1 Introduction

The excavation of shale and solid gases from deep sea and polar regions, as well as the demand for hydrogen energy have increased recently. To meet the associated industrial needs, the use of cryogenic materials is essential and the fields requiring these types of material are broadening [1,2,3,4]. The 9% Ni steels and austenitic stainless steels currently used for cryogenic materials are expensive in terms of raw materials and often cause welding defects including hot cracks, reheat cracks, and cold cracks [5,6,7,8,9,10,11,12]. Therefore, research regarding low-cost high Mn steels with excellent low-temperature properties have been recently reported [13,14,15,16,17,18,19,20,21].

High Mn steels for cryogenic applications contain large amounts of austenite stabilizing elements (Mn and C) to form austenite structure at low temperatures. In the austenitic high Mn steels, the stacking fault energy (SFE) is dependent on grain size, chemical composition, and temperature [22,23,24,25]. The observed deformation mode varies from ε-martensite to deformation twin and dislocation slip as SFE increases and the mechanical properties are typically associated with the deformation mode.

Raabe et al. reported that a fine-grain steel produced higher strain-hardening by the Hall–Petch relation in Fe-22Mn-0.6C steel and the deformation twin became saturated during the initial stage of the tensile test. Finally the tensile strength increased and ductility decreased compared to that of the coarse-grain steel [22]. Hwang et al. reported that coarse grain Fe-24.5Mn-4Cr-0.45C steel showed decreased SFE and tensile strength, with α’-martensite transformation and decreased ductility [23]. Pineau et al. reported that the deformation mode varied from slip → twin → ε-martensite as the temperature of tensile testing was decreased from 700 to 100 K for Fe-20Mn-4.1Cr-0.48C steel [25].

Welding studies of high Mn steels have also been reported. In particular, many studies have been reported regarding laser beam welds, resistant spot welds, friction stir welds, and inert gas arc welds [27,28,29,30,31,32]. However, a lack of studies have been reported regarding submerged arc welds (SAWs) with high productivity for thick plates, which are mainly used for cryogenic storage materials. Therefore, this study investigates the microstructural and tensile properties of high Mn steel SAWs as a function of the temperature. Specifically, we focus on the plastic deformation mode of the welds at 110 and 298 K, and its relationship with tensile properties.

2 Experimental Procedure

Table 1 shows the chemical composition of the base metal (BM) and filler metal. High Mn steels were produced using a blast furnace. Solution treatment for the cast ingot was performed and after the solution treatment, the ingot was hot-rolled to 20 mm, followed by controlled cooling. An undermatched filler metal with less Mn content than the BM was used to prevent the hot cracking typically generated in austenitic high Mn steels [32].

Table 1 Chemical composition (wt%) of the base and filler metals containing high Mn contents
Full size table

The BM surface was cleaned with acetone before welding and Fig. 1 shows a schematic illustration of the double-V groove for the SAW. The root-face length and face interval were 5 and 1–2 mm, respectively. A welding pass was applied to the inside groove for the 1st pass, followed by the outside groove for the 2nd and 3rd passes. Submerged arc welding was performed using a SAW machine (Lincoln electric company) and the welding conditions are listed in Table 2. The welding current was 680–720 A, welding voltage 29 V, and welding speed 0.5 m/min.

Fig. 1
figure 1

Schematic illustration of the double-V groove

Full size image
Table 2 Welding conditions to produce the SAWs
Full size table

The WM cross-sections were polished using 1 µm diamond paste after mechanical grinding. The polished surface was etched in a solution of ethanol (100 mL), hydrochloric acid (5 mL), and picric acid (2 g) and the resulting microstructure was observed by optical microscopy (OM) and scanning electron microscopy (SEM). Grain and dendrite-packet sizes were measured using the linear intercept method. The crystal structures were identified by X-ray diffraction (XRD) using a Cu-Kα target and electron back-scattered diffraction (EBSD). XRD was performed at a scan speed of 1°/min from 20 to 100°, 40 kV, and 40 mA. The diffraction data for the EBSD experiment were obtained as orientation maps with step sizes of 0.1–0.2 μm at 20 kV. The hardness distribution was measured using a Vickers hardness tester at a load of 500 g for 10 s. The strength of the SAWs was examined via tensile testing. The tensile specimens were prepared in the transverse direction of the SAWs and tensile tests performed with a standard-size specimen (ASTM E8). The extensometer was set on an Instron testing machine in a low temperature chamber and loading speed of 1 mm/min.

3 Results and Discussion

3.1 Microstructure of the BM and WM

Figure 2a, c show the cross-sectional microstructure of the BM observed with the inverse pole figure (IPF) and phase maps of the EBSD, respectively. The hot-rolled BM exhibited FCC phase with no rolling textures after solution treatment. The grain size was mainly distributed in the range of 5–15 μm and the average grain size of the recrystallized structure was ~ 12 μm (Fig. 2d). Figure 2b shows the XRD pattern of the BM and WM. The BM contained FCC single phase without precipitates, intermetallic compounds, or other phases. It was coincident with the phase identification measured from the EBSD analysis (Fig. 2a, c).

Fig. 2
figure 2

Microstructure and phase identification: a IPF image of the BM, b XRD pattern of the BM and WM, c phase map of the BM, d grain size distribution of the BM

Full size image

Figure 3 shows the cross-sectional microstructure of the SAWs. Multipass welds were achieved without defects such as cracks or voids. The fusion boundary and coarse grain heat affected zone (CGHAZ) showed no defects such as liquation cracks (Fig. 3a). Grain growth occurred near the 1st pass CGHAZ due to the input of welding heat (Fig. 3b). Figure 3c indicates the significant grain growth due to the overlapping heat input from the 1st and 2nd pass CGHAZ. Figure 3d shows an IPF map of the 2nd pass WM. The short length of columnar dendrites was mainly distributed in the range of 40–140 μm and an average short length was ~ 86 μm (Fig. 3f). Figure 3e shows an IPF map of CGHAZ near the 2nd pass boundary. The grain size was mainly distributed in the range of 5–40 μm and an average grain size was ~ 23 μm. Comparing Figs. 2a and 3d, e, the WM had a larger grain size than the BM and CGHAZ. The WM and CGHAZ contained the same FCC austenite single phase as the BM (Fig. 2b).

Fig. 3
figure 3

Macro and microstructure of the SAWs: a macrostructure of the welds, b CGHAZ microstructure near the 1st welding pass, c CGHAZ microstructure near the 1st and 2nd pass boundaries, d IPF map of the 2nd pass welds, e CGHAZ microstructure near the 2nd pass boundary (IPF map), f grain size distribution of 2nd pass welds, g grain size distribution of CGHAZ near the 2nd pass boundary

Full size image

3.2 Mechanical Properties of the BM and WM

The hardness was measured along the 1st, 2nd, and 3rd hardness lines described in Fig. 3a. Figure 4 shows the hardness distribution, where the hardness in the 1st (black line), 2nd (red line), and 3rd welding passes (blue line) were compared simultaneously. The BM exhibited an average hardness of ~ 267 Hv, which significantly decreased in the CGHAZ and welds to 205–225 Hv. The hardness values of the WM were lower than that of the BM, which was due to the use of undermatched filler metal and the large grain size of the WM. The hardness of the 2nd welding pass (red line) was ~ 23 Hv higher than that of the 1st (black line) and 3rd passes (blue line) because significant dilution of the BM into the WM occurred from the root face of the 2nd welding pass.

Fig. 4
figure 4

Distribution of Vickers hardness for each SAW pass

Full size image

Figure 5 shows the stress–strain curves of the BM and WM observed at 298 and 110 K. The ultimate tensile strength (UTS) and yield strength (YS) of the BM were 924 and 505 MPa at 298 K, which increased to 1248 and 726 MPa, respectively, at 110 K. The UTS and YS of the WM were 930 and 467 MPa at 298 K, which increased to 1150 and 617 MPa, respectively, at 110 K. The UTS and YS of the BM and WM significantly increased at 110 K. The BM exhibited mostly the same tensile properties at 110 K as at 298 K. However, the WM at 110 K showed a decrease in elongation by 20%, as compared to that observed at 298 K. Furthermore, all transverse SAWs fractured near the weld centerline regardless of temperature. The fracture position was consistent with the use of undermatched filler metal, the large grain size of the WM, and low hardness.

Fig. 5
figure 5

Tensile curves of the BM and WM measured at 298 and 110 K

Full size image

3.3 Variation of Deformation Mode for the BM and WM as a Function of Temperature

Figure 6 shows the cross-sectional microstructure of the BM observed at 1/2 thickness near the fractured area after tensile testing at 298 and 110 K. The microstructures were analysed by EBSD mapping using image quality (IQ), coincidence site lattice (CSL) boundary, and phase maps (confidence index ≥ 0.1). At 298 K, the CSL map showed a heavily deformed boundary and ∑3 boundary (deformation twin boundary) formed in the grain indicated by a red line. The BM with 24 wt% Mn showed dislocation slip and twin formation deformation modes. It was confirmed that the austenite phase represented in yellow was dominant, as shown in the phase map (Fig. 6a). As the temperature was decreased to 110 K, the tensile-fractured specimen formed a ∑3 boundary in the CSL map and ε-martensite was detected (as indicated by green) in the grain from the phase map (Fig. 6b).

Fig. 6
figure 6

Tensile-fractured microstructures of the BM at a room temperature (298 K) and b cryogenic temperature (110 K)

Full size image

Figure 7 shows the cross-sectional microstructure of the WM observed near the fractured area of the 2nd welding pass, as indicated by the red-line box. At 298 K, the CSL map indicated a bundle of ∑3 boundary and austenite phase (yellow) was present without phase transformation (Fig. 7a). At 110 K, the ∑3 boundary was more abundant than at 298 K and the phase map detected ε-martensite (green colour in Fig. 7b). Figure 7c shows the ε-martensite initiation site at 110 K occurring at the intersection of the ∑3 boundary (marked with an arrow) and ε-martensite grew along the ∑3 boundary [30]. When the microstructures near the BM and WM fractures were analysed at the same temperature (either 298 or 110 K), more ∑3 boundary was produced in the WM than in the BM.

Fig. 7
figure 7

Tensile-fractured microstructures of the WM at a 298 K and b 110 K. c Formation of ε-martensite at the ∑3 boundary junction

Full size image

Figure 8 summarizes the microstructural characteristics of the BM and WM during tensile failure observed in Figs. 6 and 7. At 298 K, the ∑3 boundary was formed in the BM and WM, and the austenite phase was present without phase transformation. Regardless of the BM and WM, the 110 K cryogenic temperature produced ε-martensite transformation and ∑3 boundaries more significantly than at room temperature (298 K).

Fig. 8
figure 8

Fraction of the ∑3 boundary, austenite and ε-martensite at 298 and 110 K for the a BM and b WM

Full size image

The SFE is known to decrease at cryogenic temperatures, transforming the deformation mode from deformation twin to a mixture of deformation twin and ε-martensite [25]. In the study, the increased UTS and similar EL for the BM at 110 K (Fig. 5) was due to activated deformation twin and ε-martensite transformation, which postponed the necking process of BM [33, 34]. On the other hand, the WM activated the deformation twin and ε-martensite more significantly at 110 K than in the BM, increasing the UTS and decreasing the EL (Fig. 5). The critical stress required for deformation twin formation is known to decrease with increasing grain size. The packet size of the WM (~ 86 µm) was coarser than the grain size of the BM (~ 12 µm). Therefore the critical stress required to produce a deformation twin should decrease for the WM as compared to the BM [35].

The critical stress is also related to the SFE of the BM and WM, and the SFE can be calculated using the linear regression analysis developed by Schramm and Reed [36]. The equation to calculate the SFE (γ) is a function of chemical composition at room temperature as follows:

$$\upgamma = - 53 + 6.2({\text{wt }}\;{\text{pct}}\;{\text{Ni}}) + 0.7({\text{wt }}\;{\text{pct}}\;{\text{Cr}}) + 3.2({\text{wt }}\;{\text{pct}}\;{\text{Mn}}) + 9.3({\text{wt }}\;{\text{pct}}\;{\text{Mo}})\;[ {\text{mJ/m}}^{2} ] .$$
(1)

The composition of the BM was determined, as listed in Table 1, and that of the WM was calculated using the dilution ratio of BM:filler metal (3:7). The BM and WM were calculated to have SFEs of 27.1 and 17.0 mJ/m2, respectively. It was confirmed that the WM with low Mn content exhibited an SFE ~ 10.0 mJ/m2 lower than that of BM. It is reported that the twinning-induced plasticity (TWIP) effect is dominant for high-Mn steels with SFE values between 20 and 40 mJ/m2 and the transformation-induced plasticity (TRIP) effect is more active for steels with SFE values below 20 mJ/m2 [37]. However, De Cooman et al. [26]. reported for the Fe-18Mn-0.6C steel that the SFE limit to produce martensite was 13 mJ/m2. Therefore, the BM and WM at 298 K reasonably exhibited a deformation mode in the deformation twin region. The WM with low SFE produced more twin than the BM regardless of temperature (Fig. 8) [25, 26]. Therefore, the BM and WM at 298 K reasonably exhibited a deformation mode in the deformation twin region. The WM with low SFE produced more twin than the BM regardless of temperature (Fig. 8) [25, 26]. The high SFE with smaller separation between partial dislocations promotes cross-slip resulting in wavy dislocation substructure, whereas the low SFE with larger separation of partials inhibits cross-slip resulting in planar slip or favors formation of twinning. Therefore, the lower SFE in the deformation twin range below 40 mJ/m2 and the larger stacking fault width, the more twins become more active for the WM.

The SFE at 110 K cannot be obtained using the Eq. (1). Alternatively, this study predicted the SFE at 110 K based on the previous study: the SFE decreases with decreasing temperature in Fe-26Mn-0.2C steels [25]. Therefore, because the SFE decreased at 110 K, the WM exhibited a small SFE and critical stress required for deformation twin activation and produced more ε-martensite at the intersection of the deformation twin than in the BM. Thus, the EL of the WM decreased to 26% because necking propagation should be significantly limited at 110 K [23, 38].

Figure 9 shows the fractography of the WM at 298 and 110 K. At 298 K, the deformation twin occurred during the tensile test, producing a typical ductile fracture with dimples (Fig. 9a). However, at 110 K the activated generation of the deformation twin produced fine dimples as compared with the 298 K fractography [39]. In addition, some quasi cleavage (QC) fractures with a flat geometry were confirmed because the ε-martensite transformation also occurred at 110 K. The formation of the QC fracture at the intersection between the martensite and/or deformation twin has also been reported in Fe-17Mn-0.3C steels [40, 41]. This study used the BM and filler with a large Mn content (~ 24 and ~ 20 wt%, respectively), producing a mixture of fine dimples and QC fractures at 110 K.

Fig. 9
figure 9

Fractography of the WM at a 298 K and b 110 K

Full size image

4 Conclusions

This study investigated the weldability, tensile properties, and microstructure of Fe-24.3Mn-3.4Cr-0.44C steels and SAW welds at room and cryogenic temperatures. The important conclusions can be summarized as follows:

  1. 1.

    The SAWs showed no defects, such as cracks or voids, and fully austenitic microstructure as in the BM.

  2. 2.

    The WM exhibited decreased hardness due to the undermatched filler metal and coarse packet size (~ 86 μm) compared to that of the BM (~ 12 μm). All WM specimens fractured near the weld centerline regardless of temperature.

  3. 3.

    Only deformation twins were observed at 298 K regardless of metal type during the tensile test. At 110 K, the deformation mode changed to a mixture of deformation twin and ε-martensite. Therefore, the YS and UTS of the BM and WM increased significantly at 110 K.

  4. 4.

    As the SFE decreased at 110 K, the WM with a small SFE and critical stress required for deformation twin formation produced more ε-martensite at the intersection of deformation twin than that of the BM. The formation of ε-martensite with deformation twins antedated necking during the tensile testing and the elongation decreased at 110 K.

  5. 5.

    At 298 K, the deformation twin formed during the tensile test, produced a typical ductile fracture with dimples. At 110 K, the activated generation of the deformation twin produced the fine dimples and some QC fractures were detected because ε-martensite transformation also occurred at 110 K.

  6. 6.

    The SAWs of high Mn steels maintained excellent low-temperature mechanical properties, including elongation, tensile strength, and yield strength values of 20%, 1150 MPa, and 617 MPa, respectively.