**1. Introduction**

**\***

Natural seawater is a biologically active medium, and it is generally acknowledged that microorganisms influence the corrosion processes of any metal or alloy immersed in a marine environment. In the case of carbon and low alloy steels, the role of sulfate-reducing bacteria (SRB) has long been recognized [1]. This has led to the hypothesis that the longterm corrosion process of carbon steel in seawater could be controlled, at least partially, by the rate of external nutrient supply that governs bacterial activity [2,3]. More generally, it has been shown that the organic molecules released by bacteria somehow influence the behavior of metals. For instance, it has been established that in some cases, extracellular polymeric substances favor the corrosion of carbon steel [4]. Conversely, it was proposed that biofilms could form along with the corrosion products, creating a protective barrier and thus decreasing the corrosion rate [5]. Enzymes, and in particular, hydrogenases that can be present in solution (i.e., out of bacterial cells), are also known to be involved [6,7].

The various molecules associated with the presence and/or activity of microorganisms can interact with the metal itself, but also with its corrosion products. One of the consequences of the activity of SRB is the formation of iron sulfide (FeS), which precipitates from Fe2+aq ions resulting from corrosion and the sulfide species produced by SRB. For this reason, FeS was identified as another major component of the corrosion product layer formed on carbon steel coupons after just 6–12 months of immersion in the water of a

**Citation:** Duboscq, J.; Vincent, J.; Jeannin, M.; Sabot, R.; Lanneluc, I.; Sablé, S.; Refait, P. Influence of Organic Matter/Bacteria on the Formation and Transformation of Sulfate Green Rust. *Corros. Mater. Degrad.* **2022**, *3*, 1–16. https:// doi.org/10.3390/cmd3010001

Academic Editors: Scott Wade and Markus Valtiner

Received: 4 October 2021 Accepted: 22 December 2021 Published: 30 December 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

seaport [8]. It could be detected locally after only 1 month [9] of immersion, and was associated with SRB in the first de-aerated regions formed inside the corrosion product layer. Due to its electronic conductivity and low overvoltage for water reduction, FeS can act as a cathodic site and promote the formation of galvanic cells [10]. Additionally, it can facilitate the influence of electroactive SRB [11].

Most studies dealing with microbiologically influenced corrosion (MIC) are focused on the interactions between metal and bacteria or between metal and species released or produced by bacteria [1]. The interactions between these species and the corrosion products, which may modify the properties of the corrosion product layer and thus the corrosion rate, have only been rarely addressed [12,13]. In a recent work, some differences were observed between the corrosion product layers obtained in artificial and natural seawater [14]. These differences were not only related to FeS, which can only form in natural environments as it requires the presence and activity of SRB, but also to sulfate green rust, GR(SO4 <sup>2</sup>−), which seemed to be favored in natural environments [14].

GR(SO4 <sup>2</sup>−) is one of the main corrosion products forming on carbon steel surfaces which are permanently immersed in natural seawater [5,8,9,15–18]; it is actually the first solid obtained from dissolved Fe species [9,15,17]. It is a mixed valence Fe(II-III) double layered hydroxide with chemical formula FeII4FeIII2(OH)12SO4·8H2O [19,20], and contains mainly (67%) Fe(II) ions. It is oxidized by dissolved O2, a process that leads to the formation of Fe(III)-oxyhydroxides such as goethite ( α-FeOOH) and lepidocrocite ( γ-FeOOH), and oxides such as magnetite Fe3O4 [21–23]. Consequently, any species that can affect the formation and transformation of GR(SO4 <sup>2</sup>−) may have an indirect influence on the corrosion process.

The present study is focused on the effect of bacteria in general (i.e., bacteria not known to induce MIC processes) on the formation of GR(SO4 <sup>2</sup>−) and the evolution of this compound in anoxic conditions. The idea was not to address the interactions between GR(SO4 <sup>2</sup>−) and the bacteria involved in the redox cycles of Fe and S, which have already been documented [12,13,24,25], but to consider bacteria having no specific link with Fe and S, the characteristic elements of GR(SO4 <sup>2</sup>−). This study aims to contribute to a better understanding of the interactions between complex bacterial communities and the overall corrosion system through the determination of possible interactions of selected bacteria with GR(SO4 <sup>2</sup>−).

First, corrosion experiments were performed in artificial and natural seawater to study the composition of the corrosion product layers in each case. These layers were characterized by X-ray diffraction (XRD) and μ-Raman spectroscopy (μ-RS). Secondly, GR(SO4 <sup>2</sup>−) was precipitated by mixing a solution of Fe(II) and Fe(III) salts with a solution of NaOH in the presence of three bacterial strains. These strains were isolated from biofilms previously formed on carbon steel coupons immersed in natural seawater [9]. The precipitates were analyzed by XRD after 1 week and 2 months of ageing at room temperature. The bacterial growth during the ageing of the precipitate was also investigated. Finally, GR was precipitated in the presence of sodium acetate to compare the effects of a small organic anion with those observed in the presence of bacteria.

#### **2. Materials and Methods**

#### *2.1. Preliminary Corrosion Experiments*

To compare the composition of the corrosion product layers formed on carbon steel in artificial and natural seawater, various S355GP steel coupons (5 cm × 5 cm × 1 cm) were exposed for 6 months in both kinds of environments. The nominal composition of this steel grade, commonly used for sea harbor sheet piles, is in wt.%: C ≤ 0.27, Mn ≤ 1.7, S ≤ 0.055, Si ≤ 0.6, Al ≤ 0.02 and Fe for the rest. The coupons were embedded in epoxy resin so that only one side (active area of 25 cm2) was exposed to seawater. The surface of this side was previously shot blasted (Sa 2.5, angular shot) and degreased with acetone.

First, three coupons were immersed in natural seawater for 6 months in the Minimes harbor, the marina of La Rochelle (Atlantic Ocean), using the experimental platform of the

LaSIE laboratory. The coupons were immersed vertically at a depth of ~20 cm (measured from the upper edge of the coupons). As the experimental platform floats on the sea surface, the immersion depth is constant. The temperature of the water close to the coupons was measured regularly. It varied from 9 ± 2 ◦C at the beginning (February) to 20 ± 2 ◦C at the end (July) of the experiment. Secondly, three other coupons were exposed to stagnant ASTM D1141 artificial seawater [26] in 10 L tanks. The seawater was renewed after 15 days, and monthly afterwards. The tanks were set in an unheated room so that the average temperature of the water increased from 12 ◦C (February) to 25 ◦C (July) during the experiment.

At the end of the experiment, the coupons immersed in the Minimes harbor were carried to the lab in a tank full of natural seawater sampled in situ with the coupons. Then, all coupons were removed from the seawater (artificial or natural) and rapidly transferred to a freezer and stored at −24 ◦C before analysis. With this procedure, already used in previous works [9,16,17], the samples can be analyzed in a frozen state so that the corrosion product layers, that contain a lot of water, can be easily handled. Due to the complexity of the corrosion product layers forming on steel in seawater [8,9,15–18], two methods were used to identify the corrosion products, namely μ-Raman spectroscopy (μ-RS) and X-ray diffraction (XRD). For each type of seawater, one coupon was analyzed by μ-RS and another by XRD.

μ-RS analysis was carried out at room temperature using a Horiba Raman spectrometer (LabRam HR evo, Horiba, Tokyo, Japan) equipped with a confocal microscope and a Peltierbased cooled charge coupled device (CCD) detector. A solid-state diode pumped green laser (wavelength = 532 nm) was used with laser power reduced to 10% (0.6 mW) of the maximum to prevent the transformation of the analyzed compounds into hematite α-Fe2O3. This transformation can take place due to an excessive heating [27,28]. The acquisition time depended on the nature of the analyzed phase, and thus, varied from 60 s to 2 min. At least 20 zones (diameter of 3–6 μm) were analyzed on the same sample using a 50× long working distance objective. The analysis was achieved without specific protection from air because the time required for analysis was short. Additionally, the samples remained wet during the procedure, which minimized oxidation.

For the XRD analysis, the whole corrosion product layer was scraped from the surface of the coupon and ground in a mortar (it was initially solid, as it was frozen) until a homogenous wet paste was obtained. This paste was then analyzed as described in Section 2.5.

#### *2.2. Preparation of Green Rust Precipitates*

Green rust compounds can be precipitated by mixing a solution of Fe(II) and Fe(III) salts with NaOH solution [19]. Based on this, a method was developed to prepare GR(SO4<sup>2</sup>−) under conditions simulating seawater, i.e., using Fe(II) and Fe(III) chlorides and adding sodium chloride and sodium sulfate to obtain a suspension with overall chloride and sulfate concentrations similar to those typical of seawater [23,29]. In the present study, this method was used once again. The concentrations of reactants are listed in Table 1, together with the distribution of the reactants in the two prepared solutions.

To obtain the GR precipitate, solution 1 (100 mL) was added to solution 2 (100 mL) and the overall 200 mL of suspension was vigorously stirred for 30 s at room temperature (RT = 21 ± 1 ◦C). After stirring, the suspension was poured into a flask, filled to the rim. The flask was hermetically sealed to avoid any oxidation by air of the precipitates during ageing periods of 1 week and 2 months. The aged precipitates were finally filtered for analysis by XRD. They were sheltered from air with a plastic membrane during filtration to avoid the oxidation of the obtained GR compounds.


**Table 1.** Concentrations of reactants used to prepare the initial green rust precipitate (mol <sup>L</sup>−1), expressed with respect to the total volume of solution (200 mL = solution 1 + solution 2), and considered bacterial strains.

1 See text for the bacterial concentrations.

The NaOH, FeCl2·4H2O and FeCl3·6H2O concentrations used here correspond to the stoichiometry of the precipitation of the sulfate GR. This reaction can be written as:

$$\text{-4Fe}^{2+} + 2\text{Fe}^{3+} + 12\text{OH}^- + \text{SO}\_4^{2-} + 8\text{H}\_2\text{O} \rightarrow \text{Fe}^{\text{II}}\_4 \cdot \text{Fe}^{\text{III}}\_2 \text{ (OH)}\\ \text{\_2SO}\_4\cdot 8\text{H}\_2\text{O} \qquad \text{(1)}$$

#### *2.3. Bacterial Strains and Culture Conditions*

Considering that the influence of microorganisms/organic matter may depend significantly on the bacterial species present, three different bacterial strains (belonging to different families of bacteria) were considered: *Pseudoalteromonas* IIIA004, *Micrococcus* IVA008 and *Bacillus* IVA016. They were previously isolated from the biofilm covering carbon steel coupons immersed for 1 week (*Pseudoalteromonas*) or 2 weeks (*Micrococcus* and *Bacillus*) in natural seawater (La Rochelle marina, Atlantic Ocean) [9]. Each strain was previously identified by sequencing the 16S rRNA gene (accession numbers in the GenBank database: KJ814569 for *Pseudoalteromonas* IIIA004, KJ814564 for *Micrococcus* IVA008 and KJ814540 for *Bacillus* IVA016) [9]). These bacteria do not belong to the families of bacteria classically described as SRB, iron oxidizing bacteria (IOB) or iron reducing bacteria (IRB). The three considered strains were cultured in aerobic conditions, and consequently, were not SRB that can only grow in anaerobic conditions (or in an environment with a low oxygen concentration).

For bacterial growth, the culture medium used, called Marine Broth, was composed of ammonium nitrate 0.0016 g <sup>L</sup>−1, anhydrous magnesium chloride 8.8 g <sup>L</sup>−1, bacteriological peptone 5 g <sup>L</sup>−1, boric acid 0.022 g <sup>L</sup>−1, anhydrous calcium chloride 1.8 g <sup>L</sup>−1, disodium phosphate 0.008 g <sup>L</sup>−1, potassium bromide 0.08 g <sup>L</sup>−1, potassium chloride 0.55 g <sup>L</sup>−1, sodium bicarbonate 0.16 g <sup>L</sup>−1, sodium chloride 19.4 g <sup>L</sup>−1, sodium fluoride 0.0024 g <sup>L</sup>−1, sodium silicate 0.004 g <sup>L</sup>−1, sodium sulfate 3.24 g <sup>L</sup>−1, strontium chloride 0.034 g <sup>L</sup>−1, yeas<sup>t</sup> extract1gL−<sup>1</sup> and ferric citrate 0.1 g L−1. The culture medium was sterilized for 20 min at 115 ◦C.

A concentrated suspension (5 mL) of the three bacteria was first prepared. For each strain, 200 mL of Marine Broth was inoculated with bacteria at 2%, from an overnight culture in Marine Broth, and incubated at 30 ◦C under constant stirring (orbital shaker, 160 rpm). After 24 h of incubation, all three bacterial suspensions were centrifuged for 20 min at 5000× *g*. The centrifugation pellet was finally set again in suspension in 5 mL of ASTM D1141 artificial seawater [26]. The final suspension of bacteria was then added to NaOH solution 1 (see Table 1).

#### *2.4. Numeration of Bacteria*

Quantification of bacteria was performed at four stages of the process: (1) after 24 h of growth independently for each strain, (2) once the bacteria had been concentrated, (3) right after mixing solutions 1 and 2, i.e., right after the formation of the GR precipitate, and (4) after the 1-week ageing period at RT (see Section 2.2) in a hermetically sealed flask. For aged samples, because of decantation, the solid phase(s) settled at the bottom of the flask. Consequently, the supernatant liquid phase and the decanted precipitate (solid phase) could be sampled and analyzed separately.

In each case, a 100 μL sample was prepared by serial dilutions (10−<sup>1</sup> to <sup>10</sup>−6) of the cell suspension in artificial seawater and inoculated on a solid culture medium composed of Marine Broth with 1.2% ( *w*/*v*) agar. After incubation at 30 ◦C in aerobic conditions, bacterial growth was evaluated by counting the number of colony forming units (CFU) (three replicates). The results are expressed in CFU mL−1.

#### *2.5. XRD Analysis of the Precipitates*

XRD analysis was performed with an Inel EQUINOX 6000 diffractometer (Thermo Fisher Scientific, Waltham, MA, USA) equipped with a CPS 590 detector that detects the diffracted photons simultaneously on a 2*θ* range of 90◦. Co-K α radiation (λ = 0.17903 nm) was used at 40 kV and 40 mA, with the XRD analysis being performed at RT with a constant angle of incidence (5 degrees) for 45 min. To prevent the oxidation of GR compounds during preparation and analysis, the wet paste obtained after filtration of the sample was mixed with a few drops of glycerol. With this procedure, the GR particles were coated with glycerol and sheltered from the oxidizing action of O2 [30]. The angular scale was calibrated using the diffraction peaks of magnetite (if present).

The crystalline phases were identified via the ICDD-JCPDS (International Center for Diffraction Data—Joint Committee on Powder Diffraction Standards) database, and the peaks indexed according to the corresponding file. The parameters of the diffraction peaks, i.e., interplanar distance, intensity and full width at half maximum, were determined via a computer fitting of the experimental diffraction patterns. The diffraction peaks were fitted with pseudo-Voigt functions to take into account the evolution of the peak profile with increasing diffraction angle. The fitting procedure was achieved using the OriginPro 2016 software (OriginLab).

μ-Raman spectroscopy was not used for the characterization of the precipitates because (i) this method is not suitable to distinguish between the various GR compounds [31], and (ii) the bacterial cells and associated organic matter mixed with the solid phases induce an important fluorescence phenomenon that makes it difficult to acquire usable data.
