Evaluation of microbial corrosion in biofuel storage tanks using split-chamber zero resistance ammetry

Split-chamber zero resistance ammetry (SC-ZRA) was used to study microbiologically influenced corrosion by aerobic chemoorganotrophic microeukaryotes isolated from biodiesel storage tanks. The magnitude and direction of electric current were measured between two shorted carbon steel electrodes, which were deployed in separate chambers connected by a salt bridge (via a SC-ZRA assembly). This approach permitted rapid screening for the corrosive activity of these previously understudied microeukaryotes. During this study, two previously understudied microeukaryotes (Byssochlamys sp. SW2 and Yarrowia lipolytica) showed increased biomass, an increase in electrochemical signal (current), and a corresponding increase in corrosion rate (weight loss). However, other previously understudied microeukaryote (Wickerhammomyces sp. SE3) showed an increase in biomass without an increase in electrochemical signal and minimal corrosion rate, indicating that the SC-ZRA technique can screen for the corrosive activity of a microorganism, regardless of overall microbial activity. This technique could be used to quickly assess the corrosive potential for a range of previously understudied microorganisms.


Introduction
Due to the United States Department of Defense Energy Initiative, the use of biodiesel blends in the United States military has increased [1,2]. Biodiesels are animal fat-or plant oil-derived fatty acid methyl esters (FAMEs) that can be mixed with petroleum diesel in blends up to 20% and used in diesel engines [3]. This blend, B20 biodiesel, is a popular blend used by the United States military [1,[4][5][6]. Despite the widespread use of the fuel, negative consequences of the switch to biodiesel have been observed [7]. The hygroscopic biodiesel absorbs water in environments with high humidity or when water is introduced by leakage or as ballast [8][9][10][11][12]. The interface that develops when water and biodiesel separate can also host extensive growth of microorganisms that use FAMEs as an electron donor and carbon source [7,8,[13][14][15][16][17][18][19].
Therefore, the biodiesel-water interface experiences extensive microbiologically influenced corrosion (MIC) as FAME biodegradation yields organic acid metabolites that acidify the water layer and enhance anodic reactions on metal surfaces and degrade passive films [14,[20][21][22]. A recent yearlong study of B20 storage tanks at two locations identified several populations of microorganisms associated with fuel fouling and MIC [23,24]. The study found a variety of bacterial taxa including Acetobacteraceae, Clostridiaceae, and Proterobacteria [23]. Abundant microeukaryotes were also found, including a filamentous fungus within the family Trichocomacea and the Saccharomycetacea family of yeasts [23,24]. However, without a way to quickly screen the large range of microorganisms, it is difficult to determine the risk of MIC caused by an individual type or consortium of microorganisms in the storage tanks [23]. As such, identification, monitoring, and prediction of microbiologically induced corrosion at the diesel-water interface in B20 storage tanks have proven difficult [16,25,26].
Electrochemical characterization of the biodiesel-water interface can be challenging because the biodiesel acts as an insulator and the liquid-liquid interface is often mechanically unstable [27,28]. Insight into electrochemical processes at diesel-water interfaces was presented by Wang et al. [29], who used carbon steel electrode arrays to determine that localized corrosion can occur near these aqueous/ non-aqueous interfaces with anodic processes on waterexposed carbon steel and cathodic processes on biodieselexposed steel. Further work by Miller et al. [30] showed the role of microbiological activities in the development of these anodic and cathodic regions of the metal surface near aqueous/non-aqueous interfaces. Here, a combination of aqueous phase acidification during FAME metabolism and penetration of water containing hyphae into the non-aqueous biodiesel layer induced the development of anodic and cathodic regions of the carbon steel that enhanced corrosion [30].
We suggest that a split-chamber zero resistance ammetry (SC-ZRA)-based approach, similar to EN used in the study of coating and hydrogen diffusion, can overcome the limitations to MIC monitoring described above and serve as a screening system to determine the risk of MIC associated with certain microorganisms or groups of microorganisms. Previous work using a split-chamber approach to assess MIC was used by Daumus [33] for the study of stainless steel corrosion in the presence of sulfate-reducing bacteria and subsequently used by Miller et al. [30,34,35] to evaluate MIC under aerobic, nitrate-reducing, and Fe(III)-reducing conditions. In this approach, two identical electrochemical cells (chambers) are separated by an ion-transport membrane. Each chamber contains an identical electrode of the same material which are electrically connected through a zero resistance ammeter (ZRA). When one of the chambers is inoculated with microorganisms, the galvanic current between the two electrodes is measured through the ZRA. This configuration mimics the microbiologically induced development of localized anodic and cathodic patches on a metal surface that leads to corrosion. The flow of electrons (difference in corrosion current between the two chambers) depends solely on the activities of microorganisms in one of the two chambers, so their influence on corrosion, as well as its extent, can be quantified. In the current work, we applied this ZRA-based approach to determine the extents of corrosion induced by organisms recovered from the aqueous-non-aqueous interface of biodiesel storage tanks [32].

Cultivation of microorganisms
Microorganisms used in these studies were obtained from Dr. Bradley Stevenson (Byssochlamys sp. SW2) and Ms. Audra Crouch (Yarrowia lipolytica, and Wickerhammomyces sp. SE3). They were isolated from biodiesel storage tanks and were selected for the current studies based on their abundances in the diesel-water interface in storage tanks and on their ability to degrade FAMEs and acidify the B20/ water interface [24,36]. Not all microorganisms selected were shown to cause extensive corrosion in batch experiments [23], to give a range of MIC capacities to test our screening technique.
Y. lipolytica and Wickerhammomyces sp. SE3 were routinely grown on tryptic soy agar or broth (TSA and TSB, respectively), consisting of tryptic soy powder (20 g/L) and bacto agar (15 g/L) as a solidifying agent for TSA. In preparation for corrosion experiments, Y. lipolytica or Wickerhammomyces sp. SE3 were grown to late log phase at room temperature with shaking at 120 rpm. Cells were harvested by centrifugation, washed with ASW, and resuspended in ASW. Y. lipolytica and Wickerhammomyces sp. SE3 growth was determined by measuring optical density of the cultures at 600 nm (OD 600 ) in a Helios UV/VIS spectrophotometer (Thermo-Fisher Scientific, Waltham, MA), and cultures were used to inoculate (10%) corrosion incubations upon reaching and OD 600 of 0.9.

Electrode sterilization
In preparation for SC-ZRA and batch experiments, carbon steel (UNS G10180) coupons were polished using progressively finer SiC papers with 240, 320, 400, and 600 grits, as described in ASTM standard E1558 [39]. Flat steel coupons (surface area of 6.48 cm 2 ) were used for batch incubations and cylindrical coupons (surface area of 5.52 cm 2 ) were used for SC-ZRA experiments. The coupons were placed at the B20/ASW interface with half of the surface area (~ 3.24 cm 2 for batch and ~ 2.7 cm 2 for SC-ZRA) placed in the ASW. After obtaining an initial mass, the carbon steel coupons were placed in either the SC-ZRA or batch set-up and air was replaced with the N 2 prior to oven sterilization at 160 °C for 4 h. This approach sterilizes the metal and minimizes alteration of the metal surface, which occurs during other sterilization approaches (e.g., autoclaving or ethanol treatment) [40].

SC-ZRA incubations
To assemble SC-ZRA incubations, two glass chambers were filled with 250 mL of filter-sterilized ASW and B20 biodiesel (referred to hereafter as B20), as shown in Fig. 1, at a ratio of 1:1 by volume (ASW: B20). B20 was obtained from Santmyer Oil Company in Dalton, OH, USA. The chambers were connected with a cation exchange membrane (CMI-7000S; Membranes International Inc.; Ringwood, NJ) that was primed in sterile 5% NaCl solution at 40 °C for 24 h prior to use. Polished and sterilized working electrodes (referred to as WE1 and WE2) were included in the two chambers, with a saturated calomel electrode (SCE) reference electrode placed in the chamber containing WE1. Carbon steel WEs were positioned at the ASW-B20 interface, such that half of each WE was in contact with ASW and half was in contact with B20. WE1-containing chambers were inoculated with microorganisms as described above. All experiments were conducted at room temperature.
Current and potential were measured using a Bio-Logic VSP 300 potentiostat/galvanostat in galvanic corrosion mode (measurements collected every 600 s) with electrodes Fig. 1 Top: Schematic of the electron flow during corrosion in the ZRA set-up. WE1 is inoculated and undergoing more extensive corrosion. Electrons formed during iron dissolution flow through the ZRA and participate in the oxygen reduction reaction on WE2. Bottom: Split-chamber zero resistance ammetry set-up. Cylindrical electrodes are placed at the B20/ water interface configured so that a positive current represented electron transfer from WE1 to WE2. Galvanic current measurements were collected at five-minute intervals during operation of the SC-ZRA experiments. During the experiments, samples from the aqueous layer were periodically collected and filter was sterilized (0.2 µm). The pH was measured in the aqueous layer. Aqueous samples were also acidified with 0.008 N H 2 SO 4 in preparation for the measurement of acetate concentration (described below). At the conclusions of the SC-ZRA experiments, steel coupons were analyzed for mass loss and pitting corrosion as described below. Biomass was determined gravimetrically based on mass retained during filtration of ASW-B20 mixtures.

Batch incubations
Batch incubations were performed to measure the corrosion rate of coupons that are not galvanically coupled to ensure that corrosion was not enhanced by connecting WE1 and WE2 through a ZRA. A 500 mL single-cell chamber was used for batch incubations, which contained 150 mL of ASW and 150 mL of B20. Sterile carbon steel coupons were hung on a four-armed glass apparatus, so that each coupon spanned the ASW-B20 interface. Incubations with microorganisms were inoculated as described above. Aqueous samples from batch experiments were collected as described above for SC-ZRA experiments, and pH and acetate were measured as described below. At the conclusion of the incubations, coupons were recovered and analyzed for mass loss and pitting corrosion as described below. Batch experiments were conducted in triplicate and incubated at room temperature.

Analytical techniques
Corrosion rates were determined by mass loss analysis method as described in ASTM method G01-03 [41]. Briefly, coupons were rinsed in deionized water, wire brushed, and immersed in Clarke's reagent (12.1 M HCl, 20 g/L antimony trioxide, and 50 g/L stannous chloride) to remove surface oxides. Coupons were then rinsed with DI water, dried, and weighed. The Clarke's reagent wash, DI water wash, and weighing were repeated until no mass was lost between wash cycles, indicating the removal of the entire oxide layer. Acetate concentration was measured by high-performance liquid chromatography (HPLC) utilizing an Agilent 1200 system (Agilent Technologies Inc., Santa Clara, CA) with a Aminex HPX-87H column (300 mm by 7.8 mm; Bio-Rad Laboratories Inc., Hercules, CA) and UV (230 nm) detector. The mobile phase was a 0.008 N H 2 SO 4 solution run at a flow rate of 0.6 mL/min. In preparation for analysis of pitting corrosion, steel electrodes were treated as described in ASTM G01-03 [41] prior to surface characterization. Carbon steel electrode surface and pitting analysis was characterized using a Keyence VK-X250 three-dimensional (3D) laser microscope (Keyence Corp., Osaka, Japan), and data were analyzed using Keyence VK-X MultiFile analyzer software. Surface profiles were generated using the Keyence laser microscope at 10 × objective magnification. Each surface profile generated 3D images with areas of approximately 1.56 mm 2 . Pits were determined by the software (VK-X250 multifile analyzer software) using a threshold of 10 µm for minimum pit depth. Pits were counted and normalized by dividing the total surface area examined.

Results and discussion
To screen for MIC by microorganisms isolated from B20 storage tanks, a series of SC-ZRA incubations were conducted with carbon steel working electrodes placed at the B20-water interface in each chamber. The SC-ZRA experiments were inoculated with a variety of microorganisms previously isolated from B20 storage tanks [24,36]. Little current was observed between electrodes in uninoculated controls (Fig. 2a) and both electrodes corroded at approximately equal rates (2.86 vs. 2.26 mpy) (Table 1), ASW pH decreased slightly from 7.19 to 5.75 in uninoculated controls, likely due to partitioning of the relatively soluble FAMEs into the aqueous layer [42,43]. Laser microscopy and SEM were performed on the surface of the coupon. Limited pitting (~ 14 pits/mm 2 ) was observed for uninoculated controls (Table 2) indicating the uninoculated controls are undergoing uniform corrosion. These observations indicate that the WE1 and WE2 electrodes were undergoing uniform corrosion at approximately equal rates in uninoculated controls due to the acidification of the ASW.
In incubations containing Y. lipolytica or Byssochlamys sp. SW2, positive current developed within three days of inoculation, as pH decreased and acetate concentration increased, indicating B20-fatty acid metabolism (Fig. 2c). Y. lipolytica or Byssochlamys sp. SW2 biomass also increased in these incubations ( Table 1). The positive current indicated electron transfer from the inoculated chamber (WE1) to the sterile chamber (WE2), as depicted in Fig. 1. Given the high O 2 concentration that biodiesel can accommodate [3,10,44] and the relatively acidic conditions induced by microbial activities (decrease in pH to ~ 3.5), the flow of electrons from inoculated to sterile chambers (positive current) indicates that the activities of Y. lipolytica or Byssochlamys sp. SW2 induce anodic conditions on WE1 and the half-reaction: While the abiotic half-reaction:  Was occurring on WE2, supporting mass loss on WE1. These patterns of current, metabolism, and corrosion are consistent with previous batch experiments examining corrosion by Byssochlamys sp. SW2 and Y. lipolytica and other organisms [30,34,35,45]. While the current direction could be interpreted as electron transfer in a microeukaryotic bioelectrochemical system (i.e., microbial fuel cell [46,47]), the microbial fuel cell systems deploy inert graphite electrodes. When used for MIC characterization, the SC-ZRA described in this work uses reactive carbon steel electrodes. Indeed, when positive current was detected, greater corrosion was detected on WE1 (Table 1), which is consistent with the R1 and R2 redox couples, as well as previous work to characterize MIC using SC-ZRA measurements [30,34,35]. SEM (Fig. 3) and laser microscopy (Fig. 4) were used to analyze the surface of coupons exposed to Byssochlamys sp. SW2 and Y. lipolytica. SEM images show biofilm formation, especially at the B20/ASW interface, for both Byssochlamys sp. SW2 and Y. lipolytica. Additionally, laser microscopy images show that incubations with Byssochlamys sp. SW2 and Y. lipolytica show enhanced pitting compared to sterile controls with pitting densities  In contrast to SC-ZRA incubations containing Y. lipolytica or Byssochlamys sp. SW2, minimal current was observed in incubations containing Wickerhammomyces sp. SE3, despite similar patterns of acetate production, pH change (final pH of 4.32), and growth ( Fig. 2 and Table 1). The SC-ZRA incubations containing Wickerhammomyces sp. SE3 suggest that microbial metabolism does not always correspond to development of current. In fact, corrosion rates, current, and pitting density in incubations containing Wickerhammomyces sp. SE3 were similar to the uninoculated control ( Fig. 2a; Table 1), even though the incubations with Wickerhammomyces sp. SE3 showed microbial activity, evidenced by pH change, acetate and biomass accumulation, and FAME depletion (SI Fig. 1).
In a previous yearlong monitoring study, Wickerhammomyces sp. SE3 and Byssochlamys sp. SW2 were frequently found together in B20 storage tanks [23,24,36]. When a consortium of Wickerhammomyces sp. SE3 and Byssochlamys sp. SW2 was added to SC-ZRA incubations, the development of current was delayed in comparison with the incubation that included Byssochlamys sp. SW2 alone (Fig. 2a). This delay could be due to a rapid initial growth and metabolism of Wickerhammomyces sp. SE3 in comparison with Byssochlamys sp. SW2 [16,45], whereby initially lower current is a reflection of the activities of Wickerhammomyces sp. SE3 alone, and the increase in current near the end of the incubation could be indicative of Byssochlamys sp. SW2 activity (Fig. 2a) and associated MIC. Indeed, similar WE mass loss patterns were observed in the SC-ZRA incubation containing Wickerhammomyces sp. SE3 and Byssochlamys sp. SW2 and the incubation that contained only Byssochlamys sp. SW2 (Table 1). Additionally, enhanced pitting occurred (pitting density of 30 pits/mm 2 ) in incubations containing both Wickerhammomyces sp. SE3 and Byssochlamys sp. SW2.
In batch incubations, metabolic patterns (Supplemental Information Tables SI1 and SI2) and corrosion rates were similar to those observed in SC-ZRA incubations. Higher corrosion rates were observed in batch incubations containing Y. lipolytica or Byssochlamys sp. SW2 compared to incubations with Wickerhammomyces sp. SE3 and sterile controls (Supplemental Information Tables SI1 and SI2),  [36], further highlighting that microbial growth is not always an indication for the risk of MIC.

Conclusion
We have presented a screening technique to study the risk of MIC caused by previously, unstudied microorganisms or group of microorganisms in complex environments like biodiesel-water interfaces. Unlike previous electrochemical techniques ( Table 2) which typically monitor for biofilm formation, our SC-ZRA technique is able to differentiate between abiotic and biotic corrosion and quickly assess the ability of these microorganisms to induce corrosion, independent of the extent of biofilm formation of biomass accumulation. Using this technique, microbial activities leading to corrosion can be assessed, rather than simply relying on the presence and/or abundance of microbial biomass, and proper mitigation techniques deployed [48][49][50][51][52]. This is the first dual-chamber technique to be tested with real-world samples and a consortium of microorganisms. Additionally, unlike other dual-chamber techniques, the results match those found in field experiments (Table 3) [36]. Future work will be done to modify and validate the laboratory-scale technique into a test kit for use in the field.