Sensitivity to Oxygen in Microbial Electrochemical Systems Biofilms.

The formation and bioelectric performance of anode biofilms in microbial electrochemical systems (MESs) are sensitive to oxygen. Investigating the temporal-spatial structure of anode biofilms will help elucidate the interfaces between oxygen and bacteria, thereby facilitating the applications of MESs in wastewater treatment and energy recovery. Here, use of optical coherence tomography, frozen sections, and a microsensor revealed that the aerobic biofilms exhibited a multilayered sandwich structure with a sparse gap between the aerobe- and amphimicrobe-enriched outer layer and the dense exoelectrogen-enriched inner layer, whereas the anaerobic biofilm consisted of only a single dense layer. Our results showed that the inner layer of aerobic anode biofilms performed electricity generation, whereas the outer layer only consumed oxygen. In this case, electron donor diffusion through the outer layer became the limiting factor in electricity generation by the bioanode. Consequently, as the anode biofilms matured, current generation decreased.

Introduction

Microbial electrochemical systems (MESs) represent promising technologies for wastewater treatment and energy recovery by using anode-respiring bacteria (ARB) as a biocatalyst to oxidize organic waste (Logan et al., 2008, Liu et al., 2005). Oxygen is a common and essential metabolic electron acceptor present in practical applications of MESs. The effect of oxygen on anode biofilms has drawn great attention because of its wide presence and selective toxicity to ARB. In the cultivation of pure-culture ARB, oxygen showed toxic effects on anaerobic ARB such as Geobacter (Caccavo et al., 1994, Lovley et al., 1993, Lin et al., 2004), but a positive effect on the growth and extracellular electron transfer (EET) of aerobic or facultative ARB such as Shewanella and Pseudomonas (Rabaey et al., 2005, TerAvest et al., 2014). In the cultivation of mixed-culture ARB, the effects of oxygen on the formation and current generation of mixed-culture anode were more complicated (Cheng et al., 2006, Oh et al., 2009, Fan et al., 2012, Quan et al., 2012, Quan et al., 2013). Directly pumping oxygen into an anode chamber led to approximately 100% voltage loss in a two-chambered microbial fuel cell (MFC) with cation exchange membrane as a separator (Oh et al., 2009). Due to the increasing oxygen crossover from the air cathode to the anode, decreasing the electrode spacing from 2 to 1 cm resulted in a considerable drop in the power density of the air cathode single-chamber MFC from 811 to 423 mW/m2 (Cheng et al., 2006). However, some studies have reported a minor effect of oxygen on the anode biofilms, and the power density of a compact MFC with a J-cloth separator between the anode and cathode, which could start up in 5 days, decreased only from 1,700 to 1,400–1,600 W/m3 when oxygen was directly pumped into the MFC chamber (Fan et al., 2012). Furthermore, Quan et al. reported that an anode biofilm could be formed under an air flow rate of 1.5 L/min and that the anaerobic-enriched biofilm showed a power generation loss of only 26% under aeration (Quan et al., 2012, Quan et al., 2013).

These discrepant results likely indicate that the anode biofilms formed under different conditions have different structures and oxygen sensitivities. Studying the spatiotemporal structure of anode biofilm during MFC start-up will facilitate a better understanding of the interfaces between oxygen and bacteria. To date, temporal-spatial structure investigations of anode biofilms have mainly involved confocal laser scanning microscopy (CLSM) and nuclear magnetic resonance (NMR) (Franks et al., 2009, Sun et al., 2015, Renslow et al., 2013a, Renslow et al., 2013b). CLSM is an optical sectioning technique that acquires fluorescence signals within the sample. However, the toxicity of fluorochromes to microorganisms limits the utility of CLSM in temporal studies of anode biofilm formation. Moreover, the inadequate permeation of fluorochromes limits the utility of CLSM for dense or thick biofilm observations. Fortunately, NMR can be employed to circumvent this problem. Renslow et al. observed anode biofilm growth and analyzed electron transfer and substrate diffusion in biofilms by NMR. However, the physical properties of these films cannot be correlated with the microbial community by NMR, and the low resolution of NMR (>10 μm) is another disadvantage when examining anode biofilms (Renslow et al., 2013a, Renslow et al., 2013b). Available analysis methods have a limited ability to correlate the anode biofilm structure to its performance. Therefore new methods for temporal-spatial studies of anode biofilms and their local structures must be developed.

Optical coherence tomography (OCT) is a new non-invasive, non-destructive, high-spatial-resolution, and real-time optical measurement technique for ophthalmologic imaging and material nondestructive testing (Huang et al., 1991, Xi et al., 2006). Compared with other morphological analysis technologies, such as scanning electron microscopy, CLSM, and magnetic resonance imaging (Franks et al., 2009, Renslow et al., 2013a, Renslow et al., 2013b), OCT allows online morphological observations and does not damage the sample; therefore, it should be suitable for investigating anode biofilm formation online.

Tissue sections have been widely employed in pathology for structural analysis. These sections can be further analyzed by microscopy or extraction techniques to evaluate the spatial distribution of microbial communities, as successfully demonstrated using aerobic and anaerobic granular sludge samples (Lemaire et al., 2008). In 2010, Franks used sections in an MES for the first and, to our knowledge, only time (Franks et al., 2010). Specifically, microarray analysis of prepared sections successfully revealed that most genes had higher transcript abundances in the inner layer (0–20 μm from anode surface of the total 60-μm biofilm) of a Geobacter sulfurreducens biofilm. However, because a microtome was used to prepare the section, it had a thickness of only 100 nm and an area of 25 mm2; thus it was only suitable for gene expression analysis. In this work, frozen anode biofilm sections with suitable size were prepared for genome analysis to investigate the structure of biofilms, especially mixed-culture biofilms.

In this study, the effects of oxygen on the formation and spatial structure of a mixed-culture anode biofilm were investigated with flowing nitrogen-oxygen mixtures (oxygen partial pressure of 0%–20%) over the anode surface. OCT was used to analyze anode biofilm formation in aerobic or anaerobic environments. Frozen sections coupled with high-throughput 16S rRNA sequencing were used to analyze the spatial community structure along the depth of the anode biofilm. The oxygen distribution along the depth of the biofilms was measured by a microelectrode. The electrochemical performance of the anode was evaluated by using linear sweep voltammetry (LSV) and electrochemical impedance spectroscopy (EIS).

Results and Discussion

MFC Startup

All MFCs were successfully started under 0%–20% oxygen partial pressure (MFC configuration shown in Figure S1). The start-up process of all the MFCs went through three stages: the lag sage (referred to as stage A), in which the cell voltage varied by only a few millivolts; the rapid growth stage (referred to as stage B), in which the cell voltage rapidly increased to 360 mV in 1 day; and the stationary stage (referred to as stage C), in which the cell voltage maintained a constant value of 365 ± 5 mV, as shown in Figure 1. The lag time increased as the oxygen partial pressure increased from 0% to 20%, corresponding to an increase in the lag time of approximately 1 day per 5% increase in the oxygen partial pressure. A similar effect of oxygen on the lag stage of MFC start-up was reported by Quan et al. (Quan et al., 2013) and Choi et al. (Choi and Chae, 2013). In their studies, the lag stage extended from several hours to several tens of hours when oxygen was introduced. However, in stage B, the cell voltages of all MFCs rapidly increased to stable voltages in a short time. The oxygen concentration had little effect on the time of stage B for OPP-0, OPP-5, and OPP-10 and only slightly lengthened the time for OPP-15 and OPP-20. The stable voltages of the MFCs in stage C decreased from 365 ± 5 mV to 310 ± 5 mV as the oxygen partial pressure increased from 0% to 20%.

Figure 1

Electricity Generation of the Anode Biofilms at Different Oxygen Partial Pressures during Startup

A, the lag stage; B, the rapid growth stage; and C, the stationary stage.

OCT Observations of Anode Biofilm Formation

During start-up, different formation processes were confirmed in the anode biofilms via in situ observations of anode biofilm formation using OCT. OCT images of the OPP-0 and OPP-20 anode biofilms collected at stages A (60 h for OPP-0 and 108 h for OPP-20), B (85 h for OPP-0 and 160 h for OPP-20), and C (180 h for OPP-0 and 250 h for OPP-20) are shown in Figure 2. For OPP-0, colonies were scattered across the electrode surface in stage A (Figure 2A), the colonies rapidly developed in the form of a hemisphere accumulated on the surface in stage B (Figure 2B), whereas at stage C, the colonies developed into a continuous biofilm with a thickness of 80 μm that covered the whole surface (Figure 2C). However, for OPP-20, a biofilm with a thickness of over 300 μm rapidly formed on the whole electrode surface in stage A (Figure 2D), whereas at stage B, the anode biofilm developed into a stratified structure with a thick outer layer (300–400 μm) and a thin inner layer (20 μm) separated by a gap (Figure 2E). At stage C, the outer layer maintained a constant thickness of 300–400 μm, whereas the inner layer increased to 30–50 μm in thickness. In addition, a thorn-like structure formed on the outer layer along the direction of the air flow (white arrow) (Figure 2F), which was probably due to biofilm detachment. When a small fragment was sloughed off the biofilm, the rest of the biofilm was lifted by the air flow to form a surface cluster.

Figure 2

3D OCT Images Showing Anode Biofilm Formation during Startup

(A–F) (A) Lag, (B) rapid growth, and (C) stationary stages for the OPP-0 biofilm; (D) lag, (E) rapid growth, and (F) stationary stages for the OPP-20 biofilm. The dashed line in (F) indicates the sparse gap between the dense inner layer and the dense outer layer. The white arrows in (C) and (F) indicate the gas flow direction. The scale bar represents 200 μm.

Spatial Microbial Community Structures of Anode Biofilms

Figure 3 shows the distribution of the predominant bacterial genera (details shown in Figure S3) along the anode biofilm thickness. The OPP-0 anode biofilm community mainly consisted of anaerobic Geobacter species (GenBank: FW306010.1) (Figure 3A). The relative abundances of the Geobacter species decreased along the biofilm thickness direction from the interior (electrode surface side, 80%) to the surface (solution side, 22%), whereas the relative abundances of the other predominant bacterial species gradually increased along the biofilm thickness direction from the interior to the surface. For example, the relative abundances of the Azoarcus (GenBank: AF011343.1), Dechloromonas (GenBank: AY945919.1), Aminiphilus (GenBank: GQ181552.1), and Arcobacter (GenBank: FJ968634.1, GenBank: FJ968635.1) species increased from 2% to 14%, from 3% to 10%, from 3% to 7%, and from 1% to 8%, respectively. Among them, Azoarcus species are facultative anaerobic bacteria owing to their nitrogen fixation abilities (Reinhold-hurek et al., 1993), both Dechloromonas (Wolterink et al., 2005) and Aminiphilus (Diaz et al., 2007) are anaerobic bacteria, and Arcobacter (Fedorovich et al., 2009) are facultative aerobic bacteria (microaerobic, favoring growth under an oxygen partial pressure of 3%–10%). Thus the OPP-0 biofilm consisted of only one layer of predominantly anaerobic and facultative anaerobic bacteria, which is consistent with the OCT observation.

Figure 3

Spatial Distributions of the Anode Biofilm Microbial Communities

(A) OPP-0, (B) OPP-5, and (C) OPP-15 samples. See also Figure S3. A distance of 0 μm corresponds to the graphite plate surface and the last obtained frozen section.

When oxygen was introduced into the system, no new species of electrochemical active bacteria were found in all the anode biofilms, and significant changes were not observed in other dominant bacterial species or their relative abundances in the interior biofilm, but significant changes were observed in the dominant bacterial species and their relative abundances along the thickness direction from the interior to the surface. In the interior biofilm, the dominant bacterial species and their relative abundances were similar. For example, the relative abundances of the dominant bacterial species Geobacter, Dechloromonas, Desulfovibrio, and Aminiphilus in the electrode surface of OPP-0 (80%, 2.7%, 1.7%, and 3%) and OPP-5 (77%, 1.3%, 1.6%, and 3.3%) were quite similar. However, along the thickness direction from the interior to the surface, for the OPP-5 anode biofilm, the relative abundances of Geobacter species (GenBank: FW306010.1) decreased from 77% (5 μm) to 56% (50 μm) and then to 1.5% ± 1% (>100 μm) (Figure 3B) along the thickness direction of the biofilm, whereas the relative abundances of the Arcobacter (GenBank: FJ968634.1, GenBank: FJ968635.1), Acinetobacter (GenBank: HQ670708.1), Pseudomonas (GenBank: HM451438.1), and Azoarcus (GenBank: AF011343.1) species were 1% at a biofilm thickness of 5 μm and increased to 3.5%, 13%, 2%, and 3% at a thickness of 50 μm, respectively, and then to 29% ± 2%, 29% ± 1%, 7% ± 2%, and 11% ± 2% at a thickness of over 100 μm, respectively. Among them, Azoarcus and Arcobacter (which favor growth under an oxygen partial pressure of 3%–10%) are facultative bacteria, and Acinetobacter (Juni, 1972) and Pseudomonas (Palleroni, 2010) are generally strictly aerobic bacteria. When the oxygen partial pressure was increased to 15%, the relative abundances of the Geobacter species in the biofilm decreased from 67% to 0% along the thickness direction from 0 to 50 μm and then remained at 0 over 50 μm (Figure 3C), whereas the relative abundances of the Pseudomonas species rapidly increased from 13% to 60% along the thickness direction from 0 to 50 μm and then remained at 60% ± 10% over 50 μm. For other predominant bacterial species in the OPP-15 biofilm, the relative abundances along the thickness direction from 0 to 100 μm of the strictly aerobic bacteria Comamonas (Tago and Yokota, 2004) increased from 1% to 6% ± 2% and of the Arcobacter and Acinetobacter species increased from 0% to 3% ± 1% and from 3% to 9% ± 2%, respectively. Both OPP-5 and OPP-15 had similar relative abundances of Azoarcus species (10%–15%) in the outer layer of the biofilm along the thickness direction over 100 μm. Thus the anode biofilm that formed under aerobic conditions presented a dual-layer structure. The anaerobic Geobacter species dominated the inner layer, whereas aerobic genera such as Acinetobacter, Pseudomonas, and Comamonas and facultative genera such as Arcobacter dominated the outer layer. The dual-layer structure was consistent with the OCT observations.

Oxygen Distributions in Anode Biofilms

The dissolved oxygen (DO) distributions in the anode biofilms are shown in Figure 4 (replication shown in Figure S4). For OPP-0, the DO concentration was less than 0.3 mg/L throughout the entire biofilm, which is consistent with the results of previous studies (Chang et al., 2014, Zhang et al., 2013). For the reactors bubbled with oxygen, the DO concentration in the solution corresponds to the saturation oxygen concentration under each oxygen partial pressure, with 1.7 mg/L for OPP-5 and 5.4 mg/L for OPP-15, whereas the DO concentration in the biofilm presented a gradient distribution along the thickness direction. For OPP-5, the DO concentration maintained a constant value of 1.7 mg/L in the solution phase, gradually decreased from 1.7 mg/L at the biofilm surface (360 μm from the electrode surface) to 0 mg/L at a distance of 200 μm from the electrode surface, and then was maintained at 0 mg/L at distances from 200 to 0 μm away from the electrode surface. When the bubbled oxygen partial pressure increased to 15% (OPP-15), the saturated DO concentration in the solution phase increased to 5.4 mg/L, whereas the DO concentration in the biofilm gradually decreased from 5.4 mg/L at the biofilm surface (260 μm from the electrode surface) to 0 mg/L at a distance of 100 μm and was then maintained at 0 mg/L at distances from 100 to 0 μm away from the electrode surface. The thickness of the biofilm in the anaerobic environment (0 mg/L DO) decreased as the bubbling oxygen concentration increased.

Figure 4

Oxygen Distributions in the OPP-0, OPP-5, and OPP-15 Anode Biofilms

The distance 0 μm represents the graphite anode surface. See also Figure S4.

Formation and Function of Dual-Layer Anode Biofilms

During start-up, oxygen greatly extended the lag stage but had a minor effect on the rapid growth stage (Figure 1). The variation in MFCs' start-up was attributed to the effects of oxygen on anode performance rather than cathode performance as the oxygen permeated from anode to cathode was found to have a minor influence on the cathode performance. The cathode potentials of all the MFCs were measured at the three stages (the lag stage, the rapid growth stage, and the stationary stage) of start-up process. The variations of cathode potential were less than 0.005 V for all the MFCs after the anode chamber was aerated with a gas mixture of each oxygen partial pressure. The OCT observations of OPP-20 showed that a 300-μm-thick biofilm formed on the electrode surface in the lag stage when the cell voltage was only a few millivolts (Figure 2D). Interestingly, the anode structure became stratified as the electricity generation rapidly increased. These results most likely indicate that the inner layer performed electricity generation, whereas the outer layer provided a suitable environment for exoelectrogen growth on the electrode surface or electron transfer from the exoelectrogens to the electrode. The difference in the terminal electron acceptor of the outer and inner layers probably resulted in the formation of loose gap between the outer and inner layers. It was previously shown that a Geobacter biofilm had a higher density near the electrode surface when the terminal electron acceptor was the anode (Jain et al., 2011, Renslow et al., 2013b). In contrast, biofilms in which oxygen acts as the terminal electron acceptor have higher densities at the biofilm-solution interface (far from the electrode surface).

Further investigation of the spatial microbial community structures showed that the inner layer was dominated by the anaerobic Geobacter species, whereas the outer layer was dominated by aerobic genera such as Acinetobacter, Pseudomonas, and Comamonas and facultative genera such as Arcobacter. Geobacter species (FW306010.1) present in the inner layer have been reported to have strong electricity generation ability (Bond and Lovley, 2003). The Arcobacter species (FJ968634.1, FJ968635.1) present in the outer layer were also reported to be electrochemically active (Fedorovich et al., 2009), whereas the Acinetobacter (HQ670708.1), Pseudomonas (HM451438.1), and Comamonas (AB021357.1) species were not reported to be electrochemically active. In an aerobic environment, Arcobacter, Acinetobacter, Pseudomonas, and Comamonas are more likely to transfer electron to oxygen rather than to electrode, without electricity generation. Thus the spatial microbial community structures also suggested that the inner layer performed electricity generation, whereas the outer layer consumed oxygen. In this case, an anaerobic environment is most likely necessary for the growth and electricity generation of exoelectrogens. The increasing lag stage with the oxygen partial pressure in the bubbled gas was due to the increased time required to grow the thick aerobic biofilm, which is necessary for the development of an anaerobic environment on the electrode surface.

To further confirm that only the anaerobic bacteria in the inner-layer biofilm were responsible for electricity generation, gentamicin and metronidazole were selected to inhibit the activity of outer- and inner-layer bacteria, respectively. Metronidazole is known for its bactericidal activity, mainly against anaerobic bacteria and methanogenic archaea (Martin et al., 1972, Löfmark et al., 2010, Dione et al., 2015). However, metronidazole has minimal effects against aerobic and facultative aerobic bacteria (Martin et al., 1972, Dione et al., 2015). Gentamicin, in contrast, displays bactericidal activity against gram-negative aerobes, but generally not against anaerobic bacteria (Pacific and Marchini, 2017). The effective antibiotic activity of gentamicin against Acinetobacter (Anstey et al., 1992), Pseudomonas (Dione et al., 2015), and Arcobacter (Córdoba-Calderón et al., 2017) has been verified. As shown in Figure 5, after metronidazole was added to the OPP-5 and OPP-15 MFCs, electricity generation ceased in 10 h and the cell voltage could not be recovered when the anode solution was refreshed. In contrast, when gentamicin was added to OPP-10 and OPP-20, the cell voltages of the MFCs only decreased by less than 10 mV in one cycle and the cycle time increased from 7 h to 12 h. These results further suggest that only the anaerobic bacteria in the inner layer were responsible for electricity generation.

Figure 5

Cell Voltage Evolution over Batch Cycles for Reactors with and without the Addition of Antibiotics

For OPP-5 and OPP-15, 120 mg/L metronidazole was added in cycles 1 and 3. For OPP-10 and OPP-20, 48 mg/L gentamicin was added in cycles 1 and 3. All antibiotics were removed in cycle 2.

Electrochemical Properties of Anode Biofilms

The electrochemical properties of the anode biofilm were evaluated by using the LSV. Figure 6A shows the LSV results with 1.5 g/L acetate; all the anode biofilms showed a significant increase in the current density at −0.48 V (versus Ag/AgCl) and reached a peak current at −0.35 V. For the anode biofilm enriched in the anaerobic environment (OPP-0), the peak current density was 2.44 A/m2, whereas for the anodes enriched in aerobic environments, the peak current densities decreased significantly. The peak current densities of OPP-5, OPP-10, OPP-15, and OPP-20 were 1.34 A/m2, 0.89 A/m2, 0.55 A/m2, and 0.32 A/m2, respectively. A linear relationship was observed between the natural logarithm of the peak current density and the oxygen partial pressure.

Figure 6

Electrochemical Properties of the Anode Biofilms Formed at Different Oxygen Partial Pressures

(A) Currents generated by the anode biofilms formed at different oxygen partial pressures in 50 mM PBS containing 1.5 g/L acetate.

(B) Linear relationships between the natural logarithms of the charge transfer resistance (ln(Rct)) and diffusion resistance (ln(Rd)) and the oxygen partial pressure. Control was a polished graphite plate anode without biofilm. The inset in (A) shows the linear relationship between the natural logarithm of the peak current density (ln(Imax)) and the oxygen partial pressure. The inset in (B) shows the Nyquist plots of the EIS spectra and the equivalent circuit. The error bars represent standard error of duplicate replicates.

The Nyquist plots (Figure 6B) obtained from the EIS test of anode in MFCs was usually fitted using a two-time-constant model, Rs(RctQ1) (RdQ2), for resistance estimation (Hou et al., 2014, Yang et al., 2017). In this model, Rs represents the anode resistance, Rct represents the charge transfer resistance, Rd represents the diffusion resistance, and Q1 and Q2 are the constant phase angle elements. As shown in Figure 6B, the solution resistance Rs of all the MFCs was approximately 19.5 Ω. In contrast, the charge transfer resistance Rct and the diffusion resistance Rd increased with increasing oxygen partial pressure used during anode start-up, and both showed a linear relationship between the natural logarithm value and the oxygen partial pressure. Compared with OPP-0, OPP-20 had Rct values that were 3.5 times higher (252 Ω versus 55.9 Ω) and Rd values that were 249 times higher (50 Ω versus 0.2 Ω).

The EIS results indicated that the decreased electricity generation of the MFCs with aerobic anode biofilms was mainly due to their higher diffusion and charge transfer resistances. Increasing oxygen had a much greater effect on Rd than on Rct, indicating that higher diffusion resistance is the key factor affecting anode performance. The natural logarithm of Rd exhibited a linear relationship with the oxygen partial pressure, indicating that the diffusion of electron donors to or protons from the biofilm was strongly affected. This limited diffusion will strongly inhibit the growth and electrogenesis of the inner layer and cause cell death and biofilm detachment (Renslow et al., 2013a, Sun et al., 2015), thereby leading to a decrease in the inner-layer thickness. The limited growth of exoelectrogens will result in an increase in charge transfer resistance. Regardless of the specific form of EET, which can be either conduction- or diffusion-based (mediated) mechanisms or a combination of mechanisms, the limited growth of exoelectrogens will lead to increased electrical resistance by reducing either the quantity of electron shuttles or the quantity of components that exhibit metal-like conduction (Strycharz-Glaven et al., 2011).

In this study, we revealed for the first time that an anaerobically enriched anode biofilm consisted of a single-layer structure and an aerobically enriched anode biofilm consisted of a dual-layer structure with the inner layer dominated by anaerobic bacteria and the outer layer dominated by aerobic bacteria. These results were confirmed by in situ OCT observations together with high-throughput 16S rRNA sequencing coupled with frozen sections. It was found that the function of the inner layer of aerobically formed anode biofilm was electricity generation, whereas the function of the outer layer was as a barrier for oxygen to realize anaerobic environment for the inner layer. However, thick outer layer will hinder the diffusion of substrate and proton in the biofilm, resulting in the reduction of electricity generation. These results implied that high power density in aerobic condition could be only generated with an oxygen-tolerant anode biofilm with thin outer layer and thick inner layer. This is significantly importance for developing the compacted electrode assembly air cathode MFCs, which had been considered as ideal configuration for high power generation (Fan et al., 2012). Future works should focus on developing methods to regulate and control the formation of anode biofilm with this structure.

In addition, the combined technology is useful for investigation of bioanodes. OCT provided a useful tool for performing in situ observations of the formation process of an anode biofilm and identifying the relationship between biofilm structure and electricity generation. Further studies of biofilm formation and structural changes during the long-term operation of MFCs under different conditions by using OCT coupled with frozen section sequencing will help clarify why an anode biofilm could maintain a constant thickness during long-term operations. This information will facilitate the development of highly aerotolerant anodes for high-performance MFC and multifunctional biofilms for efficiently removing the chemical oxygen demand and other pollutants in wastewaters.

Limitations of the Study

Here, we showed how the presence of oxygen influences biofilm formation on a microbial anode and the consequences in electricity production. Performance of the cell is affected by choices such as high DO and distance of electrodes, which were not explored in the present study. We obtained the results with duplicate tests, which in general were not sufficient to have reproducible results. The two-layer structure of the aerobic anode biofilms provides a model for developing aerotolerant anodes. However, the performance of these anodes should be further improved, possibly by increasing substrate diffusion through them. Further efforts would be focused on the construction of high-performance aerotolerant anodes to facilitate the applications of MESs in wastewater treatment and energy recovery.

Methods

All methods can be found in the accompanying Transparent Methods supplemental file.