Corrosion, Microstructural Evolution and Non-Destructive Monitoring of High-Strength Low-Alloy Steels Under Multiparametric Marine Exposure

Abstract

High-strength low-alloy (HSLA) steels in marine environments suffer from microbiologically influenced corrosion (MIC) and hydrogen-assisted degradation. This study investigates the synergistic effects of sulfate-reducing bacterial biofilms, mechanical stress, and seawater chemistry on HSLA AH36 steel using electrochemical, microstructural, and magnetic Barkhausen noise (MBN) monitoring. Under multiparametric exposure (80% yield strength tensile stress, Desulfovibrio vulgaris, 28 days), biotic samples exhibited sustained 1.88× corrosion acceleration despite 86% sulfate depletion. Magnetic Barkhausen noise RMS amplitude (MBN_RMS) peaked at day 7 (612 ± 38 mV/mm) at pit depths of only 20–50 μm, detecting subsurface hydrogen damage before macroscopic failure. Quantitative correlations (R² ≥ 0.99) between MBN_RMS and cumulative mass loss revealed distinctive linear relationships in abiotic conditions and nonlinear cubic polynomials in biotic conditions, providing a non-destructive signature diagnostic of hydrogen-assisted MIC. Directional anisotropy analysis (parallel vs. perpendicular fields) showed that hydrogen-induced damage produces isotropic magnetic signatures (anisotropy ratio: 1.27 → 1.15), enabling discrimination between hydrogen embrittlement and stress-controlled degradation. The integration of portable MBN measurements with electrochemical monitoring establishes a quantitative framework for real-time structural health assessment and predictive maintenance of HSLA steels in maritime applications.

1. Introduction

High-strength low-alloy (HSLA) steels underpin maritime infrastructure—ship hulls, offshore platforms, and subsea pipelines—due to their mechanical strength, weldability, and cost-effectiveness [1]. Yet, their durability faces synergistic threats in marine environments, where electrochemical corrosion, microbial activity, and mechanical loading interact [2,3]. Safeguarding structural integrity demands not only understanding these coupled degradation pathways but also deploying robust, non-destructive monitoring technologies capable of detecting early-stage damage [4,5,6].

Marine corrosion arises primarily from chloride ion attack, which compromises passive films and triggers localized pitting [7]. Microbiologically influenced corrosion (MIC) exacerbates this process: sulfate-reducing bacteria (SRB) such as Desulfovibrio vulgaris form biofilms that acidify local pH, produce hydrogen sulfide (H₂S), and accelerate anodic dissolution while promoting hydrogen permeation [8,9,10,11,12,13]. Recent evidence shows that specific bacterial strains induce severe pitting and microstructural degradation underestimated by abiotic models [14,15]. SRB-driven H₂S acts as a cathodic depolarizer, shifting open-circuit potentials to more negative values and sustaining elevated corrosion rates even under oxygen-depleted conditions [16,17].

Operational marine structures endure cyclic mechanical loads and residual stresses that synergize with corrosive and microbial factors, lowering thresholds for stress corrosion cracking (SCC) and corrosion fatigue [3]. Steel microstructure—grain boundaries, phase constituents (martensite, bainite), and dislocation density—governs electrochemical stability and hydrogen trapping efficiency [1,15]. Consequently, materials degrade not solely through mass loss but via critical microstructural evolution, necessitating monitoring techniques sensitive to subsurface changes.

Traditional methods—weight-loss coupons, open-circuit potential logging—fail to capture dynamic stress-microstructure interactions in real time [18]. Magnetic Barkhausen noise (MBN), a non-destructive testing (NDT) technique for ferromagnetic materials, detects irreversible domain wall motion sensitive to stress (magnetoelastic effect) and microstructural features (dislocation/grain boundary pinning) [19,20,21]. While MBN successfully detects grinding burns and residual stresses in manufacturing [22], its application for in situ monitoring of corrosion-induced degradation in complex marine environments—particularly under multiparametric exposure (chemical + microbial + mechanical)—remains underexplored [23].

Despite extensive literature on individual degradation mechanisms, few studies address multiparametric HSLA steel exposure, where chemical, microbial, and mechanical factors act simultaneously [8,9]. Critically, no prior work has quantitatively correlated advanced NDT methods like MBN with microstructural evolution driven by combined SRB biofilms and mechanical stress. This study fills that gap by integrating electrochemical sensing, microstructural analysis, and MBN measurements to elucidate synergistic MIC and stress effects on HSLA AH36 steel. Specifically, we establish—for the first time—quantitative correlations between MBN signatures and hydrogen-assisted damage accumulation under SRB-dominated conditions (80% yield strength tensile stress, anaerobic D. vulgaris, 28 days at 25 ± 2 °C simulating shallow marine/ballast tank environments; future parametric studies should explore broader thermal ranges, 10–35 °C, to capture deep-sea and tropical extremes, as SRB metabolic rates exhibit temperature sensitivity [24]. By demonstrating that MBN detects subsurface hydrogen embrittlement before macroscopic failure, this work advances predictive maintenance strategies and durable design criteria for next-generation maritime structures, contributing to sustainable materials engineering and enhanced structural reliability [23,25].

2. Materials and Methods

2.1. Material Characterization and Specimen Preparation

The material investigated in this study is a high-strength low-alloy (HSLA) steel, specifically grade AH36, widely utilized in ship hull construction and offshore platforms [26,27]. The chemical composition (wt.%) of the steel was determined using optical emission spectroscopy (Thermo Scientific ARL iSpark Plus, Thermo Fisher Scientific Inc., Waltham, MA, USA) and is listed in

Table 1. Chemical composition of the investigated HSLA steel (wt.%).

Element	C	Mn	Si	P	S	Cr	Ni	Mo	Cu	Fe
wt.%	0.16	1.40	0.35	0.015	0.005	0.05	0.03	0.01	0.04	Bal.

.

Rectangular coupons with dimensions of 100 mm × 20 mm × 5 mm were machined from the rolled plate. Prior to exposure, the working surfaces were ground sequentially with silicon carbide (SiC) papers up to 1200 grit, degreased with acetone and ethanol in an ultrasonic bath, and dried under a nitrogen stream. For electrochemical measurements, a copper wire was soldered to the non-working face, and the samples were embedded in epoxy resin, leaving a working area of 1 cm² exposed to the electrolyte. All specimens were sterilized by UV irradiation (254 nm, 30 min) before immersion to prevent abiotic contamination.

2.2. Microorganisms and Culture Media

The microbiologically influenced corrosion (MIC) experiments utilized Desulfovibrio vulgaris (ATCC 7757), a representative strain of sulfate-reducing bacteria (SRB) [16]. The bacteria were cultured in modified Postgate’s medium C, consisting of 0.5 g KH₂PO₄, 1.0 g NH₄Cl, 4.5 g Na₂SO₄, 0.06 g CaCl₂·6H₂O, 0.06 g MgSO₄·7H₂O, 6.0 g sodium lactate, 1.0 g yeast extract, and 0.3 g sodium citrate per liter of filtered artificial seawater (ASTM D1141) [28]. The pH was adjusted to 7.2 using 1 M NaOH. The medium was deoxygenated by purging with sterile N₂ gas for 45 min and autoclaved at 121 °C for 20 min prior to inoculation.

2.3. Multiparametric Marine Exposure Setup

2.3.1. Mechanical Stress Application

The HSLA steel specimens were loaded into four-point bending fixtures tailored to apply a constant elastic tensile stress equivalent to 80% of the material’s yield strength (σ_y = 355 MPa). The stress level was verified using strain gauges (Tokyo Sokki Kenkyujo TML FLA-5-11, Tokyo Sokki Kenkyujo Co., Ltd., Tokyo, Japan) attached to the tensile surface. Figure S1a illustrates the four-point bending configuration, with the specimen positioned such that the central section experiences uniform tensile stress on the lower surface, while the upper supports and loading points create the bending moment.

The strain gauge was positioned at the midpoint of the tensile surface (Figure S1b) to monitor and verify the applied stress, ensuring that the 284 MPa target stress (80% σ_y) was maintained throughout the 28-day exposure.

2.3.2. Immersion Conditions

The stressed assemblies were immersed in 2 L glass bioreactors containing the specific test medium (Figure S2). Two environmental conditions were tested: (i) Abiotic Control: Sterile artificial seawater; and (ii) Biotic Exposure: Inoculated medium with D. vulgaris (10⁶ cells/mL). The reactors were maintained at 25 ± 2 °C under anaerobic conditions. This temperature was selected to represent typical subsurface marine environments and to optimize SRB metabolic activity, as Desulfovibrio species exhibit maximum growth rates in the range of 20–37 °C [29,30]. Temperature control was achieved using a thermostatically regulated water bath, with continuous monitoring via submerged thermocouples.

2.3.3. Duration

The exposure tests were conducted for durations of 7, 14, and 28 days to monitor the temporal evolution of degradation. At each time point, triplicate specimens were retrieved from both abiotic and biotic reactors for comprehensive characterization.

2.4. Electrochemical Measurements

Electrochemical monitoring was performed using a Gamry Reference 600 potentiostat (Gamry Instruments, Warminster, PA, USA) in a standard three-electrode cell configuration (Figure S3). The HSLA steel specimen served as the working electrode, a platinum mesh (surface area 4 cm²) as the counter electrode, and a Saturated Calomel Electrode (SCE, 0.241 V vs. SHE) as the reference. The electrodes were positioned with the working electrode centrally located, the counter electrode placed approximately 3 cm away to ensure uniform current distribution, and the reference electrode positioned close to the working electrode surface (~2 mm) via a Luggin capillary to minimize ohmic drop (i_R) effects. All measurements were conducted at the open-circuit potential (OCP) after stabilization (>24 h).

The corrosion potential was taken directly as the stabilized open-circuit potential (OCP) value, as OCP closely approximates E_corr in systems exhibiting mixed anodic–cathodic corrosion kinetics under seawater exposure. The corrosion current density (i_corr) was calculated from Linear Polarization Resistance (LPR) measurements using the Stern–Geary equation: i_corr = (B/R_p), where R_p is the polarization resistance and B = 39 mV is the Stern–Geary constant (βa = βc = 120 mV/decade), appropriate for mixed kinetics involving anodic iron dissolution and cathodic oxygen reduction or sulfide reduction in seawater environments [31,32]. LPR scans were conducted within ±10 mV vs. OCP at a scan rate of 0.166 mV/s to ensure non-destructive measurements.

Full Tafel polarization curves (±250 mV vs. OCP) were not acquired during the 28-day multiparametric exposure to minimize surface perturbation and maintain the integrity of the long-term corrosion monitoring. LPR provides reliable, instantaneous i_corr estimates under these conditions and is validated by excellent agreement with gravimetric corrosion rates (Table S9, Figure S6; <20% discrepancy) [31,32].

The following measurements were performed:

• Open Circuit Potential (OCP): Monitored continuously for the first 24 h and daily thereafter to assess thermodynamic stability (Table S4).
• Linear Polarization Resistance (LPR): Conducted on days 1, 3, 7, 14, and 28 (Table S5).
• Electrochemical Impedance Spectroscopy (EIS): Performed at OCP with 10 mV RMS perturbation over 10 mHz–100 kHz (10 points/decade). Data analyzed using ZView software (version 4.0, Scribner Associates, Inc., Southern Pines, NC, USA) to fit equivalent electrical circuits (EECs) described in Section 3.4 (Table S6).

Representative Tafel polarization behavior from similar HSLA steel–seawater systems confirms the appropriateness of B = 39 mV for the observed mixed kinetics [32].

2.5. Magnetic Barkhausen Noise (MBN) Measurements

Non-destructive magnetic evaluation was carried out using a Rollscan 350 system (Stresstech Oy, Finland) equipped with a bipolar sensor (Figure S4). The MBN technique is based on detecting voltage pulses generated by irreversible magnetic domain wall movements during the magnetization of ferromagnetic materials [33].

• Excitation: A sinusoidal magnetic field was applied at a magnetizing frequency of 125 Hz and a magnetizing voltage of 5 Vpp.
• Data Acquisition: The MBN signal was filtered in the frequency band of 70–200 kHz using an analog bandpass filter to isolate Barkhausen emissions from background electromagnetic noise.
• Parameters: The Root Mean Square (RMS) value of the MBN signal (MBN_RMS) was extracted as the primary feature, expressed in mV/mm (normalized by sensor coil length). Measurements were taken at three distinct locations on the tensile surface of the stressed specimens, with the magnetic field applied both parallel (0°) and perpendicular (90°) to the rolling direction to evaluate magnetic anisotropy and stress-induced texture changes [21,34,35].

2.6. Surface and Microstructural Characterization

2.6.1. Sample Preparation

Post-exposure, corrosion products and biofilms were fixed with 2.5% glutaraldehyde in phosphate-buffered saline (PBS, pH 7.2) for 4 h at 4 °C to preserve biofilm architecture, then dehydrated in a graded ethanol series (50%, 70%, 90%, 100% for 15 min each step), and finally critical-point dried using CO₂ to prevent biofilm collapse.

2.6.2. Field Emission Scanning Electron Microscopy (FE-SEM)

Surface morphology and biofilm architecture were examined using Field Emission Scanning Electron Microscopy (FE-SEM, FEI Inspect F50, Thermo Fisher Scientific Inc., Waltham, MA, USA). Samples were sputter-coated with 5 nm gold-palladium (Au-Pd, 60:40 ratio) to enhance conductivity. Imaging was performed at accelerating voltages of 10–15 kV, with working distances of 10–15 mm, and magnifications ranging from 500× to 10,000×. Both secondary electron (SE) and backscattered electron (BSE) detectors were employed to visualize surface topography and compositional contrast, respectively.

2.6.3. Energy Dispersive X-Ray Spectroscopy (EDS)

The elemental distribution of corrosion products and biofilm constituents was analyzed via Energy Dispersive X-ray Spectroscopy (EDS, Bruker Quantax 200 system with XFlash 6|30 detector integrated with FE-SEM). EDS analysis was performed at 15 kV accelerating voltage with 10–20 s acquisition times per point. Spectral data were processed using Bruker Esprit 2.0 software with ZAF (atomic number, absorption, fluorescence) corrections for quantitative elemental analysis. Elemental mapping was conducted to visualize the spatial distribution of key elements (Fe, O, S, Cl, C) across biofilm–corrosion product–steel interfaces.

2.6.4. X-Ray Diffraction (XRD) Analysis

Post-exposure corrosion products were analyzed for phase composition using X-ray diffraction (XRD). Rust layers were gently scraped from the steel surface using a sterile plastic spatula to obtain bulk powder samples, minimizing substrate contamination. Cross-sections for depth-resolved analysis were prepared by mounting epoxy-embedded specimens and polishing to expose the steel–rust interface, followed by light etching with 1% nital to delineate layers.

Bulk XRD patterns were recorded using a Bruker D8 Advance diffractometer (Bruker AXS GmbH, Karlsruhe, Germany) (Cu Kα radiation, λ = 1.5406 Å, 40 kV, 40 mA) in θ–2θ configuration over 5–60° 2θ at 0.02° steps and 2 s/step. Phase identification was performed with HighScore Plus software (version 4.5, Malvern Panalytical, Almelo, The Netherlands) referencing the ICDD PDF-4+ database.

For layer-specific analysis, grazing incidence XRD (GI-XRD) was conducted at incidence angles 0.5–3° (penetration ~0.1–1 μm) to probe outer/inner layers selectively. Measurements targeted triplicate samples from each condition (abiotic/biotic) and time point (1, 3, 7, 14, 28 days). Instrumental broadening was calibrated with NIST SRM 660b LaB6 standard [36]. Peak positions and relative intensities were used to distinguish lepidocrocite (γ-FeOOH), goethite (α-FeOOH), akaganeite (β-FeOOH), magnetite (Fe₃O₄), mackinawite (FeS), pyrrhotite (Fe1–xS), and green rust phases.

2.7. Experimental Design Summary

The experimental matrix encompassed two primary exposure conditions (abiotic and biotic), four time points (3, 7, 14, 28 days), and multiple characterization techniques (electrochemical, magnetic, and microstructural).

Table 2. Summary of experimental design matrix.

Parameter	Condition	Technique	Measurement Frequency
Environment	Abiotic/Biotic	—	—
Temperature	25 ± 2 °C	Thermocouple	Continuous
Applied Stress	284 MPa (80% σy)	Strain gauge	Initial verification
Exposure Duration	3, 7, 14, 28 days	—	Discrete time points
OCP	Both conditions	Potentiostat	Continuous (day 1), daily
LPR	Both conditions	Potentiostat	Days 1, 3, 7, 14, 28
EIS	Both conditions	Potentiostat	Days 1, 7, 14, 28
MBN	Both conditions	Rollscan 350	Days 1, 3, 7, 14, 28
Morphology	Both conditions	FE-SEM, EDS	Days 3, 7, 14, 28

summarizes the experimental design.

3. Results

3.1. Experimental Conditions & Medium Characterization

Anaerobic conditions (dissolved oxygen < 0.5 mg/L) were maintained throughout the 28-day exposure in both abiotic and biotic bioreactors, confirmed by continuous optical oxygen monitoring (Table S1, Supplementary Materials). The biotic reactor exhibited progressive acidification from pH 7.2 ± 0.1 (day 0) to pH 6.8 ± 0.2 (day 28), consistent with organic acid production during Desulfovibrio vulgaris metabolism and sulfate reduction [29]. The abiotic control maintained a stable pH of 7.2 ± 0.1 throughout exposure.

Sulfate depletion in the biotic reactor confirmed sustained SRB metabolic activity: sulfate concentration decreased from 4.5 ± 0.2 g/L (day 0) to 1.2 ± 0.2 g/L (day 28), representing 73% sulfate consumption (Table S2, Figure S5). The consumption rate was highest during days 0–14 (3.3 mmol/L·week), declining to 0.85 mmol/L·week during days 14–28, consistent with lactate depletion and sulfide inhibition kinetics reported for D. vulgaris under closed-system conditions [37]. The abiotic control showed no sulfate depletion (4.5 ± 0.2 g/L throughout), confirming sterility.

Viable cell counts in the biotic reactor increased rapidly from the inoculum concentration of 10⁶ cells/mL to 2.5 ± 0.4 × 10⁸ cells/mL by day 7, indicating exponential-phase growth (Table S3). Cell viability plateaued between days 7–28 (2.2–2.8 × 10⁸ cells/mL), consistent with entry into the stationary phase due to resource limitation. Field-emission scanning electron microscopy (FE-SEM) confirmed progressive biofilm development (Figure 1): sparse cell clusters at day 3 (thickness 5 ± 2 μm) evolved into densely packed, multi-layered biofilms by day 28 (thickness 52 ± 8 μm) with extensive extracellular polymeric substance (EPS) matrix encasing rod-shaped cells morphologically consistent with Desulfovibrio (length 2–5 μm, diameter 0.5–1 μm). The abiotic control remained biofilm-free throughout exposure.

3.2. Open Circuit Potential (OCP) and Electrochemical Stability

3.2.1. Temporal Evolution of OCP

OCP measurements revealed distinct trajectories between abiotic and biotic conditions (Figure 2, Table S4). Abiotic OCP stabilized at 314 ± 10 mV SCE by day 28 after initial ennoblement from 287 mV (day 1). Biotic OCP shifted 144 mV more negative to 458 ± 18 mV SCE, reflecting biofilm-induced cathodic depolarization.

3.2.2. Kinetic Phases & Noise Analysis

Abiotic OCP exhibited rapid stabilization (dE/dt < 1 mV/day post-day 7) with low noise (σ = 10 mV), consistent with oxygen-diffusion control. Biotic OCP showed three distinct phases: colonization (days 1–3, ΔE = 80 mV), growth (days 3–14, ΔE = 75 mV), and stationary (>day 14), with elevated noise (σ = 25 mV) due to biofilm heterogeneity and localized sulfide production (Figure 2, Table S4). The biotic potential noise reflects spatial variability in biofilm coverage and HS concentration across the electrode surface.

3.3. Linear Polarization Resistance (LPR) and Corrosion Rates

3.3.1. Instantaneous Corrosion Rates from LPR

LPR measurements (B = 39 mV) showed abiotic i_corr decline from 152 μA/cm² (day 1) to 65 μA/cm² (day 28, v_corr = 0.75 mm/year) (Figure 3a, Table S5). Biotic i_corr plateaued at 118–134 μA/cm² (v_corr = 1.36 mm/year day 28), 1.8× higher despite O₂ depletion (Figure 3b, Table S5).

The LPR-derived i_corr values (Table S5) are in excellent agreement with gravimetric corrosion rates (Table S8, Figure S6), confirming the reliability of this non-destructive approach for long-term monitoring. The Stern–Geary constant B = 39 mV is validated for mixed kinetics in SRB-influenced seawater systems [31,32].

3.3.2. Acceleration Factor

Cumulative MIC acceleration factor (AF = biotic/abiotic): 1.38 (day 3) → 1.88 (day 28), sustained despite 86% sulfate depletion (Table S6). Peak AF = 1.82 (day 7) coincides with maximum SRB activity (viable cells 2.5 × 10⁸/mL).

3.4. Electrochemical Impedance Spectroscopy (EIS)

EIS measurements were performed at open circuit potential using a 10 mV RMS perturbation over 10⁵–10⁻² Hz. At day 1, both abiotic and biotic specimens exhibited a single, almost ideal semicircle in the Nyquist plots, characteristic of a charge-transfer controlled process with one dominant time constant. The corresponding charge-transfer resistances were comparable (abiotic: 1220 ± 95 Ω·cm², biotic: 1150 ± 88 Ω·cm²), indicating similar initial electrochemical behavior before significant biofilm development.

With increasing immersion time, the impedance response diverged markedly between conditions. In the abiotic medium, the Nyquist semicircle diameter increased monotonically, and Rct rose from 1220 Ω·cm² (day 1) to 1810 Ω·cm² (day 7), 2180 Ω·cm² (day 14), and 2590 Ω·cm² (day 28), consistent with progressive formation of a more resistive corrosion product layer and oxygen-reduction slowdown (Figure 4, Table S6). In contrast, the biotic condition showed a strong decrease in Rct to 450 ± 35 Ω·cm² (day 7), 380 ± 30 Ω·cm² (day 14), and 290 ± 22 Ω·cm² (day 28), together with distorted arcs and an additional low-frequency feature, indicating a more active and heterogeneous interface governed by biofilm and sulfide-mediated reactions.

To quantify these behaviors, the spectra at early exposure (day 1) were fitted with a single-time-constant equivalent circuit (Model 1: R_s − [R_ct ∥ CPE₁]), representing solution resistance and a non-ideal double-layer at the steel–electrolyte interface. At later times (days 7, 14, 28), especially in the biotic medium, a second time constant was required (Model 2: R_s − [R_ct ∥ CPE₁] − [R_bf ∥ CPE₂]) to account for the additional contribution of the biofilm/porous corrosion layer (R_bf, CPE₂). Schematic diagrams of Model 1 and Model 2 are shown in Figure 5, while the fitted parameter values are summarized in Table S6.

The extracted parameters highlight opposite trends in abiotic and biotic systems. In the abiotic case, increasing R_ct and relatively stable interfacial capacitance indicate gradual passivation and limited surface roughening. In the biotic case, decreasing R_ct by approximately a factor of four, combined with increasing biofilm resistance R_bf from 350 ± 28 to 680 ± 50 Ω·cm² and higher interfacial capacitance, points to a roughened, strongly heterogeneous surface where hydrogen sulfide-driven cathodic depolarization dominates over any diffusional hindrance imposed by the biofilm (Table S7). Overall, the EIS results corroborate the LPR and OCP findings by confirming that sulfate-reducing biofilms simultaneously lower the charge-transfer barrier and promote localized, MIC-controlled corrosion.

3.5. Corrosion Rate & Mass Loss

Figure 6 presents the temporal evolution of cumulative mass loss and the corresponding average corrosion rates derived from gravimetric measurements for abiotic and biotic exposures. Detailed numerical values are provided in Table S8, while a direct comparison between gravimetric and LPR-derived corrosion rates is given in Figure S6.

In the abiotic medium, cumulative mass loss increased from 0.42 ± 0.05 mg/cm² at day 3 to 0.98 ± 0.11 mg/cm² at day 7, 1.89 ± 0.21 mg/cm² at day 14, and 3.28 ± 0.35 mg/cm² at day 28 (Figure 6, Table S8). The corresponding average corrosion rate v_avg decreased from 1.28 ± 0.15 mm/year (day 3) to 1.03 ± 0.12 mm/year (day 7), 0.98 ± 0.11 mm/year (day 14), and 0.91 ± 0.10 mm/year (day 28). This monotonic decline is consistent with the decrease in i_corr obtained from LPR and the progressive increase in R_ct obtained from EIS, and reflects a transition toward an oxygen-diffusion-limited, quasi-steady-state corrosion regime in sterile artificial seawater.

In the biotic medium, mass loss was systematically higher at all exposure times, reaching 0.58 ± 0.07 mg/cm² at day 3, 1.78 ± 0.20 mg/cm² at day 7, 3.42 ± 0.38 mg/cm² at day 14, and 6.15 ± 0.68 mg/cm² at day 28 (Figure 6, Table S8). The corresponding average corrosion rates remained in the range 1.7–1.9 mm/year over the full 28-day period (1.77 ± 0.21, 1.87 ± 0.21, 1.77 ± 0.20, and 1.71 ± 0.19 mm/year at days 3, 7, 14, and 28, respectively), indicating that the biotic system did not exhibit the same long-term deceleration observed in the abiotic control. The microbiologically influenced corrosion (MIC) acceleration factor, defined as the ratio of biotic to abiotic average rate, therefore increased from approximately 1.38 at day 3 to 1.82 at day 7, 1.81 at day 14, and 1.88 at day 28 (Table S8). The highest acceleration coincides with the period of maximum sulfate-reduction rate and nearly complete biofilm coverage, corroborating the role of Desulfovibrio vulgaris metabolism in sustaining elevated corrosion kinetics.

The gravimetric results are in good agreement with the electrochemical measurements. Table S9 and Figure S6 show that, for both abiotic and biotic conditions, the average corrosion rates calculated from mass loss closely follow the trends in instantaneous rates obtained from LPR, with differences remaining within the combined experimental uncertainty. In the abiotic system, both methods converge to long-term rates close to 1.0 mm/year or lower, characteristic of oxygen-limited corrosion in synthetic seawater. In the biotic system, both techniques consistently indicate a persistent MIC-induced acceleration of approximately 40–90% relative to the abiotic baseline over 28 days. This internal consistency between gravimetric, LPR, and EIS data strengthens the conclusion that sulfate-reducing biofilms shift the rate-controlling mechanism from diffusion-limited oxygen reduction to a biofilm-mediated, sulfide-controlled corrosion process.

3.6. Surface Morphology, Pitting and Corrosion Product Characterization

Scanning Electron Microscopy (SEM) and Surface Topography

Plan-view FE-SEM imaging after 28 days revealed distinct differences in surface morphology between abiotic (Figure 7) and biotic conditions (Figure 8). Abiotic specimens exhibited a relatively continuous corrosion product layer with shallow, sparsely distributed pits and limited surface roughening (Figure 7). In contrast, biotic specimens showed a markedly more heterogeneous surface, with thick, cracked deposits, locally detached scales, and numerous under-deposit pits distributed across the tensile surface (Figure 8). These observations are consistent with the higher mass loss and corrosion rates measured electrochemically in the presence of Desulfovibrio vulgaris (Section 3.3, Section 3.4 and Section 3.5).

Cross-sectional FE-SEM micrographs confirmed that localized attack was significantly more severe in the biotic medium (Figure 8). Abiotic samples displayed a compact inner corrosion product layer of approximately uniform thickness and pits with modest penetration into the steel substrate. Biotic samples, however, exhibited a multilayered surface structure, with a porous outer deposit, a denser inner rust layer, and deep, sharply profiled pits extending well below the original surface. Quantitative pit-depth and pit-density statistics extracted from multiple cross-sections are summarized in Tables S10 and S11 and demonstrate that both the maximum and average pit depth, as well as pit number density, are substantially higher under biotic conditions than in sterile seawater.

Elemental distributions within these surface layers were assessed by EDS mapping on representative cross-sections (Figure 9). In the abiotic condition, iron and oxygen signals were dominant in the corrosion product region, with low chlorine intensity and negligible sulfur. In the biotic condition, EDS maps revealed localized enrichment of sulfur and chlorine near the steel–corrosion product interface and within discrete regions below the outer porous layer, indicating that sulfur- and chloride-bearing species are concentrated close to the actively corroding interface. Oxygen was broadly associated with both the corrosion products and the biofilm matrix. Because EDS cannot detect hydrogen and provides only semi-quantitative estimates of light elements such as oxygen, these results are interpreted qualitatively; no direct identification of specific crystalline phases or hydrogen-containing species is attempted.

Taken together, the morphological and elemental observations support a mechanism in which sulfate-reducing bacterial activity promotes under-deposit and localized corrosion, leading to deeper and more numerous pits and a more complex, layered corrosion product structure than in the abiotic control. These features are fully consistent with the electrochemical signatures of cathodic depolarization and the MIC-induced acceleration of corrosion documented in Section 3.2, Section 3.3, Section 3.4 and Section 3.5.

3.7. Magnetic Barkhausen Noise Response

Magnetic Barkhausen Noise measurements were used to monitor microstructural degradation and stress evolution in HSLA AH36 steel during exposure. The root-mean-square (RMS) amplitude of the MBN signal, MBN_RMS, was recorded in both the rolling-direction (0°) and transverse (90°) magnetization configurations, under identical excitation conditions (Figure 10). In the abiotic medium, MBN_RMS increased moderately from its initial value and then stabilized after 14–28 days, reflecting gradual development of a corrosion product layer and associated changes in near-surface stress and dislocation density, without evidence of abrupt damage accumulation (Figure 10). In contrast, in the biotic medium MBN_RMS exhibited a pronounced peak at day 7 (≈600 mV/mm) followed by a partial decrease at day 14–28, temporally coincident with maximum MIC acceleration and the onset of deep pitting (Figure 10). This transient maximum suggests that MBN responds sensitively to the combined effect of hydrogen uptake, localized plasticity around pits, and microstructural softening during the most aggressive phase of sulfate-reducing bacterial activity.

A quantitative comparison between MBN_RMS and cumulative mass loss demonstrates that MBN provides a consistent, non-destructive indicator of overall degradation. For the abiotic condition, a linear regression describes the relationship between MBN_RMS and cumulative mass loss with high coefficient of determination (R² = 0.99; Figure 11a), indicating that progressive, largely uniform corrosion and associated microstructural changes translate into a nearly proportional increase in the MBN signal.

In the biotic condition, the dependence of MBN_RMS on mass loss is distinctly nonlinear (Figure 11b). Over the limited set of available time points (n = 4), a cubic polynomial was fitted to the data, but this curve should be regarded solely as an empirical fit to guide the eye through the measured values, rather than as evidence of an underlying cubic law. Nonlinear and even non-monotonic correlations between MBN features and the stress–microstructure state have been reported previously in steels subjected to bending deformation and complex residual stress profiles [38,39]. In this context, the key outcome is not the specific polynomial order, but the fact that the biotic system exhibits a qualitatively different trajectory that mirrors the three MIC regimes identified electrochemically (initiation, peak activity, stationary phase), with the highest MBN amplitudes occurring at relatively modest pit depths (20–50 μm), well before extensive section loss is observed. This supports the use of MBN as an early-warning indicator for subsurface damage associated with microbiologically influenced corrosion and hydrogen-assisted degradation.

Directional MBN measurements further clarify the role of hydrogen and localized corrosion on the magnetic response. Under abiotic conditions, the ratio of MBN_RMS measured parallel and perpendicular to the rolling direction remains greater than unity throughout exposure, consistent with the retained texture and stress anisotropy of the rolled plate. Under biotic conditions, this anisotropy ratio progressively decreases toward unity as exposure time increases, indicating that the damage induced by MIC and hydrogen uptake tends to randomize the local stress state and domain-wall pinning landscape. The trend toward an almost isotropic MBN response at later times is consistent with widespread microstructural degradation around pits and within the near-surface volume, rather than being confined to a particular crystallographic or stress direction. Taken together, these results indicate that MBN can (i) distinguish between chemically controlled corrosion and MIC, (ii) track the transition between kinetic regimes identified by electrochemical methods, and (iii) provide a non-destructive signature of early hydrogen-assisted damage in HSLA steel.

3.8. Corrosion Product Phase Composition and MIC Mechanisms

X-ray diffraction (XRD) and grazing-incidence XRD (GI-XRD) were employed to elucidate the phase composition and stratification of corrosion products formed on HSLA AH36 steel under abiotic and biotic conditions. XRD analysis was performed on bulk corrosion products scraped from specimens retrieved at days 3, 7, 14, and 28, while GI-XRD was conducted on polished cross-sections at the same time points to resolve the outer and inner layer composition separately (incidence angles 0.5–3°, penetration depth ~0.1–1 μm). Representative bulk XRD patterns for days 7 and 28 are shown in Figure 12; the complete dataset for all time points is provided in Figures S7 and S8. Phase identification was performed using HighScore Plus software with reference to the ICDD PDF-4+ database [card numbers: lepidocrocite γ-FeOOH (PDF 00-008-0098), goethite α-FeOOH (PDF 00-029-0713), akaganeite β-FeOOH (PDF 00-034-1266), magnetite Fe₃O₄ (PDF 00-019-0629), mackinawite FeS (PDF 00-015-0037), pyrrhotite Fe_1−xS (PDF 00-029-0726), green rust sulfate Fe₆(OH)₁₂SO₄ (PDF 00-047-1775)].

In abiotic artificial seawater, the phase evolution of the corrosion product layer directly explains the electrochemical trends reported in Section 3.2, Section 3.3, Section 3.4 and Section 3.5. At days 3 and 7, bulk XRD is dominated by lepidocrocite (γ-FeOOH), with subordinate contributions from goethite (α-FeOOH), akaganeite (β-FeOOH), and minor magnetite (Fe₃O₄). This lepidocrocite-rich, porous outer layer is consistent with the relatively low R_ct values measured by EIS at day 7 (1810 ± 140 Ω·cm²) and the higher instantaneous corrosion rates from LPR at early exposure (i_corr = 87 ± 11 μA/cm² at day 7), as the poorly protective rust permits continued access of oxygen and chloride to the metal surface. By days 14 and 28, the relative intensity of γ-FeOOH reflections decreases while α-FeOOH and Fe₃O₄ peaks become more pronounced, indicating rust maturation and formation of a denser inner layer. This mineralogical transition directly corresponds to the monotonic increase in R_ct (2180 → 2590 Ω·cm²) and the concurrent decline in corrosion rate (0.98 → 0.91 mm/year from gravimetry), confirming that the development of a goethite/magnetite-rich inner barrier layer is the primary mechanism responsible for the long-term deceleration of abiotic corrosion. GI-XRD confirms this bilayer architecture (Figure 13): outer layers are enriched in γ-FeOOH and β-FeOOH, while inner layers show α-FeOOH and Fe₃O₄ dominance, consistent with the dense, stratified oxide films observed in FE-SEM cross-sections (Table S11, 10–15 μm thickness at day 28) and the shallower pit morphology (<20 μm, Table S10).

In the biotic medium, the XRD phase evolution follows a fundamentally different trajectory that provides direct mineralogical explanation for the sustained MIC acceleration documented in Section 3.3, Section 3.4 and Section 3.5. Already at day 3, broad reflections attributable to mackinawite (FeS) are detected alongside γ-FeOOH and Fe₃O₄, consistent with the onset of H₂S production confirmed by sulfate depletion data (0.3 g/L consumed, Table S2) and the first measurable shift in OCP (−385 mV SCE vs. −305 mV abiotic, Table S4). By days 7–14, mackinawite intensity decreases as pyrrhotite (Fe_1−xS) becomes the dominant sulfide phase. This sulfide transformation is mechanistically significant: it coincides precisely with the period of maximum MIC acceleration (AF = 1.82 at day 7), maximum viable cell count (2.5 × 10⁸ cells/mL), and the sharpest decrease in R_ct (450 → 380 Ω·cm²). Unlike the protective α-FeOOH/Fe₃O₄ assemblage that develops abiotically, iron sulfides are electronically conductive and non-passivating, maintaining an active interface beneath the biofilm and preventing the charge-transfer resistance recovery observed in the abiotic system. By day 28, pyrrhotite dominates the inner layer while green rust phases persist in the outer biofilm-adjacent zone, consistent with the fluctuating redox conditions within the SRB biofilm matrix and the near-complete surface coverage (≥95%, Table S11). An unassigned reflection at ~13.4° 2θ was detected in biotic specimens at days 14–28; its assignment requires additional compositional analysis (e.g., WDS or TEM-EDS) to confirm the presence of minor phases. Critically, this sulfide-rich inner layer correlates spatially with the deep pit regions identified by FE-SEM (150–250 μm maximum depth, Table S10) and the S/Cl enrichment confirmed by EDS at the metal–corrosion product interface (Figure 9), providing direct mineralogical evidence that iron sulfide accumulation drives and sustains localized anodic dissolution.

The GI-XRD stratification in biotic specimens (Figure 13) further clarifies the mechanism of hydrogen uptake and its connection to the MBN response described in Section 3.7. The outer biofilm-adjacent layer, rich in mackinawite and green rust, represents the zone of active H₂S generation via SO₄²⁻ reduction, while the inner pyrrhotite/Fe₃O₄-rich layer, accumulating progressively from day 7 onward, constitutes the region where the reaction Fe⁰ + H₂S → FeS + H₂ produces atomic hydrogen at the metal surface. This inner sulfide accumulation therefore represents the primary source of hydrogen available for subsurface uptake, directly explaining why MBN_RMS peaks at day 7 (612 ± 38 mV/mm) when mackinawite-to-pyrrhotite transformation is most active, and why the MBN anisotropy ratio decreases toward unity (1.27 → 1.15) as the inner sulfide layer thickens and hydrogen-induced microstructural damage becomes spatially distributed. The correlation between sulfide phase stratification, pit depth evolution, and MBN response thus establishes a consistent, multi-technique picture of MIC-driven degradation in which mineralogy, electrochemistry, and non-destructive magnetic signatures are mechanistically linked rather than independently observed.

4. Discussion

The present study elucidates the synergistic degradation mechanisms affecting HSLA AH36 steel under combined mechanical stress, seawater chemistry, and sulfate-reducing bacterial activity [40,41,42,43], with a particular emphasis on the role of corrosion product phase evolution in controlling pitting susceptibility, hydrogen ingress, and non-destructive detectability.

The electrochemical results demonstrate that in abiotic artificial seawater, the steel surface gradually develops a quasi-passivating state characterized by increasing charge-transfer resistance and declining corrosion rates. This behavior is consistent with the progressive formation of a multilayered rust scale, where an initially porous outer layer gives way to a denser inner region, as confirmed by GI-XRD stratification showing γ-FeOOH/β-FeOOH dominance in the outer zone transitioning to protective α-FeOOH/Fe₃O₄ closer to the metal interface. Such rust architectures are well-documented for low-alloy steels in marine immersion and atmospheric exposure, where goethite and magnetite accumulation reduces ionic transport and stabilizes the system against further uniform corrosion [42,44].

In biotic conditions, however, SRB biofilms fundamentally alter this trajectory by establishing localized chemical gradients that promote pitting and sustain high corrosion kinetics [43,45,46]. The GI-XRD evidence of mackinawite and green rust in the outer biofilm-adjacent layer, evolving to pyrrhotite in the inner pit-proximal zone, directly implicates H₂S-mediated mineralogy in MIC acceleration. Mackinawite formation via Fe⁰ + H₂S → FeS + H₂ not only consumes metal but generates atomic hydrogen at the steel surface, which is further facilitated by the cathodic activity of conductive sulfides. This establishes a positive feedback loop: anodic pitting supplies Fe²⁺ for sulfide precipitation, while sulfides enhance local cathodes, deepen pits (up to 250 μm vs. <20 μm abiotic), and promote H-uptake, as evidenced by the early MBN_RMS peak and isotropic signatures [40,42,43].

The gravimetric and pitting data corroborate these mineralogical insights. Abiotic mass loss decelerates as the protective inner α-FeOOH/Fe₃O₄ layer thickens, yielding shallow, uniform pits. Biotic conditions, conversely, exhibit persistent high rates due to non-barrier sulfides that maintain active interfaces beneath biofilms, with EDS-confirmed S/Cl enrichment aligning precisely with the pyrrhotite-rich inner zone where sharp pits initiate and propagate. Biofilm-corrosion interactions amplify this: EPS matrices trap H₂S/acids, creating occluded anodes, while sulfate depletion (86%) fails to halt MIC because sulfide transformation (mackinawite → pyrrhotite) recycles FeS cathodes, decoupling corrosion from bulk sulfate availability—a hallmark of SRB resilience in closed systems [40,44,45].

Similar patterns of localized MIC beneath deposits and biofilms have been reported for carbon steels exposed to SRB or mixed anaerobic consortia, where under-deposit regions act as anodes and biofilm-covered surfaces accumulate FeS and associated sulfide phases. Selective attack of weld zones and base metal in SRB-containing media, exacerbated under organic carbon starvation, has also been demonstrated for high-strength pipeline steels [46,47], reinforcing the view that SRB-driven MIC is strongly localized and highly sensitive to microstructural and environmental heterogeneities.

The sustained MIC acceleration observed here must also be viewed in the broader context of hydrogen uptake and embrittlement in high-strength steels exposed to marine environments. In situ permeation measurements on AISI 4135 steel in different marine zones have shown that diffusible hydrogen content is positively correlated with corrosion rate and is significantly higher in splash and immersion zones than in atmospheric exposure [48]. Other studies have demonstrated that environmental factors such as temperature, chloride deposition, and the presence of calcareous deposits strongly influence hydrogen absorption and subsequent embrittlement of high-strength low-alloy steels in seawater [43,47]. These findings, together with the well-established increase in hydrogen embrittlement susceptibility with strength level and microstructural heterogeneity in high-strength fasteners and marine steels [47,49], suggest that the combination of SRB biofilms, sulfide-rich corrosion products, and applied tensile stress employed in the present study provides favorable conditions for hydrogen ingress and hydrogen-assisted degradation in AH36.

MBN measurements provide a non-destructive window into these subsurface processes. In abiotic conditions, MBN_RMS changes only modestly and remains anisotropic, reflecting incremental modifications to near-surface stress and dislocation density as uniform corrosion removes the grinding-affected layer, a behavior consistent with previous MBN studies on corroded low-alloy steels [50,51,52]. The abiotic linear MBN_RMS–mass loss correlation reflects uniform rust-induced stress/microstructural changes. Under biotic conditions, however, MBN_RMS peaks at day 7 (612 ± 38 mV/mm), correlating nonlinearly with mass loss (R² = 0.99 cubic fit) and reducing anisotropy ratio to ~1.15, indicating isotropic hydrogen-assisted damage around pits. This early sensitivity (pit depths 20–50 μm) positions MBN for predictive maintenance. Such signatures—enhanced MBN with reduced anisotropy—have been associated with the formation of three-dimensional dislocation networks, microcrack precursors, and hydrogen-assisted damage in high-strength steels subjected to combined mechanical loading and corrosive environments [47,48,52]. Biotic nonlinearity and anisotropy collapse signal hydrogen embrittlement around sulfide-accelerated pits, detectable before macroscopic failure. This phase-resolved understanding elevates the work beyond phenomenology, linking mineral evolution directly to H₂S-pitting-H-uptake synergies.

From a non-destructive evaluation standpoint, the present work extends earlier applications of MBN to corrosion and stress monitoring in low-alloy and structural steels. Previous studies have shown that MBN parameters such as RMS amplitude and full width at half maximum can be linked to corrosion extent and stress redistribution in corroded plates, wires, and rebars [50,51,53]. Other investigations have demonstrated that magnetizing conditions and steel microstructure critically affect MBN sensitivity, especially in high-strength low-alloy steels [52,54]. The present results confirm that, with appropriate calibration, MBN can track both uniform corrosion (approximately linear relation to mass loss in abiotic media) and MIC-induced localized degradation (nonlinear response in biotic media), while also providing directional information that helps distinguish between predominantly stress-controlled degradation and more isotropic, hydrogen-assisted damage. This combination of sensitivity and directional discrimination positions MBN as a promising addition to the toolbox of magnetic and electromagnetic NDT methods for corrosion and integrity monitoring [50,53,54].

At the same time, some important limitations of the present study must be acknowledged. First, the interpretation that MBN detects subsurface hydrogen-assisted damage is supported only indirectly, through correlation with electrochemical indicators, MIC acceleration factors, and pit morphology. Unlike dedicated hydrogen permeation or thermal desorption spectroscopy studies [47,48], no direct measurements of hydrogen content, trap occupancy, or fracture surfaces were performed. Second, the experiments were conducted at a single temperature in a controlled artificial seawater with a pure SRB culture, whereas real marine environments exhibit variable temperatures, complex microbial consortia, flow and deposition processes, and intermittent oxygen ingress [44,55]. Third, the mechanical loading was uniaxial and constant, while operational structures experience multiaxial, often cyclic stresses, which strongly influence both hydrogen embrittlement susceptibility and MBN response [47,52]. Finally, the exposure time of 28 days captures biofilm establishment and early maturation, but not multi-year biofilm succession or repair processes that may further modify MIC and hydrogen uptake [46,56].

Limitations include single temperature/stress and indirect H-quantification; future studies should incorporate permeation tests and field validation. Nevertheless, the XRD/GI-XRD validation of sulfide stratification as the driver of biotic pitting and H₂S damage establishes a robust framework for MBN-based predictive monitoring of MIC in maritime HSLA structures.

Despite these limitations, the integration of electrochemical, gravimetric, microstructural, and MBN data provides a coherent mechanistic picture that is consistent with current MIC and hydrogen embrittlement literature. The three-phase MIC scenario inferred here—initiation, peak activity, and stationary sulfide-rich regime—aligns with recent work on SRB biocorrosion under varying carbon source availability and headspace conditions, in which electron uptake from metal can dominate over classical metabolite-mediated corrosion under starvation or under-deposit conditions [42,45,46,56]. The observation that corrosion rates remain high even as sulfate becomes depleted is in agreement with studies showing that sulfide-enriched biofilms and FeS-rich corrosion products can act as long-lived cathodic catalysts, sustaining MIC after active growth has slowed [43,44]. Within this framework, the MBN response—especially its early sensitivity to microstructural changes at modest pit depths—emerges as a particularly valuable indicator for early damage detection in HSLA steels exposed to MIC-prone marine environments. Future work that couples in situ hydrogen monitoring, long-term exposures in natural seawater, and field trials on operational structures will be needed to translate these laboratory-scale insights into robust, MBN-based predictive maintenance strategies.

5. Conclusions

Under multiparametric marine exposure (Desulfovibrio vulgaris biofilms, 80% yield strength tensile stress, 28 days), HSLA AH36 steel exhibited sustained MIC acceleration (factor 1.88) characterized by cathodic depolarization (ΔOCP = −144 mV SCE, Rct ↓4×), deep under-deposit pits (150–250 μm), and sulfide-rich inner rust layers (pyrrhotite/mackinawite via GI-XRD). Magnetic Barkhausen noise (MBNRMS) detected early subsurface hydrogen damage (day 7 peak, 612 mV/mm) with isotropic signatures (anisotropy ratio ~1.15), providing quantitative correlations (R = 0.99) that distinguish hydrogen-assisted MIC from abiotic degradation. This integrated electrochemical–microstructural–NDT framework establishes MBN as a portable tool for real-time structural health monitoring and predictive maintenance of maritime HSLA steels.