**1. Introduction**

The corrosion of metals immersed in the marine environment is a well-known issue, which can also be accelerated by the activities of a diverse range of microbes in a process known as microbially influenced corrosion (MIC) [1,2]. An example of MIC is the rapid failure of metal structures, typically in the form of localised corrosion around low tide level, which is referred to as accelerated low water corrosion (ALWC) [3,4]. ALWC has been reported on steel structures (e.g., sheet piling) in many ports and harbors worldwide. The reduction in thickness of steel pilings, which can be up to several mm/yr, is one of indications used to diagnose ALWC [5]. In the UK alone, the short-term cost of ALWC was estimated to be ~£250 million [3]. Methods to minimise ALWC include the application of cathodic protection, coatings and the use of special grade steel. Further work, however, is required in order to obtain a better understanding of the complex roles and types of microbes involved in ALWC.

Biofilm formation is a biological process that takes place in humid/aquatic environments, and involves a series of developmental steps. In the natural environment these intricate microbial structures typically contain a diverse range of microbes with complex interactions [6,7]. In relation to ALWC, different microbes are likely to play specific roles in the overall corrosion process, including the development of a biofilm on the steel structure, nutrient provision/cycling, as well as producing and maintaining an anaerobic environment. Specific microbes are responsible for the corrosion process [5,8,9]. Sulfate reducing bacteria (SRB), e.g., *Desulfovibrio desulfuricans*, are probably the best known bacterial group in relation to microbial corrosion and ALWC [10,11]. The SRB corrosion process is typically

**Citation:** Phan, H.C.; Blackall, L.L.; Wade, S.A. Effect of Multispecies Microbial Consortia on Microbially Influenced Corrosion of Carbon Steel. *Corros. Mater. Degrad.* **2021**, *2*, 133–149. https://doi.org/10.3390/ cmd2020008

Academic Editor: Philippe Refait

Received: 1 February 2021 Accepted: 18 March 2021 Published: 25 March 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/).

associated with a black biofilm on the surface and localised corrosion underneath [12]. In addition to SRB, sulfur oxidising bacteria (SOB) have been implicated in ALWC, and the presence of both microbial groups will contribute to the sulfur cycle and subsequent corrosion [5,8,13–17].

While there have been a large number of studies of MIC in the laboratory, most of these have failed to replicate the accelerated corrosion rates observed in the field. One of the possible reasons for this is that most tests use single microbial strains in the experiments (most commonly SRB), which is very different to the mixed microbial species found in biofilms associated with accelerated corrosion in the real environment [9,18–21]. Another possible reason is that the testing arrangemen<sup>t</sup> used (e.g., test media, oxygen levels and test duration) may influence the processes (abiotic and biotic) required for the accelerated corrosion.

Various laboratory MIC experiments have employed undefined combinations of microbes (e.g., [5,22–25]). While this might be more similar to the field situation, the largely unknown combinations of microbes make replication impossible and the roles of any individual species cannot be determined. Several MIC studies using known microbial strains in various combinations have been undertaken, and have showed how the strains collectively affect corrosion (see Table S1). In some of the latter, the multispecies combinations have led to increased corrosion and in others decreased corrosion.

The mechanisms responsible for MIC are hypothesised to include transfer of electrons from the metal of interest to the SRB and/or biochemicals produced as part of metabolic processes [26,27]. Both require the presence of a suitable electron donor and electron acceptor. The artificial supply of electron donors could facilitate increased corrosion compared to real world scenarios since seawater typically has relatively low levels of organic carbon sources [28]. Several researchers have shown that different test media used in MIC studies can affect the level of corrosion [29–31]. Additionally, specific strains of SRB (e.g., *Desulfopila corrodens* and *Desulfovibrio* sp. strain IS7) have been linked to the direct electron uptake from an iron/steel surface [26] and have also been detected in field studies of ALWC [9,17,23].

This research adds to the understanding of the effect of multispecies microbial consortia in relation to microbial corrosion such as ALWC. The work involved exposing marine grade steel to an ALWC biofilm from a marine environment, pure culture isolates from an ALWC biofilm or adjacent locations, *D. desulfuricans* or to an uninoculated media (i.e., negative control). The hypothesis of this work was that the combinations of microbes would produce increased rates of corrosion compared to tests with single isolates or an uninoculated control.

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

#### *2.1. Steel Test Coupons*

Marine grade DH36 carbon steel [32] was chosen for this research due to its widespread use in shipbuilding and offshore structures. The metal coupon dimensions were 25 mm × 25 mm with a thickness of 6 mm. A 2 mm diameter hole was drilled in each coupon near the middle of one of the edges (3–4 mm from the edge) to allow coupons to be hung vertically in the aqueous corrosion tests. The top and bottom surfaces of the coupons were prepared using an automatic grinding machine (Tegramin-25, Struers, Cleveland, OH, USA) with a series of silicon carbide papers i.e., 320, 500 and 1200 grit. The edges were manually ground with 320 grit silicon carbide paper. After grinding, the coupons were cleaned by immersing in acetone in an ultrasonic bath for 8 min, rinsed thoroughly with distilled water and then absolute ethanol, and finally dried under warm air. Coupon weight was measured using a high accuracy analytical balance (MS205DU, readability 0.01 mg, Mettler Toledo, Mississauga, Canada) shortly before the corrosion tests.

#### *2.2. Bacteria and Test Media Details*

The four marine isolates were *Halomonas korlensis*, *Bacillus aquimaris*, *Prolixibacter bellariivorans* and *Sulfitobacter pontiacus*. Their presumed phenotypes cover a range of potential functions relevant to oxygen tolerance, biofilm formation and MIC. *H. korlensis* is a denitrifier/nitrate–nitrite reducing facultative aerobe [33] that was isolated from marine sediment [21]. *B. aquimaris* is a nitrate/metal ion utilising aerobe found in marine environments/biofilms/corrosion steel surfaces [34,35] that was isolated from an orange ALWC tubercle [36]. *P. bellariivorans* is an iron-relating/sugar-fermenting anaerobe [37,38] that was isolated from an orange ALWC tubercle [36]. Finally, *S. pontiacus* is an obligate aerobe which belongs to the Rhodobacteraceae group which are noted for surface association [6] and can reduce thiosulfate and sulfite to sulfate [39]. It was isolated from a microbial test kit (used to determine ALWC microbial functions) that was inoculated with an ALWC tubercle sample.

SRB are frequently associated with microbial corrosion (including ALWC), and a culture collection strain of an SRB, *D. desulfuricans* (ATCC 27774), was used. In addition to the five pure cultures, a consortium of microbes found in an orange tubercle from a suspected case of ALWC on steel sheet piling was also used in one of the test conditions [36].

Pure cultures of the marine isolates (stored at –80 ◦C in glycerol), were freshly grown on marine agar (MA, BD Difco, Franklin Lakes, NJ, USA) plates in an aerobic environment. These cultures were then used to inoculate 45 mL volumes of autoclaved (121 ◦C, 1 atm for 16 min) half-strength marine broth (MB12 , BD Difco, Franklin Lakes, NJ, USA), which were incubated aerobically at 25 ◦C for 72 h. Since MB has relatively high nutrient levels, it was diluted to better mimic seawater. Plate counts on MA were performed for all culture broths using serial ten-fold dilution and triplicate plating.

*D. desulfuricans* was grown in tryptic soy broth (Merck, Darmstadt, Germany), supplemented with sodium lactate (4 mL/L) and magnesium sulfate (2 g/L). The medium was adjusted to pH 7.5 prior to autoclaving (121 ◦C, 1 atm for 16 min), after which 0.5 g/L of 0.22 μm filter-sterilised ferrous ammonium sulfate was added. *D. desulfuricans* was incubated in an anaerobic jar at 25 ◦C for 72 h, with a gas generator sachet (AnaeroGen, Oxoid, Basingstoke, Hampshire, UK) and indicator (Resazurin, Oxoid, Basingstoke, Hampshire, UK). A black culture broth was obtained after incubation, which is an indication of ferrous sulfide production due to sulfate reduction. Plate counts of bacteria were performed anaerobically on modified tryptic soy agar (TSA+, prepared as for the modified tryptic soy broth but with 1.5% agar included prior to autoclaving).

An orange tubercle from an ALWC setting had been sampled previously and was stored frozen at −80 ◦C. The ALWC consortia inoculum for this work was prepared by adding 2.5 mL of thawed inner and 2.5 mL of outer tubercle layers to 45 mL of sterile MB1/2 broth (10% inoculation), and incubated aerobically at 25 ◦C for 72 h. All liquid cultures were gently shaken (~10 s) once a day to help stimulate microbial growth.

Corrosion tests were performed in 500 mL sterilised glass bottles containing 350 mL of sterile MB12 with each containing a sterile test coupon suspended by Nylon string. The coupons and Nylon string were sterilised by immersion in absolute ethanol (Merck) for 30 min, then drying in a sterile dry bottle. Subsequently coupons were placed close to a Bunsen flame (10–15 cm) for 1–2 min to remove any excess ethanol prior to starting the corrosion tests. For the tests with pure cultures or orange tubercle material, 2 mL volumes of their 72 h culture broths were added and mixed thoroughly. A total of six test configurations were carried out (Table 1). All of the tests were nominally aerobic except for the test with *D. desulfuricans* (T6) for which anaerobic conditions were obtained by placing the test bottles in individual anaerobic bags together with AnaeroGen sachets and resazurin indicators (Oxoid). All the treatments were conducted statically with 6 individual replicates. All bottles were incubated in a Thermoline Scientific, Australia incubator at 25 ◦C.


**Table 1.** Relative abundance of microbes inoculated in laboratory corrosion tests and incubated at 25 ◦C in aerobic or anaerobic conditions.

#### *2.3. Corrosion Tests*

The duration of the corrosion tests was 8 weeks. Testing six replicates allowed a range of different studies to be performed using separate coupons. Periodic photos of test bottles were taken to record any changes in media or test coupons.

A total of 40 mL of fresh sterile MB12 was added into each test bottle fortnightly (i.e., fed-batch culture) to help maintain microbial viability during the study period. Sampling of the corrosion tests for bacterial counts was undertaken at weeks 2, 4 and 8 of incubation. This involved pipetting 10 mL of the culture (5 mL from near the test medium surface and 5 mL near the bottom of the medium) before nutrient amendment. Bacterial counts were obtained using MA plates for aerobic (T1–T5, Table 1) and TSA+ plates for anaerobic (T6, Table 1) corrosion test using serial ten-fold dilutions and duplicate plating. pH was measured at the beginning of incubation and then after 4 and 8 weeks of incubation. Aliquots of liquid cultures were stored in −20 ◦C for use in subsequent microbial analyses.

After 8 weeks of incubation, all coupons were removed from the test bottles. Shortly after removal, a gentle stream of sterile MilliQ water was run over the coupons to remove any planktonic cells from the biofilms. Pictures of the coupons were then taken. Certain coupon biofilms/corrosion products were examined by scanning electron microscopy (SEM), X-ray powder diffractometery (XRD) or energy-dispersive X-ray spectroscopy (EDS). For these, sterile MilliQ water was used to gently wash the coupons three times, after which they were air-dried in a physical containment level 2 cabinet for 2 h and then stored in a desiccator. For SEM of coupon biofilms, one coupon per test condition underwent fixation with glutaraldehyde solution (2.5 vol %) then dehydration via an ethanol series (25%, 50%, 75%, 90% and 100%) [40]. Coupons for steel surface analyses (SEM, 3D optical profilometery, weight loss), underwent a series of cleaning steps to remove biofilms/corrosion products. Individual coupons were sonicated in sterile MilliQ water for 2 min (the fluid from this was stored at −20 ◦C for subsequent microbial analyses), then in Clark's solution [41] for 2 min, followed by rinsing with water and ethanol, drying under warm airflow and then storing in a desiccator.

#### *2.4. Metal Coupon and Microbial Analysis*

SEM (Zeiss SUPRA 40VP-25-38, Oberkochen, Germany) images were taken of the biofilms and the steel surfaces of separate fixed and cleaned coupons for each of the test conditions using a range of magnifications (150× to 12,000×). One coupon from each of the six treatments was used for EDS (Zeiss SUPRA 40VP-25-38, Oberkochen, Germany and INCA software, version 4.13) to provide semi-quantitative analyses of the biofilms/corrosion products formed on the coupons. Intensity with weight percent (wt.%) of chemical elements was averaged from two locations of the biofilm on each coupon surface. One uncleaned coupon from each of the six treatments was used for XRD (Bruker D8 Advance, Germany). Compositions of surface products were scanned with a copper line focus X-ray tube producing Kα radiation from a generator operating at 40 kV and 30 mA.

Three cleaned coupons from each of the six treatments were used to obtain average corrosion rates via weight loss using the method described in reference [41]. These coupons were also used for 3D optical profiling (Bruker Contour GT-K1) to obtain quantitative information about corrosion morphology. Each metal coupon was scanned over the entire surface (using 5× objective), and any areas with localised pitting were subjected to more detailed, higher resolution analysis. Pit depth (μm) and relative volume (μm3) for pits greater than 10 μm depth were recorded and analysed using VISION 64 ™ software (Bruker), and then pit density was calculated by number of pit/mm2.

For corrosion morphology analysis, the pitting depth, relative pit volume and corrosion rates were statistically compared for significant differences among the treatments (*p* < 0.05) using one-way ANOVA and Tukey's post-hoc test (IBM® SPSS ™ Statistics version 25).

Microbial analysis was based on agar plate counts and 16S rRNA gene metabarcoding. Relative abundance was determined by a sum of the aerobic and anaerobic counts from the liquid in each treatment. For metabarcoding, liquid samples of T2, T4 and T5 experiments (Table 1) obtained at week 2 and week 8, together with biofilm suspensions obtained following sonication from week 8 samples were sent to a commercial test facility (AGRF, Adelaide, Australia) for DNA extraction and Illumina MiSeq sequencing. The primer pairs used were 341F (CCTAYGGGRBGCASCAG) and 806R (GGACTACNNGGGTATCTAAT) for amplification of the V3–V4 region of 16S rRNA genes. Data were processed using Quantitative Insights Into Microbial Ecology (QIIME 1.9) followed by METAGENassist [36].
