Linear Friction Stir Welding of Medium Carbon Steel at Low Temperature

Abstract

Linear Friction Welding (LFW) is a solid-state joining process in which a joint is obtained through the relative motion of two components under a high contact load. The most important factor of this conventional method is to obtain a fresh surface at the interface by expelling the weld interface as flash. In this study, medium carbon steel was welded by LFW at a low frequency, low amplitude and high applied pressure. As a result of the temperature measurements and microstructure observations, the maximum temperature of the weld plane was confirmed to be below the A₁ transformation temperature, and martensitic transformation was suppressed at the weld interface. The key concept of this method is applying a large strain deformation to the interfaces to recrystallize at a lower temperature which is different from the conventional LFW.

1. Introduction

In recent years, the global warming issue has attracted more and more attention all over the world. Weight savings and strengthening of transportation equipments have been taken as effective strategies to reduce the CO₂ emissions. Medium and high carbon steels can provide high strength and high ductility at a low cost by increasing the carbon content and microstructure control. Therefore, the medium and high carbon steels are promising structural materials which may satisfy the requirements of weight-saving and high-strength strategies for transportation equipments. However, for the conventional fusion welding of the steels with the carbon content over 0.3 mass%, since the welding temperature is quite high, the transformation to martensite, which is brittle, will occur during the post-weld cooling, which deteriorates the mechanical properties of the joint.¹⁾ Therefore, it is important to establish a welding technique which can be performed at low temperature without deteriorating the mechanical properties of the joint. Recently, many studies have been conducted on the solid-state joining processes in which the welding can be performed at a low temperature, so that the transformation to martensite that occurs during cooling can be suppressed to avoid embrittlement of the joint.^2,3,4)

Linear Friction Welding (LFW) is one of the solid-state joining processes, in which a joint is obtained through the relative motion of two components under a high contact load. The first patent of the LFW process was applied in 1929.⁵⁾ Although the description of this patent was vague, the patent specialized for the LFW equipment was licensed in 1969.⁶⁾ After the LFW equipment was manufactured in The Welding Institute (TWI) in the early 1980s, the development of this LFW technique was progressed mainly focusing on the aerospace-applied titanium and nickel alloys.^{7,8,9,10,11,12,13)} Based on this background, there is currently no effective basic patent for the LFW process.¹⁴⁾ Therefore, LFW has a more preferable condition for expanding its applications in comparison to other new joining methods such as friction stir welding (FSW).

The LFW process has several attractive characteristics, such as that no tool is necessary for the welding, an extremely short welding time, a high dimensional accuracy of the joint with a high reproducibility,¹⁵⁾ and applicability to the non-axial symmetrical objects. Furthermore, it is known that the material that is softened during welding is expelled to the outside from joint interface as flash in the LFW process, which is similar to that in a rotary friction welding (RFW) process.^16,17) On the other hand, this LFW process has mainly been applied to the bulk-shaped components,^{18,19,20,21,22,23)} but hardly applied to thin plates.

In this study, the LFWed joints were examined in order to confirm the new potential of this LFW process for medium and high carbon steels. However, because the welding temperature increases to the γ single-phase temperature range during the conventional LFW process for the medium and high carbon steels, the martensitic transformation will occur in the joint during cooling as already mentioned, thus making the joint brittle. Therefore, it is considered that if the welding can be performed at a temperature lower than the A₁ point (723°C) to inhibit the α to γ phase transformation during welding, the martensitic-transformation-induced embrittlement can be avoided and a sound weld joint can thus be obtained. In a previous study, the authors successfully fabricated the FSW steel joints with high toughness regardless of the carbon content by performing the welding at a temperature below the A₁ point to suppress the brittle martensitic transformation.^24,25,26) However, for the FSW of steels, tool life remains as a big issue,^27,28,29,30) and there are still many problems to be overcome for its practical applications.

In this study, the objective was to develop a new joining method combining the LFW process with the low-temperature (below A₁) welding process, that is, to develop a linear friction stir welding (LFSW) method in which a large strain can be introduced into the weld interface based on the principle of the LFW process. Specifically, we are developing a technique to join materials via recrystallization achieved by a large strain induced at low temperature, which is different from the conventional one in which the main joining mechanism is the formation of fresh surfaces at the joint interface caused by the expelling of softened materials to the outside from joint interface as flash. By applying this novel technique, it is possible to weld carbon steels at a temperature below the A₁ point without using a tool such as the rotating tool used in FSW.

2. Experimental Procedure

In this study, a JIS-S45C steel plate with the dimensions of 5 mm × 25 mm × 62 mm was used as the base metal. The 5 mm × 25 mm surface was used as the joint surface. The chemical composition of the base metal is shown in Table 1. The microstructure of the base metal consists of ferrite and pearlite as shown in Fig. 1.

Table 1. Chemical composition of S45C.

	Chemical composition (mass%)
Material	C	Si	Mn	P	S
S45C	0.48	0.23	0.77	0.02	0.00

Fig. 1.

SEM micrograph of the base metal.

Equation (1) shows the friction heat input, q, generated during the LFW¹⁴⁾   

$$q=v_t\times \mu P=2\pi fA cos(2\pi ft)\times \mu P$$(1)

in which v_t is the sliding velocity, f is the frequency, A is the amplitude, t is the welding time, μ is the friction coefficient, and P is the applied pressure. The welding temperature during the LFW, which is associated with the friction heat input, is thus known to be mainly affected by the frequency, amplitude and applied pressure according to the equation.

The welding conditions used in this study are shown in Table 2. The LFW processing was conducted under varying frequencies with a constant applied pressure and varying the applied pressures with a constant frequency in order to investigate the effect of the processing conditions on the welding temperature. Based on the obtained finding, the LFW was performed under a condition in which a low welding temperature can be achieved. Figure 2 shows the appearance of the LFW equipment and the schematic illustration of the specimen with both the oscillating and forging directions.

Table 2. Linear Friction Welding conditions of this study.

Parameter	Conditions
Frequency (Hz)	15, 25, 50
Amplitude (mm)	± 1, ± 2
Applied pressure (MPa)	50, 100, 200, 300
Control method	Time control

Fig. 2.

Appearance of the equipment for the Linear Friction Welding and schematic illustration of the specimen. (Online version in color.)

The temperature at the joint surface during the LFW was measured by an infrared camera (CPA-T640, Chino). The cross-section specimen was cut from the obtained joint as illustrated in Fig. 3. The macrostructure and microstructure of the cross-section specimen were observed by an optical microscope (OM) and a scanning electron microscope (SEM) with electron backscatter diffraction (EBSD) analysis (JSM-7001FA, JEOL), respectively. The Vickers hardness test was also performed on the cross-section specimen at a load of 0.98 N for a dwell time of 15 s. For the macrostructure observation, the specimen was mechanically polished with abrasive papers followed by a 1 μm diamond finish, then chemically etched in a 3% nital solution for 8–10 s. For the microstructure observation, the specimen was mechanically polished in a similar manner, then subjected to an electro-polishing in a solution composed of 90% acetic acid and 10% perchloric acid under an applied potential of 20–30 V for 10–20 s. Furthermore, the tensile tests were performed under a crosshead speed of 0.02 mm/s at room temperature in order to evaluate the tensile properties of the obtained joints.

Fig. 3.

Schematic illustration of the sampling method. (Online version in color.)

3. Results and Discussion

3.1. Effect of Welding Parameters on the Welding Temperature

Figure 4 shows the optical macrographs of the cross sections of the joints fabricated under various frequencies with the constant applied pressure of 200 MPa. Because of the symmetric characteristic of the joint microstructure, only half of the cross sections of the joints are shown here. In all the obtained joints, flashes formed on the outside of the joints along the oscillating direction, and no defects or unbonded regions are observed in the entire joint areas. The white etched region observed in the joint fabricated at the frequency of 50 Hz shows a decrease in the area fraction with the decreasing frequency. This result suggests that the microstructures of the joints change with the decreasing frequency.

Fig. 4.

Optical micrographs of the joints welded under various frequency conditions at 200 MPa applied pressure and ± 2 mm amplitude.

SEM micrographs of the center and edge parts of the joints fabricated under various frequencies with a constant 200 MPa applied pressure are shown in Fig. 5. Martensite is observed in all the obtained joints. These regions where martensite is observed are found to correspond to the white etched regions shown in Fig. 4. Thus, the area fraction of the martensite observed in the joints obviously decreases with the decreasing frequency, and most of the area shows a microstructure composed of fine ferrite and spheroidal cementite at the joint fabricated at 15 Hz. This result suggests that the martensite formation can be suppressed in the joints by decreasing the frequency.

Fig. 5.

SEM micrographs at the center and edge parts of the joints welded under various frequency conditions at 200 MPa applied pressure and ± 2 mm amplitude.

Figure 6 shows the peak welding temperature of the joint surfaces measured by the infrared camera as a function of the frequency and the applied pressure. It is found that the peak welding temperature of the joint surfaces decreases with the decreasing frequency. This is considered to be attributed to a decrease in the heat input due to a decrease in the relative velocity during the friction, which is also in agreement with the phenomenon predicted from Eq. (1). However, the peak welding temperature is found to decrease with the increasing applied pressure, which is in contrast with the Eq. (1). In the LFW process, after the two materials come into contact under a specific pressure, the vibration starts and the temperature increases due to the frictional heat. After starting the oscillation, the strength of the interface material decreases with the increasing welding temperature. When the strength of the material at the joint interface falls below the applied pressure, the material starts to deform and it is expelled as flash. Therefore, when the applied pressure is increased, the peak welding temperature deceases because the material at the joint interface can be deformed at the lower temperature. Based on these observations, it is concluded that the welding temperature decreases as the applied pressure increases in the LFW process, in which the joining is achieved only by the friction between the materials without the introduction of external energy such as the rotating tool in the FSW process. Thus, the processing condition with a low sliding velocity and a high applied pressure is known to be effective in order to achieve a low-temperature LFW processing.

Fig. 6.

Maximum temperature of the joint surface during the LFW. (Online version in color.)

3.2. Verification of Low Welding Temperature Caused by the New Welding Method

The LFW process under a low-velocity high-pressure condition with a frequency of 15 Hz, an amplitude of 1 mm, and an applied pressure of 300 MPa was performed in order to verify the achievement of the low welding temperature. Optical and SEM micrographs of half-cross-sections of the joints are shown in Fig. 7. Both the center and edge regions of the joint show a microstructure consisting of fine ferrite and spheroidal cementite. It has been reported that the spheroidal-cementite microstructure can be obtained when the microstructure containing ferrite and pearlite subjected to plastic deformation is annealed at a temperature below the A₁ point.^24,25,26) It is thus indicated that the joining at a temperature lower than the A₁ point without martensitic transformation can be achieved by the LFW. This low-temperature joining method below the A₁ point is considered to be a very effective method for joining high carbon steel in which remarkable embrittlement occurs due to the formation of martensite. The peak welding temperature measured by an infrared camera under this LFW condition was 716°C, which was confirmed to be lower than the A₁ point.

Fig. 7.

Optical and SEM micrographs of half-cross-sections of the joints welded under low temperature condition at 300 MPa applied pressure, 15 Hz frequency and ± 1 mm amplitude: (a) SEM micrograph at edge, (b) optical micrograph and (c) SEM micrograph at center.

3.3. Microstructure and Mechanical Properties of the Joints

Both the center and the edge regions of the joint show an average grain size of around 1 μm, which is much smaller than that of the base material. This microstructure formation is due to the recrystallization which is caused by the temperature rise and large plastic deformation that occurs in the joint. This result is similar to our previous FSW results.^25,26) Figure 8 shows the inverse pole figures (IPF) of the center and edge regions of the joint fabricated at the low-temperature condition. In the center region of the joint, the {101} plane is remarkably observed and the {111} plane is only partially observed, while in the edge region of the joint, the {111} plane is predominantly observed in addition to the {101} plane.

Fig. 8.

High magnification IPF of half-cross-sections of the joints welded under low temperature condition at 300 MPa applied pressure, 15 Hz frequency and ± 1 mm amplitude: (a) center and (b) edge.

The corresponding pole figures of the center and edge regions of the obtained joint are shown in Fig. 9. A strong texture of D1 and D2 was observed in the center region of the joint, while in the edge region of the joint, the texture of J and J was also partially observed in addition to the texture of D1 and D2 similar to that of the center region of the joint. These textures can be identified as the simple shear texture of the BCC metal, which suggests that the shear deformation occurs at the joint interface and the {112} <111> dislocation slip predominantly occurs during the low-temperature LFW. In addition, it is considered that the welding temperature is lower at the edge region compared to that of the center region of the joint, while the cooling rate at the edge region is higher. It has been reported that, during the recrystallization of the post-annealing of the deformed carbon steel, the textures of D1, D2, J and J can be retained when the cooling rate is high, while the textures of J and J would disappear as they become the nucleus of recrystallization when the cooling rate is low.³¹⁾ Therefore, it is considered that the difference in texture between the center and edge regions of the joint is a result from the difference in the cooling rate.

Fig. 9.

Iron-Alpha Pole Figure of the joints welded under low temperature condition at 300 MPa applied pressure, 15 Hz frequency and ± 1 mm amplitude: (a) center and (b) edge.

Figure 10 shows the Vickers hardness distributions of the LFWed joints along the central axis and the joint interface. Symbol “△” shows the result of the condition at which the highest welding temperature was obtained with a frequency of 50 Hz, an amplitude of 2 mm and an applied pressure of 200 MPa, while symbol “〇” shows the result of the low-temperature condition having a frequency of 15 Hz, an amplitude of 1 mm and an applied pressure of 300 MPa. At the high temperature condition, a hardened region with a hardness of around HV 620 was observed in the center region of the joint. This value is close to the hardness of martensite formed in the quenched S45C,³²⁾ and the width of the hardened region also corresponds to the width of the white etched martensite region observed in Fig. 4. Based on the hardness distribution along the joint interface, it can be concluded that the martensite is formed throughout the entire joint interface under the high temperature condition. On the other hand, although the hardness of the joint fabricated at the low temperature condition is slightly higher than that of the base material, noticeable hardened regions as described above are not seen in both the forging direction and the oscillation direction. Since the microstructure composed of fine ferrite and spheroidized cementite can also be observed in the vicinity of the joint interface as shown in Fig. 7, it is clarified that the martensitic transformation can be suppressed and the joint hardness can be lowered by using the low temperature condition.

Fig. 10.

Vickers hardness profiles of LFWed joints: (a) along the central axis, and (b) along the interface. (Online version in color.)

Figure 11 shows the appearance and the fracture surface of the tensile fractured joint fabricated at a temperature below the A₁ point (300 MPa applied pressure, 15 Hz frequency and 1 mm amplitude). The obtained joint shows a tensile strength of 766 MPa, which is similar to that of the base metal, and an elongation of 15%. The tensile fracture site of this joint fabricated at low temperature is seen to be the base material rather than the joint which is slightly stronger than the base material due to the microstructural refinement. Moreover, the fracture surface of the joint shows a dimple morphology, which suggests that this low-temperature joining method is obviously more superior compared with the conventional joining method fabricated in which the joint shows a brittle fracture at the joint interface.

Fig. 11.

(a) Appearance and (b) fracture surface of the joint after the tensile strength test. (Online version in color.)

During the joining under the condition of a low velocity and high applied pressure found in this study, the formation of martensite in the joint, which cannot be avoided in the conventional LFW, can be completely suppressed. Since this method can be used for the carbon steels regardless of the carbon content, it is considered to be a novel effective joining method for the medium and high carbon steels.

4. Conclusions

In this study, a LFSW process, in which LFW was applied in order to introduce severe plastic deformation into the joint interface at a low temperature to achieve a joining by recrystallization, was performed on medium carbon steel. The obtained findings are as follows:

(1) The martensite formation was successfully suppressed at the joint interface by decreasing the welding temperature to a level lower than the A₁ point by reducing the frequency and increasing the applied pressure.

(2) Under the processing condition with low velocity and high applied pressure (300 MPa applied pressure, 15 Hz frequency and 1 mm amplitude), the microstructure consisting of ferrite and spheroidized cementite was observed throughout the entire joint, which indicated that the peak welding temperature was below the A₁ point.

(3) A BCC simple shear texture was observed at the joint interface fabricated at the low-temperature condition. This suggested that the shear deformation occurred and the {112} <111> dislocation slip predominantly occurred at the joint interface during the low-temperature LFW.

(4) Under the high-temperature condition, the hardness remarkably increases in the entire joint due to the martensite formation, while for the low-temperature condition, a noticeable increase in the hardness is not seen over the entire joint due to the suppression of martensite formation. The joint fabricated at the low-temperature condition fractured at the base material during the tensile test and showed a dimple morphology on the fracture surface.

Based on the above results, it is concluded that a novel joining method can be established by which the carbon steels can be welded regardless of the carbon content and the tool life issue in FSW can be totally ignored.

Acknowledgments

This study was partly supported by the New Energy and Industrial Technology Development Organization (NEDO) under the “Innovation Structural Materials Project (Future Pioneering Projects)”.

References

• 1)   L. P.  Connor,  R. L.  O’Brien and  W. R.  Oates: Welding Handbook, Vol. 4, American Welding Society, Miami, (1998), 11.
• 2)   H.  Fujii,  R.  Ueji,  Y.  Morisada and  H.  Tanigawa: Scr. Mater., 70 (2014), 39.
• 3)   T.  Miura,  R.  Ueji,  H.  Fujii,  H.  Komine and  J.  Yanagimoto: Mater. Des., 90 (2016), 915.
• 4)   S.  Yoon,  R.  Ueji and  H.  Fujii: J. Mater. Process. Technol., 229 (2016), 390.
• 5)   W.  Richter: Creation of an adhesive bond between plaques of tool steel and their supports in the manner of welding or soldering, Germany Patent DE477084, (1929) (in German).
• 6)   R.  Maurya and  J.  Kauzlarich: Bonding apparatus - friction welding by reciprocal motion, UK Patent GB1161800 A, (1969).
• 7)   A.  Vairis and  M.  Frost: Wear, 217 (1998), 117.
• 8)   P.  Wanjara and  M.  Jahazi: Metall. Mater. Trans. A, 36 (2005), 2149.
• 9)   C.  Mary and  M.  Jahazi: Adv. Mater. Res., 15–17 (2007), 357.
• 10)   C.  Mary and  M.  Jahazi: Adv. Eng. Mater., 10 (2008), 573.
• 11)   J.  Romero,  M. M.  Attallah,  M.  Preuss,  M.  Karadge and  S. E.  Bray: Acta Mater., 57 (2009), 5582.
• 12)   E.  Dalgaard,  P.  Wanjara,  J.  Gholipour,  X.  Cao and  J. J.  Jonas: Acta Mater., 60 (2012), 770.
• 13)   T. J.  Ma,  X.  Chen,  W. Y.  Li,  X. W.  Yang,  Y.  Zhang and  S. Q.  Yang: Mater. Des., 89 (2016), 85.
• 14)   I.  Bhamji,  M.  Preuss,  P. L.  Threadgill and  A. C.  Addison: Mater. Sci. Technol., 27 (2010), 2.
• 15)   H.  Kuroki,  K.  Nezaki,  G.  Wakabayashi and  K.  Nakamura: IHI Giho, 53 (2013), 45 (in Japanese).
• 16)   M.  Kimura,  M.  Choji,  M.  Kusaka,  K.  Seo and  A.  Fuji: Sci. Technol. Weld. Join., 10 (2005), 406.
• 17)   M.  Kimura,  Y.  Ohtsuka,  M.  Kusaka,  K.  Seo and  A.  Fuji: Q. J. Jpn. Weld. Soc., 23 (2005), 577 (in Japanese).
• 18)   A.  Vairis and  M.  Frost: Mater. Sci. Eng. A, 292 (2000), 8.
• 19)   F.  Schröder,  R. M.  Ward,  A. R.  Walpole,  R. P.  Turner,  M. M.  Attallah,  J. C.  Gebelin and  R. C.  Reed: Mater. Sci. Technol., 31 (2015), 372.
• 20)   W. Y.  Lia,  S. X.  Shia,  F. F.  Wang,  T. J.  Ma,  J. L.  Lia,  D. L.  Gao and  A.  Vairis: Int. J. Therm. Sci., 67 (2013), 192.
• 21)   T. J.  Ma,  W. Y.  Li and  S. Y.  Yang: Mater. Des., 30 (2009), 2128.
• 22)   A. R.  McAndrewa,  P. A.  Colegrove,  A. C.  Addison,  B. C. D.  Flipo,  M. J.  Russell and  L. A.  Lee: Mater. Des., 87 (2015), 1087.
• 23)   W. Y.  Li,  F. F.  Wang,  S. X.  Shi,  T. J.  Ma, J. L. Li and A. Vairis: J. Mater. Eng. Perform., 23 (2014), 4010.
• 24)   T.  Furuhara,  T.  Mizoguchi and  T.  Maki: ISIJ Int., 45 (2005), 392.
• 25)   H.  Fujii,  L.  Cui,  N.  Tsuji,  M.  Maeda,  K.  Nakata and  K.  Nogi: Mater. Sci. Eng. A, 429 (2006), 50.
• 26)   L.  Cui,  H.  Fujii,  N.  Tsuji and  K.  Nogi: Scr. Mater., 56 (2007), 637.
• 27)   T.  Miyazawa,  Y.  Iwamoto,  T.  Maruko and  H.  Fujii: Sci. Technol. Weld. Join., 17 (2012), 213.
• 28)   Y. S.  Sato,  Y.  Nagahama,  S.  Mironov,  H.  Kokawa,  S. H. C.  Park and  S.  Hirano: Scr. Mater., 67 (2012), 241.
• 29)   M.  Collier,  R.  Steel,  T. W.  Nelson,  C.  Sorensen and  S.  Packer: Mater. Sci. Forum, 426 (2003), 3011.
• 30)   R.  Ohashi,  M.  Fujimoto,  S.  Mironov,  Y. S.  Sato and  H.  Kokawa: Sci. Technol. Weld. Join., 14 (2009), 221.
• 31)   W. B.  Hutchinson: Int. Mater. Rev., 29 (1984), 25.
• 32)   G.  Krauss: Mater. Sci. Eng. A, 273–275 (1999), 40.