# Fileset

[Kadowaki_2026_J._Electrochem._Soc._173_091502.pdf](https://mdr.nims.go.jp/filesets/c3c942ad-d2d1-4be7-9dad-d62b86d2c3ae/download)

## Creator

[Mariko Kadowaki](https://orcid.org/0000-0002-8988-3545), [Masahiro Yamamoto](https://orcid.org/0009-0002-1861-7757), [Hideki Katayama](https://orcid.org/0000-0001-7947-4687), [Arkapol Saengdeejing](https://orcid.org/0000-0001-8739-3262), [Taichi Abe](https://orcid.org/0000-0002-5065-0939), [Ryoji Sahara](https://orcid.org/0000-0003-0788-2985), [Kotaro Doi](https://orcid.org/0000-0002-5204-1088), [Yoshiharu Murase](https://orcid.org/0000-0001-7390-851X), [Sachiko Hiromoto](https://orcid.org/0000-0003-4666-6708), [Yusuke Tsutsumi](https://orcid.org/0000-0002-9483-1256)

## Rights

[Creative Commons BY Attribution 4.0 International](https://creativecommons.org/licenses/by/4.0/)

## Other metadata

[Key Role of Magnetite in Rust Layer of Carbon Steel for Cathodic Reactions and Hydrogen Permeation](https://mdr.nims.go.jp/datasets/abe63462-e673-495d-82dd-f8314d558fba)

## Fulltext

Key Role of Magnetite in Rust Layer of Carbon Steel for Cathodic Reactions and Hydrogen PermeationJournal of TheElectrochemical Society      OPEN ACCESSKey Role of Magnetite in Rust Layer of CarbonSteel for Cathodic Reactions and HydrogenPermeationTo cite this article: Mariko Kadowaki et al 2026 J. Electrochem. Soc. 173 091502 View the article online for updates and enhancements.You may also likeDirect Reuse of Degraded LiFePO4Electrodes with LiBH4 as ChemicalLithiation ReagentFeifan Li, Chen Chen, Zhen Zheng et al.-Electrochemical Behavior of ScandiumIons in the LiCl-KCl Eutectic MeltMaxim V. Erzhenkov, Andrey Yu.Nikolaev, Olga V. Grishenkova et al.-Breaking the Bottleneck: Pt–Ni Nanoalloysas Next-Generation Catalysts for OxygenReduction in Proton Exchange MembraneFuel CellsAmirhossein Amarlou, ShahrzadTorkashvand, Mobina Ghadri et al.-This content was downloaded from IP address 144.213.253.16 on 13/05/2026 at 08:13https://doi.org/10.1149/1945-7111/ae6483https://iopscience.iop.org/article/10.1149/1945-7111/ae6401https://iopscience.iop.org/article/10.1149/1945-7111/ae6401https://iopscience.iop.org/article/10.1149/1945-7111/ae6401https://iopscience.iop.org/article/10.1149/1945-7111/ae6401https://iopscience.iop.org/article/10.1149/1945-7111/ae6401https://iopscience.iop.org/article/10.1149/1945-7111/ae6373https://iopscience.iop.org/article/10.1149/1945-7111/ae6373https://iopscience.iop.org/article/10.1149/1945-7111/ae62adhttps://iopscience.iop.org/article/10.1149/1945-7111/ae62adhttps://iopscience.iop.org/article/10.1149/1945-7111/ae62adhttps://iopscience.iop.org/article/10.1149/1945-7111/ae62adhttps://pagead2.googlesyndication.com/pcs/click?xai=AKAOjssmwxSTze3YzicG7OMm9u4W5vcbXW7gr9uBYFQhnEF7GMfPwUXbMN6sjunWCZEasz7JbzTxQ36RKPya5asBOSSPLRQRlhvoMcli8rIRjQrrEZD3WBB7UGqRDTQ8O-C9I4oRwNGkCo6jkVJZ8gZVAXAlH56E_FIROK1bZ5GruzEpkyJbQ62I-5xI8mc9jSlMlvWeyg6Qs7CEt6Iia7t3S-pLGmcTMkLg7_r2BzVkgzJ9JlXP_qPyb3v4TOopTnhd_sy1dnYmSFYKs0XjumM3gkpOkU6brMjUy3nwwfBzOGpVxdTt439Dv2Qa-4z0LpAbZNVFbWq3rHJl6sNsytuPk_klOjOot_plPuLKY3pX7bxG0yRh4t5xtHRxZAQ&sig=Cg0ArKJSzNHxx92NPzgT&fbs_aeid=%5Bgw_fbsaeid%5D&adurl=https://www.el-cell.com/products/test-cells/force-test-cells/pat-cell-solid/%3Fmtm_campaign%3DIOP-banner%26mtm_kwd%3DPAT-Cell-Solid%26mtm_source%3Dbanner%26mtm_cid%3D2026Key Role of Magnetite in Rust Layer of Carbon Steel for CathodicReactions and Hydrogen PermeationMariko Kadowaki,*,z m Masahiro Yamamoto, m Hideki Katayama,* m Arkapol Saengdeejing, mTaichi Abe, m Ryoji Sahara, m Kotaro Doi, m Yoshiharu Murase, m Sachiko Hiromoto, m andYusuke Tsutsumi* mNational Institute for Materials Science, Tsukuba 305-0047, JapanThis study investigates the influence of rust layer composition on cathodic reactions and hydrogen permeation behavior of SM490carbon steel exposed to NaCl-containing environments. Rust layers were formed on the SM490 specimens by repeated wet/drycycles using NaCl-containing droplets. Polarization measurements revealed that rust formation and growth enhanced the cathodicreaction. The reduction reaction of rust components, particularly FeOOH, was identified as one of the predominant contributors tothis enhancement. Notably, the results indicated that, in addition to the reduction of rust components, the hydrogen evolutionreaction could actively proceed on Fe3O4. Because hydrogen evolution occurs less readily on FeOOH, an increase in the Fe3O4fraction within the rust layer should promote hydrogen evolution reactivity and consequently enhance hydrogen permeation. Thiscorrelation was confirmed by electrochemical hydrogen permeation tests. These findings provided new insights into the relationshipbetween rust layer composition and hydrogen permeation behavior, thereby advancing mechanistic understanding of hydrogen entryprocesses in practical steels under atmospheric corrosion conditions.© 2026 The Author(s). Published on behalf of The Electrochemical Society by IOP Publishing Limited. This is an open accessarticle distributed under the terms of the Creative Commons Attribution 4.0 License (CC BY, https://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse of the work in any medium, provided the original work is properly cited. [DOI: 10.1149/1945-7111/ae6483]Manuscript submitted March 12, 2026; revised manuscript received April 16, 2026. Published May 6, 2026.High-strength steels are extensively utilized as structural mate-rials in automobiles and transportation systems owing to theirexcellent mechanical properties. However, hydrogen embrittlementremains a critical problem that significantly limits structural relia-bility by causing a reduction in ductility and localized brittle fracture.In recent years, intensive research efforts have been devoted toelucidating the mechanisms governing hydrogen permeation andembrittlement in steels.1–3Structural steels are frequently exposed to atmospheric environ-ments, where the formation of rust layers on their surfaces is inevitable.The presence of rust and corrosion products is known to significantlyinfluence the susceptibility of steels to hydrogen permeation.4–11Zakroczymski et al. reported that long, uninterrupted cathodic polariza-tion of iron in 0.1 M NaOH resulted in the metal surface becomingmore prone to both hydrogen absorption and corrosion, and that thistreatment leads to the formation of a corrosion product layer which doesnot hinder hydrogen absorption.4 Sudha et al. conducted hydrogencharging on DP and IF steels in 0.1 M NaOH, and demonstrated that theeffect of the rust layer depends on its structure; it can promote hydrogenuptake when it generates catalytic iron particles during reduction, butmay also hinder hydrogen permeation by acting as a physical barrier.11While these findings were obtained in alkaline environments, there arealso reports under conditions closer to atmospheric corrosion, such asNaCl solutions. For example, Akiyama et al. investigated hydrogenpermeation behavior in boron-bearing and AISI 4135 steels during wet/dry cyclic corrosion tests using NaCl solutions and reported a strongcorrelation between rust formation and enhanced hydrogen permeation.7Similarly, Li et al. demonstrated that corrosion products formed ondispersion-strengthened-high-strength steel during seawater wet/drycycles increased hydrogen permeation and embrittlement sensitivity,particularly in welded joints.9 The promotion of hydrogen permeationby rust has been attributed to several factors, including local acidifica-tion beneath the rust layer7,8,10 and restricted oxygen transport causedby rust coverage.9Despite extensive research into the effect of rust layers onhydrogen permeation in alkaline and NaCl environments, the specificrole of rust characteristics—including phase composition—has yet tobe clearly understood. Under atmospheric corrosion conditions, rustlayers generally consist of FeOOH (α, γ, β, etc.) and Fe3O4.12–17However, the relative fraction of these components varies consider-ably depending on environmental conditions (temperature, relativehumidity, atmospheric pollutants, etc.), steel composition, and corro-sion morphology.18,19 Clarifying how such compositional variationsinfluence hydrogen permeation is essential for evaluating hydrogenembrittlement risk and improving the durability of structural steels inservice environments.Hydrogen permeation initiates through electrochemical hydrogenevolution reaction occurring on steel or rust layer surfaces. A fraction ofthe generated hydrogen by this reaction permeates into the steel andcontributes to hydrogen embrittlement. Although various factors, suchas steel surface adsorption characteristics and microstructure,20–22 affecthydrogen permeation behavior, the hydrogen evolution reactivity isparticularly one of the most critical factors governing the amount ofhydrogen permeating into the steel.23 Therefore, understanding thepropensity for hydrogen evolution is essential for assessing the risk ofhydrogen permeation and embrittlement.The presence of rust is known to substantially alter the electro-chemical behavior of steel.24–30 Previous studies have discussed theinfluence of rust on anodic reactions on steel surfaces;27,28 however,systematic understanding of cathodic reactions remains insufficient.In particular, information regarding the hydrogen evolution reactionson rust constituents is still scarce. Although rust layers composed ofFeOOH (α, γ, β, etc.) and Fe3O4 are commonly formed underatmospheric environments,12–17 few researchers have quantitativelyevaluated differences in hydrogen evolution activity between rustedand bare steel surfaces or identified the specific rust phase respon-sible for such differences.The objective of this study was to investigate the influence of rustlayer composition on cathodic reactions and hydrogen permeationbehavior of SM490 carbon steel exposed to NaCl-containing environ-ments. Rust layers were formed by repeating wet/dry cycles with NaCl-containing droplets, followed by polarization measurements. To isolatethe effect of Fe3O4, a Fe3O4-dominant specimen was fabricated via sparkplasma sintering and subjected to identical electrochemical evaluation.Based on these investigations, the role of Fe3O4 in promoting hydrogenevolution and associated hydrogen permeation was systematicallyexamined using electrochemical hydrogen permeation tests.Experimental MethodsSM490 specimens and wet/dry cycling.—Commercial SM490carbon steel sheets (thickness: 2 mm) were used in this study, and thezE-mail: KADOWAKI.Mariko@nims.go.jp*Electrochemical Society Member.Journal of The Electrochemical Society, 2026 173 091502 aaahttps://orcid.org/0000-0002-8988-3545https://orcid.org/0009-0002-1861-7757https://orcid.org/0000-0001-7947-4687https://orcid.org/0000-0001-8739-3262https://orcid.org/0000-0002-5065-0939https://orcid.org/0000-0003-0788-2985https://orcid.org/0000-0002-5204-1088https://orcid.org/0000-0001-7390-851Xhttps://orcid.org/0000-0003-4666-6708https://orcid.org/0000-0002-9483-1256https://creativecommons.org/licenses/by/4.0/https://creativecommons.org/licenses/by/4.0/https://doi.org/10.1149/1945-7111/ae6483https://doi.org/10.1149/1945-7111/ae6483mailto:KADOWAKI.Mariko@nims.go.jphttps://crossmark.crossref.org/dialog/?doi=10.1149/1945-7111/ae6483&domain=pdf&date_stamp=2026-05-06chemical composition is listed in Table I. For all experiments excepthydrogen permeation tests, the sheets were cut into couponsmeasuring 15 mm × 15 mm. For hydrogen permeation tests, speci-mens with dimensions of 30 mm × 30 mm were prepared. Prior toexperiments, specimen surfaces were mechanically polished using1200-grit SiC paper followed by polishing with 1 μm diamond paste.Rust layers were subsequently formed by subjecting the specimens torepeated wet/dry cycles using NaCl-containing droplets, followingthe procedure reported our previous study.31 The SM490 specimenswere placed horizontally in a temperature- and humidity-controlledchamber maintained at 25 °C and 60% relative humidity. During theinitial cycle, a 500 μl droplet containing 1 wt% NaCl was depositedonto the specimen surface. Under these conditions, the dropletcompletely dried within approximately 2.5 h, leaving NaCl residueon the surface. Subsequently, a 500 μl droplet of distilled water wasadded to re-form the NaCl-containing droplet. This wet/dry processwas repeated for 5, 10, or 20 cycles.Preparation of Fe3O4 sintered specimen.—A Fe3O4 specimenwas fabricated by spark plasma sintering of commercially availableFe3O4 powder (Kanto Chemical Co., Inc., purity 95%). Sintering wasperformed using a LABOX-110 spark plasma sintering system(Sinter Land Inc.) under vacuum conditions of approximately10 Pa. Although Fe3O4 is the main component in the commercialpowder, minor impurities such as bcc Fe or Fe2O3 may be presentdue to phase transformations when the oxygen partial pressureduring the powder manufacturing process deviates slightly from theexact equilibrium oxygen concentration of Fe3O4 (0.5714 at%).32 Ifbcc Fe exists as an impurity in the commercial powder, heating tohigh temperature during sintering may induce phase transformation(Fe+Fe3O4→4FeO), thereby reducing the Fe3O4 fraction in theobtained sintered specimen. Similarly, the presence of Fe2O3 mayalso promote phase transformation that decrease Fe3O4 fraction. Tomaximize the Fe3O4 fraction, it is necessary to perform sintering attemperatures at which such phase transformations cannot occur, evenif these impurities are present in the commercial powder. Based onthese considerations, the sintering temperature was determined basedon the phase diagrams shown in Fig. 1.Figure 1 shows the calculated temperature-pressure phase dia-gram from the thermodynamic assessment of the Fe–O binarysystem.33 The equilibrium oxygen concentration of Fe3O4 is0.5714 at%. In Fig. 1a, the oxygen concentration is slightly belowthe Fe3O4 equilibrium (0.571 at% O), where the two-phase region ofbcc Fe and Fe3O4 at lower temperature. It corresponds to the casethat bcc Fe coexists with Fe3O4 as an impurity in the powder. If suchpowder is used for sintering, at 10 Pa (sintering condition of thispaper, indicated by a red dashed line in Fig. 1), the region whereFe3O4 and bcc Fe coexist stably extends up to 559 °C. This meansthat when bcc Fe is present as an impurity, phase transformation canbe avoided below 559 °C at 10 Pa. In contrast, if it was heated above559 °C, the phase transformation between Fe and Fe3O4(Fe+Fe3O4→4FeO) occurs, resulting in a decreased fraction ofFe3O4 in the obtained sintered specimen. Figure 1b shows the similarphase diagram where the oxygen concentration is slightly higher thanthe Fe3O4 equilibrium (0.572 at% O). This condition correspondsto the case where Fe2O3 coexists with Fe3O4 as an impurity.If this powder is used for sintering, at 10 Pa, Fe3O4 and Fe2O3 cancoexist stably at temperatures up to 1105 °C. In other words,the phase transformation can be avoided below 1105 °C. It ispreferable to perform sintering at the highest temperature to ensuresufficient densification, while preventing aforementioned phasetransformations. Therefore, the sintering temperature in this studywas determined to be 550 °C.The powder was loaded into a cylindrical graphite die with adiameter of 15 mm. Under an applied pressure of 5.3 kPa, thetemperature was increased from room temperature to 550 °C over30 min, and maintained at this temperature for 4 h. The specimenwas subsequently cooled to room temperature.Polarization measurements.—Potentiodynamic polarizationmeasurements for the SM490 specimens (with and without rustlayers) were conducted using a conventional three-electrode cell witha platinum counter electrode and an Ag/AgCl (3 M KCl) referenceelectrode. All potentials reported in this study are referenced to thiselectrode. The measurements were performed in naturally aerated1 wt% NaCl solution at room temperature using an EC Frontier VM3electrochemical cell. The SM490 specimen was positionedTable I. Chemical compositions of SM490 (mass%).C Si Mn P S Fe0.15 0.20 0.84 0.02 0.05 Bal.Figure 1. Calculated temperature-pressure phase diagrams from the thermo-dynamic assessment of the Fe-O binary system. (a) The oxygen concentrationis set as 0.571 at%, corresponding to the case that bcc Fe coexists with Fe3O4as an impurity in the powder. (b) The oxygen concentration is set as 0.572 at%, corresponding to the case that Fe2O3 coexists with Fe3O4 as an impurity inthe powder.Journal of The Electrochemical Society, 2026 173 091502horizontally, and the cell was sealed against the surface using an O-ring, providing an exposed area of 1 cm2. Prior to polarizationmeasurements, specimens were stabilized at the open circuit poten-tial for 10 minutes. Potential scans were subsequently conducted at ascan rate of 20 mV min−1, in accordance with our previousstudies.34–36Potentiostatic polarization measurements were also conducted forSM490 specimens (with and without rust layers) and the Fe3O4specimen fabricated by spark plasma sintering. These measurementswere carried out in 1 wt% NaCl solution at −1.0 V for 7 h, at roomtemperature under naturally aerated conditions. The cell configura-tion, including the counter and reference electrodes, was identical tothat used for potentiodynamic measurements. For the SM490 speci-mens, the connection to the cell was performed in the same manneras described above for potentiodynamic polarization. For the Fe3O4specimen, to prevent solution ingress into the specimen if surfacepores were present, the specimen was suspended over the solution,exposing only the outermost surface to the solution. Prior toimmersion, all areas of the Fe3O4 specimen except the outermostsurface were coated with epoxy resin to prevent contact with thesolution.X-ray diffraction analysis.—X-ray diffraction (XRD) analysis wasperformed to determine the phase compositions of rust layers on theSM490 specimens before and after the potentiostatic polarization, andto characterize the sintered Fe3O4 specimen. Measurements werecarried out using a BRUKER Advanced X-ray Solutions D2 Phaserwith Cu-Kα radiation, a step width of 0.02° and a scan rate of 1° min−1.Microscopic observation.—For the rust layers on the SM490specimens, surface appearance before and after potentiostatic polar-ization measurements was examined using an optical digital micro-scope (KEYENCE, VHX 5000). Cross-sectional observations of rustlayers were performed by a scanning electron microscope (SEM,HITACHI High-Tech SU5000) operated at an accelerating voltage of15 kV. In addition, for the sintered Fe3O4 specimen, the surfaceappearance and topography were analyzed using a 3D microscope(KEYENCE, VR Series One-shot 3D).Hydrogen permeation test.—Electrochemical hydrogen permea-tion tests were conducted using a Devanathan-Stachurski cell.37,38The SM490 specimen was positioned horizontally between thecathodic polarization cell (commonly referred to as hydrogencharging cell) and the detection cell, as illustrated in Fig. 2. Therusted surface served as the cathodic polarization side. The hydrogendetection side was coated with nickel plating. Nickel plating wascarried out in an electrolyte containing NiSO4·6H2O (250 g l−1),NiCl2·6H2O (45 g l−1), and H3BO3 (40 g l−1) at 60 °C undergalvanostatic polarization at −3 mA cm−2 for 180 s.Two types of specimens with different rust compositions wereprepared: (1) the specimen subjected to 10 wet/dry cycles with 1 wt%NaCl-containing droplets resulting in a FeOOH-dominant rust layer,and (2) the specimen subjected to additional potentiostatic polarizationat −1.0 V for 2.5 h in 1 wt% NaCl solution after the 10 wet/dry cyclesresulting in a Fe3O4-dominant rust layer. After rust formation,specimens were stored in air at room temperature for at least 24 hprior to Ni plating and the hydrogen permeation test. This exposureperiod allowed any hydrogen that may have permeated the steel duringabove rust formation processes to diffuse out; therefore, it did notaffect the results of the hydrogen permeation test.The hydrogen detection side was potentiostatically polarized at0.14 V in 0.1 M NaOH solution. Although Hg/HgO was used as thereference electrode for the detection cell, all potentials reported inthis paper were converted to and are presented relative to the Ag/AgCl reference electrode. After the background current reached asteady value below 0.07 μA cm−2, a 1 wt% NaCl solution (adjustedto pH 3 with HCl) was introduced into the cathodic polarization cell.Previous studies have reported that the hydrogen permeation currentin rusted steel is around 0.1 to 0.2 μA cm−2.39 In this study, weensured that the background current was sufficiently lower than thesevalues. To observe the hydrogen permeation behavior clearly, anacidic solution (pH 3) was used in this experiment. The cathodicpolarization side (rusted surface) was potentiostatically polarized at−1.0 V. During the hydrogen permeation tests, the cathodic current,which reflects the hydrogen evolution reaction and other cathodicreactions, was measured on the cathodic polarization side. Theanodic current, corresponding to the oxidation of permeatedhydrogen, was measured on the detection side.Results and DiscussionCharacterization of rust layers formed during wet/dry cycles.—Optical microscopy images of the SM490 specimen surfaces after5, 10, and 20 wet/dry cycles using droplets containing 1 wt% NaClare shown in Figs. 3a–3c. In all cases, the specimen surfaces wereentirely covered with rust layers. The corresponding XRD patterns ofthese specimens are presented in Fig. 4. Diffraction peaks attribu-table to α-FeOOH, γ-FeOOH, and Fe3O4 were observed for allspecimens, indicating that the rust layers consisted primarily of thesephases. Generally, FeOOH appears as red rust, whereas Fe3O4exhibits a black coloration. Because the rust layers of all specimensexhibited reddish coloration in the optical images (Fig. 3), it isconsidered that FeOOH is the dominant constituent. The cross-sectional SEM images of the rust layers are shown in Fig. 5. The rustlayer thickness increased progressively with increasing number ofwet/dry cycles, confirming continuous rust growth during cyclicexposure.As a side note, for the specimen after 5 cycles in Fig. 4, the Fe signal(around 2θ = 44.5°) was shifted toward the high-angle side, which islikely due to the influence of surface roughness. According to previousstudies, surface roughness can affect XRD peak positions.40 As thenumber of wet/dry cycles increases, the rust layer becomes morehomogeneous, and this peak shift is no longer observed.Enhancement of the cathodic reaction by the presence of rustlayer.—To clarify changes in electrochemical behavior associatedFigure 2. Schematic illustration of a Devanathan-Stachurski cell used forhydrogen permeation tests.Journal of The Electrochemical Society, 2026 173 091502with rust formation, potentiodynamic polarization measurementswere conducted in 1 wt% NaCl solution for the SM490 specimensbefore and after wet/dry cycling, as shown in Fig. 6.As discussed in the Introduction, the present study primarilyfocuses on cathodic polarization behavior. For the specimen withoutrust (before wet/dry cycles), diffusion-limited oxygen reduction wasobserved in the potential range from −0.6 to −0.9 V, exhibiting alow current density of approximately 2× 10−5 A cm−2. At potentialsbelow −0.9 V, the cathodic current increased logarithmically withdecreasing potential, corresponding to hydrogen evolution reaction.These results represent typical cathodic behaviors of carbon steel inNaCl solutions.41 In contrast, specimens subjected to wet/dry cyclingexhibited significantly higher cathodic current densities over theentire cathodic potential region compared with the rust-free spe-cimen. Furthermore, below −0.7 V, the magnitude of cathodiccurrent slightly increased with increasing number of cycles. Theseresults clearly indicate that rust formation and growth promotecathodic reaction.Although the main focus of this study is cathodic behavior, it isnoteworthy that the formation of the rust layer also affected theanodic current in Fig. 6. Compared to the specimen without rust, theanodic current increased after 5 cycles while it decreased at 10 and20 cycles. Similar phenomena have been reported in previousstudies.27,28 Based on these studies, the increase in anodic currentduring the initial cycles is attributed to the cracking in the rust layer,which facilitates the transportation of the corrosive electrolytes (suchas Cl−) to the steel. In contrast, at later cycles, the formation of aprotective inner last layer is considered to suppress the anodicreaction.27 Although anodic processes are not examined in detailhere, the observed trend is consistent with established corrosionbehavior of rusted steels.Potentiostatic polarization behavior of the rust layer.—Based onthe potentiodynamic polarization results shown in Fig. 6, thepresence of rust was found to significantly enhanced cathodicreactions. To further clarify the cathodic processes occurring onrust layers, potentiostatic polarization measurements were conductedat −1.0 V in 1 wt% NaCl solution. A sufficiently negative potentialof −1.0 V was selected to minimize the influence of competinganodic reactions.Figure 7 shows the time variation of the cathodic current duringpotentiostatic polarization at −1.0 V. For the specimen before wet/Figure 3. Optical microscopy images of rust layers formed on SM490specimens after repeating wet/dry cycles: (a) after 5 cycles, (b) after 10cycles, and (c) after 20 cycles.Figure 4. XRD patterns of rust layers formed on SM490 specimens afterrepeating wet/dry cycles.Journal of The Electrochemical Society, 2026 173 091502dry cycling (without rust), the cathodic current remained nearlyconstant at approximately 3 × 10–5 A cm−2 throughout the measure-ment period, with no significant changes observed. Specimenssubjected to wet/dry cycling exhibited higher cathodic currentsthan the specimen without rust, and the current increased withincreasing number of cycles. These results demonstrate that rustformation and growth markedly enhances cathodic reaction,consistent with the potentiodynamic polarization behavior shown inFig. 6.Focusing on time-dependent behavior, rusted specimens after thewet/dry cycles initially exhibited high current values exceeding1× 10–3 A cm−2. The current subsequently decreased and reached tothe order of 10–4 A cm−2. After this decline, the current valueremained relatively stable and maintained the order of 10–4 A cm−2until the end of polarization. Specimens subjected to a larger numberof wet/dry cycles exhibited slower current decay and higher steady-state current values.Reduction reaction of rust components.—Figure 8 presentsoptical microscopy images of the specimen surfaces after potentio-static polarization in Fig. 7. Compared with surfaces prior topolarization (Fig. 3), the rust layers exhibited a distinct color changefrom reddish to black following polarization. The correspondingXRD patterns of the polarized specimens are shown in Fig. 9.Figure 5. Cross-sectional SEM images of rust layers formed on SM490specimens after repeating wet/dry cycles: (a) after 5 cycles, (b) after 10cycles, and (c) after 20 cycles.Figure 6. Potentiodynamic polarization curves of SM490 specimens withrust (after repeating wet/dry cycles) and without rust in 1 wt% NaCl solution.Solid and dashed lines show the cathodic and anodic polarization curves,respectively.Figure 7. Time variation of cathodic current during potentiostatic polariza-tion at −1.0 V for SM490 specimens with rust (after repeating wet/drycycles) and without rust in 1 wt% NaCl solution.Journal of The Electrochemical Society, 2026 173 091502Compared with the data before polarization (Fig. 4), diffractionpeaks associated with FeOOH became significantly weaker, whereaspeaks corresponding to Fe3O4 were markedly intensified afterpolarization. These results indicate a decrease in FeOOH contentaccompanied by an increase in Fe3O4 within the rust layer duringpolarization. Because Fe3O4 exhibits black coloration, this composi-tional transformation is consistent with the optical observations inFig. 8.FeOOH is known to undergo electrochemical reduction accordingto the following reaction:42–45+ + + [ ]+3FeOOH H e Fe O 2H O 13 4 2Therefore, it is reasonable to conclude that this reduction reactionproceeds during potentiostatic polarization in Fig. 7. The time-dependent decrease in cathodic current observed in Fig. 7 can beattributed to the progressive reduction and consumption of FeOOHwithin the rust layer. In other words, as the amount of reducibleFeOOH decreases with polarization time, the contribution of thisreduction reaction diminishes, resulting in the observed decline incathodic current. Notably, specimens subjected to fewer wet/drycycles exhibited faster current decay, which is consistent with theirsmaller amount of rust and correspondingly lower FeOOH content.Figure 10 shows cross-sectional SEM images of the specimens afterpotentiostatic polarization in Fig. 7. Compared to Fig. 5 (beforepolarization), the rust layer thickness decreased slightly due topolarization, but the difference was minimal overall.Implication of additional cathodic reactions attributed to Fe3O4.—As shown in Fig. 7, after the cathodic current decreased andreached steady-state values, specimens with rust layers continued toexhibit higher cathodic currents than the rust-free specimenthroughout the polarization period. At this stage, surface observa-tions (Fig. 8) and XRD results (Fig. 9) indicate that the rust layerconsisted predominantly of Fe3O4. Because Fe3O4 is stable and is notfurther reduced under this condition,46 the sustained cathodic currentobserved at steady state cannot be attributed to continuous reductionof rust components. This finding suggests that in addition to thereduction reaction of rust components, other cathodic reactionsoccurred on the rust layer.Figure 8. Optical microscopy images of rust layers formed on SM490specimens after repeating wet/dry cycles—(a) after 5 cycles, (b) after 10cycles, and (c) after 20 cycles— after potentiostatic polarization at −1.0 V asshown in Fig. 7.Figure 9. XRD patterns of rust layers after potentiostatic polarization at−1.0 V as shown in Fig. 7.Journal of The Electrochemical Society, 2026 173 091502Figure 11 shows the relationship between the steady-statecathodic current (measured at 25 ks in Fig. 7) and the rust layerthickness. The thickness was determined from cross-sectional SEMobservations (Fig. 10). In Fig. 10, although the thickness was notuniform across different locations, the measurements at five differentpoints for each specimen were averaged and are plotted in Fig. 11.Since the data for all specimens are aligned almost linearly, a clearcorrelation between steady-state cathodic current and rust layerthickness was observed. Because the rust layer was primarilycomposed of Fe3O4 at this time, the thickness of the rust layer canbe regarded as the amount of Fe3O4, or the available Fe3O4 surfacearea participating in cathodic reactions. These results thereforeindicate that cathodic reactions other than the FeOOH reductionreaction are closely associated with Fe3O4.Among possible cathodic reactions on Fe3O4, the contribution ofoxygen reduction is expected to be limited. Oxygen reduction isdiffusion-limited, and typically produces maximum cathodic currenton the order of 10–5 A cm−2,41 which are significantly lower than thesteady-state cathodic currents observed in Fig. 7. Furthermore, Fe3O4is a stable compound and is not further reduced.46 The electrolytewas NaCl and contained no additional reducible species.Accordingly, the dominant cathodic reaction occurring on Fe3O4 isreasonably attributed to the hydrogen evolution reaction.Fe3O4 has been reported as an electrocatalytically active materialfor hydrogen evolution reaction in energy-conversion systems suchas water electrolysis and fuel cells owing to its spinel structure andfavorable electronic conductivity.47,48 Although most previousstudies have been conducted under strongly alkaline and elevated-temperature conditions, to the best of our knowledge, there is noinformation regarding the hydrogen evolution reactivity of Fe3O4under neutral or acidic pH and room temperature, which areconditions similar to the atmospheric corrosion environment as-sumed in this study. On the other hand, the results obtained in thispaper suggest that the hydrogen evolution reaction can activelyproceed on Fe3O4 even under such conditions.Preparation of Fe3O4 sintered specimens.—As discussed above,in addition to the reduction of rust components, other cathodicreactions occur on the rust layer, and these reactions are likely to bepredominantly associated with Fe3O4. However, the electrochemicalbehavior of rust layers formed during wet/dry cycling reflects thecombined response of multiple phases, including Fe3O4 and FeOOH.Consequently, direct evaluation of the intrinsic electrochemicalcontribution of Fe3O4 from rusted specimen alone is difficult.Therefore, to isolate the electrochemical contribution of Fe3O4, aFe3O4-dominant specimen was fabricated by spark plasma sinteringof commercial Fe3O4 powder. The detailed sintering conditions weredescribed earlier in Experimental Methods section.Optical microscopy images of the sintered Fe3O4 specimen areshown in Figs. 12a and 12b. The surface exhibited no visible cracksor pores, indicating that the sintered specimen possessed high densityand structural integrity. The corresponding surface hight profileFigure 10. Cross-sectional SEM images of rust layers formed on SM490specimens after repeating wet/dry cycles —(a) after 5 cycles, (b) after 10cycles, and (c) after 20 cycles— after potentiostatic polarization at −1.0 V asshown in figure 7.Figure 11. Relationship between cathodic current measured at 25 ks of thepotentiostatic polarization in Fig. 7 and the average thickness of the rust layerfor SM490 specimens subjected to wet/dry cycles.Journal of The Electrochemical Society, 2026 173 091502Fig. 12c confirms that the specimen surface was nearly flat, with onlyminor surface roughness of approximately 60 μm.The XRD pattern shown in Fig. 12d (black graph, labeled “beforepolarization”) confirms that Fe3O4 is the dominant phase in thespecimen. Minor diffraction peaks corresponding to Fe2O3 were alsodetected, which are attributed to impurities originating from thestarting powder (explained in Experimental Methods “Preparation ofFe3O4 sintered specimen” section). The intensity of the main peakof Fe3O4 (2θ = 35°) is approximately 8.6 times greater than that ofFe2O3 (2θ = 33°). Although this is only a rough estimation since noRietveld refinement or similar analysis was conducted, it is reason-able to consider that Fe3O4 is the dominant component.Cathodic polarization behavior of sintered Fe3O4 specimen.—Potentiostatic polarization measurements were performed on thesintered Fe3O4 specimen in 1 wt% NaCl solution at −1.0 V underconditions identical to those used for the rusted SM490 specimensshown in Fig. 7. The time-dependent variation of cathodic currentduring polarization is shown in Fig. 13. The Fe3O4 specimeninitially exhibited a high cathodic current of approximately4× 10−3 A cm−2, followed by a rapid decrease within approxi-mately 2 ks before reaching a steady-state value. As indicated by theXRD results in Fig. 12d, the sintered specimen contained a minoramount of Fe2O3 impurity. The high current observed during theinitial stage is therefore attributed to reduction of Fe2O3 to Fe3O4. Asthis reduction proceeds and Fe2O3 is consumed, its contribution tothe cathodic current diminishes, resulting in the observed currentdecline. The XRD results of the Fe3O4 specimen after the polariza-tion is also shown in Fig. 12d. The characteristic peaks of Fe3O4 areclearly observed both before and after polarization, indicating thatFe3O4 does not transform into other phases during the polarization.On the other hand, the peaks assigned to Fe2O3 are much lessprominent after the polarization. Although a small Fe2O3 peak at 2θ= 33° is still observed, the signals at 2θ = 24°, 49° and 54° arenearly flat and the corresponding Fe2O3 peaks have almost com-pletely disappeared after the polarizaton. These results also supprotsour interpretation that the intial high current observed in Fig. 13 wascuased by the reduction of Fe2O3.Importantly, after reaching steady-value, the cathodic current forthe sintered Fe3O4 specimen remained within the same order ofmagnitude (10−4 to 10−3 A cm−2) as that observed for rusted SM490specimens in Fig. 7. Since the rusted specimens in Fig. 7 werecomposed of multiple phases (FeOOH and Fe3O4), it was notpossible to conclusively attribute the steady-state cathodic currentsorely to Fe3O4. However, the fact that the sintered Fe3O4 specimen,which is predominantly composed of Fe3O4, exhibited a comparablesteady-state current clearly indicates that Fe3O4 itself functions as amajor active site for cathodic reactions at this time.Discussion on the effect of Fe3O4 content on hydrogen evolu-tion and permeation.—Based on the results presented so far, it issuggested that cathodic reactions occur actively on Fe3O4. AsFigure 12. (a)–(b) Optical microscopy image, (c) topographic height profile,and (d) XRD pattern of the Fe3O4 sintered specimen.Figure 13. Time variation of cathodic current during potentiostatic polariza-tion at −1.0 V for the Fe3O4 specimen in 1 wt% NaCl solution.Journal of The Electrochemical Society, 2026 173 091502discussed earlier, the dominant cathodic reaction occurring on Fe3O4is considered to be the hydrogen evolution reaction. In contrast,based on the literature, FeOOH exhibits relatively low electronicconductivity because of its crystal structure, and cannot adequatelysupply electrons required for electron transfer reactions (hydrogenevolution reaction).49 In addition, unlike FeOOH, which consistsexclusively of trivalent iron (Fe3+), Fe3O4 contains Fe2+ states thatare considered to enhance the catalytic activity of the surface forcathodic reactions.50,51Based on these considerations, it is reasonable to expect that anincrease in the fraction of Fe3O4 within the rust layer (i.e., a decreasein the proportion of FeOOH) would enhance hydrogen evolutionreactivity. Here, it should be noted that this estimated relationship isnot necessarily strictly proportional. This is because various factors,such as the distribution of phases, interface effects between Fe3O4and FeOOH, and the surface morphology of the rust layer, caninfluence electron transport and reaction pathways. On the otherhand, as a general trend, it is reasonable to assume that a higherfraction of Fe3O4 tends to increase hydrogen evolution reactivity.Previous studies have shown that the more actively hydrogenevolution occurs, the more hydrogen permeation into steel ispromoted.23 Therefore, it is estimated that increasing the Fe3O4fraction in the rust layer enhances hydrogen permeation into thesteel. To verify this interpretation, hydrogen permeation tests wereconducted using two rusted SM490 specimens possessing differentrust compositions, as described below.Influence of Fe3O4 in rust layers on hydrogen permeationbehavior.—To evaluate the influence of rust composition onhydrogen permeation behavior, hydrogen permeation tests wereconducted using two SM490 specimens possessing FeOOH-domi-nant and Fe3O4-dominant rust layers. Optical microscopy images ofboth specimen surfaces are shown in Fig. 14. First, two SM490specimens were subjected to repeated 10 wet/dry cycles with 1 wt%NaCl droplets, to form FeOOH-dominant rust layers as shown inFig. 14a. Subsequently, to increase the Fe3O4 fraction, one specimenwas subjected to additional potentiostatic polarization at −1.0 V for2.5 h in 1 wt% NaCl solution. As shown in Fig. 14b, the surface ofthis specimen was black, indicating the formation of Fe3O4-dominantrust. After these processes, the specimens were stored in air at roomtemperature for more than 24 h prior to the hydrogen permeation tests.This exposure period allowed any hydrogen that may have permeatedthe steel during above rust formation processes to diffuse out;therefore, it did not affect the results of the hydrogen permeation tests.The conditions of hydrogen permeation tests were explained earlierin Experimental Methods section. Briefly, the hydrogen detection side(Ni-plated surface without rust) was potentiostatically polarized at0.14 V in 0.1 M NaOH. After the background current stabilized below0.07 μA cm−2, 1 wt% NaCl solution (pH 3) was introduced to thecathodic polarization side (rusted surface) and the specimen waspolarized at –1.0 V. The time variations of the current on the cathodicpolarization side and hydrogen detection side are shown in Figs. 15aand 15b, respectively. In these graphs, cathodic polarization at –1.0 Vwas started at 1.5 ks, as indicated by dashed lines.In the case of cathodic polarization side (Fig. 15a), for bothspecimens, the cathodic current initially exhibits values on the orderof 10–3 A cm−2, and subsequently decreased toward approximately10–4 A cm−2. As discussed in Figs. 7–9, this initial high current isconsidered to be due to the reduction of FeOOH within the rust layer.As FeOOH was reduced and consumed, the cathodic currentassociated with the reduction of FeOOH decreased. TheFe3O4-dominant specimen exhibited a more rapid current decrease,which is reasonable because of its lower FeOOH content. For theFe3O4-dominant specimen, the cathodic current reached a steadyvalue of approximately 2× 10–4 A cm−2 around 4 ks. This steadyvalue, observed after the reduction of FeOOH had sufficientlyproceeded, was considered to result from the hydrogen evolutionreaction on Fe3O4.In the case of hydrogen detection side (Fig. 15b), the anodiccurrent began to increase at 3 ks (that is, approximately 1.5 ks aftercathodic polarization was started). This increase indicates that thehydrogen generated on the cathodic polarization side permeatedthrough the specimen and detected on the detection side. It isimportant that the Fe3O4-dominant specimen exhibited a higher rateof current increase than the FeOOH-dominant specimen. Theseresults clearly demonstrate that the composition of the rust layerinfluences hydrogen permeation behavior, and Fe3O4 enhanced thehydrogen permeation. As discussed in the previous section, Fe3O4exhibits higher hydrogen evolution activity than FeOOH, which isconsidered to be a factor for the enhanced hydrogen permeationobserved for the Fe3O4-dominant specimen.For both specimens, the detection side current reached similarvalues beyond approximately 8 ks. This convergence can bereasonably explained by progressive reduction of FeOOH duringpotentiostatic polarization at –1.0 V in the FeOOH-dominant spe-cimen. In other words, the specimen that was initially FeOOH-dominant ultimately became Fe3O4-dominant during the measure-ment, resulting in hydrogen permeation behavior comparable to thatof the specimen that was initially Fe3O4-dominant.Figure 14. Optical microscopy images of rusted surfaces of SM490 speci-mens used for the hydrogen permeation test: specimens with (a) FeOOH-dominant and (b) Fe3O4-dominant rust layers.Journal of The Electrochemical Society, 2026 173 091502Here, it should be noted that the unstable behavior observed in thecathodic current and hydrogen permeation current in Fig. 15 has alsobeen reported in previous studies, and is generally characteristic ofrusted specimens.39,52 This tendency arises because rust layers arecomposed of multiple phases, such as Fe3O4 and FeOOH, andreduction reactions (e.g., FeOOH to Fe3O4) do not proceed uni-formly across the surface, resulting in local and stepwise changes inthe rust structure that manifest as fluctuations in the current.Overall, these results demonstrate that rust layer compositionsignificantly influences the cathodic reaction and hydrogen permea-tion behavior of SM490 carbon steel. In particular, it was found thatthe Fe3O4 within the rust layer promotes hydrogen evolution reactionand consequently enhances hydrogen permeation. In actual atmo-spheric environments, the rust composition varies depending on thesteel types, and some practical steels (including weathering steels)develop rust layers with a uniquely high FeOOH content (i.e. lowFe3O4 fraction). According to the findings of this study, such steelsmay exhibit greater resistance to hydrogen permeation and embrit-tlement compared to conventional steels. Further investigation fromthis perspective will be important for enhancing the servicereliability of steel in practical applications. While this study focusedon the effect of the rust composition on the hydrogen evolutionreaction, it should be noted that other processes, such as hydrogenadsorption, diffusion, and trapping, also play a role in hydrogenembrittlement. Therefore, further investigations targeting theseadditional processes are necessary to achieve a deeper understandingof hydrogen embrittlement mechanisms.Conclusions1. Rust layers formed on SM490 carbon steel was composed ofFeOOH (α and γ) and Fe3O4. Potentiodynamic polarizationmeasurements revealed that rust formation and growth signifi-cantly enhanced the cathodic reactions.2. Potentiostatic polarization revealed that the reduction reaction ofFeOOH within the rust layer is one of the primary factorscontributing to the enhanced cathodic reaction.3. Even after the reduction reaction of FeOOH proceeded suffi-ciently, the cathodic current remained high for the rusted SM490speicmen, indicating that further cathodic reactions occurspecifically on the rust layer. It was indicated that the hydrogenevolution reaction can actively proceed on Fe3O4.4. Electrochemical hydrogen permeation tests confirmed that rustlayer composition influences hydrogen permeation behavior.The specimen possessing Fe3O4-dominant rust exhibited en-hanced hydrogen permeation compared with FeOOH-dominantrust. The higher hydrogen evolution activity on Fe3O4 isconsidered to be a key factor responsible for this behavior.AcknowledgmentsThis study was supported by the 32nd ISIJ Research PromotionGrant (Ishihara/Asada Grant) from the Iron and Steel Institute ofJapan, and JSPS KAKENHI Grant Numbers JP22K14516.ORCIDMariko Kadowaki m https://orcid.org/0000-0002-8988-3545Masahiro Yamamoto m https://orcid.org/0009-0002-1861-7757Hideki Katayama m https://orcid.org/0000-0001-7947-4687Arkapol Saengdeejing m https://orcid.org/0000-0001-8739-3262Taichi Abe m https://orcid.org/0000-0002-5065-0939Ryoji Sahara m https://orcid.org/0000-0003-0788-2985Kotaro Doi m https://orcid.org/0000-0002-5204-1088Yoshiharu Murase m https://orcid.org/0000-0001-7390-851XSachiko Hiromoto m https://orcid.org/0000-0003-4666-6708Yusuke Tsutsumi m https://orcid.org/0000-0002-9483-1256References1. L. Liu, L. Chen, and R. Case, J. Electrochem. Soc., 169, 112697 (2022).2. T. Rubben, K. Baert, T. Depover, K. Verbeken, R. I. Revilla, and I. D. Graeve,J. Electrochem. Soc., 169, 111503 (2022).3. S. J. Kim, J. Electrochem. Soc., 171, 061501 (2024).4. T. Zakroczymski and Z. Szklarska-Smialowska, J. Electrochem. Soc., 132, 2548(1985).5. T. Tsuru, Y. Huang, M. R. Ali, and A. Nishikata, Corros. Sci., 47, 2431 (2005).6. Y. Huang and Y. Zhu, Corros. Sci., 47, 1545 (2005).7. E. Akiyama, S. Li, T. Shinohara, Z. Zhang, and K. Tsuzaki, Electrochim. Acta, 56,1799 (2011).8. S. Ajito, E. Tada, A. Ooi, and A. Nishikata, ISIJ Int., 59, 1659 (2019).9. P. Li, J. Wang, M. Du, and L. Qiao, Int. J. Hydro. Ene, 48, 35862 (2023).10. H. Kakinuma, S. Hiromoto, T. Hojo, S. Ajito, M. Koyama, and E. Akiyama, Corros.Sci., 232, 112043 (2024).11. G. Sudha, M. Rohwerder, and D. Vijayshankar, Corros. Sci., 259, 113449 (2026).12. M. Morcillo, B. Chico, J. Alcántara, I. Díaz, R. Wolthuis, and D. de la Fuente,J. Electrochem. Soc., 163, C426 (2016).13. D. A. Koleva, J. H. W. de Wit, K. van Breugel, Z. F. Lodhi, and G. Ye,J. Electrochem. Soc., 154, C261 (2007).14. M. Morcillo, B. Chico, D. de la Fuente, J. Alcántara, I. O. Wallinder, andC. Leygraf, J. Electrochem. Soc., 164, C8 (2016).15. I. Suzuki, Y. Hisamatsu, and N. Masuko, J. Electrochem. Soc., 127, 2210 (1980).16. J. T. Keiser, C. W. Brown, and R. H. Heidersbach, J. Electrochem. Soc., 129, 2686(1982).17. T. E. Graedel and R. P. Frankenthal, J. Electrochem. Soc., 137, 2385 (1990).18. S. Li and L. H. Hihara, J. Electrochem. Soc., 159, C147 (2012).19. B. Wang, Y. Zhao, X. Cheng, X. Zhang, L. Wang, Z. Li, C. Liu, D. Zhang, andX. Li, Corros. Sci., 255, 113117 (2025).20. X. Han and M. Sakairi, Corros. Sci., 231, 111993 (2024).21. K. Gong, M. Wu, and G. Liu, Corros. Sci., 165, 108382 (2020).Figure 15. Time variation of the measured current during the hydrogenpermeation test for SM490 specimens with different Fe3O4 fractions in therust layer: (a) cathodic current on the cathodic polarization side, and (b)anodic current on the hydrogen detection side.Journal of The Electrochemical Society, 2026 173 091502https://orcid.org/0000-0002-8988-3545https://orcid.org/0009-0002-1861-7757https://orcid.org/0000-0001-7947-4687https://orcid.org/0000-0001-8739-3262https://orcid.org/0000-0002-5065-0939https://orcid.org/0000-0003-0788-2985https://orcid.org/0000-0002-5204-1088https://orcid.org/0000-0001-7390-851Xhttps://orcid.org/0000-0003-4666-6708https://orcid.org/0000-0002-9483-1256https://doi.org/10.1149/1945-7111/ac96adhttps://doi.org/10.1149/1945-7111/aca182https://doi.org/10.1149/1945-7111/ad5911https://doi.org/10.1149/1.2113621https://doi.org/10.1016/j.corsci.2004.10.006https://doi.org/10.1016/j.corsci.2004.07.044https://doi.org/10.1016/j.electacta.2010.09.043https://doi.org/10.2355/isijinternational.ISIJINT-2019-113https://doi.org/10.1016/j.ijhydene.2023.05.276https://doi.org/10.1016/j.corsci.2024.112043https://doi.org/10.1016/j.corsci.2024.112043https://doi.org/10.1016/j.corsci.2025.113449https://doi.org/10.1149/2.0411608jeshttps://doi.org/10.1149/1.2715313https://doi.org/10.1149/2.0131702jeshttps://doi.org/10.1149/1.2129376https://doi.org/10.1149/1.2123648https://doi.org/10.1149/1.2086948https://doi.org/10.1149/2.013204jeshttps://doi.org/10.1016/j.corsci.2025.113117https://doi.org/10.1016/j.corsci.2024.111993https://doi.org/10.1016/j.corsci.2019.10838222. H. Tian, X. Wang, Z. Cui, Q. Lu, L. Wang, L. Lei, Y. Li, and D. Zhang, Corros.Sci., 144, 145 (2018).23. M. Nagumo, H. Uyama, and M. Yoshizawa, Corros. Sci., 44, 947 (2001).24. T. Nishimura, H. Katayama, K. Noda, and T. Kodama, Corros. Sci., 42, 1611 (2000).25. K. Asami and M. Kikuchi, Corros. Sci., 45, 2671 (2003).26. Q. X. Li, Z. Y. Wang, W. Han, and E. H. Han, Corros. Sci., 50, 365 (2008).27. L. Hao, S. Zhang, J. Dong, and W. Ke, Corros. Sci., 58, 175 (2012).28. W. Chen, L. Hao, J. Dong, and W. Ke, Corros. Sci., 83, 155 (2014).29. K. Otani, T. Tsukada, and F. Ueno, Mater. Trans., 62, 763 (2021).30. C. Hayama, M. Kadowaki, Y. Murase, H. Katayama, T. Hara, Y. Hara,H. Watanabe, I. Shitanda, and M. Itagaki, ISIJ Int., 64, 1820 (2024).31. M. Kadowaki, H. Katayama, and M. Yamamoto, Corros. Sci., 218, 111190(2023).32. Inorganic Material Database (Atom Work), National Institute for Materials Science,https://crystdb.nims.go.jp/crystdb/search-phase-details?condition_value=Fe+O&pageDetails=1&page=1&history=true&phase_diagram_id=4295868358(accessed 21 January 2026).33. B. Sundman, J. Phase Equilib., 12, 127 (1991).34. M. Kadowaki, H. Katayama, and M. Yamamoto, Corros. Sci., 211, 110918 (2023).35. M. Kadowaki and H. Katayama, Corros. Sci., 222, 111379 (2023).36. M. Kadowaki, T. Moronaga, A. Nakamura, Y. Murase, T. Hashimoto, H. Katayama,K. Takanabe, and Y. Tsutsumi, J. Phys. Chem. C, 129, 15939 (2025).37. M. A. V. Devanathan and Z. Stachurski, Proc. A, 270, 90 (1962).38. M. A. V. Devanathan, Z. Stachurski, and W. Becki, J. Electrochem. Soc., 110, 886(1963).39. S. Li, E. Akiyama, T. Shinohara, K. Matsuoka, and W. Oshikawa, ISIJ Int., 53, 1062(2013).40. J. Peng, V. Ji, and W. Seiler, Mater. Sci. Forum, 490-491, 153 (2005).41. M. Kadowaki, H. Katayama, and M. Yamamoto, ISIJ Int., (In press).42. A. Kuch, Corros. Sci., 28, 221 (1988).43. M. Morcillo, I. Díaz, B. Chico, H. Cano, and D. de la Fuente, Corros. Sci., 83, 6(2014).44. X.-L. Gao, Y. Han, G.-Q. Fu, M.-Y. Zhu, and X.-Z. Zhang, Acta Metall. Sin., 29,1025 (2016).45. F. Wu, Z. Hu, X. Liu, C. Su, and L. Hao, Corros. Sci., 315, 131968 (2022).46. T. Misawa, K. Hashimoto, and S. Shimodaira, Corros. Sci., 14, 131 (1974).47. J. Zhang, X. Shang, H. Ren, J. Chi, H. Fu, B. Dong, C. Liu, and Y. Chai, Adv.Mater., 31, 1905107 (2019).48. J. Xiang et al., J. Alloys Compd., 1036, 181652 (2025).49. M. Stratmann and J. Müller, Corros. Sci., 36, 327 (1994).50. E. R. Vago, E. J. Calvo, and M. Stratmann, Electrochim. Acta, 39, 1655 (1994).51. R. Krieg, M. Rohwerder, S. Evers, B. Schuhmacher, and J. Schauer-Pass, Corros.Sci., 65, 119 (2012).52. S. Ajito, E. Tada, A. Ooi, and A. Nishikata, ISIJ Int., 61, 1186 (2021).Journal of The Electrochemical Society, 2026 173 091502https://doi.org/10.1016/j.corsci.2018.08.048https://doi.org/10.1016/j.corsci.2018.08.048https://doi.org/10.1016/S1359-6462(00)00683-7https://doi.org/10.1016/S0010-938X(00)00018-4https://doi.org/10.1016/S0010-938X(03)00070-2https://doi.org/10.1016/j.corsci.2007.06.020https://doi.org/10.1016/j.corsci.2012.01.017https://doi.org/10.1016/j.corsci.2014.02.010https://doi.org/10.2320/matertrans.C-M2021816https://doi.org/10.2355/isijinternational.ISIJINT-2024-191https://doi.org/10.1016/j.corsci.2023.111190https://crystdb.nims.go.jp/crystdb/search-phase-details?condition_value=Fe+O%26pageDetails=1%26page=1%26history=true%26phase_diagram_id=4295868358https://crystdb.nims.go.jp/crystdb/search-phase-details?condition_value=Fe+O%26pageDetails=1%26page=1%26history=true%26phase_diagram_id=4295868358https://doi.org/10.1007/BF02645709https://doi.org/10.1016/j.corsci.2022.110918https://doi.org/10.1016/j.corsci.2023.111379https://doi.org/10.1021/acs.jpcc.5c04712https://doi.org/10.1098/rspa.1962.0205https://doi.org/10.1149/1.2425894https://doi.org/10.2355/isijinternational.53.1062https://doi.org/10.4028/www.scientific.net/MSF.490-491.153https://doi.org/10.2355/isijinternational.ISIJINT-2025-351https://doi.org/10.1016/0010-938X(88)90106-0https://doi.org/10.1016/j.corsci.2014.03.006https://doi.org/10.1007/s40195-016-0472-4https://doi.org/10.1016/j.matlet.2022.131968https://doi.org/10.1016/S0010-938X(74)80051-Xhttps://doi.org/10.1002/adma.201905107https://doi.org/10.1002/adma.201905107https://doi.org/10.1016/j.jallcom.2025.181652https://doi.org/10.1016/0010-938X(94)90161-9https://doi.org/10.1016/0013-4686(94)85150-6https://doi.org/10.1016/j.corsci.2012.08.008https://doi.org/10.1016/j.corsci.2012.08.008https://doi.org/10.2355/isijinternational.ISIJINT-2020-528