Beating the bugs: roles of microbial biofilms in corrosion

1 Introduction

Microbiologically influenced corrosion (MIC) can accelerate mechanical failure of metals in a wide range of environments ranging from oil and water pipelines and machinery to biomedical devices. Although this mode of environmentally assisted corrosion has no single definition, MIC invariably includes corrosion in the presence of microbial species including bacteria and the associated biofilms produced by those bacteria. MIC has been noted to occur for a variety of metals and alloys such as iron and steel, copper, and titanium (Rao, 2012), as well as on nonmetallic materials such as concrete (George, 2012). For instance, MIC and associated biofouling are estimated to account for 20% of direct fuel pipeline integrity and reliability costs exceeding $2 billion USD per year (Flemming, 1996; Koch, Brongers, Thompson, Virmani, & Payer, 2002). Recently, this corrosion mechanism has been implicated in several rapid, high-profile failures of pipeline including a light grade crude oil pipeline failure (Al-Jaroudi, Ul-Hamid, & Al-Gahtani, 2011), a seawater pipeline failure within a few years of installation (Abraham, Kain, & Dey, 2009), leaks in a stainless steel cooling water system (Yu, Dillon, & Henry, 2010), and leaks in the Trans-Alaskan Pipeline (Egan, 2011).

Despite decades of research, the exact mechanisms by which and environments in which MIC enhances corrosion are poorly understood. Indeed, even within the same oilfield or pipeline, some locations may exhibit MIC whereas others do not (Maruthamuthu et al., 2011; Voordouw et al., 2009; Yu et al., 2010). Conventionally, this phenomenon has been studied either from a biological perspective – considering the complex nature of bacteria consortia and biofilm formation – or from a materials corrosion perspective – considering the evidence and mitigation of metal pitting under controlled conditions. Therefore, there is no consensus on how to mitigate corrosive losses due to MIC. Here, we review the most current understanding of MIC mechanisms on steel (i.e., iron-carbon or ferrous alloys), MIC mitigation strategies, and the role of the bacterial biofilm in both corrosion and mitigation response.

2 Characteristics of MIC

MIC can occur in both aerobic and anaerobic conditions, and is defined by no single characteristic trait other than corrosion in the presence of microbial species (bacteria) and their associated biofilms. Thus, MIC is considered biotic corrosion, and the terms “microbes” and “bacteria” are used interchangeably. As bacteria are ubiquitous in almost all environments (air, water, and organic fluids), the mere presence of bacteria found at sites of corrosion does not necessarily mean that the bacteria were a significant contributing factor in corrosion initiation or acceleration. MIC can occur in conjunction with other types of abiotic corrosion, further obfuscating attempts to identify the root cause of corrosion failure. Often, MIC is associated with surface pitting, which leads to more rapid corrosive failure than uniform corrosion (Bryant, Jansen, Boivin, Laishley, & Costerton, 1991; Chamritski, Burns, Laycock, & Webster, 2004; Jan-Roblero, Romero, Amaya, & Le Borgne, 2004; Miranda et al., 2006; Yuan & Pehkonen, 2009) (Figure 1). Thus, even nonferrous metals that exhibit a naturally passivating layer at the surface (e.g., copper oxides and titanium oxides) exhibit MIC when metabolic byproducts of the microbes serve to chemically reduce and thus disrupt the passivating film under anaerobic conditions; this results in pitting corrosion at the metal-biofilm interface (Rao, 2012). As many surface characteristics including roughness, charge, and hydrophilicity can modulate bacterial adhesion, it is not yet clear whether bacteria initiate the pits or adhere preferentially to pre-existing pits. Although several studies have found evidence that microbes adhere preferentially to anodic metal regions such as welded joints, scratched edges, grain boundaries (Dexter & Eashwar, 1999; Sreekumari, Nandakumar, & Kikuchi, 2001), or to previously corroded locations (Little et al., 1999), a more recent study was not able to confirm these site-specific claims (Sherar et al., 2011). When many species of bacteria adhere and proliferate on a surface, they secrete a polysaccharide-rich matrix; a biofilm includes that extracellular polysaccharide matrix and the bacterial cells within. Attempts to spatially correlate localized corrosion with the overlying biofilm, for example, using electrochemical mapping techniques, have failed due to the conductive nature of the biofilm itself (Dong, Shi, Ruan, & Zhang, 2011). Further, although pitting is a known signature of MIC, others have found no unique characteristics that distinguish the morphology of MIC pits from abiotic pits (Thomas & Chung, 1999).

Figure 1

Scanning electron microscopy images which demonstrate the growth and pitting damage caused by sulfate-reducing bacteria growth on SAE 1020 steel studs. (A) Control stud showing concentric rings from lathe cutting; scale bar, 500 μm. (B) sulfate-reducing bacterial biofilm deposit after 64 days; scale bar, 500 μm. (C) Corroded stud after biofilm removal revealing severe pitting damage; scale bar, 500 μm (Bryant et al., 1991). Reprinted with permission from the American Society for Microbiology.

At sites of suspected MIC, a so-called consortium of several bacterial species is always found – rather than a single species – although the number and types of bacteria in that consortium varies widely from location to location. Because the culture conditions of the majority of bacterial species are not well established, determining the phylogenetic makeup is difficult. However, the application of molecular techniques now allow much more detailed analysis of extracts, quantifying the types and relative numbers of species (Amann, Ludwig, & Schleifer, 1995). Several studies have performed detailed phylogenetic characterization of fluids from gas and oil pipelines (Jan-Roblero et al., 2004; Maruthamuthu et al., 2011; Rajasekar, Anandkumar, Maruthamuthu, Ting, & Rahman, 2010) and water outlet pipes at oil production facilities (Larsen et al., 2010). The categories of bacteria typically directly linked to corrosion are sulfate-reducing bacteria (SRB), iron-oxidizing bacteria, sulfur-oxidizing bacteria, nitrate-reducing bacteria (NRB), and methanogens. These bacteria may reduce the metal directly, produce corrosive metabolic byproducts, and/or produce biofilms that indirectly alter the local environment to promote corrosion (Rao, 2012).

SRB are considered one of the main culprits of biotic corrosion in anaerobic conditions. SRB are a grouping of bacteria that includes at least 220 species that produce H₂S, and use sulfates as the terminal electron acceptor (Barton & Fauque, 2009). Most SRB are considered obligate anaerobes, meaning that the cells cannot metabolize and/or replicate in the presence of oxygen, although many species can temporarily tolerate low levels of oxygen (Johnson, Zhulin, Gapuzan, & Taylor, 1997). Furthermore, anaerobic conditions capable of supporting SRB growth can be created in overall aerobic environments, due to the microniches created within the bacterial biofilm/corrosion product layer. Although SRB are the most studied and well understood of the anaerobic corrosion inducing bacteria, we note that MIC can occur in anaerobic conditions in the absence of SRB (Rajasekar et al., 2010). However, given the wide acknowledgment of SRB presence in industrial MIC contexts, and the substantial study of this bacteria grouping at the genetic and structural levels (Wall, Arkin, Balci, & Rapp-Giles, 2008), we will focus primarily on the role of this microbe type within the consortium and biofilms relevant to MIC of ferrous alloys.

3 Characteristics of MIC-associated biofilms

MIC is attributed not to planktonic bacteria but instead to adherent bacteria in the form of biofilms on surfaces. Many bacteria types adhere to and form polysaccharide-rich biofilms on material surfaces, with interactions governed strongly by the material surface properties, fluid properties, and bacteria adhesion mechanisms (Lichter et al., 2008; Lichter, Van Vliet, & Rubner, 2009). Biofilm formation follows a series of critical steps: an initial rapid and reversible bacterial attachment, a more stable, longer term attachment, bacterial replication and matrix secretion, biofilm maturation, and finally bacterial dispersal (Lichter et al., 2009; Monds & O’Toole, 2009). The main structural components of the biofilms are produced by the bacteria themselves through the secretion of extracellular polymeric substances (EPSs). EPSs are composed chiefly of polysaccharides, proteins, nucleic acids, and lipids (Vu, Chen, Crawford, & Ivanova, 2009). Biofilms can be anywhere from μm to mm in thickness. These structures can develop and mature over the course of hours, days, or months, depending on the bacterial species and environment (Figure 2). Repeated biofilm formation has the potential to occur due to pigging, an intentional mechanical delamination of corrosion products including iron sulfides within internal pipeline surfaces, or due to general biofilm rupture. In the context of fuel transport within pipelines, the sporadic flow rates and fluid composition can also disrupt and promote biofilm formation, and likewise the presence of the biofilm can affect flow trajectories. Essentially, through these biofilms the bacteria create a unique and sustaining microenvironment that can differ significantly – in terms of composition and distribution of solids, fluids, and gases – from the overall macroenvironment.

Figure 2

(A) Scanning electron microscopy images demonstrating the growth of a biofilm over the course of 60 days. All images are at 1000× except for 60 days which is 200×. Inner boxes are at 4000× (Fernandez, Diaz, Amils, & Sanz, 2008). Reprinted with permission from Springer. (B) Original image demonstrating a cross-section of the combined biofilm/corrosion product layer after 2 days of growth of Desulfovibrio vulgaris on music spring quality steel (ASTM A228).

The formation of a biofilm itself does not necessitate an increased corrosion rate and can actually lead to corrosion resistance in some circumstances. Certain bacteria within biofilms can remove corrosion-promoting agents, such as O₂, (Dubiel, Hsu, Chien, Mansfeld, & Newman, 2002) as part of the bacteria metabolism, can secrete antimicrobial agents that inhibit growth of corrosion-causing bacteria, and can generate physical barriers that protect surfaces from corrosion (Avidan, Satanower, & Banin 2010; Zuo, 2007). The particular species makeup and environmental conditions greatly influence the corrosive potential of a biofilm. This variation has obfuscated the identification of a general, causal link between bacteria presence and corrosion. Such complexity prohibits the community from drawing broad conclusions on corrosion potential, mechanisms, and rates beyond the specific conditions in which observations were made. This difficulty has prompted the development of models to better predict when MIC will be a problem (Maxwell, 2006; Skovhus, Holmkvist, Andersen, Larsen, & Pedersen, 2012), but the applicability of such models is still currently limited to specific species and environmental conditions.

Although single species studies are naturally better controlled and simpler to interpret, recognition of the significance of the interplay among the different bacterial species within a biofilm has prompted multiple-species studies. For instance, Okabe, Itoh, Satoh, and Watanabe (1999) studied the spatial distribution of SRB in an aerobic wastewater biofilm and found that although SRB were located throughout the biofilm, sulfate reduction was focused in a narrow anoxic band. Andersson, Kuttuva Rajarao, Land, and Dalhammar (2008) systematically studied initial adherence and biofilm formation with different combinations of 13 bacterial species often found in wastewater systems, and found complex protagonistic and antagonistic interactions that depended on species composition. Interestingly, Lee, Buehler, and Newman (2006) found that a dual species biofilm with an SRB and iron-reducing bacteria resulted in reduced corrosion, as compared with a biofilm comprising only SRB. Videla, Borgne, Panter, and Singh Raman (2008) also found differences in corrosion with iron-reducing bacteria and SRB in mixed culture in different environments. Overall, however, these studies illustrate the complexity of the process rather than providing concrete conclusions about MIC mechanisms.

Environmental factors such as fluid shear flow found within pipelines add another layer of complexity to biofilm development and thus to MIC in industrial settings. It is known that shearing forces can greatly influence bacterial adhesion and biofilm development in dramatic ways, regardless of the context (Rochex, Godon, Bernet, & Escudie, 2008; Whitfield, Ghose, & Thomas, 2010). For example, in other bacterial species (e.g., aerobic species such as Escherichia coli), it has been shown that adhesion is strengthened in the presence of moderate flow rates. At lower rates, the so-called catch bond mechanism is not activated, and at higher rates the binding is again disrupted; only at intermediate rates are the adhesive structures at the bacterial surface configured to bind to the surface with highest affinity and longest lifetime (Thomas, Trintchina, Forero, Vogel, & Sokurenko, 2002; Whitfield et al., 2010). For SRB, it is unknown whether similar adhesive structures and mechanisms exist. However, moderate flow rates (0.1–0.5 m/s) have been shown to change SRB biofilm morphology, and to increase biofilm rigidity and yield shear stress (Dunsmore et al., 2002; Stoodley et al., 2001). At sufficiently high flow rates (3.5 m/s), SRB adhesion and biofilm growth are inhibited (Wen, Gu, & Nesic, 2007). Qualitatively, these results are similar to more widely studied aerobic biofilms, making it possible that mechanistic insights can be shared. In a study not directly related to corrosion research, Rochex et al. (2008) found that increased shear stresses lowered the species diversity within biofilms. A more robust understanding of how flow conditions affect MIC-associated biofilm structure, diversity, and corrosive potential would allow for improved risk assessment and pipeline flow rate and flow pattern selection.

4 Mitigation techniques

Several techniques have been investigated to mitigate the effects of MIC. These include using corrosion resistant metals, protective coatings, pigging, anodic and cathodic protection, biocides, introduction of competing bacteria, and nutrient addition or removal. Some techniques such as pigging (i.e., mechanical delamination of corrosion products) and biocide addition are widely used in field conditions that include sour (anaerobic, sulfur-containing) environments such as buried pipelines. Other approaches such as sol-gel coating barriers (Akid, Wang, Smith, Greenfield, & Earthman, 2008) or precoating with purified EPSs (Stadler, Wei, Furbeth, Grooters, & Kuklinski, 2010) are still in experimental stages. One of the more recently field tested techniques – and one that relies on a robust understanding of the biofilm – is to alter the corrosive ability of biofilms by manipulating the available nutrients via addition of nitrates or removal of sulfates.

In field conditions, nitrate addition (Voordouw et al., 2009) and sulfate removal (McElhiney, Burger, Maxwell, Davis, & Walsh, 2008) have been shown to affect SRB activity and be somewhat effective in reducing or altering corrosion. Nitrate addition can allow other bacteria such as NRB to outcompete the more corrosive SRB (Figure 3) (Hubert & Voordouw, 2007; Thorstenson et al., 2002). Additionally, there is evidence that nitrate can directly inhibit SRB metabolism and proliferation (Haveman, Greene, Stilwell, Voordouw, & Voordouw, 2004; He et al., 2010). However, nitrate addition is not a definitive solution to MIC. Although this approach typically reduces sulfide production, the effect is only temporary and SRB are not killed. Thus, sulfide production rapidly returns if active nitrate treatment is terminated (Mohanakrishnan et al., 2011). Another study showed that nitrate addition could reduce uniform surface corrosion but not pitting corrosion, the latter of which is often the cited cause of pipeline failures (Schwermer et al., 2008). Further, the same nutrient manipulation protocols work in some pipelines but not others, necessitating trial and error at each location to achieve the best results (Voordouw et al., 2009). One study even found that nitrates can increase corrosion of steel (Marques, Almeida, Lins, Seldin, & Korenblum, 2012). Finally, even when corrosion inhibition is achieved, for applications such as fuel recovery and processing, the addition of such chemicals then requires subsequent removal processes for the fluid of interest; this is not a treatment localized to the metal surface, and thus the high volume addition and removal of chemical agents affects product cost and thus the decisions to employ such anti-MIC strategies. Thus, although nutrient alteration has produced intriguing results on MIC mitigation, it has yet to produce a definitive solution.

Figure 3

Measured data on the bacteria content of a floating unit at Veslefrikk, an oil production facility in the North Sea. Production began in 1989, utilizing a biocide treatment until January 1999 when it was stopped in favor of nitrate injection. A subsequent decrease in sulfate-reducing bacteria (SRB) and increase in nitrate-reducing bacteria (NRB) numbers was observed. The population of SRB was characterized using a fluorescent antibody (FA) technique as well as the “most probable number” (MPN) method. Two types of NRB were measured using the MPN method: NRB1 which are facultative anaerobes, and NRB2 which are obligate anaerobes (Thorstenson et al., 2002). Reprinted with permission from NACE International.

5 Mechanisms of MIC

Owing to the complexity of the microbe-metal interaction, the underlying mechanism of MIC remains an open question and subject to much debate. Many possible mechanisms for MIC have been proposed, and it is possible that no predominant mechanism exists. In fact, this seems likely given the incredible diversity of species and conditions in which MIC is observed. It is known that chemical, mechanical, and structural factors all play a role, both biotic and abiotic in nature, and these may all work in concert to cause and sustain MIC (Beech & Gaylarde, 1999). Thus, it must be emphasized that evidence for one hypothesis is not necessarily contradictory to another. The given environmental conditions and bacterial species present may dictate the predominant mechanism.

Among the first proposed mechanisms for MIC, the cathodic depolarization theory – also known as the classical theory – was put forth in 1934 by Kuhr and Vlugt in order to explain the unexpectedly high rate of corrosion failures encountered on buried cast iron pipelines in the Dutch countryside (von Wolzogen Kuehr & Van der Vlugt, 1964). SRB were identified as the culprit due to the ferrous sulfide corrosion products found in conjunction with these failures. In anaerobic conditions where SRB thrive, hydrogen ions typically serve as the terminal electron acceptor at the cathode in a corrosion reaction. This reduced hydrogen would adsorb onto the metal surface, polarizing it. According to the classical theory, the role of the SRB was to consume this cathodic hydrogen by means of the enzyme hydrogenase, catalyzing the recombination of the adsorbed atomic hydrogen into hydrogen gas, and thus depolarizing the cathode.

Although the classical theory was grounded in electrochemistry and found general acceptance for nearly half a century, it was ultimately overturned when new techniques showed that hydrogenase can act only on molecular hydrogen and not atomic hydrogen (Stott, 1993). In support of this finding, Cord-Ruwisch (1996) showed that NRB with an enhanced ability to consume hydrogen failed to cause noticeable corrosion. Despite these findings, however, research continues into the role of hydrogenase in MIC, often with contradictory results. Studies by Bryant et al. (1991) and Da Silva, Basséguy, and Bergel (2002) showed a positive correlation between the presence of hydrogenase and MIC, whereas King, Miller, and Wakerley (1973b) found no such link.

Following the fall of the classical theory, myriad “alternative theories” have been proposed, often attempting to explain MIC while minimizing the role of the bacteria themselves (Javaherdashti, 2011). Such theories include anodic depolarization (Crolet, 1992), the presence of a volatile phosphorous compound (Iverson & Olson, 1983), metal-binding exopolymers (Beech & Cheung, 1995), sulfide-induced stress corrosion cracking, and hydrogen-induced blistering (Edyvean, Benson, Thomas, Beech, & Videla, 1998). Here, however, we restrict our focus to the biofilm and its role in MIC. Figure 4 illustrates the mechanisms by which a biofilm can contribute to corrosion initiation and accelerated corrosion rates.

Figure 4

Mechanisms by which a biofilm can contribute to corrosion. (A) Creation of anaerobic zones; (B) concentration of corrosive chemicals; (C) concentration of ferrous ions; (D) electron conduction away from surface; (E) creation of differential aeration zones; (F) binding of corrosion promoters; and (G) disruption of passivating film.

6 Summary and outlook

Over decades of research from the perspectives of both corrosion and microbiology, significant progress has been made in understanding MIC. However, this mode of environmentally assisted corrosion remains a pervasive failure mechanism in a wide range of industrial contexts. We know more about the bacterial species and biofilms than ever before, but the complex interplay among species in extreme environments such as anaerobic fluid-metal interfaces has defied effective mitigation strategies. Renewed attention to the biofilm as a biotic coating that can potentially promote and mitigate corrosion of ferrous alloys in sour environments prompts several outstanding questions. In which ways do biotic and abiotic corrosion pits and rates differ? What combination of bacterial species and environmental factors promote MIC? How does fluid flow affect SRB biofilm stability, and how does this depend on the genetic diversity of that species? Answers to questions such as these will require new in situ assays of biotic corrosion, and will allow for improved predictions of environmental conditions susceptible to MIC. Together, these will facilitate cost-effective corrosion-risk monitoring. By approaching this challenging problem of MIC from both a materials engineering and a biological perspective, quantitative insights at the cell-material interface are poised to offer new understanding and mitigation strategies.