2019 Volume 60 Issue 9 Pages 2003-2007
To investigate the precipitation behavior of super duplex stainless steel in its weld by friction stir welding (FSW) at a low welding speed, we carried out microstructural observation and analysis. An intermetallic compound phase, σ, was observed in the heat affected zone (HAZ). The σ phase precipitated at the interfaces between δ-Fe (ferrite) and γ-Fe (austenite) grains 1–2 mm away from the stir zone (SZ)/HAZ boundary. On the other hand, the Cr2N was observed together with the belt-like γ-Fe grain aggregates in the vicinity of the advancing side (AS) of the SZ. The other intermetallic phase, χ, was also observed at a triple junction of γ-Fe grains. This demonstrates that the precipitation of the Cr2N and χ phases correlates with the transformation of δ-Fe to γ′-Fe (secondary austenite).
Cross-sectional images of the friction stirring weld of (a) optical microscopic and (b) EBSD image quality maps.
Super duplex stainless steels are one of the next-generation high durability structural materials boasting both high strength and high corrosion resistance. The microstructure of duplex stainless steels consists of ferrite (δ-Fe) and austenite (γ-Fe). The duplex stainless steels are mainly Fe–Ni–Cr ternary systems containing other elements such as Cr, Mo, W, and N to improve corrosion resistance. Although super duplex stainless steels have excellent material properties compared to other duplex stainless steels, their corrosion resistance is significantly impaired by various precipitations in the welding process, such as σ phase and Cr2N.1,2) Therefore, it is important to investigate the precipitation behavior during welding.
The present study focuses on friction stir welding (FSW) of super duplex stainless steel. The FSW is a bonding method invented at The Welding Institute (TWI) in 1991, and its process is a kind of solid phase bonding.3) Its joining part faces very little welding trouble such as grain coarsening, residual stress, deformation, porosity, and solidification crack after construction. Figure 1 shows (a) a photo of the FSW tool (Sialon), (b) a bird’s-eye view of the FSW joining, and (c) a schematic cross-sectional image of the weld. The FSW device consists of a rotation tool with a probe and a shoulder. The rotation tool is inserted into the material and is moved along a joining line with keeping a loaded state. The softened areas by frictional heat are stirred and plastically flowed, and then joined by recombining at the back of the rotation tool. The cross-sectional structure after FSW consists of a stir zone (SZ), a heat affected zone (HAZ), and a base material. The HAZ formed by the FSW can be divided into a thermo-mechanically affected zone (TMAZ), which is affected by plastic deformation, and a HAZ with no effect of the plastic flow. Note that the left and right sides of the cross-sectional image of the welds do not show a symmetrical structure in many cases. The left and right sides are distinguished as an advancing side (AS) where the directions of the tool rotation and the material joining coincide with each other (left) and a retreating side (RS) where these two directions are opposite to each other (right).
(a) Photo of FSW tool (Sialon), (b) bird’s-eye view of the FSW joining, and (c) schematic cross-sectional image of the weld.
While it can be challenging to join super duplex stainless steels with FSW, there have been several promising reports on the development of various high durability tools. According to these reports,4–7) the welds of super duplex stainless steel consist of fine grains of δ-Fe and γ-Fe. However, to the best of our knowledge, there has been little report on precipitates in the FSW welds of super duplex stainless steels. In this study, the FSW is conducted on a super duplex stainless steel at a lower welding speed than usual, and its weld part is observed in detail by electron microscopy. On the basis of the electron microscopy experiment, we clarify the precipitation behavior at the weld of the super duplex stainless steel, which is indispensable knowledge for using the welded material safely.
The super duplex stainless steel prescribed in UNS-S32750 was used for the present study. Table 1 lists the chemical composition of the super duplex stainless steel used in the present study. The nitrogen concentration, 0.24 mass% N, is higher than that of other duplex stainless steels. A rolled material 300 mm long, 100 mm wide, and 12 mm thick was prepared, and a bead-on-plate test of the FSW with a welding length of 200 mm was carried out at the center of the adjacent materials to join.
In the welding test, the rotation speed was 400 rpm, the rotation tool tilt angle was 3°, and the setting depth was 5.1 mm. The tool was made of a commercially available ceramic material, Sialon (Si–Al–O–N), with a probe length of 4.5 mm and a shoulder diameter of 18.0 mm. The welding speed was 0.42 mm/s, which is 1/4 of the proper speed of 1.7 mm/s. The welding speed is an important parameter to obtain good weld joints by FSW. For the present experimental settings and materials, no welding defect was formed by setting the welding speed to 1.7 mm/s or less. Cross-sectional specimens of the weld from the central part of the weld bead were prepared for microstructural observations. First, the sample was electrolytically etched with a 10 kmol/m3 KOH solution for macroscopic characterization using optical microscopy. Second, micrometer-scale microstructural characterizations were carried out using scanning electron microscope (SEM) combined with an electron back-scatter diffraction (EBSD) technique. Third, nanometer-scale microstructural characterizations were performed using transmission electron microscope (TEM) and scanning transmission electron microscope (STEM). A focused ion beam (FIB) sampling technique was utilized for the TEM/STEM specimen preparation. Composition was analyzed with energy dispersive X-ray spectrometry (EDX) attached to STEM.
Figure 2 shows cross-sectional views of the weld. By setting the low welding speed, 0.42 mm/s, no welding defect occurred. However, this figure shows an unknown image contrast in the SZ near the AS. This unknown image contrast showing a belt-like morphology is recognized in the optical microscopy and is notably different from the uniform image contrast of the (δ-Fe + γ-Fe) two-phase matrix. In the following EBSD analysis, areas A, B, C, and D denoted in Fig. 2 were selected for the fields of view. Area A is located at center of SZ, area B is in the belt-like structure of the SZ, area C is located at the SZ/TMAZ interface, and area D is located 1.5 mm away from the interface in the HAZ outside the AS.
Optical microscopic cross-sectional image of the weld.
Figure 3 shows image quality maps and phase maps in each area by the SEM-EBSD analysis. No precipitate was recognized in area A at center of SZ and area C at the SZ/TMAZ interface. However, the σ phase was observed in the HAZ that was 1.5 mm away from the SZ/TMAZ interface, as shown in area D. σ phase is the tetragonal intermetallic phase described as Fe–Cr(–Mo).8–11) The σ phase precipitates from the δ-Fe/γ-Fe interface. This fact coincides with the precipitation behavior of σ phase during heat treatments reported previously.4,5) The precipitation of the σ phase in the HAZ was also observed at a location about 1 mm away from the SZ/TMAZ interface at the RS.
Image quality maps and phase maps analyzed by SEM-EBSD.
The SEM-EBSD analysis of B, in which the belt-like structure in SZ was observed, shows a precipitate different from the σ phase. Specifically, the SZ shows a microstructure in which the crystalline grains are small: a few micrometers as a whole. The phase map revealed that the belt-shaped structure is an aggregate of γ-Fe in which δ-Fe does not exist, even though the surrounding matrix of the belt-shaped structure consists of the (γ-Fe + δ-Fe) two-phase structure. Furthermore, fine Cr2N precipitates are formed in the grain boundaries of γ-Fe as shown in the phase map of Fig. 3. Cr2N is a hexagonal nitride phase commonly observed in super duplex stainless steels containing significant amounts of nitrogen.12,13) It should be noted that there is a microstructure composed of Cr2N and γ-Fe phases that has a higher solid solubility of nitrogen than that of δ-Fe.
3.2 Nanometer-scale observationsFigure 4 shows STEM high-angle annular dark-field (HAADF) images of an area depicting Cr2N precipitates at which the specimen was prepared by FIB. The HAADF image in Fig. 4(a) indicates that Cr2N showing a darker image intensity precipitates at the grain boundaries of the γ-Fe matrix. Figure 4(b) is a magnified view of the square area depicted in (a). At the triple junction of the γ-Fe grains, precipitate A showing a brighter image intensity is formed with the two adjacent Cr2N particles. Although dark contrast particles can also be seen in the γ-Fe grain (as shown in Fig. 4(c)), they were wear debris of the Sialon tool used for the composition analysis by EDX. Figure 5 shows composition profiles of the main elements (Fe, Cr, Ni) at a Cr2N grain analyzed by EDX. While the profile of Cr has a high concentration in a Cr2N grain, the composition gradient in which Cr is depleted in the close area of the grain boundary is confirmed in γ-Fe grains. In the composition gradient, an increase in Fe and Ni is observed at the same time as chromium deficiency. Also, Cr2N grains does not contain Sialon elements such as Al and Si, the wear debris does not affect the precipitation.
STEM high-angle annular dark-field (HAADF) images of area showing Cr2N precipitates.
Composition profiles of main elements (Fe, Cr, Ni) at a Cr2N grain analyzed by EDX.
EDX analysis was also performed for unknown precipitates at triple junctions, and the composition was found to be slightly more Mo than the γ-Fe matrix. Figure 6 shows selected area diffraction patterns by TEM of the same field of view as the STEM observation in Fig. 6(b). A selected area diffraction pattern in Fig. 6(a) was acquired from the γ-Fe matrix, and that in Fig. 6(b) was acquired from the precipitate A at the triple junction. Indexing the diffraction pattern in Fig. 6(b) revealed that the precipitate at the triple junction is of the χ phase. χ phase is a cubic intermetallic phase having a composition of Fe36Cr12Mo10.14,15)
Selected area diffraction patterns observed by TEM.
The STEM/TEM analysis indicates the existence of aggregates of the γ-Fe matrix containing the grain boundary precipitates of Cr2N and χ phases in the SZ. This suggests that the Cr2N phase in the form of sheets precipitated at grain boundaries of the γ-Fe matrix and the χ phase precipitated at triple junctions of the γ-Fe matrix.
3.3 DiscussionKinetic behaviors of the precipitates under heating at constant temperature were calculated using JmatPro software. Figure 7 shows calculated time-temperature-transformation (TTT) diagrams for the possible precipitates in the present alloy system. The same chemical composition as the material used in the experiment (Table 1) was assumed for the calculations. While no precipitation is expected in the matrix phases of δ-Fe and γ-Fe at higher than 1100°C, multiple types of precipitations appear at 1000°C or lower. The calculated TTT diagram indicate that the σ phase is most likely to precipitate around 950°C and that the alloy should be kept at that temperature for more than 1 min to promote the σ precipitation. The maximum temperature of the SZ under welding is considered to be about 80% of the melting point of the materials to be joined,3) and is considered to reach about 1100°C in this study. The σ phase precipitated in the HAZ that was not influenced by plastic flow during the FSW process. In the weld, the region 1–2 mm away from SZ could be kept just in that temperature region close to 950°C, and the σ phase be precipitated.
Calculated time-temperature-transformation (TTT) diagrams for various precipitations.
The calculated results also indicate that Cr2N and χ phase precipitate in a short duration compared with the σ phase. Ramirez et al.16) investigated the precipitation behavior of Cr2N under heat treatment and reported the following relationship between Cr2N and γ′-Fe (secondary austenite) phases: Cr2N nucleates from the δ-Fe phase side at δ-Fe/γ-Fe interfaces. Because the Cr2N precipitation consumes nitrogen, the grain boundaries containing the Cr2N precipitates reduce nitrogen concentrations. As the nitrogen enriched in the γ-Fe matrix diffuses toward the δ-Fe/γ-Fe interfaces, the δ-Fe matrix partially transforms to austenite, which is called secondary austenite (γ′-Fe). The phase transition processes described above is summarized as
| \begin{equation*} \text{$\delta$-Fe} + \text{$\gamma$-Fe} \rightarrow (\text{$\delta$-Fe} + \text{Cr$_{2}$N} + \text{$\gamma'$-Fe}) + \text{$\gamma$-Fe} \end{equation*} |
In the present super duplex stainless steel, Cr2N was observed in the aggregates of the austenite not containing δ-Fe phase. If the phase evolution proceeds along the reactions described above, it can be assumed that Cr2N nucleated at the grain boundary of the δ phase. At the SZ, the grain sizes of the (δ-Fe + γ-Fe) matrix became smaller (to the level of micrometer) during the friction stirring process. The δ-Fe phase around the Cr2N phase could be rapidly transformed to γ′-Fe because nitrogen diffusion from the γ-Fe phase stabilizes γ′-Fe more than δ-Fe. As a result, aggregates consisting of the γ-Fe and γ′-Fe grains were formed around the Cr2N precipitates. As shown in Fig. 5, Cr depletion layers would be formed around Cr2N grains. Fe and Mo were released into the matrix where Cr is depleted when the δ-Fe phase transformed to Cr2N and γ′-Fe, and the χ phase having a composition of Fe36Cr12Mo10 was then formed at the triple junctions. The Cr depletion layer itself is considered to cause localized corrosion in a corrosion study on precipitation of σ phase.17,18)
The Cr2N and the χ phase precipitated in the vicinity of the AS in the SZ, and aggregates of γ-Fe and γ′-Fe formed with the Cr2N and χ precipitates. Park et al.18) investigated the weld of 304 stainless steel and also reported the result of formation of σ phase and Cr depletion layer in the AS of SZ. The vicinity of the AS is a place where plastic flow circulating with the rotating tool accumulates, as shown in Fig. 1(a).19,20) The belt-shaped structure (as shown in Fig. 2) is interpreted as a trace of the stirred flow. It seems that the belt-like fluid was circulated with the tool multiple times, was exposed in a heating environment for a sufficient duration, and precipitated Cr2N and γ′-Fe. On the other hand, the Cr2N and the χ phase didn’t precipitate in HAZ in which the σ phase was precipitated. It is reported that the precipitation of σ phase is accompanied by γ′-Fe same manner as the precipitation of Cr2N.11,17,21) Therefore, Cr2N is considered to be difficult to precipitate in the temperature range where the σ phase is stable. Also, the χ phase is known to transform to the σ phase after heat treatment for enough time.15) Therefore, the χ phase was not confirmed simultaneously with σ phase in this study.
The precipitation behavior of super duplex stainless steel in a weld formed by FSW at a low welding speed was investigated using optical and electron microscopy techniques. The results are shown below.
