2024 Volume 64 Issue 12 Pages 1795-1803
Thermal hysteresis in fusion welding causes significant weld deterioration in medium- and high-carbon steels. Therefore, the development of an effective alternative welding process is required. Friction stir welding (FSW) is a solid-state welding process performed in an atmosphere that reduces the risks associated with melting and solidification of metals, making it an effective alternative method. Furthermore, it facilitates a flexible in-process control of heat input, which can be achieved by controlling the welding parameters. Considering these, the authors conducted a series of studies to elucidate the characteristics of FSW for medium- and high-carbon steels, including high-strength tempered steels.
This paper presents the results of applying FSW to 1.4 GPa-grade tempered JIS-S55C steel plates. Five distinct weld types were created by varying the welding parameters, including tool rotation and welding speed. The temperature of the interface between the tool and in-process material was measured using a thermal imaging camera. The microstructure of the welds was evaluated using optical microscopy and field-emission scanning electron microscopy (FE-SEM) with an electron-backscatter diffraction (EBSD) measurement system. The mechanical properties of the welds were evaluated through Vickers hardness and tensile tests. Digital image correlation analysis was employed to analyze the local deformation during the tensile test.
In recent years, the demand for thinner and lighter structural materials for transportation equipment has increased to reduce carbon dioxide emissions. From this perspective, innovative structural materials that impart strength and ductility to medium- and high-carbon steels via quenching and tempering, without the addition of alloying elements have been investigated.1,2,3,4) However, the mechanical properties of medium- and high-carbon steels are adversely affected by the thermal hysteresis during fusion welding. Therefore, the development of an effective alternative welding process remains a significant challenge. In particular, the deterioration of the mechanical properties of medium- and high-carbon steels due to the thermal hysteresis of fusion welding5,6,7,8) necessitates the formulation of an effective alternative welding process.3,9,10)
Friction stir welding (FSW) is a solid-state welding process that can be applied in the ambient atmosphere and exhibits a high degree of welded joint flexibility, highlighting its potential as an effective welding process. Additionally, heat input can be regulated in real time by controlling two key welding parameters: tool rotation speed and welding speed.11,12,13) However, the durability of the tools used in FSW14) is a significant obstacle that limits its wider application for steel materials. Nevertheless, the development of tool materials and shapes with sufficient strength and durability for welding steel has paved the way to resolve these issues.15,16,17,18) FSW is distinguished by its lower welding temperature and superior temperature controllability compared with conventional fusion welding.
Substantial information exists on the use of transformation points to control welded joint structure in steel FSW.19,20,21,22,23,24,25) This approach is expected to yield superior mechanical properties during the fabrication of welded joints. Fujii et al. reported the study of welding parameters and microstructure and properties of FSW-welded joints for various carbon steels as an area of international interest.12,13,26,27,28,29,30) However, the application of FSW to high-strength and -ductility steel materials based on medium-carbon steels with few alloying elements remains a nascent field of study.
From this perspective, the authors conducted a series of studies31) to elucidate the characteristics of FSW joints using medium- and high-carbon steels as the base and tempered material, as classified in JIS-S35C to S55C. This paper presents the FSW of JIS-S55C steel plates tempered to 1.4 GPa. The welding parameters, namely rotation and welding speeds, were varied at 100 rpm–100 mm/min, 200 rpm–100 mm/min, 300 rpm–100 mm/min, 400 rpm–100 mm/min, and 400 rpm–400 mm/min. Five weld types with different welding temperatures and thermal hysteresis were investigated. The temperature of the interface of the welding tool and welded material during FSW was quantified using a sophisticated thermal imaging camera. The microstructure of the weld was evaluated by optical microscopy and field-emission scanning electron microscopy (FE-SEM) with an electron backscatter diffraction (EBSD) measurement system. The mechanical properties of the FSW-welded joints were evaluated through Vickers hardness and tensile tests. Furthermore, the local deformation details observed during the tensile tests were analyzed by applying digital image correlation (DIC) analysis.
A thin steel plate with a tensile strength of 1.4 GPa was used as the test steel. The chemical compositions of the test steels are listed in Table 1. As shown in Fig. 1, the test steel used in this study has a needle-like and tempered martensite-dominated microstructure. FSW was performed on the material with a plate size of 150 mm (L) × 28 mm (W) × 1.6 mm (T) under position-controlled one-pass butt welding conditions.
| C | Si | Mn | P | S | Cr | Cu | Ni | Fe |
|---|---|---|---|---|---|---|---|---|
| 0.58 | 0.25 | 0.81 | 0.017 | 0.011 | 0.05 | 0.01 | 0.01 | bal. |
The correlations between the welding parameters and weld microstructure and mechanical properties were also investigated. The tool rotation speed27) was set in the range of 100–400 rpm to control the welding temperature, whereas the welding speed13) was set in the range of 100–400 mm/min to control the cooling rate. A tungsten carbide cylindrical tool with a probe (shoulder diameter of 12 mm, probe diameter of 4 mm, and probe length of 1.4 mm) and without grooves was used. The tool-advance angle was set to 3°, and a shield gas was used to inhibit high-temperature oxidation during welding. The tool–material interface temperatures during FSW were monitored using a thermal image analysis camera (CHINO, CPA-T640) to evaluate the correlation between the welding parameters and temperatures. Table 2 lists the FSW conditions used in this study.
| Welding parameters | ||
|---|---|---|
| Tool rotation speed (rpm) | Welding speeed (mm/min) | |
| (A) | 100 | 100 |
| (B) | 200 | |
| (C) | 300 | |
| (D) | 400 | |
| (E) | 400 | |
For the microstructure of the FSW joints, specimens were obtained to observe the cross-section perpendicular to the welding direction. The macrostructure observation was performed using an optical microscope (digital microscope), and the microstructure observation was performed by FE-SEM. The surfaces of the specimens used for the microstructural evaluation were prepared by wet polishing, followed by etching for 7–10 s in a 3% nital solution at room temperature for the macrostructural observation. For the microstructural observations, the specimens were prepared by electropolishing after dry polishing in a mixed electrolytic solution of perchloric acid and acetic acid (HClO4:CH3COOH = 1:9) at room temperature for 10–30 s at a voltage of 35 V using a DC-stabilized power supply.
2.3. Evaluation of Mechanical PropertiesThe mechanical properties of the welds were evaluated by Vickers hardness and tensile tests. The Vickers hardness test was performed on a cross-section perpendicular to the welding direction at 10 mm from the center of the weld to each end (impact distance set to 0.5 mm with 41 impact points), and three measuring sections were set at 0.3 mm from the weld surface, thickness center of the weld, and 0.3 mm from the bottom of the weld. A load of 2.94 N was applied, and the load-holding time was set to 10 s.
Tensile tests were performed on the specimens along a longitudinal direction perpendicular to the weld line with the stir zone as the gauge section, as shown in Fig. 2. The top and rear surfaces were polished to eliminate the effects of burrs and thickness reduction in the FSW area, resulting in a final thickness of 1.3 mm. The test was conducted at room temperature, and the crosshead speed was set to 1.0 mm/min. The deformation behavior during the test was recorded using a high-speed camera (Photron, FASTCAM SA-X), and DIC analysis was used to evaluate the stress–strain correlation and local deformation distribution.
Figure 3 shows an example of the temperature measurement results obtained at speeds of 300 rpm and 100 mm/min. The actual FSW time is indicated by the arrows. As the temperature measurement was conducted via fixed-point observation, measurement could not be performed in some sections owing to the presence of clamps attached to the FSW machine stage, which obstructed the view during operation. Nevertheless, the temperatures observed throughout the welding process exhibited a notable degree of consistency with a slight increasing trend. The highest temperature recorded during the welding process is represented by the red symbol. The same analysis was conducted for the other welding conditions; the results are presented in Fig. 4 as the maximum welding temperatures in the Fe–Fe3C phase diagram. Thermal image analysis was conducted on the red-heated area at the outer edge of the rotating tool shoulder. The highest temperature was 1175°C at 400 rpm–100 mm/min, followed by 1108, 1051, 991, and 769°C at 300 rpm–100 mm/min, 400 rpm–400 mm/min, 200 rpm–100 mm/min, and 100 rpm–100 mm/min, respectively. This confirms the general trend that the magnitude of the welding temperature is reflected by the increase in the tool rotation speed.
Figure 5 illustrates the results of a macrostructural observation of the welded cross-section obtained using an optical microscope (digital microscope). Observations were conducted following the etching process using nital reagent. Under all conditions, the stir zone reached the rear surface and the formation of the back bead was confirmed by visual observation. All cross-sections exhibited a sound microstructure without defects, such as kissing bonds or voids. Under a low rotation speed (100 rpm–100 mm/min) and high welding speed (400 rpm–400 mm/min), the stir zone on the rear side tends to narrow. This phenomenon can be attributed to the suppression of heat input and subsequent expansion of the temperature gradient between the top and rears sides during welding. In accordance with the welding conditions established in this study, no color transition is observed, indicating either a microstructural transition or formation of band shapes,32,33) which could be attributed to tool wear in the stir zone.
Figure 6 depicts the Vickers hardness distribution of the FSW weld cross section obtained under each welding condition. The hardness distribution is presented as a two-dimensional color map, with the region from 260 to 450 HV divided into ten grades of color tone. The weld interface was placed at the center, with the right side corresponding to the advancing side (AS) the left side corresponding to the retracting side (RS). If the region sufficiently far from the stir zone is considered the base-metal region. The hardness of the base metal can be confirmed to be approximately 400 HV. At 400 rpm–400 mm/min, the hardness of the material in the center of the stir zone increased by approximately 250 HV compared with that of the base metal. Under different conditions, the stir zone tended to soften by approximately 50–100 HV compared to the base metal. Additionally, an area indicating the HAZ softening with a minimum hardness value was observed near the outer edge of the stir zone under all welding conditions. However, at 300 rpm–100 mm/min, the distribution of the softening zone narrowed than that under other conditions.
Figure 7 depicts the DIC results of the strain distribution at the onset of necking based on the image data acquired using the high-speed camera during the tensile test. Figure 8 illustrates the stress–strain diagram generated through DIC analysis. MSYS: GOM Correlate software was used for the analysis of strain distribution.
The DIC analysis results presented in Fig. 7 are organized with the R.S. positioned at the top, and the strain is represented by a color-coded scale of 12 grades. The color warmth increases as the strain in the region increased. The distribution of the regions exhibiting a local strain of more than 8% can be identified at tool rotation speeds of 300 rpm or less. However, at tool rotation speeds of 400 rpm or higher, almost no regions exhibited local strains. This tendency is also clearly demonstrated in the stress–strain diagram in Fig. 8, illustrating the distinction between the welding conditions below 300 rpm, whereby ductile fracture is observed, and those above 400 rpm, whereby brittle fracture is noted. High welding temperatures were obtained under 300 rpm–100 mm/min and 400 rpm–100 mm/min, where the joint efficiency exceeded 70% of the base metal tensile strength ratio. The welding parameters oriented toward lower temperatures, specifically the 100 rpm–100 mm/min and 400 rpm–400 mm/min conditions, which were oriented toward higher cooling rates, were not advantageous for ensuring joint efficiency.
3.5. Cross-sectional Micro-observationFigure 9 shows a cross-sectional SEM image of the center of the stir zone (center of the plate thickness at the weld interface) obtained under each welding condition. Observations were conducted after electropolishing with a mixed perchloric and acetic acid electrolyte solution. The darker coloration observed in the ferrite phase and lighter coloration observed in the bainite phase corresponded to the results at 100 rpm–100 mm/min, as discussed in Section 4.1. Under welding conditions of 300 rpm–100 mm/min, the surface was covered with two distinct microstructures: martensite and bainite. At 400 rpm–100 mm/min, the same microstructure of martensite and bainite was observed. However, the greater martensite fraction was substantiated. Under the condition of 400 rpm–400 mm/min, which was subjected to a higher cooling rate, the martensite fraction tends to dominate. Overall, martensite is observed at welding speeds of 200 rpm or higher, whereas pearlite with lamellar morphology is formed at 200 rpm. However, this tendency was less pronounced at speeds greater than 300 rpm, where bainite formation is more prevalent.
Figure 10 presents an inverse pole figure (IPF) image of the cross-section of the stir zone obtained through EBSD measurements. The measurement was conducted with an acceleration voltage of 20 kV, irradiation current of 35 nA, step size of 0.1 μm, and measurement area of 20 μm × 20 μm per field of view. The red lines indicate low-angle grain boundaries with a crystallographic orientation of 3°–15° between adjacent measurements. Black lines indicate high-angle grain boundaries with a crystallographic orientation of 15° or higher. The vertical and horizontal directions in the figure are aligned along the normal direction (ND) and transverse direction (TD), respectively. At 100 rpm–100 mm/min, fine equiaxed microstructure is extensively distributed across a broad area within the field of view, which is the only condition under which the welding temperature was measured near the A3 transformation point. In the temperature measurements of the welded materials during FSW, a temperature gradient of approximately 100°C is noted between the topmost and rear surfaces, even in the case of thin plates.28) Welding was conducted at a relatively low temperature under these conditions, resulting in the formation of a fine equiaxed microstructure owing to the dynamic recrystallization in the stir-zone microstructure. However, the contrasts shown in Fig. 9(a) indicate the presence of a two-phase microstructure, comprising ferrite (darker tone) and bainite (lighter tone). This assumption is based on the postulated transformation of these microstructures under the conditions of the two-phase region during FSW, resulting in the formation of fine ferrite and austenite grains.
Carbon enrichment34) in austenite grains is postulated to improve the quenchability35) of steel, thereby increasing the tendency for bainite transformation products. Conversely, temperatures that are sufficiently above the A3 transformation point are obtained at welding conditions of 200 rpm–100 mm/min, 300 rpm–100 mm/min, and 400 rpm–100 mm/min. The IPF images captured the recrystallized grains, demonstrating a transition in the intragrain orientation difference. These microstructures are believed to have formed because of the transformation of martensite in the base metal into austenite during welding, thereby forming dynamically recrystallized austenite grains. Furthermore, austenite was postulated to transform back into martensite during cooling. The grains of these microstructures tended to be coarser at higher rotation speeds and at the bottom, where heat removal was suppressed by the back plate, owing to the high heat input and suppressed cooling.
As illustrated in Fig. 9, these weld microstructures exhibit a common feature: a pearlite or bainite microstructure dispersed within the martensite microstructure. However, the appearance of pearlite is particularly notable under a rotation speed of 200 rpm. In contrast, bainite became more pronounced at speeds greater than 300 rpm. These observed microstructural characteristics are consistent with the hardness tendency of the stir zone at 200 rpm, which is slightly lower than that at 300 rpm or above (Fig. 6). As previously stated, welding was performed at a temperature sufficiently higher than the A3 transformation point. This temperature was selected to ensure that the base metal microstructure, which is based on a tempered martensitic microstructure, became fully austenitic during FSW. The maximum welding temperature and cooling rate influence the formation of the microstructure during welding. In other words, as these conditions have the same welding speed, the difference in the maximum welding temperature is expected to affect the microstructure formation as a difference in the cooling rate in accordance with Newton’s law of cooling (1):
| (1) |
As coefficient k in the equation is constant, the temperature change per unit time
The red ◁ in the DIC analysis image shown in Fig. 7 corresponds to the location where the specimen ruptured at the end of the test. In addition to the image captured by the high-speed camera, the specimen was retrieved after the test, and the fracture location was analyzed using basic information, including the center of the weld and gripping area. The results indicated that the specimens ruptured at 2.0, 2.5, 2.5 4.0, and 1.0 mm from the weld center under the conditions of 100 rpm–100 mm/min, 200 rpm–100 mm/min, 300 rpm–100 mm/min, 400 rpm–100 mm/min, and 400 rpm–400 mm/min, respectively.
Using the specified information as a reference, a new observation specimen was prepared from the same weld member from which the tensile specimen was obtained. The SEM observation of the area estimated to be at approximately the same coordinates as the fracture position is shown in Fig. 11. The observation was conducted after electropolishing with a perchloric acid-acetic acid mixed solution electrolyte. The estimated bonding temperature is below the A3 transformation point, as shown in Fig. 11(a). The microstructure observed throughout the field of view is composed of fine equiaxed ferrite and cementite. The dispersion of fine particles, presumed to be spherical cementite, on the grain boundaries is clearly different from that at the stir zone center.
The Vickers hardness test results indicated that a softening zone was initiated in this region at approximately 1 mm from the welding center, extending toward the outer edge of the stir zone. The formation of the softening zone is postulated to be because of the precipitation of dynamically recrystallized ferrite and spherical cementite. The welds obtained at 100 rpm–100 mm/min exhibited higher ductility than the base metal, which is likely ascribed to the formation of a soft ferrite phase. Despite miniaturization, the strength of the microstructure greatly degrades the base metal performance and joint efficiency. Conversely, the weld microstructure exhibited hardening at 400 rpm–400 mm/min, representing the maximum rotation and welding speeds among the tested conditions. This phenomenon is ascribed to the extensive martensite-dominated microstructure observed in the stir zone, as indicated by the microstructural observations shown in Fig. 11(e). Martensite in the base metal is heated during FSW and stirred in the austenite region to form dynamically recrystallized austenite, which is then transformed back into martensite during the subsequent rapid cooling process. The stir zone is dominated by hard martensite, and the characteristics of the tempered martensite structure of the base phase are completely lost.
The discrepancy in the microstructural morphology is reflected in the considerable reduction in the mechanical properties of the high-strength and ductile test steel plate, as illustrated in Fig. 8. In other words, the welds obtained in 400 rpm–400 mm/min did not exhibit any ductile fracture behavior, whereby the lowest tensile strength is obtained. As illustrated in Figs. 11(b)–11(d), the microstructures corresponding to the fracture positions at 200 rpm–100 mm/min, 300 rpm–100 mm/min, and 400 rpm–100 mm/min are identified as composite structures that are predominantly composed of pearlite with a minor presence of bainite.
Figure 12 shows the results of the fracture surface investigation after the tensile test. The fracture position of the weld does not correspond to the softest part of the weld, except for the condition of 100 rpm–100 mm/min. This can be ascribed to the softest part that could not be designated as a gauge section owing to the specimen designation. As illustrated in Fig. 12, a significant difference is noted in the fracture surface morphology of the sample obtained at 300 rpm–100 mm/min and 400 rpm–100 mm/min. The macroscopic fracture surfaces of the welds fabricated at 300 rpm–100 mm/min exhibited ductile fracture characteristics, including shear lips and dimples in the microscopic morphology. In contrast, the sample fabricated at 400 rpm–100 mm/min has a cleaved fracture surface with a macroscopic morphology and river pattern with microscopic morphology. The macroscopic and microscopic fractures obtained under 400 rpm–100 mm/min exhibit clear brittle fracture characteristics.
As illustrated in Figs. 11(c) and 11(d), the microstructures of the welds at these fracture positions were predominantly pearlite with bainite, exhibiting the characteristics of a composite microstructure. However, the densities of pearlite and bainite are considerably higher in the microstructures of the welds obtained at 300 rpm–100 mm/min, in which ductile fracture is observed. The maximum welding temperature for this condition is 1108°C, which is approximately 60°C lower than that obtained under 400 rpm–100 mm/min (1175°C). In other words, the welds obtained at 400 rpm–100 mm/min exhibit a higher degree of maximum softening and larger softened zone than those obtained at 300 rpm–100 mm/min. Conversely, the fracture location of the welds obtained at 300 rpm–100 mm/min (2.5 mm from the center) can be identified in the region of uniform hardness within the stir zone, whereas the fracture location for those obtained at 400 rpm–100 mm/min (4 mm from the center) is located near the transition zone from the stir zone to the HAZ. Considering the disparity in the distance from the center of the welds to the fracture location, the fracture location for the sample obtained under 400 rpm–100 mm/min condition is expected to have a lower temperature during the welding process and lower cooling rate than that for the sample obtained under 300 rpm–100 mm/min. Consequently, at the fracture positions of the welds obtained at 300 rpm–100 mm/min, which have a high cooling rate, a dense cementite/ferrite composite structure (lamellar pearlite structure) is formed in the stir zone. This structure is believed to be responsible for the superior strength and ductility obtained in the tensile tests, which is consistent with previous reports38,39) that demonstrated a correlation between the density of the cementite/ferrite phase interval and strength of steels. The reason for the significant discrepancy in fracture position between these two conditions remains unclear and will be elucidated in future works.
Based on the five welding parameters investigated in this study, the general trends of the microstructures and phase ratios (morphology and fractions of pearlite, bainite, and martensite) in the welded joints of the specimen steels and mechanical properties of the welds were ascertained. Setting a welding temperature in a region higher than the A3 transformation point prompted the coexistence of bainite and martensite microstructures at the center of the weld and formed a dense cementite/ferrite composite microstructure (lamellar pearlite microstructure) within the gauge section, thereby maintaining and improving the mechanical properties of welds of the same tested material. The findings of this study indicate that the joint efficiency and ductility can be maintained at a high level by forming a composite structure of dense cementite/ferrite phases (lamellar pearlite structure) in the gauge section. The weld microstructure obtained at 300 rpm–100 mm/min was effective in ensuring the joint efficiency and ductility. Therefore, the formation of such a microstructure represents an effective form of microstructural control in the design of effective microstructures to improve the joint properties of tempered medium- and high-carbon steels.
In this study, FSW was applied to JIS-S55C steel plates tempered to a 1.4 GPa equivalent grade. The following findings were obtained:
(1) FSW joints were obtained under different welding parameters. Sound butt welds of quenched and tempered 0.55C martensitic steel with a grade of 1.4 GPa were obtained under all FSW conditions.
(2) The temperature measurement and microstructural analysis results confirmed that the welding temperature was below the A3 transformation temperature at 100 rpm–100 mm/min, indicating a lower welding temperature, and above the A3 transformation temperature under all other conditions. At 400 rpm–400 mm/min, which resulted in high cooling rates, substantial martensite fraction was noted in the weld microstructure, and the hardness of the stir zone increased significantly.
(3) The microstructure of the weld obtained at 100 rpm–100 mm/min consisted mainly of fine and equiaxed grains. However, the hardness of the microstructure after grain refinement was lower than that of the base material, and fine cementite precipitated at the grain boundary at approximately 2 mm away from the stir-zone center, forming a softening zone.
(4) At 200 rpm–100 mm/min, 300 rpm–100 mm/min, and 400 rpm–100 mm/min, a composite structure of pearlite, bainite, and martensite coexisted in the weld structure. Under these conditions, the fraction of martensite increased and bainite formation was promoted at a rotation speed of 300 rpm.
(5) Among the five conditions investigated, the weld obtained under 300 rpm–100 mm/min maintained the performance of the base metal in terms of both tensile strength and elongation. Under these conditions, the fracture surface of the weld exhibited a dimpled shape, and the joint efficiency exceeded 70% of the tensile strength of the base material. The microstructure corresponding to the fracture position of the weld formed a dense dual-phase cementite/ferrite (lamellar pearlite) microstructure. The formation of such microstructures was considered an effective form of microstructural control for improving and maintaining the mechanical properties of welds in tempered medium- and high-carbon steels.
We express our gratitude to the New Energy and Industrial Technology Development Organization (NEDO) for this achievement, which was made possible through NEDO’s subcontracting work.
