The carbon steel coupons, mounted in epoxy resin and attached to a polypropylene fishing line (as described in Section 2.3), were immersed in 250 mL of sterilized test solutions (either Postgate C medium, modified Postgate C medium or Inorganic medium) in a 250 mL Duran bottle. Using aseptic techniques, 1 mL of bacterial culture was taken from a 3-day old stock culture of SRB and inoculated into the Duran bottle as pure culture. The solution was not de-aerated. A layer of paraffin was then placed on the top of the solution, with a view to maintaining anaerobic conditions once these were established over the course of the experiment. Growth of the introduced SRB strain confirmed that anaerobiosis developed within the 72 h period. An abiotic control was also set up under the same conditions. The carbon steel specimen was exposed to the bacterial culture (or abiotic conditions) for a period of 72 h or 168 h at 30°C. At least two samples were observed for each condition in order to check for the reproducibility.

The morphology of the biofilm on the carbon steel specimen was observed using the 3D Quanta FEI in ESEM mode. Biofilms were fixed using 2% glutaraldehyde solution for 1 h and then subjected to 2 washes in deionised water, each wash for 5 min. The use of relatively high pressures (∼650 Pa, or 1 mbar) and water vapor in the specimen chamber in the ESEM allowed the samples to be viewed in a hydrated state, and without the need for metallic coating that is often required in SEM sample preparation which could introduce artifacts to the biofilms. At least two samples were observed under ESEM for each condition.

FIB-SEM technique was used to investigate sub-surface features of the biofilm. FIB milling was carried out using the FEI Quanta 3D FEG instrument. As this instrument also has SEM and EDS capabilities, the subsurface structure could be viewed, and elemental composition analysed following the milling. As the sample was placed in a high vacuum chamber (∼10⁻⁶ Pa), in addition to the sample preparation steps detailed in Section 2.5.2 for ESEM, the sample was also dehydrated, and sputter coated with 3 nm of platinum coating in order to avoid charging. The surface of the specimen was first viewed under SEM to locate a site of interest in the biofilm for subsequent cross-sectional analysis. The site chosen for FIB-milling was typical and representative for the condition under study. A platinum strip, 1 µm in width and 1 µm in thickness, was applied across the length of the area of interest in order to protect the biofilm from ion beam degradation during the milling process (Blankemeier, 2011). The sample was then tilted to 52^o so that the milling could be performed using a gallium (Ga⁺) beam with the current in the range of 5 nA. Beam currents were lowered (to the pA range) for subsequent cleaning of the cross-sections for removing material re-deposited on the area during the initial rough milling.

The specimens exposed to biotic and abiotic conditions were removed after 72 h and any loosely attached bacterial cells were rinsed off with phosphate buffer solution. The specimens were left in a desiccator for at least 24 h to dehydrate. FTIR spectra were obtained using a Perkin Elmer Spectrum 100 series spectrometer. The spectral acquisition (128 co-added scans at 8 cm⁻¹ of spectral resolution in the 4,000 cm⁻¹ to 600 cm⁻¹ range) were obtained by pressing the exposed steel surface to the ATR crystal and spectral processing was done using Perkin Elmer Spectrum Express software. Background scans, consisting of the clean pressure applicator against the ATR crystal, were acquired prior to scanning of each specimen.

Postgate C medium allowed reliable and rapid growth of D. desulfuricans due to its high organic nutrient content. Pits were found sparsely throughout the carbon steel following a 72 h exposure to the SRB (Figures 3A,B). Although earlier studies have reported that pit morphology in the presence of SRB occurred as disc shaped concentration cells (Videla, 1987), the pits observed in this study were irregular-shaped and did not follow a distinctive trend (Figure 3C). Another distinctive feature in the SRB biofilm was the presence of tubercle structures (Figure 3D). When cultured in the nutrient-rich Postgate C medium, bacterial cells were not observed to be attached to the carbon steel surface. They were either well hidden within corrosion product layers or did not attach to the carbon steel due to favourable growth conditions in the bulk solution. In the latter case the iron sulphide deposited on the film would have formed due to the H₂S production by planktonic (free-floating) SRB cells. Xu et al. (Xu et al., 2023) reported that planktonic cell numbers had no impact on MIC pitting attack as they did not utilise electrons from the iron dissolution reaction for the reduction of sulphate and oxidation of organic carbon nutrients. EDS elemental analysis detects carbon, oxygen, sulphur, phosphorous, silicon, manganese and iron on 72 h SRB biofilm on carbon steel (Figure S1 in supplementary document).

Figure 3E shows the FTIR spectra of SRB biofilm on carbon steel exposed to biotic solution and the film formed during exposure to abiotic medium for a period of 72 h. A much stronger IR signal could be obtained for the SRB biofilm as compared to the abiotic Postgate C medium. The presence of organic compounds in the film can be readily identified by the absorption bands between 2,956–2,852 cm⁻¹, accounting for the CH₃ and CH₂ functional groups from the aliphatic chains in lipids (Coates, 2000; Ramesh et al., 2006). Complementary bands supporting the presence of C-H peaks were found in the region between 1,462 and 1,378 cm⁻¹, where bending vibrations of the -CH₃ and -CH₂ groups occur (Naumann and John WileySons, 2006; Ojeda et al., 2009). A broad absorption band near 3,320 cm⁻¹ in the SRB biofilm in Postgate C arose from the stretching vibration of hydroxyl groups of structurally bound water in iron oxyhydroxides, or possibly proteins (Ramesh et al., 2006). This broad peak was most prominent in the pure SRB culture biofilm than any other biofilm investigated in this study. Structurally bound water can occur during rapid precipitation of iron oxyhydroxides where chemisorbed/dissociated water molecules at the mineral surface may be trapped within the crystalline mineral. This leaves excess hydroxyl surrounding the lattice site where Fe²⁺ ion would otherwise be found. When temperatures are sufficiently low to impede the diffusion of H or OH out of the material, this excess hydroxyl level can increase and be stabilised through structural phase transformation (Pinney and Morgan, 2013).

The N-H stretch bands of amides occur in the region 3,300–3,000 cm⁻¹ and overlap with the broad peak of OH stretching. Therefore, the region between 1800 and 1,500 cm⁻¹ were used to identify proteins as it is dominated by conformation sensitive amide I and amide II bands (Naumann and John WileySons, 2006). The exact frequency of amide band vibrations is determined by the secondary structure of the protein (Meyers, 2004). The amide I (C=O) band of bacterial proteins (Dowling et al., 1987) was detected in the biofilm formed by the pure SRB culture in Postgate C medium at 1,638 cm⁻¹ as well as the amino acid side chains at 1,510 cm⁻¹ (Al-Saadi et al., 2021b) (however, this could also be due to the structurally adsorbed water, as discussed earlier). The PO₄ ^2- symmetric stretching vibration of phosphates was detected in the region (∼1,085 cm⁻¹). Vibrations due to the carbohydrate backbone also occurred in the region between 1,184 cm⁻¹ to 1,002 cm⁻¹. The “fingerprint region” refers to the spectral range 900 to 600 cm⁻¹ and contains weakly expressed bands arising from the aromatic ring vibrations of nucleotides (Naumann and John WileySons, 2006) and the CH₂ rocking vibration of lipids (Coates, 2000; Socrates, 2000; Meyers, 2004; Naumann and John WileySons, 2006).

Cross sectional analysis using FIB-SEM (Figure 4) shows the corroded surface under the SRB biofilm. Pitting appeared to be initiated under the SRB biofilm at the areas shown in Figures 4B,C. Both pit mouths in the cross section appeared to have two layers of biofilm; one layer covered the pit mouth while the other covered the whole. At the 72 h stage, pits had not yet progressed to become deep or well defined. The biofilm contained cracks and some porosity was observed adjacent to areas of compact film. Pores observed in Figures 4B,C would better allow aggressive ions such as S⁻² and Cl⁻ in the electrolyte to come in contact with the bare metal, and hence, would be expected to enhance pitting corrosion. EDS analysis in Figure S2 shows the presence of oxygen and sulphur, suggesting that iron oxyhydroxides and iron sulphide were present in the material covering the pit following 72 h immersion.

The term tubercle is defined as a small-rounded prominence and has been used previously to describe corrosion product deposits (Ray et al., 2010). The tubercle formed in the SRB biofilm Figure 4D had a build-up of biological material within the core intertwined with corrosion products. It is widely reported in literature that the area under a tubercle is deprived of oxygen and acts as an anode to the surrounding cathodic area, which leads to pitting corrosion of carbon steel (Little et al., 2010; Ray et al., 2010). Localised pitting corrosion, however, had not occurred under the deposit in Figure 4D due to the initial iron oxide layer remaining intact. Therefore, FIB-SEM technique has clearly shown that the presence of tubercles on carbon steel does not always lead to pitting corrosion underneath. The fundamental mechanisms of SRB-influenced deterioration of steel have been based on the activity of hydrogenase enzymes and/or the existence of iron sulphide species (Beech, 2003). The capacity of SRB to produce and utilize H₂ suggests that these microorganisms possess hydrogenase enzyme, which catalyses the reversible reaction in Eq. (2) (Pankhania, 1988; Beech, 2003). 2 H + + 2 e − = 2 H = H 2 (2)

The detection of hydrogenase in the biofilms can be attributed to the shift from chemo-organic to chemo-lithotrophic metabolism by some organisms present in the mixed biofilm inhabitants as a result of the low levels of consumed carbon sources; lactate or acetate remained in the recycling medium. The induction of hydrogenase activity was probably due to the ability of those organisms to use cathodically produced hydrogen as an alternative energy source (Bryant et al., 1991). The most relevant effects of SRB metabolites on corrosion (Videla, 2000) would include providing several sulphur compounds that constitute corrosive anions for steel and iron, and producing aggressive final metabolic products (sulfides, bisulfides, hydrogen sulfides) or intermediate metabolites such as thiosulfate and polythionates. Moreover, through biological and inorganic transformation paths, SRB metabolites change to sulfide anions that are corrosive to iron and steel. The characteristics and intensity of sulfide action are closely connected to the nature of the pre-existing passive film on the metal surface (Videla et al., 2004). Other ions (e.g., chlorides) in the medium enhance the participation of sulfide in the corrosion process (Videla and Characklis, 1992).

Figures 5A,B depict the ESEM micrograph of carbon steel exposed to SRB for 168 h. Extensive pitting (with a shallow pit morphology, as seen in Figure 5C) was observed. The biofilm varied in thickness with areas of high thickness containing porous deposits observed alongside tightly adherent film. The detection of sulphur in the SRB biofilm, by EDS analysis (Figure S3), suggested that over time, sulphide ions were incorporated into SRB biofilms which were initially rich in iron oxides and/or iron hydroxides. The highly nutritious conditions in the bulk media meant that the bacteria within the biofilm did not pre-require attachement to the metal surface. EDS also detected other elements present in the bacterial cell body (viz; C, O, S, and P) (Li et al., 2020). The cross-section of the 168 h SRB biofilm in Postgate C medium (Figure 5C) showed the porosity in film was greater as compared to the film observed at 72 h.

Electrochemical information on the evolving biofilm formation was monitored non-destructively using the linear polarisation technique. The polarisation resistance (R_p) monitored over a 72 h period is shown in Figure 6. The standard deviation of the R_p data for the SRB biofilm was significantly greater than that for the abiotic film. Subsequent to the 72-h reading, R_p oscillated considerably and reproducible data could not be obtained. R_p data obtained from linear polarisation technique should only be converted into corrosion rate if uniform corrosion occurs, the solution is highly conductive and no scale/deposits are present on the surface (Hilbert et al., 2007). As can be observed from earlier ESEM analysis (Section 3.2.1), the formation of a heterogeneous biofilm, developed over 72 h exposure and thereafter, meant that the LPR technique could only be used qualitatively for this study. Hence, the increase in R_p can be taken as a sign of an apparent passive film forming on the surface of the carbon steel which offered mass transfer resistance to active anodic dissolution. Newman et al. (Newman et al., 1992) observed that in high sulphide environments, low alloy carbon steel passivated to give in R_p values over 50,000 Ω/cm².

The potentiodynmic polarisation curve for carbon steel following 72 h exposure to SRB environment is shown in Figure 7A. The corrosion potential (E_corr) shifted to more positive potentials for the carbon steel coupon exposed to common Postgate C solution with SRB compared to steel immersed in abiotic solution. The corrosion current density (I_corr) of carbon steel in abiotic solution was more than 6x higher than that of steel in biotic solution. The improvement in corrosion resistance of carbon steel in biotic solution is attributed to the development of protective FeS film on carbon steel surface, which is consistent with the reported literature (Al-Saadi et al., 2017; Al-Saadi and Raman, 2019). However, consequent to application of an anodic scan the film was disrupted at around −0.6 V_SCE and active corrosion followed thereafter.

In contrast, the anodic curve of the potentiodynamic polarisation scan following 168 h exposure (Figure 7B) showed resistance to the anodic dissolution when the corrosion potential was increased positive of −550 mV. Due to the adherent film of iron sulphides which acted as a diffusive barrier to reactants, the anodic current was lower at higher overpotentials under the SRB environment as compared to the abiotic condition. The mechanism of development of the robust iron sulphide film upon application of anodic overpotential is well-covered in the literature (Al-Saadi et al., 2017; Al-Saadi and Raman, 2019; Al-Saadi et al., 2021a).

When cultured in modified Postgate C medium (Postgate C, without the organic nutrient yeast extract), the biofilms with much greater porosity content were observed (Figure 8) and a significantly greater number of bacterial cells were observed on the metal surface following 72 h exposure as seen in Figure 8B. Pitting and surface roughness are apparent throughout the biofilm. EDS (Figure S4) was taken on a colony of bacteria showing a high carbon content, confirming the presence of biological cells and EPS present in the biofilm.

In contrast to the SRB biofilm developed in Postgate C medium, the relatively slower growth of the SRB in the modified Postgate C medium resulted in thinner film and hence weak signals for bio-molecules in the IR spectra (Figure 8C). Typical thickness of this SRB biofilm was ≈500 nm as seen in Figure 8D. Ideal thickness for the ATR-FTIR is ≥1 µm. However, ATR-FTIR analysis in Figure 8C was able to discern weak signals enabling some information regarding the IR active functionality of the biofilm to be investigated. The IR signal obtained for the SRB biofilm in modified Postgate C medium showed the presence of hydrocarbons, carbohydrates, phosphates and amide groups due to proteins.

Cross-sectional images show the presence of initial porous layer of corrosion product layer which was primarily iron oxyhydroxides (Figure 8D). Videla has reported that 48–72 h biofilms formed in laboratory conditions generally contain spherules of colloidal goethite and crystalline hematite (Videla, 2001). Magnetite (Fe₃O₄) and ferrous hydroxide (Fe(OH)₂) have also been observed in previous studies (Lee et al., 1993). Iron oxyhydroxides is reported to be protective in stable films on iron (Foroulis, 1980; Jones, 1996; Revie, 2008; Sato and Sharma, 2012). FIB-SEM analysis in Figure 8D clearly shows that pitting corrosion occurred at locations of breakage in the iron oxyhydroxide film which allowed the chemical species to reach the carbon steel and cause corrosion. The breakdown of passive films is a complex problem and currently there is no single accepted mechanism. The passive iron oxide film may have broken down due to penetration (Strehblow and Marcus, 2002) or adsorption (Szklarska-Smialowska, 1986; Strehblow and Marcus, 2002) of aggressive anions from the media (such as sulphide and chloride) into the film or as a result of direct access of these ions to the metal substrate due to disruptions in the passive film (Strehblow and Marcus, 2002; Sato and Sharma, 2012). Previous studies have reported mackinawite (FeS_1-x) to be the first iron sulphide to form (King and Miller, 1977; Yuan et al., 2013) and considered responsible for breakdown in passivity when exposed to SRB (Videla, 1987).

Figure 9 shows the ESEM micrographs of carbon steel exposed for 168 h to the modified biotic Postgate C solution (Postgate C without yeast). The ESEM image in Figure 9A shows extensive pitting following 168 h exposure. Surface coverage was greater in this case (Figure 9B) than for the 72 h biofilm (Figure 8B). EDS results in Figure S5 shows high sulphur and iron contents, suggesting iron sulphide incorporation into the SRB biofilm. Rapid H₂S formation occurred when SRB was cultured in nutrient rich Postgate C medium; however, EDS analysis consistently showed a higher ratio of sulphur incorporated into the 168 h SRB biofilm in the modified Postgate C medium which contained fewer organic nutrients. This may be due to attached bacterial cells as observed in the biofilm formed in nutrient-deficient Postgate C medium, but not observed when cultured in the Postgate C medium. The attached bacterial cells were probably able to concentrate H₂S into the biofilm more effectively than planktonic cells. It is widely reported that once the iron sulphide species make contact with carbon steel, galvanic corrosion occurs where the iron sulphide behaves as the cathode while the steel becomes the anode (Lee et al., 1993; Hilbert et al., 2007). Factors causing this galvanic corrosion include nobler potentials of iron sulphides, their semi-conducting properties, low overpotential for hydrogen reduction and common defects in the iron sulphide crystal structure (Hilbert et al., 2007). Heterogeneous thickness of the biofilm can be observed in Figure 10 as well. Some areas of the film (not shown) were as thick as 10 µm and were rich in elemental sulphur. Substructure features of this biofilm include hollow spaces and shrinkage-cracking which would allow contact with ions in solution.

Change in OCP with time is shown in Figure 12A. From 0 h to 96 h, the E_corr was relatively stable (∼-726 mV). SRB growth was likely to be at exponential phase at around 72 h, as indicated by earlier results presented in Figure 1. However, E_corr was observed to increase suddenly after 96 h exposure exclusively to SRB containing medium, caused by iron sulphide precipitation. At 92 h, strong H₂S odour and heavy blackening of the solution was observed. The next E_corr reading at 120 h, after the blackening observed at 96 h, resulted in a significant increase in E_corr to −628 mV. This positive shift in E_corr of about 100 mV from the reading at 0 h may be due to the precipitation of iron sulphide on the carbon steel (Hack, 1988). Oscillations of E_corr were observed more commonly when taking OCP readings for SRB biofilm as compared to the abiotic control. This was an indication of localised corrosion in the presence of SRB, due to local rupture of the biofilm and/or transpassive dissolution (Fonseca et al., 1998).

The linear polarisation resistance measurement over the 168 h period is shown in Figure 11B. While the E_corr remained stable up to 96 h (Figure 11A), the corresponding R_p increased from 702 Ω.cm² at 0 h to 2,279 Ω.cm² at 5 h due to rapid precipitation of an initial iron sulphide film in the presence of the SRB. From 5 h to 96 h the R_p increased further as sulphides became increasingly incorporated into the biofilm. However, from 120 h to 168 h, where an increase in E_corr was observed through OCP measurement, the R_p decreased showing that the film was increasingly facilitating corrosion over time as more H₂S was accumulated in the batch culture.

The potentiodynamic polarisation scan for the carbon steel following 72 h exposure to SRB in modified Postgate C medium (without yeast) showed active anodic and cathodic behaviour (Figure 12A) due to the lack of a protective passive SRB biofilm. There was no significant difference between the anodic and cathodic regime of the carbon steel exposed to abiotic control or the SRB containing medium. This result was consistent with the OCP results at 72 h (Figure 11A) where E_corr readings were also similar for both conditions. Although R_p was higher at 72 h in the SRB environment (Figure 11B), such resistance to anodic dissolution was not observed in the potentiodynamic polarisation results, since iron sulphide precipitates were insufficient in concentration to form a FeS layer on carbon steel upon application of anodic potential scan. However, over time, the concentration of H₂S and concurrent iron sulphide formation could be observed to increase, as confirmed by the blackening of the test solution (Figure 1) and H₂S odour. As in the case with the biofilms formed in Postgate C medium, the biofilm formed in modified Postgate C resisted uniform anodic dissolution during following 168 h (Figure 12B). The decrease in anodic rate occurred due to the precipitation of iron sulphide layer at high overpotentials which imposed mass transfer limitations to further anodic dissolution of the carbon steel. Pitting potential (E_p) could be observed at −450 mV. Unlike the biofilms formed under nutritionally rich Postgate C medium, the cathodic corrosion current density was higher than the abiotic control. Bacteria may influence the cathodic reactions by consumption or production of cathodic reactants such as oxygen and hydrogen (Videla, 2001). Tang et al. linked the increased H₂S formation by the SRB to promotion of the cathodic hydrogen evolution reaction on carbon steel (Tang et al., 2010). Increased bacterial cell attachment in the modified Postgate C medium as compared to the Postgate C medium may have had a role in concentrating the H₂S close to the metal surface and thus accelerating cathodic processes.

The modified Postgate C medium contained sufficient nutrients for the growth of SRB. However, limiting the organic nutrient content available for the bacteria by removing yeast extract slowed growth and hence, slowed the iron sulphide precipitation on carbon steel. The lower organic nutrient in the media aided attraction of bacterial cells to the carbon steel surface and hence significantly increased attachment to the alloy steel surface.

SRB cells can be seen on the surface of the specimen in Figure 13A. However, due to insufficient nutrients in the inorganic medium, the bacterial cells were smaller in size (dwarf cells). In this state, the bacteria are not killed, but remain dormant until conditions become favourable (Oliver, 1993). A higher number of pits were found in the carbon steel exposed to SRB in the inorganic medium than in the other two culture media at 72 h. However, these pits were shallower than the pits formed under more nutrient sufficient conditions. Cracks in the biofilm which can likely lead to pitting corrosion were also observed (Figure 13B). EDS elemental analysis of this biofilm in Figure S6, corresponding to the area scattered with bacterial cells, detected high concentration of oxygen as well as phosphorous and inorganic magnesium.

The cross-section of a representative area of the biofilm (Figure 14A) shows a rough, corroded steel surface. The backscatter image in Figure 14B of the area seen in Figure 14A shows that the first corrosion product layer has a similar contrast to that of the bare metal. EDS analysis in Figure S7 showed that this film contains high levels of oxygen, sulphur and iron probably as a combination of iron sulphide and iron oxyhydroxide corrosion products. Shoesmith et al. showed that the H₂S absorbs at active oxide sites and may displace OH from the iron oxyhydroxide due to the sulphide ion’s greater polarisability (Sherar, 2011).

Figures 15A,B depict ESEM images of hydrated biofilm developed on carbon steel exposed for 168 h to inorganic Postgate C solution with pure SRB. High concentration of inorganic precipitates as well as biological material was found on the surface of the coupon. Bacterial cells also attached to the surface as a coping strategy for unfavourable conditions. The shallow pits observed at 72 h exposure in the inorganic medium had become covered by corrosion deposits and therefore could no longer be seen. Cross-sectional analysis of the 168 h SRB biofilm (Figure 15C) did not show well-defined deep pitting as observed under more nutritionally favourable conditions. The lower ability of the SRB biofilm to attack carbon steel may be due to the weakened enzymatic activity of the starved SRB cells (Xu et al., 2023) under the condition.

Figures 16A,B show the potentiodynamic polarization plots of carbon steel exposed for 72 and 168 h to inorganic Postgate C solution with and without SRB. It can be seen that upon 72 h of exposure, there was very little resistance to anodic dissolution and, actually, corrosion rate was accelerated compared with the abiotic control. The low sulphide level at 72 h also appeared to have accelerated anodic process in the presence of SRB than that observed under abiotic conditions (Figure 16A). In 168 h, the relatively low accumulation of sulphide produced by the SRB as well as biofilm containing bacterial cells and EPS material offered some resistance to uniform corrosion as seen by the slightly lower anodic current (Figure 16B).