# Fileset

[[xxx]Manaka_2026_J._Electrochem._Soc._173_101502.pdf](https://mdr.nims.go.jp/filesets/9af9fa0d-fbcb-4ba5-9ff9-8e1cf5e105e1/download)

## Creator

Tomoyo Manaka, [Yusuke Tsutsumi](https://orcid.org/0000-0002-9483-1256), Mitsuhiro Goto, [Mariko Kadowaki](https://orcid.org/0000-0002-8988-3545), [Yoshiharu Murase](https://orcid.org/0000-0001-7390-851X), [Hideki Katayama](https://orcid.org/0000-0001-7947-4687), Takuya Ishimoto, Takao Hanawa

## Rights

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

## Other metadata

[Corrosion Resistance Enhancement of Type 420J2 Martensitic Stainless Steel by Laser Thermal Processing](https://mdr.nims.go.jp/datasets/4dd249bf-2876-48c1-b3df-50a3c459ae5a)

## Fulltext

Corrosion Resistance Enhancement of Type 420J2 Martensitic Stainless Steel by Laser Thermal ProcessingJournal of TheElectrochemical Society      OPEN ACCESSCorrosion Resistance Enhancement of Type420J2 Martensitic Stainless Steel by LaserThermal ProcessingTo cite this article: Tomoyo Manaka et al 2026 J. Electrochem. Soc. 173 101502 View the article online for updates and enhancements.You may also likeImprovement of the Corrosion Resistancefor Martensitic Stainless Steel By LaserThermal ProcessingTomoyo Manaka, Yusuke Tsutsumi,Mitsuhiro Goto et al.-The influence of heat treatment on themicrostructure and corrosion behavior ofselective laser melted 316L stainless steelin Ringer’s solutionSeyed Mohammadali Jazaeri Moghadas,Mahdi Yeganeh, Seyed Reza Alavi Zareeet al.-Effects of laser scanning overlap rate onmicrostructure and properties of lasersurface remelting stainless steelYuanlong Chen, Xiang Li, Jinyang Liu etal.-This content was downloaded from IP address 144.213.253.16 on 16/06/2026 at 00:46https://doi.org/10.1149/1945-7111/ae697d/article/10.1149/MA2024-02151613mtgabs/article/10.1149/MA2024-02151613mtgabs/article/10.1149/MA2024-02151613mtgabs/article/10.1088/2051-672X/ac6c42/article/10.1088/2051-672X/ac6c42/article/10.1088/2051-672X/ac6c42/article/10.1088/2051-672X/ac6c42/article/10.1088/1402-4896/ac9ca5/article/10.1088/1402-4896/ac9ca5/article/10.1088/1402-4896/ac9ca5https://pagead2.googlesyndication.com/pcs/click?xai=AKAOjsth3v4LYZ1Z08jmwHX8G1dPdS9d11bUx6ggkqH2gFaH0J6lEHsNfvKcioiq_B4gr8Axpel8Rr-ElD_W-Jr4_XGvJ4GIzgYyp52DWeaJCumsDlaniEvrlomqMFEZb60jfsfZfhnwLZf2o2kxRSQRRO1_SUz3JuK2joRtIsay8zzJl3-XsHaM8UMXJZaHjbTBAatLex5_V6fhcpii8We29FODELucEfuhYuLbVlonuWcAPqAfyl3ch6JTb22VCwNpegDNL1VyKLkgN3vGYNo9ZNE_pmLEYAqJ9U0hqZ9NZjZoJjSfzNU7poKzRXDcAt-6P2FBFVoNfrOPTkMTGOgTjamGdvEmKmOjnvDEimOlWJD9fX4Kg6TDG9mDNsU&sig=Cg0ArKJSzFJvb80c9RxO&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%3D2026Corrosion Resistance Enhancement of Type 420J2 MartensiticStainless Steel by Laser Thermal ProcessingTomoyo Manaka,1,2,z m Yusuke Tsutsumi,3,4 m Mitsuhiro Goto,5 m Mariko Kadowaki,3 mYoshiharu Murase,3 m Hideki Katayama,3 m Takuya Ishimoto,1,2 m and Takao Hanawa6,7,8 m1School of Sustainable Design, University of Toyama, Toyama 930-8555, Japan2Titanium Research Center, University of Toyama, Toyama 930-8555, Japan3Research Center for Structural Materials, National Institute for Materials Science, Tsukuba 305-0047, Japan4School of Materials and Chemical Technology, Institute of Science Tokyo, Tokyo 152-8552, Japan5Fuji Koushuha Industry Co., Ltd, Sakai 590-0001, Japan6Graduate School of Medicine, Kobe University, Kobe 650-0047, Japan7Institute of Science Tokyo, Tokyo 101-0062, Japan8Graduate School of Engineering, Osaka University, Suita 565-0871, JapanA laser thermal processing method was developed to enhance the corrosion resistance of martensitic stainless steels, which typicallyoffer high hardness but poor corrosion resistance. Specifically, type 420J2 martensitic stainless steel (420J2 SS) plates weresubjected to laser irradiation, causing rapid heating and quenching that modified the microstructure. After processing, three distinctlayers were observed from the surface downward: a remelted layer, a phase-transformed layer, and the substrate. The remelted layerwas formed by localized melting from the laser heat, followed by rapid quenching and solidification. Corrosion-inducing inclusionswithin this layer melted into the matrix, and their reprecipitation was effectively suppressed. The microstructure of the remeltedlayer comprised fine needle-like features, along with martensitic and austenitic phases. Due to the dominance of the martensiticphase, the hardness of the remelted layer increased significantly. Consequently, the remelted surface exhibited improved corrosionresistance and hardness. In contrast, the microstructure and corrosion resistance of the inner layer remained similar to those of theuntreated 420J2 SS, indicating that the laser thermal processing did not adversely affect the substrate. Overall, this method providesan effective means of improving both the corrosion resistance and surface hardness of martensitic stainless steels.© 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/ae697d]Manuscript submitted January 7, 2026; revised manuscript received April 15, 2026. Published May 18, 2026.Stainless steels are widely used in applications ranging fromstructural components to medical devices. Their high corrosionresistance is largely attributed to the spontaneous formation ofpassive films on their surfaces, primarily composed of Cr oxide orCr-hydrated oxyhydroxide.1–3 However, despite this inherent resis-tance, stainless steels are susceptible to localized corrosion, such aspitting, particularly in Cl− containing environments.4Among various types of stainless steels—precipitation, hardening,ferritic, austenitic, and duplex—martensitic stainless steels exhibit thelowest corrosion resistance. This is partly due to their carbon content,which, while necessary for enhancing hardness through quenching,adversely affects corrosion resistance. Martensitic stainless steels arevalued for their high hardness, strength, and wear resistance whensubjected to appropriate heat treatments.5 These properties make themsuitable for medical tools such as surgical instruments that requiresharp cutting edges. However, achieving both high hardness and goodcorrosion resistance is inherently challenging. As a result, the limitedcorrosion resistance of martensitic stainless steels restricts theirapplicability in environments that demand both mechanical perfor-mance and long-term corrosion durability.Micron-scale sulfide inclusions, particularly manganese sulfide(MnS), are known to serve as initiation sites for corrosion in stainlesssteels.6–12 Removing these inclusions from the surface can signifi-cantly enhance corrosion resistance. Additive manufacturing (AM)techniques, especially laser powder bed fusion (LPBF), are effectivein reducing MnS content. LPBF involves melting and layering metalpowders using a laser beam,13 and studies have reported improvedcorrosion resistance in LPBF-fabricated 316 l and 420J2 stainlesssteels.14–20 The rapid melting and solidification in LPBF suppress thereformation of inclusions, contributing to enhanced corrosionperformance.14,19,20 As schematically illustrated in Fig. 1, thesuppression of inclusion reformation reduces the number of micro-galvanic sites, where anodic dissolution (Fe → Fe2+ + 2e−) occursat locally active regions and cathodic oxygen reduction (O2 + 2H2O+ 4e− → 4OH−) proceeds on the surrounding surface, therebyimproving corrosion resistance.Despite its advantages, AM has limitations when applied to large-scale industrial manufacturing. High costs, long processing times,and restricted build volumes make it more suitable for high-value,small-sized components, such as those used in biomedical oraerospace applications. In contrast, general-purpose stainless-steelproducts, including structural materials, require scalable, cost-effective, and time-efficient processing methods. Moreover, LPBFcannot be applied post-manufacture to modify existing components.To address these challenges, alternative techniques that canproduce LPBF-like microstructures with reduced MnS content atlower cost and complexity are required. One promising approachinvolves laser-based surface treatments. Ultra-rapid cooling via laserirradiation can induce specific thermal histories in the surface layer,similar to those achieved by LPBF, without relying on powder-basedprocesses. Laser surface modification of stainless steels has beenstudied previously.21–24 For example, yttrium-aluminum-garnet lasertreatments have been shown to form ferritic microstructures withhomogenized chromium distribution or δ-ferrite, both of whichimprove pitting corrosion resistance.21,22 However, the use of semi-conductor lasers to selectively modify only the extreme surface layerof martensitic stainless steels, mimicking LPBF microstructures andcorrosion resistance depending on irradiation conditions, has not beenthoroughly explored. Furthermore, the impact of irradiation conditionson microstructural evolution and corrosion resistance, as well as therole of MnS inclusions, remains poorly understood.In this study, we focused on laser thermal processing, a techniquetraditionally used for hardening and cladding, and adapted it toselectively melt a thin surface layer of martensitic stainless steel. Weinvestigated whether this approach could generate LPBF-like surfacemicrostructures with enhanced corrosion resistance. The effects oflaser thermal processing on corrosion resistance, microstructuralevolution, and hardness were systematically evaluated for both thetreated surface layer and the underlying base material.zE-mail: manaka@sus.u-toyama.ac.jpJournal of The Electrochemical Society, 2026 173 101502 aaahttps://orcid.org/0000-0001-9649-2035https://orcid.org/0000-0002-9483-1256https://orcid.org/0009-0000-2426-3233https://orcid.org/0000-0002-8988-3545https://orcid.org/0000-0001-7390-851Xhttps://orcid.org/0000-0001-7947-4687https://orcid.org/0000-0003-0081-0591https://orcid.org/0000-0003-1688-1749https://creativecommons.org/licenses/by/4.0/https://creativecommons.org/licenses/by/4.0/https://doi.org/10.1149/1945-7111/ae697dhttps://doi.org/10.1149/1945-7111/ae697dmailto:manaka@sus.u-toyama.ac.jphttps://crossmark.crossref.org/dialog/?doi=10.1149/1945-7111/ae697d&domain=pdf&date_stamp=2026-05-18ExperimentalSpecimen preparation.—A martensitic 420J2 stainless steel(420J2 SS) plate (40 × 80 × 5 mm) (Test Materials Co., Ltd) wasused in this study. The specimens were solution-treated and annealedaccording to the standard and had an HBW of 235 or less. Theirchemical composition is presented in Table I. The surfaces of thespecimens were mechanically wet ground using #150, #320, #600,and #800 waterproof abrasive SiC papers. Subsequently, the speci-mens were ultrasonicated twice in acetone and once in isopropanolfor 600 s each.The prepared 420J2 stainless steel specimens were subjected tolaser thermal processing using the following procedure. Laserirradiation was performed using a semiconductor laser oscillator(LDF5000–100). The specimens were irradiated with a beam widthof 10 mm at a scanning rate of 100 cm min−1 using laser powers of950, 1100, 1650, 2000, 2500, 3000, 3500, or 4000 W. The irradiationwas performed under atmospheric conditions in a large-sizedenclosure, especially designed for processing large structures. Thelaser unit was positioned above the specimen, scanning from edge toedge along its longitudinal direction.The width of the area exhibiting metallographic structuralchanges after one cycle of laser irradiation ranged from approxi-mately 8.0 to 10.0 mm. For electrochemical measurements, both rims—including the interfaces between laser-irradiated and non-irra-diated areas —were covered with insulating resin paint (SUNECONMask Ace S, Taiyo Chemicals and Engineering Co., Ltd) toeffectively evaluate the influence of laser irradiation on corrosionresistance. The resin adhered well to the specimen and preventedcrevice corrosion by sealing gaps at the specimen–resin interface.Therefore, only the 7.5 mm wide laser-irradiated area was exposed tothe test solution.The thermally oxidized surface layer formed during laserprocessing was removed by mechanical polishing prior to subsequentanalyses, including electrochemical measurements. The specimenswere stored under controlled conditions at a constant temperature,and the elapsed time between laser processing and polishing wasmore than several days, while that between polishing and measure-ment was within several days.Potentiodynamic polarization measurements.—Anodic polari-zation measurements were conducted to evaluate the effect of laserthermal processing on corrosion resistance. The specimen wasplaced in an electrochemical measurement cell, exposing only acircular measurement area to the solution. The original exposedarea was 0.78 cm2, but was reduced to 0.67 cm2 due to the maskingof the rims with an insulating resin paint, as described above.A Pt-coated Ti rod and an Ag/AgCl (3.33 M KCl) electrodewere used as the counter and reference electrodes, respectively.A mixed aqueous solution of 0.1 M Na2SO4 and 0.585 M NaCl,simulating seawater chloride content, was used as the test solutionat 297 ± 3 K.Each electrode was connected to a potentiostat (HABF-501A,Meiden Hokuto Corp.), a function generator (HB-111, MeidenHokuto Corp.), and a data logger (midi LOGGER GL220,Graphtec Corp.). The open-circuit potential (Ecorr) was recordedafter 600 s of immersion. A linearly increasing anodic potential scanat 20 mV min−1 was applied. Anodic polarization was initiated fromEcorr − 100 mV to ensure stable initial conditions and reproduciblemeasurements. Starting slightly below Ecorr helps to minimizefluctuations near the open-circuit potential and enables a smoothtransition to anodic polarization. The holding time at the initialpotential was minimized to avoid surface modification such ascathodic alkalization. The test was stopped when the current densityFigure 1. Schematic illustration of the corrosion mechanism on the laser-processed surface of Type 420J2 stainless steel in a neutral chloride solution, whereanodic dissolution and cathodic oxygen reduction occur at locally separated sites, and suppression of inclusion reformation reduces micro-galvanic sites andimproves corrosion resistance.Table I. Chemical composition of 420J2 SS.Element Fe Cr Ni CConcentration (ppm) Bal. 13.20 0.28 0.31Element P S Si MnConcentration (ppm) 0.017 0.003 0.48 0.42Journal of The Electrochemical Society, 2026 173 101502reached 10 mA cm−2 after a rapid increase caused by pittingcorrosion. The pitting potential (Epit) was defined as the potentialat which the current density reached 100 μA cm−2. Each experimentwas repeated five times under the same conditions to ensurereproducibility.Surface and cross-sectional observation.—The surfaces andcross-sections of the specimens before and after laser thermalprocessing were observed using optical/laser microscopy (VK-X200, VHX-5000, Keyence Corp.) and scanning electron micro-scopy (SEM) equipped with an energy-dispersive X-ray spectro-scopy (EDS) (TM4000Plus, Hitachi High-Tech Corp.). A cross-section approximately 2 mm from the specimen edge was analyzedafter laser thermal processing.The specimens were cut into small pieces and embedded in cold-mounting epoxy resin (Epofix, Struers LLC) to expose the cross-sectional surfaces. Each embedded specimen was mechanicallyground using P150, P320, P600, and P800 grit abrasive SiC papersand then buff-polished to a mirror finish using 9 and 3 μmpolycrystalline diamond and 1 μm Al2O3 suspension. For opticalmicroscopy, the specimens were chemically etched to revealmetallographic structures using a solution of 33 ml of H2O, 33 mlof ethanol, 33 ml of HCl, and 1.5 g of CuCl2. The specimens wereimmersed in the etching solution at room temperature for 10 s andthen rinsed with pure water. Before SEM/EDS analysis, the resin-embedded specimen surfaces were coated with Pt (Quick AutoCoater, SC-701AT, Sanyu Electron Co., Ltd).Microstructural characterization.—Phase compositions weredetermined using X-ray diffraction (XRD; D8 ADVANCE, BrukerCorp.) and SEM (SU5000, Hitachi High-Tech Corp.) equipped withan electron backscatter diffraction (EBSD) detector (Digi View VEBSD camera, EDAX) to assess the effect of laser thermalprocessing on the crystal structure.XRD data were acquired from the standard surface and fromsurfaces ground to depths of 50, 250, and 750 μm using a specimenprocessed at 2000 W. The surfaces were mechanically ground usingP150, P320, P600, and P800 grit abrasive SiC papers, followed bybuff-polishing with a 9 and 3 μm polycrystalline diamond, 1 μmAl2O3 suspension, and 0.06 μm colloidal silica suspension. Thespecimens were then ultrasonicated twice in acetone and once inisopropanol for 600 s each. Cu Kα radiation at 40 kV and 40 mA wasused as the X-ray source, and data were collected over a 2θ range of30–90°. Diffraction patterns were analyzed using PC software(DIFFRAC. EVA, Bruker Corp.).EBSD data were acquired from the cross-sections of specimensprocessed at 950 and 2000 W, as well as unprocessed specimens.Surface preparation followed the same procedures as used for XRDspecimens.Vickers microhardness measurement.—Vickers hardness testswere performed to evaluate the effects of the laser thermal proces-sing on the hardness of the material. The same specimens as thosedescribed in the Microstructural characterization section wereemployed. Hardness was measured using a Vickers microhardnesstester (HMV-G 20S, SHIMADZU Corp.) with a 2.942 N load and aloading time of 15 s. Measurements were taken at 12 randomlyselected points on each surface, and the average hardness wascalculated from 10 values after excluding the maximum andminimum.Results and DiscussionMaterial morphology.—The metallic luster of the laser-irradiatedregion disappeared after processing, and the surface appeared dullbrown. This discoloration and roughened texture are attributed tothermally induced oxidation caused by high surface temperaturesduring laser thermal processing in atmospheric conditions.Cross-sectional images were acquired to examine how laserthermal processing altered the microstructure beneath the surface ofthe material. Specimens were etched to enhance the visibility ofmicrostructural differences. Observations revealed distinct character-istics between specimens processed at 1100 W or lower and thoseprocessed at 1650 W or higher. Figure 2 shows optical micrographs ofthe specimens processed at 950 and 2000 W. The 950 W specimenexhibited a flat surface profile, whereas the 2000 W specimendisplayed a slight central bulge along the laser irradiation line andshallow indentations on both sides. This morphology is attributed tothe Gaussian-like energy distribution of the laser beam, whichproduces a higher power density at the center, resulting in preferentialmelting in this region. In addition, thermally induced surface tensiongradients within the molten pool generate Marangoni convection,leading to the redistribution of molten material from the hotter centralregion toward the surrounding areas. As a result, material accumulatesin the central region, forming a slight bulge, while shallow depressionsare formed near the edges.25,26 These observations indicate that thesurface remains largely unmelted at 950 W, while pronounced surfacemelting and melt pool dynamics are evident at 2000 W. Allmicrographs revealed a bright-colored phase beneath the laser-irradiated area, contrasting with a darker underlying phase. The brightlayer represents microstructural transformation under specific thermalconditions during laser irradiation. Specimens treated at 1650 W orhigher exhibited two distinct bright layers in two phases termed“Phase A” (upper part) and “Phase B” (lower part). In contrast, only asingle bright layer was observed in specimens processed at 950 andFigure 2. Optical micrographs of the cross-section of specimens subjected tolaser thermal processing at (a) 950 and (b) 2000 W.Journal of The Electrochemical Society, 2026 173 1015021100 W, indicating the presence of only Phase B. Therefore, Phase Acorresponds to the remelted region, while Phase B represents thephase-transformed zone without melting. To evaluate the spatialhomogeneity of the Phase A layer, Vickers microhardness measure-ments were performed across the processed region. Although somescatter in hardness values was observed, no position-dependent trendwas identified, suggesting that the microstructure within Phase A isnearly homogeneous. Phase depths increased with increasing irradia-tion power. The region beneath Phase B, which may have beenmoderately heated but not transformed, is hereafter referred to as“Phase C.” Table II summarizes the irradiation powers and corre-sponding layer thicknesses of Phases A and B. With increasing laserpower, the thickness of Phase A initially increases due to the higherenergy input. However, at higher powers (⩾3000 W), the thickness ofPhase A tends to saturate. This behavior is attributed to enhanced heatdissipation into the surrounding material and increased melt poolconvection, which limit further deepening of the molten region. Incontrast, Phase B is formed in regions experiencing sufficient heatingwithout melting. At the highest power (4000 W), a larger volume ofmaterial exceeds the melting temperature, resulting in the expansion ofPhase A and a corresponding reduction in the thickness of Phase B.Microstructure.—Laser thermal processing-induced microstruc-tural changes shown in Fig. 3 were analyzed using EBSD and XRD.Figure 3 displays the inverse pole figure (IPF) and phase maps of theunprocessed, 950 W-processed, and 2000 W-processed specimens.Martensitic phases typically exhibit a body-centered tetragonal (bct)structure.27,28 However, because EBSD does not distinguish betweenferrite (body-centered cubic, bcc) and martensite (bct) phases, bothappear green in the phase maps.The unprocessed 420J2 stainless steel microstructure (Figs. 3aand 3b) consisted of equiaxed ferrite grains with random orientationand small amounts of residual austenite. Figures 3c and 3d presentthe microstructure of the surface layer (Phase B) of the 950 W-processed specimen. This area corresponds to the bright layer inFig. 1a. The microstructure consisted of acicular martensite andspherical austenite, some of which was surrounded by bcc phase,suggesting partial ferritic presence. The inner layer (Phase C,Figs. 3e and 3f) retained a microstructure (area corresponding tothe dark layer in Fig. 1a) similar to the unprocessed specimen,indicating that no phase transformation occurred.Figures 3g and 3h shows the microstructure of the top surfacelayer (Phase A) of the specimen processed at 2000 W. This areacorresponds to the top of the bright layer in Fig. 1b. The top surfaceconsists of acicular martensitic and relatively coarse austeniticmicrostructures. Figures 3i and 3j present the microstructure of thebottom surface layer (Phase B) of the 2000 W-processed specimen.This area corresponds to the bottom of the bright layer in Fig. 1b.The microstructure was similar to the top surface of the 950 W-processed specimen. Figures 3k and 3l show the microstructure ofthe inner layer (Phase C) of the 2000 W-processed specimen. Thisarea corresponds to the dark layer in Fig. 1b. The microstructure issimilar to the unprocessed specimen.The XRD patterns of the 2000 W-processed specimen (Fig. 4)revealed peaks corresponding to bcc structures,29,30 confirmingferrite as the primary phase. The Phase A layer displayed broaderand weaker peaks, indicating reduced crystallinity. The full width athalf maximum (FWHM) values decreased in the order: Phase A >Phase B > Phase C and the unprocessed, with peak intensityshowing the opposite trend. The XRD peak shape and intensity ofthe Phase C layer were nearly identical to those of the unprocessedspecimen. The large FWHM peaks indicate that the Phase A layerhas a less crystalline microstructure than the original layer. Figure 4bshows the enlarged graph of the α (110) peak. An asymmetric α(110) peak was observed in Phase A, with peak splitting indicative ofa bct structure,31 confirming martensitic transformation. Thesefindings align with the EBSD phase maps in Fig. 3. Laser thermalprocessing, therefore, enables the formation of engineered micro-structural layers, several hundred micrometers deep, on specimensurfaces.In addition, MnS inclusions, which are known to influencecorrosion behavior, were observed in the unprocessed specimen,whereas they were not clearly detected in the remelted layer, asdiscussed in detail in Fig. 9.Hardness.—Martensitic stainless steels are known to hardenthrough heat treatments such as quenching.32,33 Given the rapidcooling during laser thermal processing, significant surface hard-ening is expected. Further, the Phase A layer is composed of amartensitic structure owing to the rapid cooling conditions.Therefore, the hardness before and after laser thermal processingwas investigated.Figure 5 shows the Vickers microhardness profile of the 2000 W-processed specimen. The unprocessed specimen had a hardness of175 HV, while that of Phase A exceeded 690 HV, confirmingsubstantial hardening due to martensitic transformation. This result isconsistent with the XRD diffraction pattern (Fig. 4) and the EBSDphase map (Fig. 3h), which confirms the presence of martensite inPhase A. Generally, the martensitic phase is harder than the ferriticphase owing to the presence of solidified carbon atoms in the crystallattice34 Therefore, the cooling rate of laser thermal processing issufficient to form a hardened surface in the Phase A containing themartensitic phase.The Phase B layer also showed elevated hardness compared tothe unprocessed and Phase C layers. This gradient likely buffers thehardness mismatch at the interface between Phase A and Phase Clayers, minimizing cracking. The Phase C layer exhibited hardnesscomparable to the unprocessed material.Corrosion resistance.—Anodic polarization tests were performedto investigate the effect of laser thermal processing on the corrosionresistance of stainless steel. Measurements were conducted on thesurface of the as-laser-thermal-processed (as-processed) specimen,and polished surfaces exposing individual phases (A, B, and C).The surface of the as-processed specimen was thermally oxidizedto a dull brown color. Figure 6 presents polarization curves of theunprocessed, 950 W-processed, and 2000 W-processed specimens.The unprocessed specimen exhibited a passive region with a lowcurrent density below 10−6 A cm−2, followed by a sharp currentincrease, indicating pitting corrosion. The Epit value was approxi-mately 0.1 V, and the average Epit was 0.056 ± 0.039 V, which wasthe lowest among all tested conditions.As shown in Fig. 6, several transient current spikes wereobserved prior to the rapid current increase, indicating metastableTable II. Thicknesses of newly formed layers induced by laser thermal processing.Laser irradiation power (W)950 1100 1650 2000 2500 3000 3500 4000Phase A (μm) — — 110 170 220 360 360 330−130 −210 −240Phase B (μm) 180 270 320 440 450 490 460 340−290 −370 −470 −480Journal of The Electrochemical Society, 2026 173 101502pitting events associated with repeated pit initiation and repassiva-tion. In contrast, small fluctuations in the low-current region areattributed to instrumental noise near the detection limit of thepotentiostat and do not represent metastable pitting behavior.Notably, the frequency of metastable pitting events decreases fromthe unprocessed specimen to Phase B and further to Phase A,suggesting a reduction in the number of active sites for localizedcorrosion. This behavior is consistent with the reduction or elimina-tion of MnS inclusions, which are known to act as preferentialinitiation sites for pitting corrosion.In contrast, the as-processed specimens exhibited lower Ecorrvalues than the unprocessed specimen and showed active dissolutionwithout the formation of a stable passive region. Cross-sectionalSEM observations revealed that the thermally formed oxide layer hasa thickness of approximately 2–3 μm. During polarization, dissolu-tion proceeds within this oxide layer, preventing exposure of theunderlying microstructure. As a result, passive film formation issuppressed, and active dissolution persists, leading to lower corro-sion resistance.The thermally oxidized layer is relatively thick and exhibits arough and inhomogeneous morphology, which is unfavorable for theformation of a stable passive film. Furthermore, spontaneousregeneration of a protective passive film in air is considered limitedunder such conditions. The time-dependent evolution of this surfacelayer should be investigated in future work.The thermally oxidized layer was removed by polishing, ex-posing the flat surface. After polishing, specimens processed at 950and 2000 W show improved passivation and pitting behavior, asevident in Fig. 6. The specimens processed at 2000 W exhibited ahigher pitting potential than the unprocessed specimen. As shown inFig. 3, the top surfaces of the specimens processed at 950 and 2000W correspond to Phases B and A, respectively. The higher Epit valueof the 2000 W specimen compared to the 950 W specimen indicatesthat the Phase A layer has superior corrosion resistance to the PhaseB layer. The transient peaks observed in the polarization curves areattributed to metastable pitting events, which involve repeatedinitiation and repassivation of localized corrosion sites. In contrast,the small fluctuations observed in the low current density region areattributed to instrumental noise near the detection limit of thepotentiostat and do not represent metastable pitting behavior. Thefrequency and magnitude of these events depend on the micro-structure and are influenced by the presence or reduction ofcorrosion-inducing inclusions such as MnS.Consequently, the corrosion resistance of the newly formedlayers via laser thermal processing (Phases A, B, and C) wasevaluated to assess the relationship between thermal processingand the corrosion resistance of martensitic stainless steel. The innerlayers of all processed specimens exhibited passive regions andsubsequent pitting corrosion at higher applied potentials. The Epitvalues of the Phase A and B layers were consistently higher thanthose of the unprocessed specimens. Figure 7 summarizes the Epitvalues obtained from the anodic polarization measurements for allthe specimens.The corrosion resistances of various layers in the processedspecimens subjected to laser thermal processing under identicalconditions were compared. The polarization curves of the Phase Alayer for specimens processed at 3500 W exhibit the highest Epitvalue, as presented in Fig. 8. To quantitatively compare the corrosionbehavior of each layer, the corrosion potential (Ecorr), corrosioncurrent density (icorr), and Epit are summarized in Table III. The icorrvalues were determined by the Tafel extrapolation method using thelinear regions of the anodic and cathodic branches. The resultsindicate that the Phase A layer exhibits the lowest icorr and highestEpit, whereas the Phase B layer shows a higher icorr, suggesting moreactive dissolution behavior.Each Phase A layer exhibited an evident passive region withoutany current spike, followed by a rapid increase in current owing topitting corrosion at higher applied potentials. The Phase A layersFigure 3. IPF and phase maps obtained from EBSD analysis: (a) and (b)maps of the unprocessed specimen; (c)–(f) Phase B and C maps of thespecimen subjected to laser thermal processing at 950 W; (g)–(l) Phase A, B,and C maps of the specimen subjected to laser thermal processing at 2000 W.Journal of The Electrochemical Society, 2026 173 101502under all processing conditions exhibited much higher Epit valuesthan those of the unprocessed specimens. In addition, the Phase Alayers showed almost no indications of metastable pitting, unlike theunprocessed specimen. The Phase B layer also exhibited a passiveregion accompanied by a current spike, indicating metastable pitting.Subsequently, it exhibited growth-type pitting corrosion similar tothe other specimens. The Epit values of the Phase B layer wereslightly higher than those of the unprocessed specimens. The phase Clayer exhibited corrosion behavior similar to that of the unprocessedspecimen; metastable pitting initiated near the Ecorr, and growth-typepitting corrosion was observed at potentials ranging from approxi-mately 0 to 0.1 V, comparable to the unprocessed specimen. Basedon the experimental results, the corrosion resistance of the newlyformed layers produced by laser thermal processing followed theorder: Phase A layer > Phase B layer > Phase C layer. The thermalhistory of the Phase C layer, located just below the Phase B layer,showed almost no detrimental effect on corrosion resistance. This isconsistent with the hardness evaluation results shown in Fig. 5.The EBSD measurements (Fig. 3) show that the Phase A and Blayers are martensitic and exhibit similar microstructures. However,their corrosion resistances are markedly different. This suggests thatmicrostructure is not the dominant factor governing the corrosionresistance of the 420J2 stainless steel. As described in theIntroduction, MnS inclusions are well known to be capable ofinducing corrosion in stainless steels. Although the likelihood ofMnS inclusions is low, even a single exposed MnS inclusion on thesurface can trigger corrosion.Figure 9 shows the SEM images and EDS maps of the unprocessedspecimen, the Phase A layer of the specimen processed at 3500 W,and the Phase B layer of the specimen processed at 2000 W.Inclusions containing Mn and S can be observed in the unprocessedspecimen, as shown in Fig. 9a. These MnS inclusions led to pittingcorrosion at low potentials. In contrast, no MnS inclusions can beidentified in the Phase A layer of the 3500 W processed specimen, asshown in Fig. 9b. While small MnS particles may still be present, theydo not induce pitting corrosion.33 On the other hand, MnS inclusionswith sizes of several micrometers are clearly observed in the Phase Blayer, as shown in Fig. 9c, similar to those in the unprocessedspecimen. This indicates that, unlike the remelted Phase A layer, thePhase B layer retains MnS inclusions due to the absence of meltingduring laser processing.Instead, submicron-sized spherical inclusions containing Si and Owere distributed across the entire surface. These inclusions areconsidered to originate from Si added as a deoxidizer.35–37 The wavysurface morphology shown in Fig. 1, together with the formation ofthese fine inclusions, indicates that melting and subsequent rapidsolidification occurred within a very short time during laser proces-sing of the Phase A layer. Under such conditions, the originalFigure 4. XRD patterns of the unprocessed specimen and the specimen subjected to laser thermal processing at 2000 W.Figure 5. Vickers microhardness values of the unprocessed specimen andthe specimen subjected to laser thermal processing at 2000 W.Journal of The Electrochemical Society, 2026 173 101502inclusions were dissolved in the molten pool, and only a limitedamount of impurities had sufficient time to reprecipitate as fineinclusions. Consequently, the formation of corrosion-inducing MnSinclusions is significantly suppressed in the Phase A layer, whichcontributes to its enhanced corrosion resistance. Although MnSinclusions may still exist at sizes below the detection limit of SEM,their role as dominant pit initiation sites is considered to besubstantially reduced. Instead, pitting corrosion in the Phase A layeris considered to initiate at submicron-sized inclusions such as Si-based oxides or other microstructural heterogeneities. However, dueto their small size and relatively homogeneous distribution, the localelectrochemical heterogeneity is reduced, resulting in a lowerprobability of stable pit formation and thus improved corrosionresistance.Each layer is redefined based on the observed thermal andstructural changes. Phase A is remelted by laser irradiation and istherefore defined as the remelted layer. In contrast, the Phase B layeris not melted by the laser heat; instead, it undergoes a phasetransformation. Therefore, Phase B is defined as the phase-trans-formed layer. The Phase C layers exhibit Epit values similar to thoseof the unprocessed specimens. Additionally, the Phase C layer showsXRD spectra, microstructure, and Vickers microhardness valuesalmost identical to those of the unprocessed specimen (Figs. 3–5).Therefore, although the Phase C layer is heated by laser irradiation,its microstructure remains unchanged and retains the same char-acteristics as the unprocessed material. As such, the Phase C layercan be regarded as the substrate.The optimal laser processing conditions are discussed at the endof this section. A comparison of the results obtained under identicallaser thermal processing conditions revealed that the Phase A layerexhibited significantly improved corrosion resistance. However, theEpit of the Phase A layer varied under different laser power levels.On average, Epit showed an increasing trend with increasingirradiation power up to 3500 W. Therefore, the Phase A layerprocessed at 3500 W was confirmed to exhibit the highest corrosionresistance.The Epit value of the specimen processed at 4000 W, the highestpower condition in this study, was lower than that of the specimenprocessed at 3500 W. This reduction is attributed to the formation ofcracks induced by excessive thermal stresses during rapid heatingand cooling. Figure 10 shows the optical micrographs of the cross-section of the specimen processed at 4000 W. Cross-sectionalobservations revealed cracks penetrating from the top surface intoFigure 6. Polarization curves of unprocessed and laser-thermal-processedspecimens.Figure 7. Epit values of unprocessed and laser-thermal-processed specimens for each layer and irradiation power.Journal of The Electrochemical Society, 2026 173 101502the inner layers in both the 3500 and 4000 W specimens (beforegrinding), likely due to excessive residual stresses induced by rapidcooling. The cracks in the specimens processed at 4000 W weredeeper than those in the specimens processed at 3500 W. Althoughcracks were not confirmed to act directly as initiation sites, they canform geometrical features similar to micro-pits. Such features maylocally restrict solution transport and promote the formation ofoccluded environments, thereby facilitating localized corrosion. Incontrast, the cracks in the 3500 W specimen were completelyremoved after grinding to expose the intact Phase A layer.Therefore, the specimen processed at 3500 W exhibited the highestEpit among all the specimens examined in this study.The effects of laser irradiation on corrosion resistance can besummarized as follows: higher laser irradiation output tends toimprove corrosion resistance to some extent. However, excessivepower can introduce cracks, which hinder the localized corrosionresistance of martensitic stainless steel. Therefore, an optimalcombination of irradiation power (3000–3500 W in this study) andgrinding depth is crucial for achieving maximum corrosion resis-tance through laser thermal processing.This discussion can be further summarized as follows: when thespecimens were irradiated with laser power below 1100 W, noremelting zone was formed, resulting in moderate heating andquenching. Consequently, a Phase B (phase-transformed) layer wasformed on the surface via laser thermal processing. Alternatively,when the laser power exceeded 1650 W, a remelted layer was newlyformed on the surface. The Phase A (remelted) layer and theunderlying Phase B (phase transformed) layer were formed on thesurface of the specimens through laser thermal processing. Therelationship between the mechanical properties and corrosionFigure 8. Polarization curves of the unprocessed specimen and the Phase A and Phase B layers processed at 3500 W.Table III. Summary of Ecorr, icorr, and Epit values for unprocessed and laser-processed specimens (mean ± standard deviation).Ecorr (V) icorr (A cm−2) Epit (V)Unprocessed −0.084 ± 0.0079 (7.8 ± 0.85) × 10−9 0.11 ± 0.060Phase A (3500 W) −0.13 ± 0.0056 (7.8 ± 3.0) × 10−10 0.62 ± 0.086Phase B (3500 W) −0.094 ± 0.015 (1.8 ± 0.46) × 10−8 0.16 ± 0.036Table IV. Microstructure, Vickers hardness, corrosion-inducing inclusions, and corrosion resistance of the layer formed by laser thermalprocessing.Phase Vickers hardness (HV) MnS inclusions Corrosion resistancePhase A: Remelted Martensite+Austenite 699 Absence HighPhase B: Phase transformed Martensite+Austenite 403 Presence ModeratePhase C: Substrate Ferrite+Residual austenite 190 Presence LowUntreated Ferrite+Residual austenite 175 Presence LowJournal of The Electrochemical Society, 2026 173 101502resistance achieved via laser thermal processing is presented inFig. 11 and Table IV. Laser thermal processing produced a remelted(Phase A) layer with reduced formation of corrosion-inducinginclusions and enhanced martensitization. Typically, in stainlesssteel, hardness and corrosion resistance are inversely related.However, in the remelted (Phase A) layer, both high hardness andhigh corrosion resistance were simultaneously achieved.This study demonstrated that laser thermal processing caneffectively enhance both hardness and corrosion resistance.However, certain limitations exist. Martensitic stainless steel har-dened by quenching must typically undergo additional heat treatment(tempering) before practical application. Tempering is essential tomitigate brittleness while maintaining adequate hardness. Therefore,further studies are needed to investigate the effects of temperingtreatment at 423–473 K on the corrosion resistance of laser-thermal-processed martensitic stainless steels.Figure 9. SEM images and EDS mapping images of (a) the unprocessedspecimen, (b) the Phase A layer processed at 3500 W, and (c) the Phase Blayer processed at 2000 W. MnS inclusions are clearly observed in theunprocessed specimen and the Phase B layer, whereas they are not detected inthe Phase A layer.Figure 10. Optical micrographs of the cross-section of the specimen afterlaser thermal processing at 4000 W.Figure 11. Pitting corrosion potential and Vickers hardness of unprocessedand laser thermal processed specimen at 2000 W. The Phase A layer of thelaser-thermal-processed specimen exhibits both high corrosion resistance andhigh hardness.Journal of The Electrochemical Society, 2026 173 101502ConclusionIn this study, we investigated whether the high corrosionresistance of 420J2 stainless steel—attributed to the formation ofmicrostructures with reduced MnS inclusions, as observed in LPBF—can be achieved solely through laser irradiation. The followingconclusions were drawn:1. Different microstructural layers were formed after laser thermalprocessing, including a remelted layer and a phase-transformedlayer.2. Both the remelted and phase-transformed layers were marten-sitic and exhibited high hardness.3. The remelted and phase-transformed layers also demonstratedimproved corrosion resistance compared to the unprocessedspecimen, with the remelted layer showing significantly bettercorrosion resistance.4. Higher laser irradiation power resulted in a deeper remeltedlayer. However, excessive laser power led to the formation ofsurface cracks, which adversely affected corrosion resistance.5. In the remelted layer, the formation of corrosion-inducinginclusions (such as MnS) was suppressed due to the rapidremelting and solidification induced by laser irradiation.Consequently, laser-thermal-processed stainless steel exhibitedenhanced corrosion resistance.Therefore, laser thermal processing is a promising technique forachieving both high corrosion resistance and hardness, an outcomethat is typically difficult to attain using conventional methods formartensitic stainless steels.AcknowledgmentsThis study was supported by JSPS KAKENHI (grant number:JP22KJ1204, JP23K19179, JP24K17529, JP24K01223, JP26K17673).This study was also supported by The Japan Institute of Metals andMaterials (JIMM) grants-in-aid for frontier research. This study wassupported by Institute of Light Metals (ILM) Joint Usage/ResearchGrant, Kumamoto University & University of Toyama. This work wassupported by the Division of Instrumental Analysis at University ofToyama.ORCIDTomoyo Manaka m https://orcid.org/0000-0001-9649-2035Yusuke Tsutsumi m https://orcid.org/0000-0002-9483-1256Mitsuhiro Goto m https://orcid.org/0009-0000-2426-3233Mariko Kadowaki m https://orcid.org/0000-0002-8988-3545Yoshiharu Murase m https://orcid.org/0000-0001-7390-851XHideki Katayama m https://orcid.org/0000-0001-7947-4687Takuya Ishimoto m https://orcid.org/0000-0003-0081-0591Takao Hanawa m https://orcid.org/0000-0003-1688-1749References1. J. Pan, Front. Mater., 7, 133 (2020).2. C. O. A. Olsson and D. Landolt, Electrochimic. Acta, 48, 1093 (2003).3. E. Hamada, K. Yamada, M. Nagoshi, N. Makiishi, K. Sato, T. Ishii, K. Fukuda,S. Ishikawa, and T. Ujiro, Corros. Sci., 52, 3851 (2010).4. L. Gardner, Prog. Struct. Eng. and Mater., 7, 45 (2005).5. L. D. Barlow and M. Du Toit, J. Mater. Eng. Perform., 21, 1327 (2012).6. A. Chiba, I. Muto, Y. Sugawara, and N. Hara, J. Electrochem. Soc., 160, C511 (2013).7. A. Chiba, I. Muto, Y. Sugawara, and N. Hara, J. Electrochem. Soc., 159, C341(2012).8. D. Li, F. Huang, X. Lei, and Y. Jin, Corros. Sci., 211, 110860 (2023).9. R. Ke and R. Alkire, J. Electrochem. Soc., 139, 1573 (1992).10. G. S. Frankel, J. Electrochem. Soc., 145, 2186 (1998).11. M. P. Ryan, D. E. Williams, R. J. Chater, B. M. Hutton, and D. S. McPhail, Nature,415, 770 (2002).12. J. H. Park and Y. Kang, Steel Res. Int., 88, 1700130 (2017).13. H. L. Jia, H. Sun, H. Z. Wang, Y. Wu, and H. W. Wang, Int. J. Adv. Manuf.Technol., 113, 2413 (2021).14. Y. Tsutsumi, T. Ishimoto, T. Oishi, T. Manaka, P. Chen, M. Ashida, K. Doi,H. Katayama, T. Hanawa, and T. Nakano, Addit. Manuf., 45, 102066 (2021).15. S. H. Sun, T. Ishimoto, K. Hagihara, Y. Tsutsumi, T. Hanawa, and T. Nakano, Scr.Mater., 159, 89 (2019).16. Y. Zhao, H. Xiong, X. Li, W. Qi, J. Wang, Y. Hua, T. Zhang, and F. Wang, Corros.Commun., 2, 55 (2021).17. Q. Chao, V. Cruz, S. Thomas, N. Birbilis, P. Collins, A. Taylor, P. D. Hodgson, andD. Fabijanic, Scr. Mater., 141, 94 (2017).18. M. Kazemipour, M. Mohammadi, E. Mfoumou, and A. M. Nasiri, JOM, 71, 3230(2019).19. G. Sander, S. Thomas, V. Cruz, M. Jurg, N. Birbilis, X. Gao, M. Brameld, and C.R. Hutchinson, J. Electrochem. Soc., 164, C250 (2017).20. K. Saeidi, D. L. Zapata, F. Lofaj, L. Kvetkova, J. Olsen, Z. J. Shen, and F. Akhtar,Addit. Manuf., 29, 100803 (2019).21. P. H. Chong, Z. Liu, P. Skeldon, and P. Crouse, Appl. Surf. Sci., 247, 362 (2005).22. P. H. Chong, Z. Liu, X. Y. Wang, and P. Skeldon, Thin Solid Films, 453, 388 (2004).23. Y. Tsutsumi, M. Shimabukuro, T. Manaka, M. Goto, M. Kadowaki,T. Hashimoto, M. Kawashita, T. Ishimoto, and H. Katayama, APL Mater., 14,051105 (2026).24. A. Conde, I. Garcia, and J. J. de Damborenea, Corros. Sci., 43, 817 (2001).25. T. N. Le and Y. L. Lo, Mater. Des., 179, 107866 (2019).26. B. Song, T. Yu, X. Jiang, W. Xi, and X. Lin, Int. J. Mech. Sci., 165, 105207 (2020).27. Y. Tomota, H. Tokuda, S. Torii, and T. Kamiyama, Mater. Sci. Eng. A, 434, 82(2006).28. T. W. Liu, D. H. Ping, T. Ohmura, and M. Ohnuma, J. Mater. Sci., 53, 2976 (2018).29. E. P. Yelsukov, E. V. Voronina, and V. A. Barinov, J. Magn. Magn. Mater., 115,271 (1992).30. K. Sumiyama, Y. Hirose, and Y. Nakamura, J. Phys. Soc. Jpn., 59, 2963 (1990).31. Y. Lu, H. X. Yu, and R. D. Sisson, Mater. Sci. Eng. A, 700, 592 (2017).32. P. Landgraf, J. Man, M. Cieslar, and J. Dlouhy, Metals, 11, 1071 (2021).33. W. Jiang, Y. Liu, X. Zhang, and H. Wang, Metals, 14, 322 (2024).34. B. Hutchinson, J. Hagstrom, O. Karlsson, D. Lindell, M. Tornberg, F. Lindberg, andM. Thuvander, Acta Mater., 59, 5845 (2011).35. Z. Zhan, G. Wang, R. Shi, W. Zhang, Y. Zhang, and G. Cheng, ISIJ Int., 63, 622(2023).36. K. Suzuki, S. Ban-ya, and M. Hino, ISIJ Int., 42, 146 (2002).37. L. Su, X. Li, Y. Zhang, Z. Zhang, and J. Li, J. Mater. Res. Technol., 15, 1949(2021).Journal of The Electrochemical Society, 2026 173 101502https://orcid.org/0000-0001-9649-2035https://orcid.org/0000-0002-9483-1256https://orcid.org/0009-0000-2426-3233https://orcid.org/0000-0002-8988-3545https://orcid.org/0000-0001-7390-851Xhttps://orcid.org/0000-0001-7947-4687https://orcid.org/0000-0003-0081-0591https://orcid.org/0000-0003-1688-1749https://doi.org/10.3389/fmats.2020.00133https://doi.org/10.1016/S0013-4686(02)00841-1https://doi.org/10.1016/j.corsci.2010.08.025https://doi.org/10.1002/pse.190https://doi.org/10.1007/s11665-011-0043-9https://doi.org/10.1149/2.081310jeshttps://doi.org/10.1149/2.054208jeshttps://doi.org/10.1016/j.corsci.2022.110860https://doi.org/10.1149/1.2069458https://doi.org/10.1149/1.1838615https://doi.org/10.1038/415770ahttps://doi.org/10.1002/srin.201700130https://doi.org/10.1007/s00170-021-06810-3https://doi.org/10.1007/s00170-021-06810-3https://doi.org/10.1016/j.addma.2021.102066https://doi.org/10.1016/j.scriptamat.2018.09.017https://doi.org/10.1016/j.scriptamat.2018.09.017https://doi.org/10.1016/j.corcom.2021.09.002https://doi.org/10.1016/j.corcom.2021.09.002https://doi.org/10.1016/j.scriptamat.2017.07.037https://doi.org/10.1007/s11837-019-03647-whttps://doi.org/10.1149/2.0551706jeshttps://doi.org/10.1016/j.addma.2019.100803https://doi.org/10.1016/j.apsusc.2005.01.136https://doi.org/10.1016/j.tsf.2003.11.094https://doi.org/10.1063/5.0334196https://doi.org/10.1016/S0010-938X(00)00114-1https://doi.org/10.1016/j.matdes.2019.107866https://doi.org/10.1016/j.ijmecsci.2019.105207https://doi.org/10.1016/j.msea.2006.06.067https://doi.org/10.1007/s10853-017-1731-0https://doi.org/10.1016/0304-8853(92)90069-Zhttps://doi.org/10.1143/JPSJ.59.2963https://doi.org/10.1016/j.msea.2017.05.094https://doi.org/10.3390/met11071071https://doi.org/10.3390/met14030322https://doi.org/10.1016/j.actamat.2011.05.061https://doi.org/10.2355/isijinternational.ISIJINT-2022-382https://doi.org/10.2355/isijinternational.42.146https://doi.org/10.1016/j.jmrt.2021.09.011