A total of 1 Kg of sand was collected from a high pressure (HP) separator at an Australian oil production facility experiencing sand accumulation inside the flow line. The system was under chemical treatment with a commercial imidazoline based corrosion inhibitor and a commercial chemical biocide containing Tetrakis (hydroxymethyl) phosphonium sulphate Benzyl-(C12-C16 Linear Alkyl) (THPS) and Dimethyl-Ammonium Chloride Formaldehyde (DACF). The samples were transported in a sterile anaerobic container under refrigerated conditions (4°C) to the laboratory for the recovery of native microorganisms and chemical characterization.

Upon arrival, oilfield sand was immediately inoculated in different culture media to recover the oilfield microbial community to be used in corrosion studies. For this purpose, ten grams of sand were grown in 40 mL of four (4) selective culture media to target the growth of sulphide producing prokaryotes (SPP), acid-producing bacteria (APB), iron reducing bacteria (IRB), and methanogens (MET). Culture media were prepared following the standard method NACE TM0194 (2004); SPP media was prepared following the guidelines proposed elsewhere (Suarez et al., 2019). All culture media were sparged with a gas mixture of 20% CO₂/80% N₂ for 1 h to saturate the solution, and dispensed in serum vials capped with rubber stoppers crimped sealed and autoclaved. The inoculated tubes were incubated for 21 days at 40°C. The day of the experimental set up, an aliquot of each culture media that showed positive microbial growth was adjusted to a final concentration of 1.9 × 10⁷ cells/mL in a falcon tube. The falcon tube was centrifuged at 15,000 × g for 5 min to harvest cells from each culture media, finally, the four different pellets were combined in 5 mL of sterile PBS for the inoculation of the reactors. Molecular identification of the consortium was performed using 16S rRNA gene sequencing and the results were described elsewhere (Diaz-Mateus et al., 2021).

Part of the oilfield sand was washed with ultrapure water (Milli-Q system, resistivity 18.2MΩcm) and roasted at 600°C for 3 h in a muffle furnace (Thermolyne Industrial Benchtop Muffle Furnaces, Thermo Fisher Scientific) to remove surface organic matter and used as treated sand (Tian et al., 2012). After the temperature treatment, the sand was cooled to room temperature inside the furnace; and, later stored under vacuum conditions for further characterization.

Carbon steel with elemental composition of (weight %): C (0.07–0.8), Mn (1.38–1.39), Si (0.16–0.68), S (0.01), P (0.01), Ni (0.01–0.03), Cr (0.09–0.23), Mo (0.03–0.06), Cu (0.06–0.11), V (0.02–0.06), Nb (<0.01), Ti (<0.01), Al (0.009–0.038), B (<0.0005), and Fe (balance) was cut into square coupons with a surface area of 5.29 cm² including a weld in the center of the sample. Coupons were electro-coated with epoxy (Powercron 6000CX; PPG Industrial coatings) and only one surface of the samples was wet-ground to a 600-grit finish using SiC paper, to expose it to the tested sand. Subsequently, the samples were washed, rinsed with absolute ethanol, and dried under nitrogen gas. Coupons were finally sterilized by UV irradiation (15 min at each side).

Two different corrosion scenarios were assessed using 2-litre capacity glass cells. (1) UDMC with raw untreated sand, (2) UDMC with treated sand. Five coupon replicates were used to evaluate each scenario. Coupons were placed horizontally in custom-made glass containers (Φ40 × 20 mm) designed to ensure uniform deposition of a 20 mm layer of sand on the surface of the samples (Machuca et al., 2017). 55 ± 0.2 grams of sand (treated and untreated, accordingly) were deposited on top of the carbon steel coupons and tapped to achieve the same height in the five glass containers. The reactors were then connected to a filtered-sterilized gas line (20% CO₂/80% N₂ mixture) to maintain anaerobic conditions throughout the exposure period.

The test solution used was synthetic produced water with the following composition: NaCl 462 mM, CaCl₂.2H₂O 1.5 Mm, K₂HPO₄ 2.8 mM, NH₄Cl 4.6 mM, KCl 4.6 mM, MgCl.H₂O 2.4 Mm, D-glucose 5 mM, Na₂SO₄.5H₂O 45 mM, Na₂S₂O₃4.02 mM, Na-formate 14.7 mM, Na-lactate 5.8 mM, Na-acetate 4.9 mM, 1.4 g L − 1 bactocasamino acids (BD), Na-pyruvate 4.4 mM and, 1 l of ultrapure water (Milli-Q system, resistivity 18.2MΩcm). The solution pH was buffered with sodium bicarbonate and the initial pH was 7.3 ± 0.2 reflecting pipeline in situ conditions.

Microbial consortium cells (Section “Corrosion measurements and metal surface analysis”) were added to the reactors at a final concentration of 10⁷cells mL⁻¹. The temperature and stirring of the reactors were set to 40°C ± 1°C, and 200 rpm, respectively. To maintain an active microbial consortium throughout the exposure, reactors were maintained under batch feeding (30% of test solution replenished every 4 days). The test exposure period was 4 weeks.

Statistical analysis of corrosion and adenosine triphosphate (ATP) data were conducted using SPSS 27 and PAST (V4.10). The statistical analyzes applied were selected based on the normality of the data in each variable. Shapiro–Wilk method was used to test the normality of the data in each variable. Then, one-way analysis of variance (ANOVA) with Tukey’s post-hoc means separation test was implemented to test the homogeneity of variances in each variable and identify statistically significant differences between variables normally distributed. Statistical comparison of microbial alpha diversity levels between the two sand deposits was calculated on rarefied data with a parametric t-test. Results of statistical tests were considered significantly different with value of p ≤0.05.

The sessile microbial community developed in both treated and untreated sand was dominated by fermenting microorganisms, despite the initial inoculum was heavily dominated by sulfate reducing, thiosulfate reducing and iron oxidizing bacteria, with lower abundance of fermenters. These results suggest that the fermenters present in the consortium had a stronger competitive preference for the carbon sources available in the test solution. Other authors have described similar results where the influence of carbon sources in the microbial community structure of a native sediment consortium containing iron-reducing bacteria, sulfate reducing bacteria and fermenting organisms was studied. Authors found that the microbial community shifted to a community dominated by fermenting organisms in cultures enriched with glucose and lactate as carbon sources (Lentini et al., 2012).

Significant differences were seen in the relative abundances of the predominant genera in the sessile communities developed at each condition. One of the main differences found is the increased abundance of Acetomicrobium in the treated sand, which resulted in almost the total make-up of the community. Although little is known about the genus Acetomicrobium, the few reported species belonging to this have shown higher growth rates when glucose is present in the culture medium (Soutschek et al., 1984). Though, this obligate anaerobe has wide metabolic plasticity and can ferment other substrates such as amino acids, dicarboxylic acids, and, other sugars (maltose, fructose), besides, it can use several sulphur compounds as terminal electron acceptors (Cook et al., 2018). These broad metabolic capabilities may have influenced in the establishment of this genera as the main group in the final microbial community in the treated sand deposit.

Aside from Acetomicrobium wide metabolic plasticity, the presence of residual biocide in the untreated sand might have also played a role in the significant differences evidenced in the microbial community developed in the sand samples. Although the organic compounds detected in the FITR spectra (Figure 2) cannot be identified with this technique, we know that biocide injection was an ongoing MIC mitigation strategy in the oilfield from where the sand sample was taken. Thus, it is possible that the dominant presence of Acetomicrobium in the reactor where sand depleted from organic compounds is due to the growth inhibition effect that the biocide was having on that population in the oilfield from where sand samples were obtained; and ceased when the biocide was removed from the system. Nonetheless, further analysis will be required to confirm susceptibility of Acetomicrobium to the field biocides.

Conversely, in the untreated sand, fermenters were mainly represented by five different genus (Thermovirga, Vibrio, Aminirod, Alkalibacter, and Acetomicrobium), coexisting with sulfate and thiosulfate reducers. Syntrophic interactions between fermenters and SRB have been reported in different anaerobic environments, as fermenters’ metabolic by-product H₂ can act as an electron donor for the sulfate and thiosulfate reduction by SRB and thiosulfate reducing bacteria (TRB) (Dar et al., 2008). Considering that the only difference between the untreated and treated sand was its chemical composition (as the same carbon sources and electron acceptors were supplied in the test solution), it is inferred that the heteroatoms of nitrogen, sulphur, carbon, oxygen and phosphorous present on the hydrocarbons, residual biocide and corrosion inhibitors, represented an additional source of energy and nutrients to the microbial consortium inoculated in the reactor. This enriched environments, then led to a more diverse community, which was supported by the higher metabolic rates (ATP concentration) found in the untreated sand.

Alpha diversity analysis of the microbial communities established in the treated and untreated sand at the end of the 4 weeks of immersion confirmed that differences in the two communities biodiversity were significant, and that the untreated oilfield sand allowed the development of diverse microbial communities that resembled more the microbial communities usually found in oil and gas facilities and reservoirs (Machuca Suarez and Salgar-Chaparro, 2018; Suarez et al., 2019; Nicoletti et al., 2022). It has been reported that the predominant genera in the untreated sand, Thermovirga, possess homologs of the benzyl succinate synthase gene (bssA), which codes for a benzyl succinate synthase, the key enzyme of anaerobic toluene degradation, indicating that members of this genera may play a key role as primary fermenter in the anaerobic degradation of hydrocarbons (Vigneron et al., 2017). Similarly, Aminirod (also present in higher abundance in the community) has shown the ability to act as a secondary fermenter in the degradation of hydrocarbons. This strain can ferment propionate and butyrate (metabolic by-products generated by primarily fermenters) into acetate and H₂ (Liu et al., 2021). Hence, our results support that the hydrocarbons present in the untreated sand (Table 1) were potentially being degraded by Thermovirga and Aminirod in a syntrophic metabolism. In contrast, genes related with carbohydrate, aminoacids and pyruvate metabolism were found as biomarkers in the microbial community developed in the treated sand.

Results from this study demonstrated that the chemical differences of the sand deposits tested lead to different under deposit microbial corrosion rates. Despite the differences in the relative abundances of the genera found in both treated and untreated sand, fermenting organisms, previously related to MIC were the predominant bacteria of the consortium (de Paula et al., 2014; Salgar-Chaparro et al., 2020b; Cai et al., 2021). Acetomicrobium, the dominant genus found in the treated sand microbial community (94.3%) uses the fermentation of simple sugars and amino acids as primary metabolic strategy, releasing corrosive metabolites such as acetic acid, CO₂ and H₂ to the environment; moreover, it possess the ability to reduce thiosulfate, elemental sulphur and cysteine to hydrogen sulphide (Maune and Tanner, 2012; Cook et al., 2018), which acidifies the water, causing pitting corrosion to carbon steel pipelines. Similarly, Thermovirga, the most abundant genus in the untreated sand microbial community (94.3%) is a thermophilic bacterium capable of fermenting proteins, organic acids, and single amino acids, producing ethanol, H₂, and CO₂ as metabolic by-products. Thermovirga, just as Acetomicrobium can also couple fermentation with the reduction of elemental sulphur (S₀) to hydrogen sulphide (H₂S) (Dahle and Birkeland, 2006; Duncan et al., 2009). It has to be noted that despite sulphide concentration was not measured during the test, a characteristic smell of H₂S and the presence of a black cover in the sand deposits at the end of both test (suggesting FeS formation) suggests the formation of H₂S in the tests.

The statistically significant differences in the corrosion damage generated by the two different microbial communities were likely influenced by the different metabolic rate (based on ATP) measured in the two microbial communities. A higher metabolic activity can result in higher production of corrosive metabolites. In our study, as fermenters were the most abundant microbial groups it likely that higher metabolic activity resulted in higher concentration of acids in the test solution. An increased bacterial metabolic rate based has been previously associated as the main contributing factor in the acceleration of steel corrosion (Xu et al., 2022; Zhang et al., 2022). For example, Dzierzewicz et al. (1997), reported a statistically significant relationship between H₂S release, bacterial growth rates, and, enzymatic activities rates (hydrogenases and ATPS-reductases) of D. desulfuricans with steel corrosion rates. It is worth mentioning that the synergy between higher concentrations of corrosive secondary metabolites being released by the biofilm developed in the untreated sand deposit, together with the physical barrier that the sand bed represents for the diffusion of these corrosive chemical species away from the metal surface, is very likely the main factor contributing to the high localized corrosion damaged observed in the metal surface under the untreated sand deposit.

Moreover, the higher corrosion rates found in the carbon steel exposed to untreated oilfield sand, together with the higher microbial diversity and activity found on it, suggests that MIC took place because of synergistic interactions among the different microbial species in the community. Higher relative abundances of H₂ consuming microorganisms, such as sulfate reducers (Desulfovibrio) and thiosulfate reducers (Oceanotoga, Shewanella) were found in the untreated sand deposit, in comparison with the treated sand deposit. A syntrophic interaction among these microbial groups and fermenters have been previously studied (Laanbroek and Pfennig, 1981; Finke and Jørgensen, 2008). Fermenters secondary metabolites such as H₂, can be used by sulfate and thiosulfate reducing bacteria as electron donors (Dar et al., 2008). Multispecies biofilms have been found more corrosive in comparison with single species biofilms due to the cascade of biochemical reactions that occur between taxonomically and metabolically different microorganisms (Zuo, 2007; Videla and Herrera, 2009).

Scanning electron microscope analysis demonstrated that despite the severity of the corrosion rates derived from the environment developed at the metal-deposit interface along the two sand deposits was different, the corrosion mechanisms revealed the grain boundaries in a low grade (micro-etching) in both tests, resembles the micro etching of carbon steel resulting from the standard practice of applying acid treatment previous to the microscopic examination of carbon steel (ASTM E407-07, 2007). Hence, results suggest that the UDMC mechanism is likely related with the organic acids released by bacteria as metabolic by-products, trapped in the metal-deposit interphase because of the tortuous pathways that the sand grains represent for their diffusion to the bulk solution. Most of bacterial acid metabolic byproducts are in the free acid form and are highly corrosive because their reduction coupled with the oxidation or iron is a thermodynamically favorable reaction and kinetically not retarded (Gu and Galicia, 2012).

The interaction of microorganisms with oilfield deposits is a topic that remains unexplored. In this paper, it is demonstrated that the chemical composition of one of the most commonly found deposits in oil and gas pipelines (sand), impacts the diversity, metabolic activity, and functional attributes of multispecies microbial communities, and consequently, affects the extent of under deposit microbial corrosion (UDMC). Results of this investigation provide valuable information about how microbial communities respond to different environmental conditions and how these microbiological changes impact the risk of corrosion. In addition, these results highlight the importance of including field samples in laboratory-based corrosion experiments to create systems that better simulate real field scenarios and therefore, generate more accurate corrosion risk assessments.