Microstructural control and mechanical properties in ...

Materials Science & Engineering A 636 (2015) 24–34

Contents lists available at ScienceDirect

Materials Science & Engineering A journal homepage: www.elsevier.com/locate/msea

Microstructural control and mechanical properties in friction stir welding of medium carbon low alloy S45C steel Murshid Imam n, Rintaro Ueji, Hidetoshi Fujii Joining and Welding Research Institute, Osaka University, Ibaraki 5670047, Japan

art ic l e i nf o

a b s t r a c t

Article history: Received 11 February 2015 Received in revised form 18 March 2015 Accepted 20 March 2015 Available online 31 March 2015

Critical control of the welding conditions produced a fine ferrite–martensite duplex structure with the martensite volume fraction of around 33% in friction stir welding (FSW) of medium carbon (0.45 wt% C) low alloy steel sheets. This microstructure provides the preferable combination of tensile strength and ductility. The findings were achieved to clarify the effect of the welding speed on the weldability. The fraction of low angle grain boundaries (LAGBs) decrease with increasing welding speed. At the welding speed of 400 mm/min, the top region microstructure consists of lath martensite, while the bottom region shows a fine ferrite–pearlite structure. This microstructural variation can be linked with the distribution of the peak temperatures along the thickness of the weld. Additionally, it was shown that a non-contact thermal imaging system can be used as an effective tool for the online monitoring of several kinds of weld decays. & 2015 Elsevier B.V. All rights reserved.

Keywords: Cooling rate Peak temperature Duplex structure Martensite volume fraction Strength Ductility

1. Introduction Friction stir welding (FSW) is an effective solid state joining technique [1]. The solid-state nature of this process avoids solidification problems which are common in fusion welding methods [2]. In addition, the absence of ultraviolet or electromagnetic hazards, reduced energy requirements, elimination of consumables such as rods and grinding pastes are cited as some of the additional key benefits of this process [3–5]. These advantages make FSW more attractive for joining steels in many industrial applications such as shipbuilding, bridge decking, pipe seam welding, or the applications which were interrupted due to the weldability [6]. Microstructural evolution in steels during FSW is a more complex process than that in aluminum alloys due to the occurrence of phase transformations [7–9]. Several factors including tool design, tool rotational speed (ω), welding speed (V), axial load, tool tilt angle, and plunge depth are used to control the heat generation, material flow, and resultant thermo-mechanical cycle during the FSW process. The dominant parameters of thermal cycle that affect the microstructural evolution are peak temperature and cooling rate. As shown by Imam et al. [10] and Simar et al. [11], the peak temperature in the stir zone (SZ) increases with the increasing tool rotational speed and decreases monotonically with

n

Corresponding author. Tel.: þ 81 6 6877 5111; fax: þ 81 6 6879 8689. E-mail address: [email protected] (M. Imam).

http://dx.doi.org/10.1016/j.msea.2015.03.089 0921-5093/& 2015 Elsevier B.V. All rights reserved.

the increasing welding speed. Swaminathan et al. [12] found that the peak temperature on the advancing side is higher than that on the retreating side, and the peak temperature on the advancing side decreases with the increasing welding speed. Fujii et al. [13], Ghosh et al. [14], and Manvatkar et al. [15] reported that the cooling rate increases with the increasing welding speed. These reports are consistent so that the consensus for these thermal cycle parameters seems to be well shared. In recent years, significant progress has been made in the understanding of the effect of these thermal cycle parameters on the microstructure and mechanical properties of the FSW steel joints. Fujii et al. [13] reported that the microstructure and mechanical properties of the carbon steel joints are significantly affected by the peak temperature and cooling rate. They also reported that when the stirring is performed in the ferrite– austenite two phase regions, the microstructure is refined and the highest strength is achieved. Song et al. [16] reported that the ferrite grain size decreases with the increasing cooling rate, and as a result, improved mechanical properties were obtained as compared to the base metal (BM). Sun et al. [17] observed the formation of brittle martensite and bainite phases in the SZ when the peak temperature exceeds the A1 and A3 transformation temperatures. Along similar lines, Khodir et al. [18] reported the formation of martensite in the SZ when the peak temperature exceeds the A1 transformation temperature. Chung et al. [19] reported that the fraction of martensite decreases from the top to bottom of the joint because of the difference in peak temperatures between the top and bottom regions. Xue et al. [20] reported

M. Imam et al. / Materials Science & Engineering A 636 (2015) 24–34

25

the formation of martensite and bainite phases resulting from the high cooling rate and peak SZ temperature reaching the A3 transformation temperature. Based on these studies, it is now well recognized that the formation of brittle phases during the FSW process is sensitive to the peak temperature and cooling rate. Cui et al. [21] suggested the ways to avoid the formation of martensite by performing the welding below the A1 temperature. They also reported that the cooling rate can be controlled by decreasing the welding speed. However, decreasing the welding speed may be undesirable from the point of view of weld productivity. In other words, an additional study to extend the process in order to omit the brittle phases should be necessary. It should also be noted that the benefit of the FSW is more evident with the higher carbon content, while the weldability of the fusion welding further decreases. The other challenges that need to be addressed before commercialization are FSW tool pin durability, reproducibility of process, and process certification and qualification. The process certification and qualification are done by mechanical, metallurgical, and nondestructive tests which are time intensive and expensive, and beyond reach of several industries particularly in developing countries. Therefore, this work also emphasized that the temperature monitoring is one of the efficient ways to narrow the process window by identifying the process parameters. Also, the durability of FSW tool pin for weld lengths that are sufficient for commercial development is investigated. To this end, friction stir welds of medium carbon steel sheets were obtained over a wide range of welding speeds, and for each welding condition, the peak temperature and cooling rate were monitored. It is shown that proper control of the morphology and distribution of the microstructures produces the best strength and ductility combination. The identified process parameters help us to achieve this in a cost efficient way.

Table 1 Process parameters at which friction stir butt welds were obtained.

2. Experimental procedure

paper. Furthermore, for the microhardness measurements, samples of the 35 2 mm2 weld cross-sections were polished using different grades of emery paper. The microhardness measurements on the polished weld cross-sections were then taken using Vicker's microhardness tester at 200 gf and a 10 second dwell time. The stress–strain responses of the welds were characterized by transverse tensile tests on a specimen prepared as per ASTM: E8–M11 guidelines, while, small longitudinal tensile specimens were used to obtain the responses of the stir zones (see Fig. 1). Note that the longitudinal direction is same as the plate rolling or welding direction. Uniaxial tests were conducted at room temperature by a tensile testing machine at a fixed cross head speed of 1.5 mm/min.

2 mm thick 0.45 wt% steel sheets were friction stir welded at the process parameters shown in Table 1. The welding direction coincided with the plate rolling direction. The chemical composition of the plates used is given in Table 2. Tools with a cylindrical pin profile made of tungsten carbide (WC) were used to obtain the welds. During the welding, temperatures in the stir zone (SZ) and the heat affected zone (HAZ) on the advancing side were measured using thermocouples. The locations of the thermocouples are shown in Fig. 1 and 2, respectively. In the SZ, a thermocouple was placed on the bottom surface of the workpiece, whereas, the thermocouple on the advancing side was inserted at a distance of 8 mm from the weld centerline by drilling a 0.25 mm blind hole. Furthermore, on-line monitoring of the FSW process and surface temperatures were made using a non-contact temperature measurement system. The scan was made along the welding direction. In order to investigate the microstructure of the obtained FSW joints, specimens were prepared perpendicular to the welding direction and examined by scanning electron microscopy (SEM) and the electron backscatter diffraction (EBSD) technique. To prepare a specimen for the EBSD measurements, samples were electrolytically polished in a 20 ml HClO4 þ 180 ml CH3COOH solution at 10 1C. A field emission type scanning electron microscope (FE-SEM) operated at 15 kV was used for the SEM and EBSD measurements at a step size of 0.1 μm. In addition, for the optical macroscopy and hardness measurements, 30 4 2 mm3 samples were cut such that the length of the specimen was along the weld transverse direction (TD), width was along the weld line and thickness coincided with the sheet thickness direction. The 30 2 mm2 weld cross-sections were then polished on a disc polishing machine using different grades of emery

Tool pin profile

Rotational speed (rpm)

Cylindrical 400

Welding speed Pin (mm/min) height (mm)

Axial load (kN)

Tool tilt angle (1)

100, 200, 300, 400, 500

20

3

1.8

Table 2 Chemical compositions of S45C carbon steel sheets. Element

C

Si

Mn

P

S

Fe

Weight%

0.45

0.22

0.82

0.008

0.004

Bal

Transverse tensile specimen AS Thermocouple R6

R4

Tool 2 mm

40 6 30 10

10 mm

Thermal scan location at trailing edge (TE)

RS

Rolling direction Fig. 1. Schematics of transverse and longitudinal specimen used for characterizing the tensile responses of welds, prepared according to ASTM:E8-M11 guidelines. Note that the longitudinal specimens contains properties only from the stir zone.

3. Results and discussion 3.1. Peak temperatures Fig. 2 shows typical temperature profiles and recorded peak temperatures at the bottom of the SZs as a function of the welding speed. It can be clearly seen from this figure that the cooling rate increases and the peak temperature decreases with the increasing welding speed. Table 3 lists the obtained cooling rates, peak temperatures, and volume fraction of martensite for the welding conditions shown in Table 1. Note that the cooling rates listed in Table 3 are the average cooling rate from 800 to 500 1C. However, when the peak temperature (TP) was below 800 1C, the cooling rate is defined by the average cooling rate from the peak temperature to 500 1C. Also note that the lower and upper critical temperatures of the phase transformation shown in Fig. 2 are calculated from the empirical equations suggested by Trzaska and Dobrzanski [22]. The

26


1000

1200

800 Thermocouple A3 A1

800 600 v = 400 mm/min

400

v = 300 mm/min

v = 200 mm/min

v = 100 mm/min

T peak (o C)

o

Temperature ( C )

1000

600

Tp = 1989.4 v -0.188 R2 = 0.98

400 200

200

Vm _ 70 %

0

0 50

0

100 150 Time (secs)

200

250

0

100

200

300

400

Pin Failure 500

600

Welding speed (mm/min)

Fig. 2. (a) Temperature histories obtained from thermocouples, and (b) peak temperatures in stir zone as a function of welding speeds. Note that Vm refers to the volume fraction of martensite.

Table 3 A summary of calculated heat input, thermal data and volume fraction of martensite. S. No. Rotation- Welding al speed speed (mm/ (rpm) min)

Tp T HAZ (1C) (1C)

Cooling rate (1C/s)

Heat input (J/mm)

Volume fraction of martensite in SZ (%)

1 2 3 4 5

825 740 688 649 567

64.2 76.8 98.9 108.2 –

115.2 57.5 38.5 28.8 23.1

90 88 67 33 –

400 400 400 400 400

100 200 300 400 500

510 482 446 412 377

calculated A1 and A3 transformation temperatures are 727 and 780 1C, respectively. As discussed in the introduction, the resultant peak SZ temperatures and cooling rates, in turn, depend upon the process parameters [15,21]. The process parameters controlled the heat input conditions. Considering FSW as a simple friction process, in which the torque required to rotate circular shaft relative to the plate surface under the action of an axial load, is given by [23] T ¼ 23 πμPR3 ;

ð1Þ

where T is the applied torque, μ is the friction coefficient (here μ is considered equal to 0.4), R is the radius of the rotating pin, and P is the contact pressure. Assuming an ideal condition in which the total work is converted into frictional heat, the heat input (q) during the friction stir welding is given by [24] Z TR Z R 4 q¼ ω dt ¼ ω2πμPr 2 dr ¼ π 2 μPNR3 ; ð2Þ 3 0 0 This equation shows that the heat input depends both on the applied tool rotational speed and the shoulder radius, leading to a non-uniform heat generation during the FSW process. Eq. (3) can be modified by taking welding speed and the heat loses to the backing plate and FSW tool into account (ignoring heat loss through the radiation), and it is written as follows [25]: 4 μηPNR3 Q ¼ π2 ; 3 V

ð3Þ

where Q is the heat input per unit weld length, η is the heat input efficiency (0 o η o 1), and V is the welding speed. Note that the heat input efficiency is defined as the fraction of heat that remains in the workpiece. Eq. (3) clearly shows that the Q decreases with the increasing V, thus resulting in higher peak temperatures at the lower welding speed. Furthermore, for a given tool geometry and

plunge depth, the peak temperatures in the SZ were influenced by the tool rotational speed and welding speed [26], so that Eq. (3) should be further modified. A simplified relation between the peak temperatures, Tp and FSW process parameters, ω and V, was suggested by Arbegast and Hartely [27]: 2 α Tp ω ¼K ; ð4Þ Tm v where α and K are constants, Tm is the melting point of the alloy, and Tp is the peak temperature during welding. By using Eq. (4) and based on the experimental results in this study, a simplified relationship obeying power law equation is obtained between the peak temperatures (TP) and welding speeds V at a constant tool rotational speed (see Fig. 2b), which is given below: 0:188 1 ; ð5Þ T p ¼ 1989:4 V From Eq. (5), it can be well understood that the decrease in peak temperature with the increasing welding speed is on the expected line. Note that similar observations were made by Saeid et al. [28], Esmailzadeh et al. [29], and Bilgin and Meran [30] for the friction stir welding of duplex stainless steel and AISI 430 ferritic stainless steels. Fig. 3 shows the measured temperature profiles on the workpiece surface near the trailing edge of the rotating tool at the welding speeds of 100, 200, 300, 400, and 500 mm/min. A large flash formation from the excess heat generation can be clearly seen at the welding speed of 100 mm/min. A sudden drop in the temperature profile and the appearance of hot debris at the welding speed of 500 mm/min indicates the location of the tool pin failure during the welding operation. It is worth noting that welds without any pin failure can be obtained up to a welding speed of 400 mm/min, while higher bending and shear stresses experienced by the tool pin due to the lower heat input at the welding speed of 500 mm/min are responsible for the FSW tool failure [31]. Note that 170 mm of weld length for each welding condition is obtained for the welding speed up to 400 mm/min, while tool fails after 30 mm of weld length at the welding speed of 500 mm/min. 3.2. Microstructural characterization An SEM image and an EBSD orientation color map of the base metal (BM) are shown in Fig. 4. The EBSD results are illustrated by the map colored according to the crystal orientations parallel to the sample normal direction, which is indicated by the coloring in the


8 00

8 00 start 7 00 point

Temperature ( oC)

o

Temperature ( C)

7 00 6 00 5 00

Plunging time

4 00 3 00

Flash

2 00 1 00 0

20

40

6 00 5 00 4 00 3 00 2 00

0

6 0 8 0 1 00 1 20 1 40 1 60 Time (Secs)

8 00

8 00

7 00

7 00

Temperature ( C)

6 00

o

o

Temperature ( C)

0

5 00 4 00 3 00 2 00 1 00 0

20

40 60 80 Time (Secs)

1 00

V = 200 mm/min 0

20

40

6 0 8 0 1 00 1 20 1 40 Time (Secs)

Welding direction

6 00 5 00 4 00 3 00 2 00 1 00

V = 300 mm/min 0

End point

1 00

V = 100 mm/min

1 20

27

0

V = 400 mm/min 0

20


1 00

1 20

8 00

Pin fracture

o

Temperature ( C)

7 00 6 00 5 00 4 00 3 00 2 00 1 00 0

V = 500 mm/min 0

20


1 00

1 20

Fig. 3. (a–e) Online monitoring of welds using non-contact thermal imaging system for the welding conditions considered in this study.

standard stereo triangle. The black and red lines in the EBSD map indicate the high and low angle boundaries, respectively. The threshold angle between the high and low angle boundaries is 151. The presence of ferrite and lamellar cementite within the ferrite, commonly known as pearlite, can be seen from the SEM image shown in Fig. 4a. The pearlite structure includes some amount of low angle boundaries (LAGBs), which can be clearly seen in Fig. 4b. This feature well agrees with the previous report by Koga et al. [32]. Fig. 5 shows the typical microstructures at the center of the stir zone for welds obtained at the different welding speeds. A high volume fraction of lath martensite was formed at the welding speeds of 100 and 200 mm/min. At the welding speed of 300 mm/ min, the microstructure consists of ferrite, pearlite, and

martensite. This observation suggests that the peak temperature at the center is below the A3 temperature at this welding speed. In addition, the morphology of the martensite is different between the joints formed at 100 and 200 mm/min welding speeds. This represents an increase in the grain boundaries due to the large structural refinement with the increasing welding speed. On the other hand, a fine ferrite and pearlite structure and reduction in lath martensite volume fraction can be found at the welding speed of 400 mm/min. These changes can be well linked with the peak temperatures in the stir zone. The peak temperatures correspond to the austenite single-phase region (above A3), ferrite–austenite two-phase region (below A3), and the ferrite–cementite two-phase region (below A1). The observed peak temperatures at the bottom of the stir zone at the welding speeds of 300 and 400 mm/min are

28


Fig. 4. (a) SEM microstructure, and (b) EBSD orientation color map of base metal. (For interpretation of the references to color in this figure caption, the reader is referred to the web version of this paper.)

Fig. 5. Typical microstructures at the centre of SZs for welds obtained at (a) 100 mm/min, (b) 200 mm/min, (c) 300 mm/min, and (d) 400 mm/min welding speeds.

lower than the A1 transformation temperature. However, this temperature does not demonstrate that the entire stir zone is at this temperature. Note that Cui et al. [21] reported that the cooling rate is the dominating factor for the microstructural control in high carbon steel. On the other hand, based on these results it is quite evident that the peak temperature has the dominating effect in the friction stir welding of low or medium carbon steels. Thus, friction stir welding performing at the higher welding speed is always desirable for both the microstructural control from the

decreasing peak temperature and process productivity, particularly true in low or medium carbon steels. Fig. 6 shows the EBSD orientation color maps in the stir zone for welds obtained at the different welding speeds. A clearer visualization of the crystallographic orientation can be obtained by these EBSD maps. Also the presence of typical packet and block sizes in the martensitic steel can be seen from these images. It is worth noting that when martensite is formed from austenite, it contains packets and blocks. Packets are group of laths with the same habit


29

Fig. 6. EBSD orientation color maps showing microstructures at the centre of the SZs for welds obtained at (a) 100 mm/min, (b) 200 mm/min, (c) 300 mm/min, and (d) 400 mm/min welding speeds. (For interpretation of the references to color in this figure caption, the reader is referred to the web version of this paper.)

plane, and blocks are groups of laths of the same orientation (same variants) (detailed discussion on the crystallographic properties of the martensite phase was reported by [33]). The prior austenite grain sizes are believed to be large at lower welding speed because of high peak temperature, and as a result large block and packet sizes are present in the martensite. These block and packet sizes decrease with the decreasing prior austenite grain size because of the decrease in peak temperature at lower welding speed. Consequently, block and packet sizes decrease with the increase in welding speed. At the welding speed of 400 mm/min, most of the region shows equiaxed grains, while the local misorientation change in a few grains suggests the presence of pearlite. Furthermore, in order to analyze the grain boundary characteristics, the misorientation angle distributions are shown in Fig. 7. Four important observations can be made from this figure: (i) The presence of high angle grain boundaries (HAGBs) in the BM is close to a random distribution, (ii) The presence of LAGBs at the welding speeds of 100, 200, and 300 mm/min is due to the sub-structure within the

lath martensite, while the presence of HAGBs with 601 misorientation angles is due to the martensite phase transformation (see Fig. 7b–e). Note that a special boundary is created with a special orientation relationship at the misorientation angle of 601 [34], (iii) The fraction of LAGBs decrease with the increasing welding speed, and (iv) The fraction of LAGBs are similar to the base metal for the weld obtained at the 400 mm/min welding speed, suggesting that grain refinement occurs (see Fig. 7 e) in the SZ. Along similar lines, Xue et al. [20] reported a decrease in the fraction of LAGBs at the center of the SZ during the friction stir welding of low alloy steel. The SEM microstructure obtained at the welding speed of 400 mm/min is significantly different from the other welds. Therefore, a separate microstructural study was made to observe the temperature gradient effect along the thickness direction. Fig. 8 shows the SEM microstructures of a weld cross-section obtained at the welding speed of 400 mm/min. The top region microstructures consist of martensite, ferrite, and partially deformed pearlite

30


16

16 + 57

Base metal (BM) f = 15.2 % LAGBs

12 Random

10 8 6 4

LAGBs

12 10 8 6 4 2

2

0

0 0

10

20

30

40

50

60

0

14 Number Fraction (%)

16

V = 200 mm/min

+ 62

f

30

40

50

60

Misorientation angle (degrees)

= 30.4 %

LAGBs

12 10 8 6 4

V = 300 mm/min f = 23.2 %

+ 60


16

20

10

Misorientation angle (degrees)

LAGBs

12 10 8 6 4 2

2 0

V = 100 mm/min f = 32.2 %

+ 65


Number Fraction (%)

14

0 0

10

20

30

40

50

60

0

10

Misorientation angle (degrees) 16

Number Fraction (%)

60

V = 400 mm/min f LAGBs= 14.3 %

+ 60

14

20 30 40 50 Misorientation angle (degrees)

12 10 8 6 4 2 0 0

10

20

30

40

50

60

Misorientation angle (degrees) Fig. 7. Misorientation angle distribution maps of (a) base metal (BM) and (b–e) SZs for welds obtained at (a) 100 mm/min, (b) 200 mm/min, (c) 300 mm/min, and (d) 400 mm/min welding speeds.

suggesting that the peak temperature in this region is above the A1 transformation temperature. The bottom region consists of finer grains of ferrite–pearlite and a mixture of spheriodized cementite, which is similar to the base metal suggesting that the peak temperature is below the A1 transformation temperature. In addition, the martensite volume fraction is observed higher on the advancing side than in the center of the stir zone. Note that this observation is in line with the observation made by Sato et al. [35]. Moreover, this microstructural difference on the advancing and retreating sides could also be related to the higher strain and strain

rates on the advancing side as compared to the retreating side. The higher strain and strain rate was assumed to be responsible for the higher peak temperature on the advancing side which can be thought of another possibility for its formation [36,37]. Furthermore, the microstructure on the bottom region is similar to that of the base metal except that the ferrite grain size decreases from the plastic deformation in this region. In order to understand the mechanisms of the formation of the microstructural features between the top and bottom regions of the weld obtained at the welding speed of 400 mm/min, EBSD


1 µm

1 µm

31

1 µm

RS

AS

1 mm

µm 11 µm

µm 11 µm

1 µm

Fig. 8. SEM microstructures of weld cross-section for weld obtained at the welding speed of 400 mm/min.

Top

5 µm

Centre

Bottom

111

5 µm

5 µm

001

101

Fig. 9. (a–c) EBSD orientation color maps showing the high and low angle grain boundary distributions along the thickness direction for the weld obtained at the welding speed of 400 mm/min. Note that the black lines correspond to high angle grain boundaries (HAGBs) having misorientation angles higher than 151 while the red lines correspond to low angle grain boundaries (LAGBs) having misorientation angles between 11 and 151. (For interpretation of the references to color in this figure caption, the reader is referred to the web version of this paper.)

32


800 700 600 500 400 300

600 500 400

AS

0 -30

-20

100

RS -10

0

10

20

0 -30

30

Distance from weld centre (mm)

-20

-10

0

800

V = 300 mm/min

700

10

20

30

V = 400 mm/min

700

600

Hardness (Hv)

Hardness (Hv)

RS

AS


800

500 400 300

600 500 400 300 200

200 100

BM

300 200

200 100

V = 200 mm/min

700

Hardness (Hv)

Hardness (Hv)

800

V = 100 mm/min

AS

0 -30

-20

100

RS -10

0

10

20

30


0 -30

RS

AS -20

-10

0

10

20

30


Fig. 10. Vicker's microhardness contours at weld mid-sections for welds obtained at different welding speeds and a constant tool rotational speed of 400 rpm.

orientation color maps showing the distribution of the high and low angles are shown in Fig. 9. At the top region, the presence of the lath martensite can be clearly seen. In low carbon alloys steel (0.00026–0.4%C), each block consists of lathes of two specific K–S (Kurdjumov–Sachs) variant groups which are misoriented by small angles of about 101 [34]. From Fig. 9a, the detected line boundaries between the lath martensite are around 101 (shown by red lines), which suggests the presence of these variants groups. Also from Fig. 9a, the LAGBs are parallel to the HAGBs within the lath martensite. Based on these observations, it is confirmed that the peak temperature at the top region is above the A1 temperature and the resulting microstructure is from the phase transformation. On the other hand, the bottom region shows finer ferrite grains with the misorientation angle distribution of LAGBs scattered from 11 to 81. Also, the distributions of the LAGBs are not parallel to the HAGBs. These characteristics show that severe plastic deformation occurred in the bottom region. 3.3. Mechanical properties Fig. 10 shows Vicker's microhardness plots of the weld midsections as a function of the distance from the weld center-line for welds obtained at the welding speeds of 100, 200, 300, and 400 mm/min. It can be clearly seen that the average hardness in the SZs decreases with the increasing welding speed. The average hardness obtained at the welding speeds of 100 and 200 mm/min shows the hardness of typical martensitic structures formed in the low carbon alloys steel [33]. Note that the higher peak temperature at the lower welding speed is responsible for the increase in

the austenite volume fraction, which on transformation results in a high volume fraction of martensite during cooling. Thus, the higher average hardness at the welding speeds of 100 and 200 mm/min is on expected line. Furthermore, the width of the SZs as inferred by the hardness contours also decreases with the increasing welding speed. Interestingly, there is a significant increase in hardness on the advancing side as compared to the retreating side for weld obtained at the welding speed of 400 mm/ min. Note that the formation of the hardening phase (martensite) on the advancing side is believed to be the main reason for this. Fig. 11 shows the engineering stress strain curves and bar graphs representing the ultimate tensile strength (UTS) and ductility of the specimens prepared along the transverse and longitudinal directions for welds obtained at the welding speeds of 100, 200, 300, and 400 mm/min, respectively. Note that the longitudinal specimen (01) contains properties only from the stir zone, while the transverse specimen (901) also contains the base metal properties. Also note that the base metal tensile specimens were prepared from both the transverse and longitudinal directions in order to compare them with the corresponding weld specimens. It can be clearly seen from the engineering stress– strain curves that the strain hardening responses are significantly affected by the welding speeds. Furthermore, it may be noted that since the HAZs are not affected by the welding conditions, therefore, the strain hardening responses of the transverse weld specimens are mainly governed by the shape and size of the SZs and hardness distribution across the weld cross-section (see Fig. 11a). Imam et al. [26] reported in detail about the effect of the shape and size of the SZ and hardness distribution across weld cross-


1000

6 00 5 00

V = 400 mm/min

4 00

BM 90 0

V = 300 mm/min

3 00

Elongation

800

20

600

15

400

10

200

5

2 00 Transverse weld specimen

1 00

0

0

0 0

0 .0 5

0 .1

0 .1 5

0 .2

0 .2 5

BM 90 0 100

0.30

Engineering strain 1800

V = 100 mm/min

1600

1200 V = 300 mm/min

800

V = 200 mm/min

600 400

BM 0 0

200

Longitudinal weld specimen

0 0

0.05

0.10

0.15

0.20

300

400

0.25

25

1800 1600

V = 400 mm/min

1400

1000

200


UTS (MPa)

Engineerin stress (MPa)

25 UTS

Elongation (%)

V = 200 mm/min

UTS

Elongation 20

1400 1200

15

1000 800

10

600 400

Elongation (%)

V = 100 mm/min

UTS (MPa)

Engineerin stress (MPa)

8 00 7 00

33

5

200 0

0.30

Engineering strain

BM 0 0

100

200

300

400

0


Fig. 11. (a, c) Engineering stress–strain curves of transverse and longitudinal weld specimens, and (b, d) bar graphs representing ultimate tensile strength (UTS) and % elongation for joints obtained at different welding speeds. Note that the uniform elongation is the engineering strain at the load maximum.

section on the ductility of the FSW joint. In addition, the strain hardening responses of the longitudinal specimens are significantly affected by the volume fraction of the martensite (see Fig. 11c). Interestingly, strain hardening response of the weld obtained at the 400 mm/min welding speed is quite different from the other welds. At this welding speed, microstructure has a large amount of ferrite which is typically soft and lower work hardening rate. As shown in the bar graphs (see Fig. 11b and d), the optimum strength and ductility are obtained at the welding speed of 400 mm/min in both the transverse and longitudinal specimens. These results clearly show the advantage of the ferrite based duplex structure in the weld of medium carbon steel.

4. Conclusions Parameters controlling the microstructural evolution were investigated during the friction stir welding of medium carbon low alloy S45C steel sheets. It was shown that despite making any compromise with the weld productivity, improved strength and ductility can be achieved. The following conclusions are drawn from this study:

bottom region microstructure consists of finer ferrite grains. This microstructural variation can be linked to the peak temperatures differences between the top and bottom regions. A non-contact thermal imaging system was seen to be an effective tool for the online monitoring of the FSW process such as tool failure, excessive flash formation, and discontinuous welds.

Acknowledgements This study was supported by the New Energy and Industrial Technology Development Organization (NEDO). The Global COE Program and a Grant-in-Aid for Science Research from the Japan Society for Promotion of Science are also acknowledged. The materials were provided by Nippon steel and Sumitomo Metal Co., Ltd. The support provided by the students of this laboratory is also greatly appreciated. I am also grateful to Dr. Y.F. Sun, Mr. Miura, and Mr. Yokochi for their guidance and cooperation. References

The volume fraction of martensite and average hardness in the

SZs decrease with the increasing welding speed. The optimum strength and ductility combination is achieved when the volume fraction of martensite is about 33% with the remaining ferrite area. The peak temperatures decrease and cooling rates increase with the increasing welding speed. The fraction of the LAGBs decrease with the increasing welding speed. At the welding speed of 400 mm/min, the stir zone on the top region shows the formation of lath martensite, while the

[1] W.M. Thomas, E.D. Nicholas, J.C. Needham, M.G. Church, P. TempleSmith, C.J. Dawes, GB Patent Application no. 9125978.9, 1991. [2] R. Nandan, T. Debroy, H.K.D.H. Bhadeshia, Prog. Mater. Sci. 53 (2008) 980–1023. [3] P.L. Threadgill, A.J. Leonard, H.R. Shercliff, P.J. Withers, Int. Mater. Rev. 54 (2) (2009) 49–93. [4] M. Imam, V. Racherla, K. Biswas, Mater. Des. 64 (2014) 675–686. [5] M. Imam, V. Racherla, K. Biswas, J. Int. Adv. Manuf. Technol. (2014) 1–15. http: //dx.doi.org/10.1007/s00170-014-6617-6. [6] C.A. Maltin, L.J. Nolton, J.L. Scott, A.I. Toumpis, A.M. Galloway, Mater. Des. 64 (2014) 614–624. [7] T.J. Lienert, W.L. Stellwag, B.B. Grimmett, R.W. Warke, Weld. J. 82 (2003) 1s–9s. [8] R.S. Mishra, Z.Y. Ma, Mater. Sci. R. Eng. 50 (2005) 1–78.

34


[9] L. Cui, H. Fujii, N. Tsuji, K. Nakata, K. Nogi, R. Ikeda, M. Matsushita, ISIJ Int. 47 (2) (2007) 299–306. [10] M. Imam, K. Biswas, V. Racherla, Mater. Des. 52 (2013) 730–739. [11] A. Simar, Y. Brechet, De.B. Meester, A. Denquin, C. Gallais, T. Pardoen, Prog. Mater. Sci. 57 (2012) 95–183. [12] S. Swaminathan, K. Oh-ishi, A.P. Zhilyaev, C.B. Fuller, B. London, M. W. Mahoney, T.R. Mcnelley, Metall. Mater. Trans. A 41A (2010) 631–640. [13] H. Fujii, L. Cui, N. Tsuji, M. Maeda, K. Nakata, K. Nogi, Mater. Sci. Eng. A. 429 (2006) 50–57. [14] M. Ghosh, Md.M. Husain, K. Kumar, S.V. Kailas, J. Mater. Eng. Perform. 22 (2013) 3890–3901. [15] V. Manvatkar, A. De, L.-E. Svensson, T. Debroy, Scr. Mater 94 (2014) 36–39. [16] K.H. Song, H. Fujii, K. Nakata, Mater. Des. 30 (2009) 3972–3978. [17] Y.F. Sun, Y. Konishi, M. Kamai, H. Fujii, Mater. Des. 47 (2013) 842–849. [18] S.A. Khodir, Y. Morisada, R. Ueji, H. Fujii, Mater. Sci. Eng. A 558 (2012) 572–578. [19] Y.D. Chung, H. Fujii, R. Ueji, N. Tsuji, Scr. Mater. 63 (2010) 223–226. [20] P. Xue, Y. Komizo, R. Ueji, H. Fujii, Mater. Sci. Eng. A 606 (2014) 322–329. [21] L. Cui, H. Fujii, N. Tsuji, K. Nogi, Scr. Mater. 56 (2007) 637–640. [22] J. Trzaska, L.A. Dobrzanski, J. Mater. Process. Technol. 192–193 (2007) 504–510. [23] O.T. Midling, O Grong, Acta Metall. Mater. 42 (1994) 1595–1609, 1611–22. [24] O. Frigaard, O. Grong, O.T. Midling, Metall. Mater. Trans. A 32A (2001) 1189–1200. [25] H. Lombard, D.G. Hattingh, A. Steuwer, M.N. James, Eng. Fracture Mech. 75 (2008) 341–354.

[26] M. Imam, K. Biswas, V. Racherla, Mater. Des. 44 (2013) 23–34. [27] W.J. Arbegast, P.J. Hartley, In: Proceedings of the Fifth International Conference on Trends in Welding Research. Pine Mountain, GA, USA, June 1–5, 1998, p. 541. [28] T. Saeid, A. Abdollah-zadeh, H. Assadi, F.M. Ghaini, Mater. Sci. Eng. A 496 (2008) 262–268. [29] M. Esmailzadeh, M. Shamanian, A. Kermanpur, T. Saeid, Mater. Sci. Eng. A 561 (2013) 486–491. [30] M.B. Bilgin, C. Meran, Mater. Des. 33 (2012) 376–383. [31] R. Rai, A. De, H.K.D.H. Bhadeshia, T. DebRoy, Sci. Technol. Weld. Join. 16 (4) (2011) 325–342. [32] N. Koga, N. Nakada, T. Tsuchiyama, S. Takaki, M. Ojima, Y. Adachi, Scr. Mater. 67 (2012) 400–403. [33] G. Krauss, Mater. Sci. Eng. A 273-275 (1999) 40–57. [34] S. Morito, X. Huang, T. Furuhara, T. Maki, N. Hansen, Acta Mater. 54 (2006) 5323–5331. [35] Y.S. Sato, T.W. Nelson, C.J. Sterling, Acta Mater. 53 (2005) 637–645. [36] Y.C. Chen, H. Fujii, T. Tsumura, Y. Kitagawa, K. Nakata, K. Ikeuchi, K. Matsubayashi, Y. Michishita, Y. Fujiya, J. Katoh, J. Nucl. Mater. 420 (2012) 497–500. [37] S.H.C. Park, Y.S. Sato, H. Kokawa, K. Okamoto, S. Hirano, M. Inagaki, Scr. Mater. 49 (2003) 1175–1180.