwer A, Barnes SJ, Altenkirch J, Johnson R, Withers PJ.
references Friction stir welding of HSLA-65 steel: Part II. The influence
of weld speed and tool material on the residual stress
distribution and tool wear. Metall Mater Trans A Phys Metall
Mater Sci 2012, http://dx.doi.org/10.1007/s11661-011-0643-x.
[1] Gibson BT, Lammlein DH, Prater TJ, Longhurst WR, Cox CD,
[20] Young J, Field D, Nelson T. Material flow during friction stir
Ballun MC, et al. Friction stir welding: process, automation,
welding of HSLA 65 steel. Metall Mater Trans A Phys Metall
and control. J Manuf Process 2014;16:56–73,
Mater Sci 2013, http://dx.doi.org/10.1007/s11661-013-1684-0.
http://dx.doi.org/10.1016/J.JMAPRO.2013.04.002.
[21] Wei L, Nelson TW. Influence of heat input on post weld
[2] Bhadeshia HKDH, DebRoy T. Critical assessment: friction stir
microstructure and mechanical properties of friction stir
welding of steels. Sci Technol Weld Join 2009,
welded HSLA-65 steel. Mater Sci Eng A 2012,
http://dx.doi.org/10.1179/136217109X421300.
http://dx.doi.org/10.1016/j.msea.2012.06.057.
[3] Mishra RS, Ma ZY. Friction stir welding and processing. Mater
[22] Reynolds AP, Tang W, Posada M, Deloach J. Friction stir
Sci Eng R Rep 2005;50:1–78,
welding of DH36 steel. Sci Technol Weld Join 2003;8:455–60,
http://dx.doi.org/10.1016/J.MSER.2005.07.001.
http://dx.doi.org/10.1179/136217103225009125.

Please cite this article in press as: Cunha PH, et al. Effect of welding speed on friction stir welds of GL E36 shipbuilding steel. J Mater Res Technol.
2018. https://doi.org/10.1016/j.jmrt.2018.07.014
JMRTEC-470; No. of Pages 11
ARTICLE IN PRESS
j m a t e r r e s t e c h n o l . 2 0 1 8;x x x(x x):xxx–xxx 11

[23] Cater S, Martin J, Galloway A, McPherson N. Comparison [30] Arbegast WJ, Hartley PJ. Friction stir weld technology
between friction stir and submerged arc welding applied to development at lockheed Martin Michoud space system – an
joining DH36 and E36 shipbuilding steel. In: Frict. Stir Weld. overview. ASM Proc Int Conf Trends Weld Res 1998.
Process VII. 2013. p. 47–58, [31] Stevenson R, Toumpis A, Galloway A. Defect tolerance of
http://dx.doi.org/10.1002/9781118658345.ch6. friction stir welds in DH36 steel. Mater Des 2015;87:701–11,
[24] Azevedo J, Infante V, Quintino L, dos Santos J. Fatigue http://dx.doi.org/10.1016/J.MATDES.2015.08.064.
behaviour of friction stir welded steel joints. Adv Mater Res [32] Ramesh R, Dinaharan I, Kumar R, Akinlabi ET. Microstructure
2014;891–892:1488–93, and mechanical characterization of friction stir welded high
http://dx.doi.org/10.4028/www.scientific.net/AMR.891-892.1488. strength low alloy steels. Mater Sci Eng A 2017;687:39–46,
[25] Toumpis A, Galloway A, Cater S, McPherson N. Development http://dx.doi.org/10.1016/j.msea.2017.01.050.
of a process envelope for friction stir welding of DH36 steel – [33] Lee S-J, Sun Y, Fujii H. Stacking-fault energy, mechanical
a step change. Mater Des 2014, twinning and strain hardening of Fe–18Mn–0.6C–(0,1.5)Al
http://dx.doi.org/10.1016/j.matdes.2014.04.066. twinning-induced plasticity steels during friction stir
[26] Thomas WM, Threadgill PL, Nicholas ED. Feasibility of welding. Acta Mater 2018;148:235–48,
friction stir welding steel. Sci Technol Weld Join 1999, http://dx.doi.org/10.1016/j.actamat.2018.02.004.
http://dx.doi.org/10.1179/136217199101538012. [34] Zakerinia H, Kermanpur A, Najafizadeh A. Color
[27] Al-Moussawi M, Smith AJ, Young A, Cater S, Faraji M. metallography; a suitable method for characterization of
Modelling of friction stir welding of DH36 steel. Int J Adv martensite and bainite in multiphase steels. Int J ISSI 2009.
Manuf Technol 2017, [35] Imam M, Ueji R, Fujii H. Microstructural control and
http://dx.doi.org/10.1007/s00170-017-0147-y. mechanical properties in friction stir welding of medium
[28] Al-Moussawi M, Smith AJ. Defects in friction stir welding of carbon low alloy S45C steel. Mater Sci Eng A 2015;636:24–34,
steel. Metallogr Microstruct Anal 2018;7:194–202, http://dx.doi.org/10.1016/j.msea.2015.03.089.
http://dx.doi.org/10.1007/s13632-018-0438-1. [36] Toumpis A, Galloway A, Molter L, Polezhayeva H. Systematic
[29] Cho J-H, Boyce DE, Dawson PR. Modeling strain hardening investigation of the fatigue performance of a friction stir
and texture evolution in friction stir welding of stainless welded low alloy steel. J Mater 2015;80:116–28,
steel. Mater Sci Eng A 2005, http://dx.doi.org/10.1016/j.matdes.2015.04.046.
http://dx.doi.org/10.1016/j.msea.2005.03.002.

Please cite this article in press as: Cunha PH, et al. Effect of welding speed on friction stir welds of GL E36 shipbuilding steel. J Mater Res Technol.
2018. https://doi.org/10.1016/j.jmrt.2018.07.014