# Fileset

[_ISIJINT-2023-437.pdf](https://mdr.nims.go.jp/filesets/f730bbf1-1361-4c07-9c8c-432fb6a5aab3/download)

## Creator

[Fumiyoshi Yoshinaka](https://orcid.org/0000-0003-0534-7815), [Yusuke Tsutsumi](https://orcid.org/0000-0002-9483-1256), [Tomoya Nagira](https://orcid.org/0000-0002-2415-0623), [Susumu Takamori](https://orcid.org/0000-0003-3422-9391), [Satoshi Emura](https://orcid.org/0000-0001-5789-6408), [Takahiro Sawaguchi](https://orcid.org/0000-0002-9405-002X), [Hideki Katayama](https://orcid.org/0000-0001-7947-4687), Terumi Nakamura, Yasuhiko Inoue, Susumu Motomura, Atsumichi Kushibe

## Rights

[Creative Commons BY-NC-ND Attribution-NonCommercial-NoDerivs 4.0 International](https://creativecommons.org/licenses/by-nc-nd/4.0/)

## Other metadata

[Effects of the Cr and Ni concentrations on the fatigue and corrosion resistances of Fe−Mn−Cr−Ni−Si alloys for seismic damping applications](https://mdr.nims.go.jp/datasets/37c2d062-f493-4831-9f35-4efa75bafcbb)

## Fulltext

1ISIJ International, Vol. 64 (2024), No. 7, pp. 1–9https://doi.org/10.2355/isijinternational.ISIJINT-2023-437* Corresponding author: E-mail: Sawaguchi.takahiro@nims.go.jp© 2024 The Iron and Steel Institute of Japan. This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs license (https://creativecommons.org/licenses/by-nc-nd/4.0/).1.  IntroductionVibration-controlled structures are a design concept where the seismic energy input to a building is consumed via a damping element (i.e., seismic damper) to reduce the vibration and prevent serious damage to the principal structural members.1) Hysteretic dampers utilize an energy absorption and dissipation mechanism associated with cyclic elastoplastic deformation, and they are typically made from steel, known as steel damper.2) Because steel dampers work by cyclic plastic deformation, they need to be made from steel with a lower yield point than the steel used in principal structural members. However, the repeated plastic deforma-tions mean that the service life of the damper depends on its resistance to low-cycle fatigue. For high-rise buildings or long-span bridges subjected to long-period and long-term ground motions from giant earthquakes, steel dampers made from general low-yield-point steel are generally considered Effects of the Cr and Ni Concentrations on the Fatigue and Corrosion Resistances of Fe–Mn–Cr–Ni–Si Alloys for Seismic Damping ApplicationsFumiyoshi YOSHINAKA,1)  Yusuke TSUTSUMI,1)  Tomoya NAGIRA,1)  Susumu TAKAMORI,1)   Satoshi EMURA,1)  Takahiro SAWAGUCHI,1)*  Hideki KATAYAMA,1)  Terumi NAKAMURA,1)  Yasuhiko INOUE,2)  Susumu MOTOMURA2) and Atsumichi KUSHIBE2)1)  National Institute for Materials Science, 1-2-1 Sengen, Tsukuba, Ibaraki, 305-0047 Japan.2)  Takenaka Corporation, 1-5-1, Otsuka, Inzai, Chiba, 270-1395 Japan.(Received November 21, 2023; Accepted March 4, 2024; Advance online published March 14, 2024; Published May 15, 2024)Previous research has shown that Fe–Mn–Cr–Ni–Si alloys offer excellent low-cycle fatigue resistance via reversible bidirectional transformation between face-centered cubic (FCC) γ-austenite and hexagonal closed-packed (HCP) ε-martensite. The alloy shows superior low-cycle fatigue life and is used for seismic damping applications, but there have been concerns over their resistance to highly corrosive environ-ments. In this study, Fe–15Mn–aCr–bNi–4Si alloys were prepared with different Cr and Ni concentrations to evaluate the effects on the fatigue and corrosion resistances: Z1 with (a, b) =  (14, 10.1), Z2 with (a, b) = (12.5, 8.8), Z3 with (a, b) =  (11, 7.5), Z4 with (a, b) =  (9.5, 6.1), and Z5 with (a, b) =  (8, 4.8). Z2 had the longest fatigue life. The alloy showed Gibbs free energy difference between γ-austenite and ε-martensite phases close to the ideal of zero and the α′-martensitic transformation was suppressed well, which agreed with the design criteria for achieving bidirectional transformation-induced plasticity. The developed alloys showed superior corrosion resistance in seawater. Local pitting corrosion was observed that was attrib-uted to the high Mn concentration of the alloys, although this was greatly mitigated by adjusting the Cr and Ni concentrations, especially with Z1 and Z2.KEY WORDS:  fatigue; steel; microstructure; martensitic transformation; corrosion resistance.insufficient as a countermeasure. Therefore, research has been ongoing to develop steel alloys with excellent low-cycle fatigue resistance.3–6)In the literature, the relationship between the low-cycle fatigue life Nf and total strain range Δεt is generally accepted to not strongly depend on the type of steel.7) Recently, however, researchers have reported that some advanced high-strength steels (AHSSs) and high-entropy alloys (HEAs) have superior low-cycle fatigue resistance owing to the transformation-induced plasticity (TRIP) or twinning-induced plasticity (TWIP) effects.8–15) We previously reported that a new TRIP mechanism called bidirectional TRIP (B-TRIP)16,17) where the reversible bidirectional transformation between face-centered cubic (FCC) γ-austenite and hexagonal closed-packed (HCP) ε-martensite is effective at improving the low-cycle fatigue life.5,18) The phase transformation is responsible for the Schockley partial dislocation, and its highly reversible nature may retard the accumulation of fatigue damage. The importance of reverse motion of dislocations, including _ISIJINT-2023-437https://creativecommons.org/licenses/by-nc-nd/4.0/https://orcid.org/0000-0003-0534-7815https://orcid.org/0000-0002-9483-1256https://orcid.org/0000-0002-2415-0623https://orcid.org/0000-0003-3422-9391https://orcid.org/0000-0001-5789-6408https://orcid.org/0000-0002-9405-002Xhttps://orcid.org/0000-0001-7947-4687https://orcid.org/0000-0002-0582-9067ISIJ International, Vol. 64 (2024), No. 7©  2024  ISIJ 2reverse transformation, has been emphasized in literature.19) Shams et al.15) and Bahadur et al.20) reported the bidirec-tional transformation of HEAs under fatigue loading. We utilized B-TRIP to develop a Fe–15Mn–10Cr–8Ni–4Si B-TRIP alloy; it demonstrated excellent fatigue resistance with Nf >  10 000 cycles at a total strain amplitude of 1%.4) This alloy has already been applied in a high-rise building and an international exhibition hall.21,22) However, this alloy has a high solidification cracking susceptibility, and special care is required for welding. Therefore, the alloy was modi-fied to Fe–15Mn–11Cr–7.5Ni–4Si for reduced solidification cracking susceptibility with comparable fatigue resistance.6)In addition to fatigue, another important mechanism for the damage and deterioration of steel structures is corro-sion.23–26) In particular, steel dampers in a corrosive envi-ronment, such as buildings or bridges near the shore or at sea, must have high corrosion resistance in addition to the general application of antirust paint.27,28)The objective is to develop a new alloy with exceptional fatigue durability, whilst focusing on the impact of alloy-ing elements that intimately associated with corrosion resistance; Cr and Ni. Fe–15Mn–aCr–bNi–4Si alloys were prepared where the Cr and Ni concentrations (a and b, respectively) were systematically varied so that the Creq/Nieq ratio was approximately equal to that of the Fe–15Mn–11Cr–7.5Ni–4Si alloy that exhibited Ferrite-Austenite (FA) mode solidification that can reduce the solidification crack-ing susceptibility.6) Cr and Ni both increase the corrosion resistance of stainless steels; Cr forms a passive film, and Ni reinforces the stability of the film.23,29) Experiments were performed to investigate the effects of Cr and Ni on the phase transformation, fatigue, and corrosion behaviors of the prepared alloys.2.  BackgroundThe chemical composition of a B-TRIP steel alloy needs to satisfy the following conditions to ensure its fatigue resistance and weldability. First, a high fatigue resistance requires balancing the thermodynamic stability of γ-austenite and ε-martensite to allow for bidirectional transformation. This condition can be written by setting the Gibbs free energy difference between γ-austenite and ε-martensite ΔGγ→ε as close to zero as possible (i.e., ΔGγ→ε ≈ 0). ΔGγ→ε is calculated as follows:  � � �G G Gchem mag� � � � � �� � �� � ...................... (1)Further details are presented in the previous work.30)Next, the formation of body-centered cubic (BCC) α′-martensite should be prevented.5) This is because the α′-martensitic transformation involves a volume change and thus is irreversible while the ε-martensitic transforma-tion to ε-martensite is an almost volume-constant process. In B-TRIP alloys, the γ-austenite phase is sufficiently stable against α′-martensite to prevent the γ→α′-martensitic transformation. However, two-stage γ→ε→α′-martensitic transformation can occur during fatigue deformation, albeit slight.To ensure weldability, the solidification mode should be FA to reduce the susceptibility to solidification cracking.6) This is the same policy used for stainless steels.31,32) To con-trol the solidification mode, the ratio of Cr equivalent (Creq) to Ni equivalent (Nieq) can be used as a design parameter. Creq and Nieq are calculated as follows:33)Cr wt %Cr wt %Mo wt %Si wt %Nbeq � � � �. . . . . .1 5 0 5 ... (2)  Ni wt %Ni wt %Mn wt %Ceq � � �. . . .0 5 30 .......... (3)In situ observation of the solidification process during welding was conducted by using synchrotron radiation X-ray imaging at SPring-8.34,35) Fe–15Mn–10Cr–8Ni–4Si exhibited austenite (A) mode solidification and cracking while Fe–15Mn–11Cr–7.5Ni–4Si showed FA mode solidi-fication and no crack propagation.3.  MethodsDifferential scanning calorimetry (DSC) was used to characterize the thermal martensitic transformation of the prepared specimens. Axial strain-controlled fatigue tests were conducted to examine the low-cycle fatigue resistance. The deformation microstructure was analyzed by using X-ray diffraction (XRD), electron backscattering diffraction (EBSD), energy-dispersive X-ray spectroscopy (EDS), and a ferrite scope to evaluate the fatigue mechanism in terms of the microstructure. The corrosion resistance was investi-gated by electrochemical and non-electrochemical methods: an anodic polarization test and salt spray test, respectively.3.1.  MaterialsAlloys were prepared with a nominal composition of Fe–15Mn–aCr–bNi–4Si according to weight percentage: Z1 with (a, b) =  (14, 10.1), Z2 with (a, b) =  (12.5, 8.8), Z3 with (a, b) =  (11, 7.5), Z4 with (a, b) =  (9.5, 6.1), Z5 with (a, b) =  (8, 4.8). Z3 had the same composition as the previ-ously reported material.6) The alloy materials were melted in an argon environment to obtain a square frustum-shaped ingot that was forged at 1 100°C, rolled at 500°C, and heat-treated at 1 000°C for 1 h followed by water quenching. Table 1 lists the actual chemical compositions of the alloys, which were measured by inductively coupled plasma atomic emission spectrophotometry (CIROS-120EOP, Rigaku Cor-poration, Japan). Z1 had the highest concentrations of Cr and Ni, followed by Z2 and so on. The contributions of Cr and Ni to corrosion resistance appeared to decrease with increasing concentrations. The Creq/Nieq ratios as calculated by Eqs. (2) and (3) had a range of 1.13–1.15 for all alloys with Z3 having the lowest value and Z5 the highest. As noted above, synchrotron radiation measurement at SPring-8 Table 1.  Chemical compositions (wt%) of the Fe–15Mn–aCr–bNi–4Si alloys [(a, b) =  (14, 10.1), (12.5, 8.8), (11, 7.5), (9.5, 6.1), (8, 4.8)].Alloy Fe Mn Cr Ni Si Creq/NieqΔGγ→ε  [J/mol]ΓSFE [mJ/m2]Z1 Bal. 15.1 14.0 10.0 4.06 1.14 34.4 21.04Z2 Bal. 15.0 12.5 8.74 4.06 1.14 −13.4 18.21Z3 Bal. 15.0 11.0 7.53 4.00 1.13 −61.0 15.39Z4 Bal. 15.0 9.48 6.12 4.08 1.15 −117.1 12.07Z5 Bal. 14.9 8.00 4.83 4.11 1.15 −172.7   8.78ISIJ International, Vol. 64 (2024), No. 7©  2024  ISIJ3revealed that Z3 solidified in FA mode.34,35) Synchrotron radiation measurements were also performed on Z5 to confirm whether all alloys in this study solidified in FA mode. Figure 1 shows the (a) synchrotron radiation X-ray radiograph and (b), (c) diffraction patterns obtained as Z5 solidified during spot welding. Z5 was observed to solidify in FA mode based on the following features: primary colum-nar δ-ferritic dendrites grew from the weld pool boundary to the center, and the secondary equiaxed γ-austenitic dendrites nucleated in front of the columnar dendrites. Therefore, Z1–Z5 was concluded to all solidify in FA mode.Table 1 presents the ΔGγ→ε values at room temperature (25°C) as calculated by Eq. (1). ΔGγ→ε decreased with decreasing Cr and Ni concentrations, which indicates an increase in the stability of ε-martensite against γ-austenite. Table 1 also presents the stacking fault energy (SFE) at room temperature, which is calculated as follows:  � SFE � ��2 2� �� � � ��G / ...................... (4)The SFE is calculated from the molar surface fraction on the γ plane ρ, ΔGγ→ε, and the interfacial surface energy between the γ- and ε-phases σγ /ε. A detailed description of the SFE and the calculation of ΔGγ→ε are presented in the previous paper.30) The effect of Si on the γ→ε-martensitic transformation in the SFE or ΔGγ→ε calculation is given by the equation of Dumay et al.36) Si has a nonlinear effect on the transformation and does not significantly affect the transformation temperature. However, it plays an important role in improving the plastic fatigue life.37)3.2.  ExperimentsDSC (TA Instrument Q-2000) was used to evaluate the martensitic transformation temperatures. Specimens were cut into dimensions of 3 mm ×  3 mm ×  0.4 mm with low-speed cutter. The measurement program was set as follows. Specimens were fully austenitized to 400°C and cooled to −150°C to measure the martensite start (Ms) and martensite finish (Mf) temperatures. Then they were heated to 400°C to measure the austenite start (As) and austenite finish (Af) temperatures. The cooling and heating rates were both 20°C/min.3.2.1.  Low-cycle Fatigue TestAxial strain-controlled fatigue tests were conducted at room temperature by using a hydraulic servo fatigue test-ing machine (MTS 370, MTS Systems Corporation, USA) equipped with a 250 kN load cell (661.22H-01, MTS Sys-tems Corporation, USA). The gauge length of the extensom-eter (632.26F-20, MTS Systems Corporation) was 8 mm. A triangular test waveform was used with a strain rate of 0.4%/s and total strain amplitude of 1%. The specimens were machined so that the loading axis was parallel to the rolling axis. All tests were continued until specimen failure.3.2.2.  Microstructural AnalysisMicrostructural analyses were conducted before and after fatigue tests via XRD (SmartLab, Rigaku Corpora-tion, Japan), field-emission scanning electron microscopy (JEOL-7900F, JEOL Corporation, Japan) equipped with an EBSD detector (DVC5, TSL Corporation, Japan) and an EDS detector (JED-2300, JEOL Corporation, Japan), and a ferrite scope (FMP30, Fischer Technology Incorporated, USA). Specimens were cut from the as-heat-treated plate for analysis of the initial microstructure and from the parallel part of the fractured specimen for analysis of the deformed microstructures. The specimens were polished with water-proof papers up to #1200 followed by electropolishing by a solution of 70% perchloric acid (30 mL) with 99.5% ethanol (470 mL) for 600 s at 18 V and 15°C. The electropolishing removed approximately 45 μm of the thickness.3.2.3.  Corrosion Resistance(1)  Anodic Polarization TestThe anodic polarization test was performed by using a potentiostat (HABF-501A, Hokuto Denko Corp., Japan) connected to a function generator (HB-111, Hokuto Denko, Japan) with an analog cable. A silver–silver chloride (Ag/AgCl) electrode in a saturated KCl solution and a platinum electrode were used as the reference and counter electrodes, respectively. The specimens were cut by using a fine cutter into small pieces with dimensions of 10 mm ×  10 mm, and the surfaces were wet-ground by using SiC water-resistant abrasive paper (KOVAX Corp., Japan). The surfaces were then finished by using #800 grit papers. The specimens were fixed in a polytetrafluoroethylene holder with an O-ring. The exposed area contacting the electrolyte was 0.35 cm2 (6.7 mm in diameter). After the specimens were immersed in artificial seawater (SF-1, Osakayakken Co. Ltd., Japan) at 25°C under the open-air condition, the open circuit potential (OCP) was recorded for 30 min. Then, a gradient anodic potential was applied at a constant sweep rate of 20 mV min −1 from the initial potential of −100 mV from the final OCP value. The measurement was stopped when the current density limit of 1 mA cm −2 was recorded.(2)  Salt Spray TestThe salt spray test (SST) was performed according to the following procedure. Specimens were cut by a fine cutter into sheets approximately 1 mm in thickness, and the cross-sections were wet-ground with SiC water-resistant abrasive paper (KOVAX Corp., Japan). The surface was finished by using #800 grit paper. The sides and backside surfaces of the specimens were coated with insulating resin paint, which left a top surface area of 10 mm ×  30 mm to ensure that only the target surface was in contact with the salt solution mist. The specimens were placed in the chamber of a salt spray testing machine (STP-30, Suga Test Instruments Co., Fig. 1.  Solidification sequence of Z5 during spot welding: (a) syn-chrotron radiation X-ray radiograph and diffraction pat-terns detected during (b) columnar dendrite growth and (c) equiaxial dendrite nucleation. (Online version in color.)ISIJ International, Vol. 64 (2024), No. 7©  2024  ISIJ 4Ltd., Japan) at a 70° horizontal angle. The test solution was 50 g L −1 NaCl aqueous solution, the test temperature was 35°C, and the amount of salt spray injected in the testing chamber was adjusted within the range of 18.75 ±  6.25 μL h −1 cm −2 (1.5 ±  0.5 mL h −1 for the 80 cm2 aperture area of the horizontally placed collection vessel). After 24 h from the start of the test, the specimens were temporarily taken out, gently rinsed with pure water, and then dried quickly. Surface observation and morphology measurements were performed, followed by an additional 24 h of SST. The sur-face characterization was performed using a 3D optical/laser microscope (VR-3200, KEYENCE Corp., Japan).4.  Results4.1.  Differential Scanning CalorimetryThe martensitic transformation temperatures of Ms, As, and Af were identified by using DSC. Figure 2 shows the (a) colling- and (b) heating-DSC curves. Table 2 lists the martensitic transformation temperatures. Our recent work confirmed that the transformation between γ-austenite and ε-martensite occurred during in-situ cooling XRD measure-ments in Fe-15Mn-10Cr-8Ni-4Si alloy.30) Similarly, the same type of transformation, not involving α′-martensite, may occur in the present Fe–15Mn–aCr–bNi–4Si alloys at each transformation temperature measured by DSC. As shown in Fig. 2 and Table 2, the martensitic transformation temperatures increased with decreasing Cr and Ni concen-trations or decreasing ΔGγ→ε and ΓSFE. Z2 showed a small peak for only the heating curve, and Z1 did not show mar-tensitic transformation peaks for both the cooling and heat-ing curves. The Ms values of Z1 and Z2 were estimated by extrapolation of the trend for Z3–Z5 as −120 and −80°C, respectively. In the low temperature region, however, γ-austenite may be stabilized by the effect of Néel point,34) resulting in lower Ms. If this is the case, sufficient driving force to continue the transformation cannot be obtained and the corresponding peak becomes weaker. Z4 and Z5 had rel-atively high Ms compared to Z1–Z3, and Z5 had a Ms over room temperature at 55.9°C. Therefore, reducing the Cr and Ni concentrations facilitated the martensitic transformation.4.2.  Initial MicrostructureThe microstructure after heat treatment (i.e., initial microstructure) was examined by XRD, EBSD, and a ferrite scope. Figure 3 shows the XRD spectra for Z1–Z5. Some of the identified peaks were very weak and corresponded to ε-martensite in Z5 with high 2θ. Z1–3 had fully austenitic structures while Z4 and Z5 contained ε-martensite because of their high Ms. Figure 4 shows the wide-field EBSD phase maps. As expected from the XRD results, Z1–Z3 had fully austenitic microstructures while Z4 and Z5 had a dual-phase microstructure comprising mostly γ-austenite with ε-martensite fractions of 3.5% and 25.6%, respectively. The grain sizes were measured counting high angle bound-aries (>  15°) including the twin and γ /ε boundaries in the EBSD phase maps as 84.3 μm for Z1, 77.3 μm for Z2, 70.2 μm for Z3, 70.6 μm for Z4, and 57.2 μm for Z5. The specimens were examined by using the ferrite scope, but no α′-martensite was detected in any alloy.4.3.  Low-cycle Fatigue TestFigure 5 shows that the fatigue lives Nf of the alloys strongly depended on their chemical compositions. Z2 had the highest Nf of 22 750 cycles at a total strain amplitude of 1%, which was twice that of Z3 (10 450 cycles) devel-oped in the previous study. Figure 6 shows the relationship Fig. 2.  DSC measurements during (a) cooling and (b) heating. (Online version in color.)Table 2.  Martensitic transformation temperature of the Fe–15Mn–aCr–bNi–4Si alloys [(a, b) =  (14, 10.1), (12.5, 8.8), (11, 7.5), (9.5, 6.1), (8, 4.8)].Alloy Ms [°C] As [°C] Af [°C]Z1 ND ND NDZ2 ND ND NDZ3 −47.1   59.1   94.9Z4 16.1   96.0 122.2Z5 55.9 121.4 150.0 Fig. 3.  X-ray diffraction spectra of Z1–Z5 after heat treatment. (Online version in color.)ISIJ International, Vol. 64 (2024), No. 7©  2024  ISIJ5between Nf and ΔGγ→ε at room temperature. The results for Fe–15Mn–10Cr–8Ni–4Si, which was tested at different tem-peratures to obtain different ΔGγ→ε, are also shown.30) The same tendency was observed in the Nf–ΔGγ→ε relationship for both Fe–15Mn–10Cr–8Ni–4Si at different temperatures and for Z1–Z5 at room temperature: Nf took its maximum value near ΔGγ→ε =  0, gradually decreased with decreasing ΔGγ→ε to approximately 5 000 cycles in the negative ΔGγ→ε regime, and rapidly decreased to below 5 000 cycles with increasing ΔGγ→ε in the positive ΔGγ→ε regime.4.4.  Microstructural AnalysisThe deformation microstructure that developed after fatigue fracture (i.e., post-fatigue microstructure) was investigated. The XRD measurements revealed that the post-fatigue microstructure comprised retained γ-austenite and deformation-induced martensite, as shown in Fig. 7. Some of the identified peaks were very weak. Z1 exhibited strong peaks for γ-austenite and weak peaks for ε-martensite. α′-martensite was detected in Z4 and Z5 but not in Z1–Z3.Figure 8 shows the EBSD phase maps. In Z1, ε-martensite formed locally along the loading or rolling direction, as shown in Fig. 8(a). Figures 8(b)–8(e) show the correspond-ing EDS maps of Mn, Cr, Ni, and Si for Z1, where a brighter color corresponds to a higher concentration, and indicates Fig. 4.  Phase maps obtained by electron backscattering diffraction of (a) Z1, (b) Z2, (c) Z3, (d) Z4, and (e) Z5 after heat treatment. The field of view was 1.6 mm ×  1.6 mm, and the step size was 2 μm. (Online version in color.)Fig. 5.  Fatigue lives of Z1–Z5 at a strain ratio of −1 and total strain amplitude of 1%. (Online version in color.)Fig. 6.  Relationship between the fatigue life and Gibbs free energy difference between γ-austenite and ε-martensite ΔGγ→ε for Z1–Z5. The results obtained for Fe–15Mn–10Cr–8Ni–4Si at −20°C–120°C are also shown for com-parison.30) (Online version in color.)ISIJ International, Vol. 64 (2024), No. 7©  2024  ISIJ 6strong rolling segregation in Mn, Cr, and Ni. The distribu-tion trends of Mn and Ni were in good agreement, while the Cr distribution was opposite. The rolling segregation varied a local ΔGγ→ε and may cause the local γ→ε-martensitic transformation in the area with low ΔGγ→ε Figs. 8(c) and 8(d) show that the retained austenite in Z2 and Z3, respec-tively, was island-shaped and tended to be elongated along the rolling direction, which may also be attributed to the segregation. However, the effect of segregation is outside the scope of the present study. Figure 9 shows the area fractions of each phase measured on the EBSD maps in Fig. 8. The fraction of ε-martensite increased with decreasing Cr and Ni concentrations, and Z4 and Z5 contained over 80% ε-martensite. Z2 and Z3 exhibited relatively large fractions of ε-martensite but also contained considerable amounts of retained austenite. Figure 9 show the fraction of α′-martensite measured by the ferrite scope. The fraction of α′-martensite increased with the Cr and Ni concentra-tions. No α′-martensite was detected in Z1. While XRD and EBSD did not identify any α′-martensite in Z2, the ferrite scope detected a small amount. Considerable amounts of α′-martensite were detected in Z4 and Z5.4.5.  Corrosion ResistanceFigure 10 shows the polarization curves of the specimens together with the reference carbon steel (JIS SM490A) in artificial seawater. The specimens had much nobler OCP values than the carbon steel. In addition, the current den-sity of the specimens around the OCPs was apparently suppressed. These results indicate that the carbon steel was actively dissolved without forming any protective layer under the anodic polarization condition while the specimens were passivated. The base current density at the passive potential region was on the order of 10 −6 A cm −2, which is representative behavior of typical passivation-type corrosion-resistant alloys.38) However, unstable current behavior indicating the transient occurrence of metastable Fig. 7.  X-ray diffraction spectra of Z1–Z5 after fatigue fracture. (Online version in color.)Fig. 8.  Phase maps obtained by electron backscattering diffraction and Mn maps obtained by energy-dispersive X-ray spectroscopy after fatigue fracture: (a) EBSD and (b) Mn, (c) Cr, (d) Ni, and (e) Si EDS maps for Z1 and EBSD maps for (f) Z2, (g) Z3, (h) Z4, and (i) Z5. The EBSD had a field of view of 800 μm ×  800 μm and a step size of 1 μm. (Online version in color.)Fig. 9.  Phase fractions in Z1–Z5 after fatigue fracture: (a) area fractions measured in the electron backscattering dif-fraction phase maps of Fig. 8 and (b) volume fractions of α′-martensite measured by the ferrite scope. The step size dependence of the phase fraction in the EBSD measurement in the present microstructure was confirmed to be small. (Online version in color.)ISIJ International, Vol. 64 (2024), No. 7©  2024  ISIJ7pitting corrosion and immediate re-passivation was fre-quently observed. At a potential of 0–0.2 V, breakdown of the passive film occurred, and stable corrosion pit propaga-tion occurred. These results indicate that the specimens had superior corrosion resistance compared to carbon steel due to their passivation nature even in a seawater environment.The changes in appearance of the specimens before and after SST are shown in Fig. 11(a). The pale-green insulat-ing resin coating surrounding the area of the test surface demonstrated excellent adhesion and durability even after 48 h of testing. All specimen surfaces showed the formation of rust after 24 h of SST. For the carbon steel, a red-brown rust layer broadly developed and covered the surface. In contrast, the specimen surfaces showed local corroded sites, and the rust seemed to flow downward from the specimens placed in a standing position. Figure 11(b) shows the topo-graphic profiles at the same observed locations. The rusted areas of the specimens had relatively concave pit shapes, and the rust precipitated around the pit. This indicates that these specimens were loosely passivated in the SST envi-ronment and that they exhibited localized corrosion. The appearance of these specimens remained almost unchanged after an additional 24 hours of SST. The corrosion pits may have been re-passivated by rinsing and drying for observa-tion, so they were never reactivated in the additional testing period. In addition, these specimens showed almost no signs of new corrosion pits on their surfaces. In contrast, the car-bon steel exhibited further corrosion during the additional testing period.The change in surface roughness was measured for a quasi-quantitative evaluation of the degree of corrosion. The results are presented in Table 3. The surface roughness of the specimens before SST was generally at the same level, although there was only slight variation in the submicron order among the specimens. These differences were prob-ably introduced not only by the intrinsic fine profile factor, but also by the macroscopic specimen shape like as the warpage and the waviness. The surface roughness tended to increase with the length of the testing period. It is due to the corrosion phenomena forming both the concave and convex areas: initiation/propagation of the corrosion pits and deposition of the corrosion products, respectively. The specimens showed a much smaller increase in roughness than the carbon steel. In addition, the increase in rough-ness was suppressed during the additional 24-h testing period compared to the initial 24-h period. The degree of the roughness increase was smaller among specimens with higher Cr and Ni concentrations. Therefore, these elements were concluded to help enhance the corrosion resistance of Fig. 10.  Polarization curves of Z1–Z5. The curve for carbon steel (JIS SM490A) is also shown for comparison. (Online ver-sion in color.)Fig. 11.  Appearance of specimens before and after salt spray test. (Online version in color.)Table 3.  Changes in arithmetic surface roughness, Sa (μm) of the specimens before and after the salt spray test.Alloy Z1 Z2 Z3 Z4 Z5 Carbon steelBefore SST 0.6 0.5 0.9 0.4 0.5   0.4After SST (24 h) 0.9 0.8 1.3 1.9 2.9 27.7After SST (48 h) 1.1 1.0 1.5 2.3 3.6 52.7ISIJ International, Vol. 64 (2024), No. 7©  2024  ISIJ 8the developed alloys.5.  DiscussionThe fatigue life Nf of the Fe–Mn–Cr–Ni–Si alloys strongly depended on the Cr and Ni concentrations. Nf was 3 197 cycles for Z5 and 22 750 cycles for Z2. The difference in Nf can be explained by the deformation mechanism based on ΔGγ→ε. As shown in Fig. 6, the relationship between Nf and ΔGγ→ε showed the same trend for Z1–Z5 as for the previously developed alloy Fe–15Mn–10Cr–8Ni–4Si at various temperatures:30) Nf is maximum at ΔGγ→ε ≈ 0, gradu-ally decreases with decreasing ΔGγ→ε in the negative ΔGγ→ε regime, and rapidly decreases with increasing ΔGγ→ε in the positive ΔGγ→ε regime. These results indicate that both the effects of chemical composition and test temperature on the Fe–Mn–Cr–Ni–Si alloy system can be evaluated by using ΔGγ→ε to characterize the deformation mechanism. The effects of the Cr and Ni concentrations on the fatigue life can thus be discussed together using ΔGγ→ε rather than individually. Z2 exhibited the longest Nf of over 20 000 cycles, which was twice that of the previously developed Z3 (over 10 000 cycles), which itself also had an extraordi-nary long Nf compared to conventional steels (about 2 000 cycles).6) Z2 and Z3 had ΔGγ→ε of −13.4 and −61.0 J/mol, respectively. A negative near-zero ΔGγ→ε value is favorable for the B-TRIP mechanism.18,30) The high Nf of both Z2 and Z3 indicates the effectiveness of the B-TRIP mechanism at increasing fatigue resistance. Z2 has a ΔGγ→ε value closer to zero than Z3, so it has a smaller potential that must be overcome to reversely transform from ε-martensite to γ-austenite. The smaller fraction of ε-martensite in Z2 com-pared to Z3 shown in Fig. 9(a) implies the frequent occur-rence of the reverse transformation. The easier occurrence of the B-TRIP mechanism in Z2 may significantly mitigate damage accumulation during cyclic loadings. However, it is possible that the low amount of ε-martensite in Z2 compared to that in Z3 is a result of the stabilization of γ-austenite and suppression of the γ→ε martensitic transformation due to the increase in Mn content, as well as the reverse transfor-mation. It is also worth noting that the moderate suppression of the γ→ε transformation in Z2, as compared to Z3, may have contributed to the improvement in fatigue durability in Z2. Further research is needed to fully understand the effect of the γ→ε transformation or the formation and deformation of ε-martensite on fatigue life. Additionally, the fraction of α′-martensite was larger in Z3 than in Z2. In the γ→ε martensitic transformation, the reverse ε→γ transformation occurs due to the inverse motion of the Schockley partial dislocation, but this reverse transformation mechanism does not operate in the γ→α′ martensitic transformation because of a large amount of slip introduced by the lattice-invariant deformation. Additionally, sufficient driving force for the reverse transformation cannot be obtained since the thermal equilibrium temperature in the γ→α′ martensitic transfor-mation is higher than that in the γ→ε martensitic trans-formation. Therefore, the large fraction of α′-martensite in Z3 can explain the reduced Nf compared to Z2, where the α′-martensitic transformation was suppressed well.Z1 exhibited a smaller Nf than Z2 due to the deviation of its ΔGγ→ε value, which is not conducive to the B-TRIP mechanism. However, it still had a greater Nf than that reported for conventional steels. Mughrabi pointed out that the occurrence of cross-slip reduces the reversibility of dis-location motions because the reverse glide of dislocations cannot occur on the same slip plane with the forward glide once cross-slip has occurred.19) Z1 had a low ΓSFE of 21.04 mJ/m2, which indicates that dislocations can easily extend and cross-slip is suppressed. The high reversibility of the extended dislocation motion may be responsible for the relatively long Nf of Z1 even though it does not follow the B-TRIP mechanism.The dominant deformation mechanism in Z4 and Z5 might be stress-assisted ε-martensitic transformation, which can be attributed to their high Ms. The large fractions of ε-martensite as shown in Figs. 8 and 9 also indicate the high tendency for ε-martensitic transformation. The gradual decrease in Nf with decreasing ΔGγ→ε in the negative ΔGγ→ε regime (Z2–Z5) may be attributed to the reduction in the frequency of bidirectional transformation. Z4 and Z5 had a lower ΓSFE than Z2 and Z3, which implies that their disloca-tion motion was more reversible. Meanwhile, ε-martensite is considered a brittle phase that causes premature frac-ture.39,40) The poor tendency for reversible ε→γ transfor-mation and the detrimental effect of ε-martensite may be responsible for the reduced Nf of Z4 and Z5 compared to Z2 and Z3.As shown in Figs. 10 and 11, the developed alloys exhib-ited much better corrosion resistance than carbon steel due to passivation. The developed alloys demonstrated localized corrosion behavior in an artificial seawater: breakdown of the passivated film and localized pitting corrosion. Stainless steel is defined as containing more than 10.5% Cr.41) It is interesting that the compositions of Z4 and Z5 in this study did not satisfy this requirement, although they showed local-ized corrosion behavior due to the spontaneous formation of passive film. The pitting corrosion may occur at MnS-type inclusions caused by the high Mn concentration of 15 wt%. In the case of stainless steels processings, the Mn and S concentrations are limited to improve the pitting corrosion resistance.42) In fact, detrimental effect of high content of Mn on corrosion resistance of stainless steel was reported.43) On the other hand, the addition of Si to stainless steels has also been found to improve corrosion resistance.44,45) Therefore, the superior corrosion resistance of the speci-mens with relatively lower Cr and high Mn content, such as Z4 and Z5, is considered to be due to the enhancement effect on passive film formation by Si. In our earlier study, Nishimura46) reported that the passivation films formed on the steel with similar compositions to those in this study consisted of two-layer structure with an Si-enriched inner layer. While pitting corrosion occurred during the SST, the increase in the number of pits seemed to saturate after 24 h. Additionally, most of the corrosion pits were re-passivated, which prevented the corrosion from proceeding further. The number of pits tended to decrease with increasing Cr and Ni concentrations. Z1 and Z2 had the first and second highest Cr and Ni concentrations and exhibited fewer corrosion pits as well as a smaller change in surface roughness than the previously developed Z3. Therefore, the developed alloys demonstrated pitting corrosion that can be attributed to the high Mn concentration, but presence of Si and increasing ISIJ International, Vol. 64 (2024), No. 7©  2024  ISIJ9the Cr and Ni concentrations helped improve the corrosion resistance.6.  ConclusionThe fatigue and corrosion resistant alloy were newly developed by optimizing Cr and Ni concentrations, by adapting the design concept for fatigue resistant alloy based on bidirectional TRIP (B-TRIP). Compared with the previously developed damping steel Fe–15Mn–10Cr–8Ni–4Si and Fe-15Mn-11Cr-7.5Ni-4Si (Z3), newly developed 15Mn–12.5Cr–8.7Ni–4Si alloy (Z2), slightly increasing the Cr and Ni concentrations optimized ΔGγ→ε to increase Nf twofold. Increasing the Cr and Ni concentrations simul-taneously increased the corrosion resistance. During SST, the developed alloys exhibited pitting corrosion that can be attributed to the high Mn concentration, so their applica-tion in harsh corrosive environments should be carefully considered. However, the improved corrosion resistance of Z2 compared to Z3 indicates that increasing the Cr and Ni concentrations can extend the range of applications of fatigue-resistant steel alloys. Future work will involve improving the corrosion resistance of these alloys, such as by reducing the Mn concentration, while maintaining the excellent fatigue resistance and weldability.AcknowledgmentsWe would like to acknowledge the assistance provided by the Materials Manufacturing and Engineering Sta-tion, NIMS, for material processing. We also acknowl-edge the support of a NEDO Leading Research Program (JPNP14004) from the New Energy and Industrial Technol-ogy Development Organization (NEDO), Japan. A part of this work was supported by a Grant-in-Aid for Early-Career Scientists (21H01659, 23K13227) from the Japan Society for the Promotion of Science (JSPS), Japan.REFERENCES1)  B. Spencer, Jr. and S. Nagarajaiah: J. Struct. Eng., 129 (2003), 845. https://doi.org/10.1061/(ASCE)0733-9445(2003)129:7(845)2)  D. R. Teruna, T. A. Majid and B. Budiono: Adv. Civil. Eng., 2015 (2015), 1. https://doi.org/10.1155/2015/6317263)  C. Fang, W. Wang, Y. Ji and M. C. Yam: J. Const. Steel. Research., 184 (2021), 106817.4)  I. Nikulin, T. Sawaguchi, A. Kushibe, Y. Inoue, H. Otsuka and K. Tsuzaki: Int. J. Fatigue., 88 (2016), 132. https://doi.org/10.1016/j.ijfatigue.2016.03.0215)  F. Yoshinaka, T. Sawaguchi, S. Takamori, S. Emura and Y. Inoue: Int. J. Fatigue, 171 (2023), 107581. https://doi.org/10.1016/j.ijfatigue. 2023.1075816)  F. Yoshinaka, T. Sawaguchi, S. Takamori, T. Nakamura, G. Arakane, Y. Inoue, S. Motomura and A. Kushibe: Scripta Mater., 197 (2021), 113815. https://doi.org/10.1016/j.scriptamat.2021.1138157)  K. Hatanaka: JSME International Journal. Ser. 1, Solid Mechanics, Strength of Materials, 33 (1990), 13. https://doi.org/10.1299/jsmea1988.33.1_138)  T. Hilditch, H. Beladi, P. Hodgson and N. Stanford: Mater. Sci. Eng. A., 534 (2012), 288. https://doi.org/10.1016/j.msea.2011.11.0719)  Y. Wen, H. Peng, D. Raabe, I. Gutiérrez-Urrutia, J. Chen and Y. Du: Nature Com., 5 (2014), 4964. https://doi.org/10.1038/ncomms596410)  C. Shao, P. Zhang, R. Liu, Z. Zhang, J. Pang and Z. Zhang: Acta Mater., 103 (2016), 781. https://doi.org/10.1016/j.actamat.2015.11.01511)  C. Shao, P. Zhang, R. Liu, Z. Zhang, J. Pang, Q. Duan and Z. Zhang: Acta Mater., 118 (2016), 196. https://doi.org/10.1016/j.actamat. 2016.07.03412)  W. Tasaki, T. Sawaguchi, I. Nikulin, K. Sekido and K. Tsuchiya: Mater. Trans., 57 (2016), 639. https://doi.org/10.2320/matertrans.MBW20150313)  M. W. Klein, M. Smaga and T. Beck: Metals, 8 (2018), 832. https://doi.org/10.3390/met810083214)  K. Lu, A. Chauhan, M. Walter, A. S. Tirunilai, M. Schneider, G. Laplanche, J. Freudenberger, A. Kauffmann, M. Heilmaier and J. Aktaa: Scripta Mate., 194 (2021), 113667. https://doi.org/10.1016/j.scriptamat.2020.11366715)  S. A. A. Shams, J. W. Bae, J. N. Kim, H. S. Kim, T. Lee and C. S. Lee: J. Mater. Sci. Technol., 115 (2022), 115. https://doi.org/10.1016/j.jmst.2021.10.01016)  W. Lu, C. H. Liebscher, G. Dehm, D. Raabe and Z. Li: Adv. Mate-rial., 30 (2018), 1804727. https://doi.org/10.1002/adma.20180472717)  F. Yoshinaka, T. Sawaguchi, S. Takamori and S. Emura: Materi. Sci. Eng. A, 833 (2022), 142583. https://doi.org/10.1016/j.msea.2021. 14258318)  T. Sawaguchi, Y. Tomota, F. Yoshinaka and S. Harjo: Acta Mater., 242 (2023), 118494. https://doi.org/10.1016/j.actamat.2022.11849419)  H. Mughrabi: Metall. Mater. Trans. B, 40 (2009), 431. https://doi.org/10.1007/s11663-009-9240-420)  F. Bahadur, R. Jain, K. Biswas and N. Gurao: Int. J. Fatigue, 155 (2022), 106545. https://doi.org/10.1016/j.ijfatigue.2021.10654521)  T. Sawaguchi, T. Maruyama, H. Otsuka, A. Kushibe, Y. Inoue and K. Tsuzaki: Mater. Trans., 57 (2016), 283. https://doi.org/10.2320/matertrans.MB20151022)  Y. Inoue, A. Kushibe, K. Umemura, Y. Mizushima, T. Sawaguchi, T. Nakamura, H. Otsuka and Y. Chiba: J. Arc. Review, 4 (2021), 76. https://doi.org/10.1002/2475-8876.1219323)  D. A. Jones: Principles and Prevention, Corrosion, 2 (1996), 168.24)  R. Ebara: Procedia Eng., 2 (2010), 1297. https://doi.org/10.1016/j.proeng.2010.03.14125)  M. Amirifard, A. Z. Hanzaki, H. Abedi, N. Eftekhari and Q. Wang: J. Mater. Res. Technol., 17 (2022), 1672. https://doi.org/10.1016/j.jmrt.2022.01.10926)  V. Urban, V. Krivy and K. Kreislova: Proc. Eng., 114 (2015), 546. https://doi.org/10.1016/j.proeng.2015.08.10427)  K. J. Kere and Q. Huang: J. Bridge Eng., 24 (2019), 04019007. https://doi.org/10.1061/(ASCE)BE.1943-5592.000136128)  M. H. Hebdon and J. T. Provines: J. Perform. Const. Facil., 34 (2020), 04020108. https://doi.org/10.1061/(ASCE)CF.1943-5509.000149829)  H. Iken, R. Basseguy, A. Guenbour and A. B. Bachir: Electro. Acta, 52 (2007), 2580. https://doi.org/10.1016/j.electacta.2006.09.01330)  T. Sawaguchi, I. Nikulin, K. Ogawa, S. Takamori, F. Yoshinaka, Y. Chiba, H. Otsuka, Y. Inoue and A. Kushibe: Acta Mater., 220 (2021), 117267. https://doi.org/10.1016/j.actamat.2021.11726731)  S. Katayama, T. Fujimoto and A. Matsunawa: Correlation among Solidification Process, Microstructure, Microsegregation and Solidifi-cation Cracking Susceptibility in Stainless Steel Weld Metals (Mate-rials, Metallurgy & Weldability), 14 (1985). https://doi.org/10.18910/ 1094232)  J. Lippold: Solidification Behavior and Cracking Susceptibility of Pulsed-laser Welds in Austenitic Stainless Steels, Welding Journal Including Welding Research Supplement, 73 (1994), 129s.33)  P. Guiraldenq and O. Hardouin Duparc: Metall. Res. Technol., 114 (2017), 613. https://doi.org/10.1051/metal/201705934)  T. Nagira, T. Nakamura, F. Yoshinaka, T. Sawaguchi, Y. Aoki, M. Kamai, H. Fujii, A. Takeuchi and M. Uesugi: Scripta Mater., 216 (2022), 114743. https://doi.org/10.1016/j.scriptamat.2022.11474335)  T. Nagira, T. Nakamura, T. Kimura, F. Yoshinaka, T. Sawaguchi, T. Yamashita, Y. Aoki and H. Fujii: Quarterly Journal of the Japan Welding Society, 41 (2023), 1s. https://doi.org/10.2207/qjjws.41.1s36)  A. Dumay, J.-P. Chateau, S. Allain, S. Migot and O. Bouaziz: Sci. Eng. A, 483 (2008), 184. https://doi.org/10.1016/j.msea.2006.12.17037)  I. Nikulin, T. Sawaguchi, K. Ogawa and K. Tsuzaki: Acta Mater., 105 (2016), 207. https://doi.org/10.1016/j.actamat.2015.12.00238)  Y. Tsutsumi, S. Bartáková, P. Prachár, S. Migita, H. Doi, N. Nomura and T. Hanawa: J. Electro. Soc., 159 (2012), C435. https://doi.org/10.1149/2.045210jes39)  M. Koyama, T. Sawaguchi and K. Tsuzaki: Metall. Mater. Trans. A, 43 (2012), 4063. https://doi.org/10.1007/s11661-012-1220-740)  F. Yoshinaka and T. Sawaguchi: Int. J. Fatigue, 130 (2020), 105271. https://doi.org/10.1016/j.ijfatigue.2019.10527141)  ISO 15510: 2014, Stainless steels -- Chemical composition.42)  A. Chiba, I. Muto, Y. Sugawara and N. Hara: J. Electro. Soc., 160 (2013), C511. https://doi.org/10.1149/2.081310jes43)  R. F. A. Jargelius-Pettersson: ISIJ Int., 36 (1996), 818. https://doi.org/10.2355/isijinternational.36.81844)  W. T. Tsai, Y. N. Wen, J. T. Lee, H. Y. Liou and W. F. Wang: Surf. Coat. Tech., 34 (1988), 209. https://doi.org/10.1016/0257-8972(88)90113-245)  R. Nishimura, K. Yamakawa, J. Ishiga, Y. Matsumoto and H. Nagano: Mater. Chem. Phys., 54 (1998), 289. https://doi.org/10.1016/S0254-0584(98)00104-746)  T. Nishimura: Mater. Trans., 55 (2014), 871. https://doi.org/10.2320/matertrans.M2014029