Anode Modification with Fe2O3 Affects the Anode Microbiome and Improves Energy Generation in Microbial Fuel Cells Powered by Wastewater

This study investigated how anode electrode modification with iron affects the microbiome and electricity generation of microbial fuel cells (MFCs) fed with municipal wastewater. Doses of 0.0 (control), 0.05, 0.1, 0.2, and 0.4 g Fe2O3 per the total anode electrode area were tested. Fe2O3 doses from 0.05 to 0.2 g improved electricity generation; with a dose of 0.10 g Fe2O3, the cell power was highest (1.39 mW/m2), and the internal resistance was lowest (184.9 Ω). Although acetate was the main source of organics in the municipal wastewater, propionic and valeric acids predominated in the outflows from all MFCs. In addition, Fe-modification stimulated the growth of the extracellular polymer producers Zoogloea sp. and Acidovorax sp., which favored biofilm formation. Electrogenic Geobacter sp. had the highest percent abundance in the anode of the control MFC, which generated the least electricity. However, with 0.05 and 0.10 g Fe2O3 doses, Pseudomonas sp., Oscillochloris sp., and Rhizobium sp. predominated in the anode microbiomes, and with 0.2 and 0.4 g doses, the electrogens Dechloromonas sp. and Desulfobacter sp. predominated. This is the first study to holistically examine how different amounts of Fe on the anode affect electricity generation, the microbiome, and metabolic products in the outflow of MFCs fed with synthetic municipal wastewater.


Introduction
Microbial fuel cells (MFCs) convert energy from biomass into electricity using electroactive bacteria (electrogens, EEs). To date, most research has been conducted at the laboratory scale because of the cells' low power output, making it challenging to transfer MFC technology to commercial-scale systems [1]. As far as the authors know, the largest MFC tested to date had a volume of 1200 L [2].
The most important factor for the efficiency of MFCs is the electrode material. Carbon materials that possess good electrical conductivity, such as carbon felt (CF), carbon fabric, or carbon paper, are most commonly used [3]. In recent years, the focus has been on improving these materials by coating their surfaces with metal oxides, which increases the surface roughness, thereby improving the adhesion of bacteria and increasing electron transfer at the anode [4,5]. Iron oxide, for example, improves the power output of MFCs by stimulating the activity of microorganisms and by improving electrical conductivity. Iron oxide (III) is an attractive material for anode electrode modification because of its properties to improve the surface structure of the anode and generate electrical energy in the MFC. Studies report that modifying the anode electrode with Fe 2 O 3 changes the surface from smooth to rough, improving the microorganisms' attachment [6]. Additionally, the hydrophobic carbon surface can be hydrophilized by coating with Fe 2 O 3 due to the hydrogen bond formed between the Fe 2 O 3 oxygen and the water molecule [7]. The advantages of choosing this the electron transfer resistance. After modification, the power generated in the MFC increased by 385%, 170%, and 130% for the CF-, CC-, and G-based electrodes, respectively. In addition, in the MFCs with modified electrodes, over 80% of organic compounds were removed from the effluent.
Although adding Fe promotes the formation of the anode's biofilm and accelerates MFC start-up, high doses of Fe may reduce cell efficiency. Zheng et al. [21] investigated how different doses of Fe 3 O 4 in a medium affect energy generation and the anodic microbiome in an MFC. A dose of 4.5 g/L provided the highest power (391.11 ± 9.4 mW/m 2 ) and the highest share of electroactive bacteria of the genus Geobacter sp. in the microbiome. In a different study, 100 µM FeSO 4 improved the performance of an MFC so that it produced a maximum voltage of 0.55 V, but higher doses reduced the voltage to 0.47 V. Iron salts at a dose of 200 µM decreased Geobacter sp. abundance from 49.3% to 24.4% [22]. Similarly, another study found that 200 µM Fe(III) increased an MFC's electrochemical activity and current efficiency and decreased the anode overpotential. However, excess amounts of Fe(III) (1000 and 2000 µM) competed with the anode for electrons and inhibited the electrochemical activity of biofilms, thus lowering the power density [23]. Mechanistic investigations showed that Fe 3 O 4 increased the conductivity of the fermented sludge, providing a better conductive environment for the anaerobic microbes. A Fe(II)/Fe(III) redox cycle was present in the fermentation system with Fe 3 O 4 , which likely increased electron transfer (ET) efficiency [24].
An important aspect of the large-scale use of MFC is the durability of the anode electrodes. Since they come into contact with the water environment and microorganisms, which may cause them to swell, materials with hydrophobic properties should be used. The surface of the anode electrodes should be rough enough to increase the adhesion of organisms, but not too rough so as not to cause the growth of pollutants resulting from the prolonged use of the electrodes [25]. Yaquoob et al. [26] studied the durability of composite graphene oxide (GO) electrodes made from biomass and the same GO with TiO 2 . The longevity of the fabricated electrodes was 85 days. The increasing voltage on both electrodes indicated stable colonization of bacterial species on the surface of the anodes. Later, the voltage suddenly dropped due to cell death. The second cycle with fresh inoculum also showed an increasing voltage trend to 170 mV, which was due to the fresh inoculum source and organic substrate enhancing the bacterial respiration process to actively oxidize the substrate. The introduction of TiO 2 nanoparticles increased the lifetime of the composite anode, resulting in better interaction between bacterial cells and anode, smooth anode surface, and higher anode conductivity. In addition, a visualization test was performed after the completion of the reaction (85 days), and the electrodes were found to be mechanically stable and in excellent condition. Gnana Kumar et al. [27] tested a CC composite anode with reduced graphene oxide (rGO), poly(3,4-ethylenedioxythiophene) (PEDOT), and iron oxide (Fe 3 O 4 ) nanorods. The authors analyzed electrode lifetime in an open circuit voltage (OCV) as a function of time under a constant load mode with an external resistance of 510 Ω. Bare CC had a lower lifetime due to the reduction of the exposed surface after repeated cycles, due to the large number of water particles on the surface as well as the bare mass CC. The composite of rGO and Fe 3 O 4 prevented shrinkage of the active sites during repeated cycles and provided rapid stability of rGO/Fe 3 O 4 /CC parameters. Among the tested catalysts, the rGO/PEDOT/Fe 3 O 4 composite was characterized by high physical and electrochemical strength and high electrical conductivity, which effectively prevented the destruction of the electrocatalytic activities due to strong π-π interactions and Fe-S coordination bonds between the active carbon support and the conductive polymer and the conductive polymer and Fe 3 O 4 nanorods, respectively. The maximum OCV of 0.45 V was maintained for 200 h, and the concrete OCV was maintained for 600 h in three cycles.
Introducing Fe into an MFC system, as in the studies cited above, requires regular monitoring of Fe concentration in the effluent and carries the risk of elevated Fe concentrations in the MFC effluents. In contrast, covering the anode with Fe not only ensures direct contact between the microorganisms and Fe, but the fact that the biofilm covers the anode minimizes the risk of Fe contamination of the wastewater.
Therefore, the objective of this study was to holistically determine the influence of an MFC's anode modification with different doses of Fe 2 O 3 on the power generation, the microbiome, and the conversion of organic matter. The use of a low-cost method for anode modification and the use of wastewater make this proposed solution economical, thus increasing its potential for implementation. Furthermore, the molecular results obtained greatly expand the knowledge of which microorganisms play a key role in biofilm formation and electricity generation when Fe is present on the anode.

Experimental Set-Up
The dual-chamber MFCs were made of plexiglass. The anode and cathode chambers had active volumes of 2 L. An 8 × 8 cm Nafion 117 proton-exchange membrane (PEM) (Dupont) was used as a separator between the chambers. Before the membranes were placed in the MFCs, they were soaked in acetone for 15 min, then rinsed in distilled water, soaked in 1M HCl for 30 min, and rinsed again with distilled water. Because of membrane clogging, the membranes were cleaned in 1M HCl and rinsed with distilled water once per week (for details, see Nosek and Cydzik-Kwiatkowska [28]).
The anodes were made of carbon felt (10 × 20 × 0.3 cm) connected to a stainless-steel wire. Before use, the carbon felt was sonicated in an ultrasonic bath (InterSonic, 15 min) to remove impurities. An MFC with an unmodified anode was used as a control (MFC control ). In the remaining MFCs, Fe 2 O 3 was deposited on the anode in doses of 0.05 g (MFC 0.05Fe , 1.25 g Fe 2 O 3 /m 2 ), 0.1 g (MFC 0.1Fe , 2.5 g Fe 2 O 3 /m 2 ), 0.2 g (MFC 0.2Fe , 5 g Fe 2 O 3 /m 2 ), and 0.4 g (MFC 0.4Fe , 10 g Fe 2 O 3 /m 2 ). For deposition, Fe 2 O 3 (Chempur) was suspended in 100 mL of distilled water in an ultrasonic bath for 15 min. Next, the anode was placed in a crystallizer, quenched with a Fe 2 O 3 slurry, autoclaved (121 • C, 1.1 Bar, Classic Prestige Medical 210001), and dried at 80 • C.
The anode chambers were inoculated with 100 mL of a 1:1 (v/v) mixture of fermentation sludge from the municipal wastewater treatment plant in Olsztyn (Poland) and a laboratory methane fermentation reactor. The MFCs were supplied with synthetic municipal wastewater [29], and sodium acetate was used as a source of organics in the amount of 400 mg COD/L. The anode chamber was sealed to prevent air access, and the chamber contents were mixed at a speed of 100 rpm/min. The cathode chamber was aerated by an air diffuser (20 mL/min). The composition of the catholyte was 75 mL of phosphate buffer and 3 g of NaCl in 2 L of distilled water, and it was replaced once per week.
Initially, the MFCs were left for 3 days in open circuit mode to adapt the biomass to the environmental conditions in the MFCs and to support microbial colonization of the electrodes [30]. After 3 days, the contents of the anode chambers were replaced with fresh portions of wastewater. After this time, the MFCs were operated with an external resistance of 1200 Ω. This value was chosen based on preliminary experiments [31]. The operational cycle of the MFCs was 48 h; after this time, half of the anode chamber volume was replaced. The experiment was conducted for 18 cycles (36 days), and stable MFC operation was observed after 5 cycles. In the effluent, the concentration of COD, NH 4 -N, volatile fatty acids (VFAs) [32], pH, and alkalinity (TitroLine, Donserv) were determined. The VFA composition was determined using a Varian CP-3800 chromatograph [33]. For spectroscopic characterization of the surfaces of pure carbon felt after sonication in the ultrasound batch and carbon felts modified with different doses of Fe 2 O 3 , a Quanta FEG 250 Scanning Electron Microscope (SEM) equipped with Bruker XFlash 6010 Energy-dispersive X-ray spectrometer (EDX) was used.

Electrochemical Analyses
The polarization and power curves were determined according to Watson and Logan [34] using a True-RMS multimeter, varying the external resistance of the cell in the range of 75-7200 Ω. Voltage changes were recorded every minute using a 6600 Counts PC-LINK data acquisition unit. The current was calculated from the external resistance using Ohm's law. Cyclic voltammetry (CV) was performed with a three-electrode system: an anode as a working electrode, a platinum countercurrent electrode, and an Ag/AgCl reference electrode with a constant potential of 0.197 mV (Gammry Instrument Interface 1010E). CV measurements were conducted at a scan rate of 20 mV/s. The electrochemical behavior of the biofilm-anode system was tested in the 14th cycle of MFC operation, 25 h after reactor feeding (stable current generation) with the catholyte present in the cathode chamber. The MFCs were operated, and the measurements were performed at room temperature.

Molecular Analyses
Genomic DNA was extracted from 100 µg of inoculum and biomass from the anode surface using a FastDNA SPIN Kit for Soil (MP Biomedicals). Samples from the anode surface were collected in the experiment's 3rd, 10th, and 18th cycles. The purity and concentration of the DNA were measured using a NanoDrop spectrometer (Thermo Scientific). The DNA was amplified using a 515F/806R primer set (5 -GTGCCAGCMGCCGCGGTAA-3 /5 -GGACTACHVGGGTWTCTAAT-3') targeting the hypervariable V4 region of bacterial and archaeal 16S rDNA genes [35]. The amplicons were sequenced using the MiSeq platform (Illumina) at Research and Testing Laboratory (USA). The reads were analyzed bioinformatically [36] and deposited in the NCBI Sequence Read Archive (BioProject PRJNA822890).

Statistical Analyses
The results from each MFC's last five cycles of operation were statistically analyzed (p < 0.05 considered significant, Statistica 13.3, StatSoft). One-way analysis of variance (ANOVA) was used, followed by Tukey's test (HSD). For statistical and metagenomic analysis of microbiome data, MicrobiomeAnalyst [37,38] was used (p < 0.05). Due to the fact that, in complex microbial communities, bacteria with a low abundance may be of great importance, the number of reads was not normalized before the calculation of diversity indices [39].

Analysis of the Electrode Surface
The photos show a pure carbon felt that was subjected to sonication ( Figure 1a) and Fe 2 O 3 -modified anodes (Figure 1b-e). The mass fractions of iron that were detected (by means of a combined SEM/EDX methodology) on the electrodes were 5.5%, 6.9%, 10.7%, and 14.2% when Fe 2 O 3 was dosed on the carbon felt in amounts of 0.05, 0.1, 0.2, and 0.4 g, respectively. Small differences in weight percentages of Fe between electrodes may be caused by limitations in the SEM/EDX technique, as this analysis shows only information on the surface of the electrode. Thus, this method was used as a qualitative confirmation of the Fe deposition, not the quantitative one. Analysis of SEM shows that iron has been deposited on the GF anode electrodes. The photos (Figure 1b-e and Figure S1) show Fe 2 O 3 particles deposited on GF fibers. Compared to the bare electrode (Figure 1a), the fibers have a rougher surface, which significantly increases the specific surface area of the electrodes [40] and enhances and prolongs the bioadhesion of microorganisms to the surface of modified anode materials [41].

Electricity Generation
In all tested MFCs, regardless of the iron dose, the voltage was generated from the first cycle of operation. The highest voltages were obtained immediately after the addition of the substrate, which may have resulted from rapid proton flow through the membrane, but during the first 2-3 h, the voltages dropped rapidly. In MFC0.05Fe ( Figure 2b) and MFC0.1Fe (Figure 2c), the voltage decreased until c.a. 24 h of the cycle and increased afterward. The highest average voltages in the cycle (125 ± 48.5 mV) were observed in MFC0.05Fe and were about 3 times higher than in the control. The average voltages obtained in MFC0.05Fe and MFC0.1Fe were significantly higher than those in the other MFCs ( Figure S2). In the control, MFC0.2Fe, and MFC0.4Fe, the voltage gradually decreased during the cycle to 20-50 mV at the cycle end ( Figure 2). The voltage was lowest in MFCcontrol, and was also the least stable, as indicated by large standard deviations. Zheng et al. [21] added Fe3O4 to MFC with a medium. Their studies showed that the output voltage at the highest dose of 18 g/L was the lowest of all MFC tested. This suggests that the voltage shifts at higher doses could be due to the accumulation of Fe occupying the attachment site of electroactive bacteria. SEM/EDX analysis showed that the anode electrodes with higher iron doses had greater roughness. High electrode surface roughness increases the likelihood of polymerization and fouling of the electrode surface as well as electrode poisoning after prolonged use [25], which can lead to lower output voltages. High Fe concentrations can have an inhibitory effect on energy generation by reducing substrate biodegradation [23,42] or by increasing internal resistance of MFC, which was confirmed in our studies (internal resistance was higher at doses of 0.2 and 0.4 than at 0.05 and 0.1 g Fe2O3/the entire surface of the electrode). Power output over 18 cycles is shown in Supplementary Materials Figure S3.

Electricity Generation
In all tested MFCs, regardless of the iron dose, the voltage was generated from the first cycle of operation. The highest voltages were obtained immediately after the addition of the substrate, which may have resulted from rapid proton flow through the membrane, but during the first 2-3 h, the voltages dropped rapidly. In MFC 0.05Fe ( Figure 2b) and MFC 0.1Fe (Figure 2c), the voltage decreased until c.a. 24 h of the cycle and increased afterward. The highest average voltages in the cycle (125 ± 48.5 mV) were observed in MFC 0.05Fe and were about 3 times higher than in the control. The average voltages obtained in MFC 0.05Fe and MFC 0.1Fe were significantly higher than those in the other MFCs ( Figure S2). In the control, MFC 0.2Fe , and MFC 0.4Fe , the voltage gradually decreased during the cycle to 20-50 mV at the cycle end ( Figure 2). The voltage was lowest in MFC control , and was also the least stable, as indicated by large standard deviations. Zheng et al. [21] added Fe 3 O 4 to MFC with a medium. Their studies showed that the output voltage at the highest dose of 18 g/L was the lowest of all MFC tested. This suggests that the voltage shifts at higher doses could be due to the accumulation of Fe occupying the attachment site of electroactive bacteria. SEM/EDX analysis showed that the anode electrodes with higher iron doses had greater roughness. High electrode surface roughness increases the likelihood of polymerization and fouling of the electrode surface as well as electrode poisoning after prolonged use [25], which can lead to lower output voltages. High Fe concentrations can have an inhibitory effect on energy generation by reducing substrate biodegradation [23,42] or by increasing internal resistance of MFC, which was confirmed in our studies (internal resistance was higher at doses of 0.  The electrochemical behavior of the control and Fe-modified electrodes with and without biofilm was tested by cyclic voltammetry in a working solution. Figure S4 shows the cyclic voltammograms recorded for all examined electrodes over a potential range of −0.50 to 0.50 V vs. SCE, which were obtained at room temperature with a sweep rate of 20 mV/s. The voltammograms of Fe-modified electrodes without biofilm did not exhibit any well-defined cathodic and anodic features related to iron behavior. However, the current of the cathodic peak corresponding to the hydrogen evolution reaction increased slightly with the addition of Fe. Nevertheless, the current increase was not strictly correlated with the amount of added iron, which could be caused by insufficient electric contact between the base electrode and the deposited iron oxide. On the other hand, after biofilm formation, the current density of the Fe-modified electrodes was significantly higher than that of the control with biofilm and the Fe-modified electrodes without biofilm. Additionally, the current values positively correlated with the amount of Fe that was deposited (188, 152, 51, and 43 mA with 0.4, 0.2, 0.1 and 0.05 g of Fe).
Furthermore, the CV profiles of the electrodes with 0.4 and 0.2 g of Fe2O3 exhibited two broad reversible features centered at -0.32 and 0.38 V vs. SCE. Therefore, those anodic peaks are most likely related to the process of iron oxidizing to Fe(II) and (III), while the cathodic peaks are associated with the reduction of those oxidized species [43]. Therefore, the changes observed after biofilm formation are most likely due to the following The electrochemical behavior of the control and Fe-modified electrodes with and without biofilm was tested by cyclic voltammetry in a working solution. Figure S4 shows the cyclic voltammograms recorded for all examined electrodes over a potential range of −0.50 to 0.50 V vs. SCE, which were obtained at room temperature with a sweep rate of 20 mV/s. The voltammograms of Fe-modified electrodes without biofilm did not exhibit any well-defined cathodic and anodic features related to iron behavior. However, the current of the cathodic peak corresponding to the hydrogen evolution reaction increased slightly with the addition of Fe. Nevertheless, the current increase was not strictly correlated with the amount of added iron, which could be caused by insufficient electric contact between the base electrode and the deposited iron oxide. On the other hand, after biofilm formation, the current density of the Fe-modified electrodes was significantly higher than that of the control with biofilm and the Fe-modified electrodes without biofilm. Additionally, the current values positively correlated with the amount of Fe that was deposited (188, 152, 51, and 43 mA with 0.4, 0.2, 0.1 and 0.05 g of Fe).
Furthermore, the CV profiles of the electrodes with 0.4 and 0.2 g of Fe 2 O 3 exhibited two broad reversible features centered at −0.32 and 0.38 V vs. SCE. Therefore, those anodic peaks are most likely related to the process of iron oxidizing to Fe(II) and (III), while the cathodic peaks are associated with the reduction of those oxidized species [43]. Therefore, the changes observed after biofilm formation are most likely due to the following phenomena: a better electrical connection between iron oxide particles and the base electrode provided by the biofilm and the increased biofilm formation at higher doses of Fe. The areas under the CV curves ( Figure S2b obtained in MFC 0.2Fe and MFC 0.4Fe ) were larger than in the other MFCs, which may indicate a higher charge capacity. This may be due to the fact that the charge capacity of the electrode is proportional to the electrode surface, which indicates that the surface area increases after modification [44], which is also visible in SEM images. Figure 3a shows the power curves from the 14th cycle of MFC operation. The highest power density of 1.39 mW/m 2 was obtained in MFC 0.1Fe and MFC 0.2Fe . This power density was 2.8 times higher than that obtained in MFC control , and 1.7 and 5.8 times higher, respectively, than the values obtained in MFC 0.05Fe and MFC 0.4Fe . Yang et al. [45] studied an anode with a carbonized Shewanella sp. biomass that produced a nanocomposite iron oxide/carbon catalyst. Their results showed that coating the anode with the nanocomposite increased the cell's power 3.5 times compared to that of a cell with a pure CF anode. The power was higher in the cell with the modified anode because, compared to the control, the surface roughness of the anode was higher with lower charge transfer resistance. Liu et al. [23] found that, in an MFC fed with sodium acetate and increasing iron doses in the substrate, the cell power density decreased from 0.95 W/m 2 at 200 µM Fe(III) to 0.59 W/m 2 at 2000 µM Fe(III). In MFC 0.4Fe in our study, with the highest dose of Fe 2 O 3 used for anode modification, the highest power achieved in the cell was two-fold lower than the value achieved with the control anode. The low power in MFC 0.4Fe could be due to an excess of iron at the anode that decreased the quality of the anode. Yang et al. [45] observed that Fe(III) aggregated on the anode can compete for electrons with the anode and inhibit the electrochemical activity of biofilms. phenomena: a better electrical connection between iron oxide particles and the base electrode provided by the biofilm and the increased biofilm formation at higher doses of Fe.
The areas under the CV curves ( Figure S2b obtained in MFC0.2Fe and MFC0.4Fe) were larger than in the other MFCs, which may indicate a higher charge capacity. This may be due to the fact that the charge capacity of the electrode is proportional to the electrode surface, which indicates that the surface area increases after modification [44], which is also visible in SEM images. Figure 3a shows the power curves from the 14th cycle of MFC operation. The highest power density of 1.39 mW/m 2 was obtained in MFC0.1Fe and MFC0.2Fe. This power density was 2.8 times higher than that obtained in MFCcontrol, and 1.7 and 5.8 times higher, respectively, than the values obtained in MFC0.05Fe and MFC0.4Fe. Yang et al. [45] studied an anode with a carbonized Shewanella sp. biomass that produced a nanocomposite iron oxide/carbon catalyst. Their results showed that coating the anode with the nanocomposite increased the cell's power 3.5 times compared to that of a cell with a pure CF anode. The power was higher in the cell with the modified anode because, compared to the control, the surface roughness of the anode was higher with lower charge transfer resistance. Liu et al. [23] found that, in an MFC fed with sodium acetate and increasing iron doses in the substrate, the cell power density decreased from 0.95 W/m 2 at 200 µM Fe(III) to 0.59 W/m 2 at 2000 µM Fe(III). In MFC0.4Fe in our study, with the highest dose of Fe2O3 used for anode modification, the highest power achieved in the cell was two-fold lower than the value achieved with the control anode. The low power in MFC0.4Fe could be due to an excess of iron at the anode that decreased the quality of the anode. Yang et al. [45] observed that Fe(III) aggregated on the anode can compete for electrons with the anode and inhibit the electrochemical activity of biofilms.  The polarization curves showed that modification of the anode with iron reduced the cell's internal resistance, most likely via better electron transfer due to the presence of iron particles. The highest internal resistance of 1029 Ω was recorded in MFC control , while the lowest internal resistance (184.9 Ω) was recorded in MFC 0.1Fe (Figure 3). Increasing the iron dose used for anode modification from 0.1 to 0.4 g gradually increased the internal resistance of the cell to 397 Ω in MFC 0.4Fe . In a study by Mohamed et al. [6], electrode position of iron (200 mM FeCl 3 working solution) on CF, G, and CC electrodes caused the total internal resistance of an MFC to decrease by 2.0, 1.9, and 1.4 times, respectively. In addition, electrolytic deposition improved the wettability and increased the porosity and biocompatibility of the surfaces.

COD Removal
The elimination of organic compounds in the MFC is carried out by their microbial decomposition, the products of which are carbon dioxide, hydrogen protons, and electrons, the acceptor of which is the electrode. The decomposition of organic compounds can be illustrated by the example of acetic acid as follows: Complex organic compounds must be converted to monosaccharides or other lowmolecular-weight compounds [46]. Therefore, the high biodiversity of microorganisms is important in MFC systems using complex effluents. Organic nitrogen in wastewater is converted to NH 4 -N, which is removed in the MFC mainly by migration through the EMF and volatilization in the cathode chamber [47] or by precipitation in the form of struvite, cattiite ((Mg 3 (PO 4 ) 2 ·22H 2 O)) or in the presence of iron-vivianite (Fe 3 (PO 4 ) 2 ·8H 2 O) [48]. In anaerobic environments, in the presence of Fe(III), NH 4 + is an electron donor and is oxidized to NO 2 − by reduction of Fe(III) to Fe(II) according to the formula [49] (Equation (1)) The average COD removal efficiency was above 70% for all reactors, indicating that the anaerobic biofilm on the anode was well developed and active. Low iron dosages favored more stable and effective COD removal compared to MFC control (Figure 4a). In MFC 0.05Fe and MFC 0.1Fe , COD was stably removed with efficiencies of 83.96 ± 10.1% and 85.5 ± 14%, respectively (Figure 4b,c). The mean COD concentrations in the treated wastewater from MFC 0.05Fe and MFC 0.1Fe were 29.5 ± 14.9 and 30.8 ± 7.6 mg COD/L, respectively, and were significantly lower than concentrations in the other MFCs. Better COD removal can be explained by the iron-modified anode having a higher surface area, which results in better biocompatibility and adhesion of microorganisms [4,50]. However, our results suggest that if the Fe dose is too high, the stability of COD removal decreases (Figure 4d,e). For example, in MFC 0.4Fe , the range of COD in wastewater was between 29 and 141 mg/L. Mohamed et al. [51] reported that Fe(III) at a dose of 40 mg/L in wastewater treatment systems decreases COD's removal efficiency. At high concentrations of Fe(III) in the environment, Fe(III) penetrates bacterial cells in large numbers and reduces enzyme activity or generates toxic free radicals that damage cell structures [52,53].
The addition of iron facilitated the maintenance of a neutral pH in the MFCs. The mean pH of the effluent from MFC control (pH 8.3) was significantly higher than that of the MFCs with iron-modified anodes (pH 7.7, 7.9, 7.8, and 7.8, for MFC 0.05Fe , MFC 0.1Fe , MFC 0.2Fe , and MFC 0.4Fe , respectively). Although the pH was lower in the MFCs with iron-modified anodes, it increased during the cycle, indicating that the rate of transport of protons through the PEM was slower than the rate of their production in the anode chamber. The efficiency of an MFC is optimal when a constant pH is achieved in the anode chamber, and the rate of proton production in the anode chamber is equal to the rate of their consumption at the cathode [54]. At the same time, the effluents from MFC 0.2Fe and MFC 0.4Fe had significantly lower alkalinity than that from MFC control . The oxidation of organic compounds usually produces more protons than bicarbonate ions, which lowers the pH and alkalinity and may harm the anode biofilm [55]. The lower alkalinity that was observed in the MFCs modified with the highest Fe doses may decrease conductivity, increase internal resistance, and decrease the efficiency of electricity generation [56]. alkalinity and may harm the anode biofilm [55]. The lower alkalinity that was observed in the MFCs modified with the highest Fe doses may decrease conductivity, increase internal resistance, and decrease the efficiency of electricity generation [56].
The efficiency of NH4-N removal in all MFCs was about 37%, and the NH4-N concentrations in the effluents did not differ significantly.

Chromatographic Analyses of VFAs
In our study, the differences in the effluent compositions indicated that the presence of Fe on the anode affected the metabolic conversions in the MFC. To illustrate the changes in VFA composition during wastewater treatment in MFCs, chromatographic analyses were performed ( Figure 5). Acetic acid comprised the largest share of VFAs in the substrate (86%), while the remaining 14% were other acids, mostly propionic acid (7%). Yu et al. [57] investigated the influence of various substrates on energy production and did not note any changes in the VFA profile in an acetate-fed rector-acetic acid was the only VFA in the effluent. In our study, in contrast, acetic acid was present only in small amounts (from 0.2% in MFC0.1Fe to 3.0% in MFCcontrol), while propionic and valeric acid proportions significantly increased in comparison to their proportions in the raw wastewater. In the outflow from MFCcontrol, propionic and valeric acids constituted 58% and 23% of VFAs, respectively, while hexanoic and iso-hexanoic acids accounted for 12%. As the iron dose was increased, the proportion of propionic acid also increased, while the proportions of The efficiency of NH 4 -N removal in all MFCs was about 37%, and the NH 4 -N concentrations in the effluents did not differ significantly.

Chromatographic Analyses of VFAs
In our study, the differences in the effluent compositions indicated that the presence of Fe on the anode affected the metabolic conversions in the MFC. To illustrate the changes in VFA composition during wastewater treatment in MFCs, chromatographic analyses were performed ( Figure 5). Acetic acid comprised the largest share of VFAs in the substrate (86%), while the remaining 14% were other acids, mostly propionic acid (7%). Yu et al. [57] investigated the influence of various substrates on energy production and did not note any changes in the VFA profile in an acetate-fed rector-acetic acid was the only VFA in the effluent. In our study, in contrast, acetic acid was present only in small amounts (from 0.2% in MFC 0.1Fe to 3.0% in MFC control ), while propionic and valeric acid proportions significantly increased in comparison to their proportions in the raw wastewater. In the outflow from MFC control , propionic and valeric acids constituted 58% and 23% of VFAs, respectively, while hexanoic and iso-hexanoic acids accounted for 12%. As the iron dose was increased, the proportion of propionic acid also increased, while the proportions of valeric, caproic, and iso-caproic acids decreased. Some microorganisms are able to form storage compounds in the form of polyhydroxyalkanoate (PHA) during the anaerobic degradation of organic compounds, especially VFA. Under aerobic conditions, PHAs are used as a carbon source, e.g., during biological phosphorus removal [58]. Wang et al. [59] proposed an anaerobic pathway for the synthesis and degradation of PHAs. The type of VFA determines the composition of stored PHA, which in turn affects the types of VFA produced during PHA degradation. Polyhydroxybutyrate (one of the forms of PHA) is degraded to acetate and butyrate, and polyhydroxyvalerate (PHV) is degraded to propionate, acetate, and valerate. In our study, in the biofilm microorganisms such as Zooglea sp., Dechloromonas sp., Acidovorax sp., and Hydrogenophaga sp. (see Section 3.5), capable of PHA accumulation, were observed. Due to the prevailing anaerobic conditions, it is possible that they converted them to organic acids present in the wastewater, i.e., mainly propionic and valeric acids, which would indicate that the reserve substance was stored mainly in the form of PHV.
valeric, caproic, and iso-caproic acids decreased. Some microorganisms are able to form storage compounds in the form of polyhydroxyalkanoate (PHA) during the anaerobic degradation of organic compounds, especially VFA. Under aerobic conditions, PHAs are used as a carbon source, e.g., during biological phosphorus removal [58]. Wang et al. [59] proposed an anaerobic pathway for the synthesis and degradation of PHAs. The type of VFA determines the composition of stored PHA, which in turn affects the types of VFA produced during PHA degradation. Polyhydroxybutyrate (one of the forms of PHA) is degraded to acetate and butyrate, and polyhydroxyvalerate (PHV) is degraded to propionate, acetate, and valerate. In our study, in the biofilm microorganisms such as Zooglea sp., Dechloromonas sp., Acidovorax sp., and Hydrogenophaga sp. (see Section 3.5), capable of PHA accumulation, were observed. Due to the prevailing anaerobic conditions, it is possible that they converted them to organic acids present in the wastewater, i.e., mainly propionic and valeric acids, which would indicate that the reserve substance was stored mainly in the form of PHV. These results indicate that microbial metabolism in MFCs leads to the formation of VFAs other the acetic acid; this process has been observed, for example, during acetate conversion under anaerobic conditions [60]. Moreover, propionic acid that is present in wastewater may not be metabolized in MFCs. During anaerobic fermentation, the conversion of butyric acid and propionic acid to acetic acid does not occur spontaneously, due to the high Gibbs free energy of the reaction, resulting in an accumulation of the two acids [61]. The presence of propionate as a metabolic product of wastewater conversion in MFCs inhibits the activity of acidogenic bacteria and methanogens [62], which can eliminate electron loss during methane fermentation in MFCs.

Microbial Structure of Inoculum and Anode Biofilm in the MFCs
A total number of 498,906 readings was obtained after sequencing. The lowest number of readings was recorded in MFC0.4Fe in the 10th cycle of MFC operation, and the highest in MFC0.1Fe in the 10th cycle of operation ( Table 1). The flattening of the rarefaction curves ( Figure S6) indicates that the depth of the sequencing was sufficient. However, the OTU number for the tested samples was quite low and varied from 66 to 117. For comparison, the average number of OTUs for biofilms from the anode modified with Fe2O3 in an acetate-fed dual-chamber MFC was about 600 [63]. In our study, many unclassified microorganisms were detected, which may explain the low number of OTUs. These results indicate that microbial metabolism in MFCs leads to the formation of VFAs other the acetic acid; this process has been observed, for example, during acetate conversion under anaerobic conditions [60]. Moreover, propionic acid that is present in wastewater may not be metabolized in MFCs. During anaerobic fermentation, the conversion of butyric acid and propionic acid to acetic acid does not occur spontaneously, due to the high Gibbs free energy of the reaction, resulting in an accumulation of the two acids [61]. The presence of propionate as a metabolic product of wastewater conversion in MFCs inhibits the activity of acidogenic bacteria and methanogens [62], which can eliminate electron loss during methane fermentation in MFCs.

Microbial Structure of Inoculum and Anode Biofilm in the MFCs
A total number of 498,906 readings was obtained after sequencing. The lowest number of readings was recorded in MFC 0.4Fe in the 10th cycle of MFC operation, and the highest in MFC 0.1Fe in the 10th cycle of operation ( Table 1). The flattening of the rarefaction curves ( Figure S6) indicates that the depth of the sequencing was sufficient. However, the OTU number for the tested samples was quite low and varied from 66 to 117. For comparison, the average number of OTUs for biofilms from the anode modified with Fe 2 O 3 in an acetate-fed dual-chamber MFC was about 600 [63]. In our study, many unclassified microorganisms were detected, which may explain the low number of OTUs. The ACE index in MFC control and the MFCs operated with the two lower iron doses increased with time, but in the reactors with modified anodes, the diversity slightly decreased during the 10th cycle of reactor operation (Table 1). In MFC 0.2Fe and MFC 0.4Fe , the ACE changed more dynamically, indicating that the presence of Fe changed the metabolism and biodiversity of the anode biofilm. In the third cycle of reactor operation (In MFC 0.2Fe and MFC 0.4Fe ), the ACE dropped significantly from 92 in the inoculum to about 70. However, the species diversity then increased with time, and at the end of the experiment, it was nearly two times higher than during the third cycle of MFC operation.
Throughout the experiment, Proteobacteria predominated in the microbial communities in the individual MFCs. The share of Proteobacteria increased in relation to the dose of iron used for the anode modification (Figure 6), reaching a value above 75% of all identified microorganisms in MFC 0.4Fe . Proteobacteria members are electroactive bacteria crucial for extracellular electron transport [64]. They are involved in sludge hydrolysis and short-chain fatty acid production and are the main consumers of acetate, propionate, and butyrate during fermentation [65], which explains their high abundance in MFCs fed with acetate-containing wastewater. In a study in which Fe 3 O 4 was dosed to the medium in a single-chamber MFC, the numbers of Proteobacteria in the biomass were also high, followed by Desulfobacterota and Bacteroidota [21]. The second most abundant group in the present study was Bacteroidetes (up to 11% in MFC 0.05 ). A high percentage of unclassified bacteria (up to 44% in MFC 0.05 ) indicated that many yet undiscovered bacteria played an important role in energy production in MFCs.
A common feature of the microbiome in all MFCs was the presence of Acidovorax sp. during reactor start-up (Figure 7). Its share ranged from 3 to 18.5% depending on the MFC, and then, in subsequent cycles, it decreased to <0.5%; only in the control did its share amount to 1.3% in the 18th cycle. Acidovorax sp. are capable of efficient extracellular polymeric substances (EPS) production (over 150 mg/L in R2A agar medium [66]), and probably participated in anode colonization and the formation of an anode biofilm in the initial stage of MFC operation. A common feature of the microbiome in all MFCs was the presence of Acidovorax sp. during reactor start-up (Figure 7). Its share ranged from 3 to 18.5% depending on the MFC, and then, in subsequent cycles, it decreased to <0.5%; only in the control did its share amount to 1.3% in the 18th cycle. Acidovorax sp. are capable of efficient extracellular polymeric substances (EPS) production (over 150 mg/L in R2A agar medium [66]), and probably participated in anode colonization and the formation of an anode biofilm in the initial stage of MFC operation.    A common feature of the microbiome in all MFCs was the presence of Acidovorax sp. during reactor start-up (Figure 7). Its share ranged from 3 to 18.5% depending on the MFC, and then, in subsequent cycles, it decreased to <0.5%; only in the control did its share amount to 1.3% in the 18th cycle. Acidovorax sp. are capable of efficient extracellular polymeric substances (EPS) production (over 150 mg/L in R2A agar medium [66]), and probably participated in anode colonization and the formation of an anode biofilm in the initial stage of MFC operation.   One of the most frequently identified EEs is Geobacter sp., which is able to convert energy from the decomposition of organic compounds with electron transfer to the anode using nanowires [67]. Liu et al. [22] observed a decrease in the share of Geobacter sp. in the bacterial community with an increasing concentration of FeCl 3 in a medium from 200 to 1000 µM. Zheng et al. [21] investigated the effect of Fe 3 O 4 added to the medium in a single-chamber MFC. Their results show that the lowest tested dose of 4.5 g/L led to an enrichment of the anode biofilm in Geobacter sp. (17.4% for control and 31.5% for MFC with the tested Fe 3 O 4 dose), but higher doses decreased their share. Our study shows that Geobacter sp. was present in all tested MFCs, with the highest proportion (6.3%) being observed in the MFC control . In the remaining MFCs, the share of Geobacter sp. ranged from 1.3 to 3.3%. The results indicate that the percentage of Geobacter sp. in the biofilm did not directly translate into energy production in MFCs, which was highest in MFC 0.1Fe . The numerical abundance of microorganisms in biofilms cannot be assumed a priori to correlate to the capacities of these species to produce power [68]. In MFCs with lower Geobacter sp. abundances, a greater role was played by the synergistic interactions between microorganisms. In co-cultures with EE microbes, other bacteria may facilitate the current generation by removing chemicals or by producing a substrate for the current generation [69]. The lower proportion of Geobacter sp. in the presence of iron ions may result from the fact that these microorganisms do not appear to significantly reduce crystalline forms of Fe(III) [70]. Growing these microorganisms in media with crystalline forms of Fe(III) led to the enrichment of methane-producing cultures, while the addition of weakly crystalline Fe(III) resulted in the successful enrichment of Geobacter sp. [67]. The other study focused on the analysis of protein expression in Geobacter sulfurreducens showed that in the presence of Fe(III) oxide, about 76% of proteins were less abundant than in the presence of Fe(III) citrate due to the slower rate of bacterial metabolism and growth with an insoluble electron acceptor. Most of these proteins were involved in metabolic processes such as electron transport (13 c-type cytochromes) or were structural proteins for electrically conductive pili (PilA, [71]).
As environmental conditions in MFCs with iron-modified anodes did not favor Geobacter sp., in these MFCs, electricity generation was supported by other microorganisms. At two lower doses of iron on the anodes, EEs belonging to the genera Pseudomonas and Oscillochloris were identified in large numbers. Oscillochloris sp. is a type of filamentous bacteria, classified as anoxygenic phototrophs, that binds carbon through the reductive pentose phosphate cycle [72]. Phototrophic bacteria of various types are capable of generating energy [73][74][75]. In the MFC 0.1Fe , Pseudomonas sp. comprised up to 70% of bacteria in the anode biofilm. Pseudomonas sp. produces several electrochemically active phenazine derivatives, such as phenazines-1-carboxylic acid, pyocyanin, oxychlororaphin, and pyorubin, which can act as redox mediators in the MFC [76]. The high abundance of Pseudomonas sp. in MFC 0.1Fe may explain the high power obtained in this cell because electrons produced by other bacteria could be transferred to the anode via mediators produced by Pseudomonas sp. In high concentrations, the mediators produced by Pseudomonas sp. can be toxic to other species, thus giving Pseudomonas sp. an environmental advantage in competition with other bacterial species [69]. The energy production by pure cultures of Pseudomonas sp. is lower than that of biofilms with a high proportion of Geobacter sp.; however, this lower energy production by Pseudomonas sp. may be compensated by their high abundance in biofilm. Moreover, competition in biofilms involves many factors, and the ability to conduct current may not be the only or main reason for the predominance of high-power-producing bacteria. For example, some EEs bacteria in a mixed culture of the anode in MFC are capable of higher power production than after isolation from this biofilm and cultivation in the pure culture [68].
In MFC 0.05Fe and MFC 0.1Fe , Rhizobium sp., Thiovirga sp., Longilinea sp., and Bacteroides sp. occurred in the anode biofilm, which all have been reported to be involved in energy generation in MFCs. The Rhizobium anhuiense strain is an effective EE and generated a maximum voltage of 635 mV and output power of 1.07 mW/m 2 in a glucose-fed dual-chamber MFC operated in an open circuit [77]. Thiovirga sp. is a type of chemolithoautotrophic bacteria, and in MFCs, they play an essential role in sulfide metabolism and COD removal [78]. Thiovirga sp. shows good resistance to high metal concentrations; however, our study indicates that, although Fe presence favors Thiovirga sp. growth in anodic biofilm, this genus prefers lower Fe concentrations in the environment [79,80]. Longilinea sp. decomposes hydrocarbon into acetate and H 2 [81]. Yang et al. [82] reported that the presence of Fe 2 O 3 at the anode promoted the reproduction of Longilinea sp., which could release intracellular electrons through the metabolism of substrates or intermediates [83]. Bacteroides sp. were one of the main bacteria producing electricity in the constructed wetland MFC [84]. It efficiently produced electricity by degradation of organic substances, preferably acetate [85], as well as by transferring electrons that reduce Fe(III) [86].
At the two highest doses of iron at the end of the study, biomass was mostly occupied by Desulfobacter sp., Dechloromonas sp., and Thiobacillus sp. Dechloromonas sp. have been reported to both generate electricity in MFCs [87,88] and play an important role in the conversion of ferrous ions in Fe-contaminated soils or waters [89,90]. Li et al. [91] observed that during municipal wastewater treatment in MFC coupled with an up-flow in denitrification biofilter abundance of Dechloromonas sp. ranged from 1.29% to 9.52%. The authors reported a negative correlation between the abundances of Dechloromonas sp. and Geobacter sp., which play a predominant role in denitrification and anodic respiration. This observation indicates their similar respiring and metabolizing processes and, thus, potential competing relationships. Our studies indicate that their abundance also depends on the iron content in the environment. Desulfobacter sp. are a type of EE [92] using acetate as an electron donor and sulfate, sulfite, and thiosulfate as final electron acceptors [93]. Desulfobacter sp. generated electric power of 370-1400 mW/m 2 in the concentration range from 550 to 1270 mg COD/(L·h) [94]. Thiobacillus sp. was present in both the anode and cathode biofilm [95,96] with abundances in anode biofilms as high as 24.5% during the treatment of wastewater with high sulfur content [97].
In MFC 0.4Fe at the beginning of the experiment, Proteobacteria constituted as much as 96% of the entire bacterial community, and the Zoogloea sp. share was close to 70%. The predominance of Zoogloea sp. in the biofilm from MFC 0.4Fe may result from the fact that these microorganisms can produce large amounts of EPS. Production of protein-and polysaccharide-rich EPS plays a vital role in cell protection under metal stress conditions [98,99], as functional groups from EPS form complexes with metal ions protecting the cell surface [100]. Metal ions may also be sorbed on the cell surface of Zoogloea sp. by electrostatic interactions such as van der Waals forces [101,102]. A decrease in the abundance of Zoogloea sp. in the course of the experiment indicated bacterial succession maturation of the anodic biofilm. Another bacterial group that preferred high iron concentrations was Flavobacterium sp. (up to 10.1% of the bacterial community in MFC 0.4Fe ). Previous research indicated that Flavobacterium sp. predominated in acetate-fed MFC [103] and was an important player in electricity production [104,105].
The presence of Fe 2 O 3 on the anode was positively correlated with the abundances of Methylovresatilis sp. and Fluviicola sp., whose share on the MFC 0.4Fe anode reached 1.2% and 1.5%, respectively ( Figure S5). Methyloversatilis sp. is a type of facultative methylotroph that can grow on a variety of C 1 and multi-carbon compounds [106] and has been reported in MFCs [107]. Our study indicated for the first time that the presence of Methyloversatilis sp. in the environment is stimulated by increasing Fe concentrations. It may be explained by the important role of Fe-containing enzymes in Methyloversatilis sp. [108] as well as by the involvement of this genus in iron conversions in the environment [109]. The positive effect of the addition of ferrous ions (FeII) on Fluviicola sp. abundance was observed in up-flow anaerobic sludge blanket digestion reactors in which these bacteria conducted NO 2 reduction together with Fe(II) [110].

Conclusions
Bioelectrochemical systems that recover energy from wastewater and waste can play an important role in the energy industry's future. In this study, to improve electrical power generation in MFCs, various doses of Fe 2 O 3 were used to modify the anode surface. In MFC 0.1Fe, the power and cell resistances were the best, 2.8 times higher and 5.6 times lower than in the control. Organics removal from wastewater was more stable in MFCs with anodes modified with lower Fe 2 O 3 doses. Although the main source of carbon in wastewater was acetate, propionic and valeric acids predominated among the VFAs in the MFC effluents; the proportion of propionic acid to other VFAs increased with increasing iron dose. EPS producers, such as Zoogloea sp. and Acidovorax sp., were abundant during reactor start-up, facilitating microbial colonization of the anode and the development of anode biofilms. The addition of iron increased the diversity of the bacterial community. Our study indicated that EE composition was strongly affected by the presence of iron on the anode. The abundance of EE Geobacter sp. was 2-3% on the anodes modified with Fe 2 O 3 in comparison with 6% in the control MFC. At the two lower doses of Fe 2 O 3 , the abundance of EEs belonging to Oscillochloris sp. and Pseudomonas sp. increased, while at the two highest Fe 2 O 3 doses, EEs belonging to Dechloromonass sp. and Desulfobacter sp. predominated in the biofilm. To improve the operation of the MFC, more research is needed to evaluate the effects of iron dose on the anode electrode, because too much iron can negatively affect the current generation and the anode biofilm, and researchers often focusing on the modification itself, ignoring the microbiological aspect. This research could help scientists further develop MFCs through low-cost methods, but more resistant materials need to be developed for industrial applications, as the technologies need to work in the long term.  Figure S2. Statistical differences in the voltages obtained in the individual reactors (ANOVA; Tukey's HSD post hoc test), p < 0.05, * significantly higher than in the remaining MFCs. Figure S3. Power output within 18 cycles for (a) MFC control , (b) MFC 0.05Fe , (c) MFC 0.1Fe , (d) MFC 0.2Fe , (e) MFC 0.4Fe . Figure S4. CV for (a) abiotic anodes, (b) with biofilm. Figure S5. Refraction curves; the number after the reactor name indicates the cycle in which sampling was performed. Figure S6.  Data Availability Statement: The datasets used and/or analyzed during the current study are available from the corresponding author upon reasonable request.

Conflicts of Interest:
The authors declare that they have no conflict of interest.