Bioelectrochemically enhanced autotrophic Feammox for ammonium removal via the Fe(II)/Fe(III) cycle

Abstract

Autotrophic anaerobic ammonium oxidation coupled to Fe(III) reduction (Feammox) is a potential technology for removing ammonium from low-C/N wastewater, but it requires a continuous supply of Fe(III) source. To reduce the supply, a microbial electrolysis cell (MEC) was employed to allow iron recycling in Feammox under different voltages (0.2 V, 0.6 V, and 1.0 V). Results showed that the optimal voltage was 0.6 V, with a maximum efficiency for ammonium oxidation of 71%. The ammonium oxidation rate achieved 2.5 ± 0.1 mg N L⁻¹ per day, which was 3 times that of conventional Feammox. Cyclic voltammetry confirmed that ammonium oxidation and iron redox occurred on the anode. The bacterial population had a unique evolutionary direction at 0.6 V, with Geobacteraceae becoming the dominant family. Positive interactions between nitrogen-related bacteria and iron-related bacteria enhanced the autotrophic Feammox process. This study will further advance Feammox in the treatment of ammonium-containing wastewater.

---

Water impact

Autotrophic Feammox has attracted a lot of interest in the treatment of low-C/N wastewater. However, incessant supplementation with Fe(III) sources is required for Feammox in engineering. This study adopted an MEC to drive the iron cycle and promote the ammonium oxidation rate of Feammox. The results revealed the potential mechanism of the iron cycle and bacterial interactions, providing valuable research ideas and technological references for Feammox in engineering.

Introduction

Anaerobic ammonium oxidation coupled with Fe(III) reduction, termed Feammox, is a novel and promising approach to remove ammonium, which is of great interest by virtue of its new metabolic pathway and low cost. Feammox was first defined as the bioprocess by which iron reduction bacteria oxidized ammonium to nitrite using Fe(III).¹ Subsequently, the reactions of the Feammox process were proposed using Fe(OH)₃ as the iron source [eqn (1)–(3)], which was based on thermodynamic calculations and isotope labeling.² Feammox has been observed in various surroundings in succession, such as paddy soil,³ intertidal wetlands,⁴ lake sediment⁵ and groundwater.⁶ The findings indicate the widespread presence of Feammox in natural environments. Feammox processes likewise occurred in activated sludge and anaerobic sludge by supplementing with Fe(III) compounds,^7,8 which implies the possible existence of Feammox populations in wastewater treatment systems and provides new inspiration for ammonium removal.

NH₄⁺ + 3Fe(OH)₃ + 5H⁺ → 3Fe²⁺ + 9H₂O + 0.5N₂ Δ_rG_m = −245 kJ mol⁻¹ (1)

NH₄⁺ + 6Fe(OH)₃ + 10H⁺ → 6Fe²⁺ + 16H₂O + NO₂⁻ Δ_rG_m = −164 kJ mol⁻¹ (2)

NH₄⁺ + 8Fe(OH)₃ + 14H⁺ → 8Fe²⁺ + 21H₂O + NO₃⁻ Δ_rG_m = −207 kJ mol⁻¹ (3)

The rate of ammonium oxidation has a strong positive correlation with Fe(III) reduction in the Feammox process (Pearson correlation coefficient 0.97, P < 0.01).⁹ This means that good ammonium removal by Feammox requires a large amount of Fe(III) sources. The Fe(III) resources of Feammox are diverse, including FeCl₃, Fe₂O₃, Fe₃O₄, Fe(OH)₃, ferrihydrite, and goethite.^10,11 However, they are always limited in wastewater treatment systems. A sufficient supplement of Fe(III) is necessary to sustain Feammox, which undoubtedly increases the cost of wastewater treatment, making it hard to utilize directly at a large scale. The recycling of iron is one of the most applicable ways to solve this problem, with the interconversion of Fe(III) and Fe(II). Fe(III) can be reduced to Fe(II) by the iron reduction bacteria of Feammox; thus, the point of the iron cycle is the oxidation of Fe(II) to Fe(III). Common methods for oxidizing Fe(II) involve supplying O₂ and coupling to a bioprocess. Within the tolerance range of Feammox to oxygen, constant micro-aeration or intermittent aeration were adopted to provide a low-oxygen environment, which achieved Fe(II) oxidation and further enhanced Feammox performance.^12,13 In addition, the most frequently mentioned bioprocess is nitrate-dependent Fe(II) oxidation (NDFO). Fe(II) is oxidized by nitrate, coupled with Feammox to implement the iron cycle by supplying nitrate.¹⁴ However, these methods demand rigorous control over oxygen and incessant input of nitrate, which makes Feammox complicated and uneconomic in practice. Therefore, other alternatives to the iron cycle are worth exploring.

By inserting a pair of polarized inert electrodes, the redox environment for electroactive microbes can be regulated by an external power source, while the system is known as a microbial electrolysis cell (MEC).¹⁵ MEC is an ideal technology for superseding O₂ and nitrate due to its low cost, long-term sustainability, simple operation and environmental friendliness. Although Feammox in the MEC has been studied, the role of the iron cycle in the system has rarely been studied. For example, the ammonium oxidation rate of Feammox was improved with continuous supplementation by Fe₂O₃, which might be the effect of the iron cycle in the MEC system.¹⁶ Therefore, exploring the role of the MEC in the iron cycle is essential for understanding the mechanism of the improvement, which has not yet been investigated. MEC may stimulate microbes involved in the iron cycle or directly participate in the redox reaction of iron on the electrodes.

Here, electrochemically assisted Feammox (shortened to E-Feammox) was constructed, and the feasibility of in situ electron supplementation to the Feammox microbial community was demonstrated. The role of the iron cycle and functional bacteria under the different voltages were investigated, and the mechanism was studied according to microbial electrochemical tests, biofilm morphology and microbial community analysis.

Materials and methods

Experimental setup

The E-Feammox system was conducted in a single-chamber membrane-less MEC (100 mL) to prevent anodic acidification and cathodic alkalinization (Fig. S1†). Carbon felt (3.0 cm in diameter and 2.0 mm in thickness) was used as the anode and placed flat on the bottom of the reactor to support iron precipitates and bacteria. The cathode was a 304 stainless steel mesh (3.0 cm in diameter, 100 mesh) facing the anode. Different voltages were applied using wireless data acquisition equipment developed in our laboratory, which can collect the current generated from the MEC system under a constant voltage.¹⁷

Inoculation and operation

The Feammox bacteria employed in the present study were cultivated from paddy soil for up to 4 years, and their electron donor was Fe(OH)₃.¹⁸ First, 10 g of paddy soil and 80 mL of medium were added to an anaerobic bottle and incubated in an opaque shaker (150 rpm, 35 °C). The precipitate was injected into the new medium through centrifugation (6000 rpm, 10 min), resulting in continuous enrichment of the iron precipitates in the medium during the domestication process. Since Feammox bacteria attached to the surface of the iron precipitates were difficult to separate, the mixture of bacteria and iron was inoculated into the anode after centrifugation. A single-chamber MEC was filled with 80 mL of culture solution and 1 g (dry weight) of the mixture. The biomass of the inoculant was approximately 5 mg protein per L. The media contained 0.27 g L⁻¹ of NH₄Cl, 1.2 g L⁻¹ of NaHCO₃, 0.60 g L⁻¹ of KH₂PO₄, 0.40 g L⁻¹ of MgCl₂·6H₂O, 0.10 g L⁻¹ of CaCl₂·H₂O, 0.040 g L⁻¹ of L-cysteine hydrochloride monohydrate (for oxygen elimination), 2.5 mL of trace minerals and 0.50 mL of vitamin solution, as described by.¹⁹ The media solution was flushed by N₂/CO₂ with a ratio of 4 : 1 (V/V) for 40 minutes to remove oxygen.

The voltages were set at 0.2 V, 0.6 V and 1.0 V, respectively, to MECs, which were lower than the oxygen evolution potential of carbon felt according to a linear sweep voltammetry (LSV) test (Fig. S2†). The experimental controls included abiotic MECs (ck-0.2, ck-0.6 and ck-1.0, corresponding to the different voltages) and a biological control without voltage (ck-bio). All MECs were performed in duplicate at room temperature, avoiding light. When sampling, N₂ bags were used for replacement to prevent O₂ permeation and to maintain the pressure balance.

Chemical and electrochemical measurement

The liquid samples were filtered using 0.22 μm membranes and later detected immediately. The concentrations of NH₄⁺-N, NO₂⁻-N and NO₃⁻-N were measured according to standard methods.²⁰ The hydroxylamine (NH₂OH) content was detected on the basis of previous literature.²¹ Iron concentrations were determined using the phenanthroline spectrophotometry method at 510 nm. After vacuum drying the sample, X-ray diffraction (XRD) (Ultima IV, Rigaku, Japan) was performed in the 2θ range of 10–90° for iron species. The valences of the iron precipitates were analyzed by X-ray photoelectron spectroscopy (XPS) (ESCALAB 250Xi, Thermo Scientific, USA).

To determine E-Feammox reactions on the anode, cyclic voltammetry (CV) was conducted using a potentiostat (Autolab PGSTAT 302 N, Metrohm, Switzerland) in a three-electrode system. Ag/AgCl (0.197 V vs. SHE) with saturated KCl was used as the reference electrode. The CV curve was scanned from −1.0 V to 1.0 V at a rate of 1.0 mV s⁻¹.²² The procedure for measuring oxidation peak currents (I_pa) involved drawing a tangent line along the curve to the peak bottom, and a perpendicular line from the peak top to tangent line. The height of this line represented the I_pa value. Charge and current were used to assess the electron transfer to the anode during Feammox, which were calculated as in eqn (4),²³

C = ΔNH₄⁺·n·v·F (4)

where C is the theoretical amount of charge in Coulombs (C), ΔNH₄⁺ is the oxidation amount of ammonium (M), n is the number of electrons produced from ammonium oxidation (e.g. for NH₄⁺ → NO₂⁻, n = 6), v is the volume of liquid in the MEC (L), and F is the Faraday constant (96 485.33 C/M).

Biofilm and bacterial analysis

Confocal laser scanning microscopy (CLSM) (LSM880, Zeiss, Germany) was used to assess the biofilm morphology of the different MECs. The active and inactive cells of the samples were identified through staining with a LIVE/DEAD BacLight viability kit (L7007, Molecular Probes, Invitrogen detection technology).

Biofilms attached to the anode were collected at the end of the experiment to analyze the bacterial community under the different voltages. Bacterial analysis used an Illumina MiSeq sequencing platform by Biomarker Technologies (Beijing, China). The V3–V4 region of the 16S rRNA gene was amplified with primers 338F (5′-ACTCCTACGGGAGGCAGCA-3′) and 806R (5′-GGACTACHVGGGTWTCTAAT-3′). The operational taxonomic units (OTUs) were identified at 97% similarity level. OTU sequences were annotated using a conservative OTU filtration 0.0050% threshold. Clean read assembly was performed using the DADA2 package. The within-habitat diversity and between-habitat diversity of the bacterial community were examined through alpha and beta diversity analysis. The correlation network between genera was constructed by Spearman rank correlation analysis (p-value < 0.050) and exhibited with Gephi 0.10.1. Functional classification statistics of COG (clusters of orthologous groups of proteins) were conducted using the reconstruction of unobserved states (PICRUSt2) to predict the potential functions of E-Feammox bacteria.

Results and discussion

Ammonium removal under different voltages

Rates of ammonium oxidation were relatively slow in the first 15 days when different voltages (0.2 V, 0.6 V and 1.0 V) were applied to the MECs inoculated with Feammox (Fig. 1). Ammonium was not removed in all MECs. This poor ammonium oxidation could be due to the absence of an Fe(III) source in the system during the start-up phase.

Fig. 1 Variation of nitrogen concentrations (lines) and removal efficiency (areas) in biotic and abiotic (ck) MECs at 0.2 V, 0.6 V, 1.0 V and no voltage control (ck-bio). Solid lines: NH₄⁺-N, dotted lines: NO₂⁻-N, dashed lines: NO₃⁻-N.

The oxidation of ammonium gradually accelerated after 15 days. The ammonium oxidation rates at 0.2 V, 0.6 V and 1.0 V were 2.0 ± 0.2 mg N L⁻¹ per day, 2.5 ± 0.1 mg N L⁻¹ per day and 2.1 ± 0.2 mg N L⁻¹ per day, which were 2.4–3.0 times that of conventional Feammox [no voltage, 0.82 mg N L⁻¹ per day].¹⁸ While the concentration of ammonium was constant in ck-bio, indicating that the ammonium oxidation by E-Feammox was promoted in the MECs. The highest ammonium oxidation rate was achieved at 0.6 V, followed by 1.0 V and 0.2 V, with anodic potentials (vs. SHE) of 0.64 ± 0.01 V, 0.77 ± 0.1 V and 0.22 ± 0.2 V. These potentials were also within the range (−0.2 to 0.8 V vs. SHE) reported in other literature for anodic ammonium oxidation.^24–26 As reported in previous studies, the bacteria growing on the anode were impacted by the electrode potential.²⁷ Low potential was not sufficient to support the metabolism of the functional bacteria, while a high potential could destroy the structure of the cell membrane.²⁸ Therefore, in this study, 0.6 V was a suitable voltage for the growth of E-Feammox bacteria. On the 36th day, ammonium concentrations under different voltages were 2 ± 0.2 mM, 2 ± 0.1 mM and 2 ± 0.4 mM, with corresponding removal efficiencies of 59 ± 2%, 71 ± 4% and 61 ± 6%. In contrast to biotic MECs, abiotic ck-0.2, ck-0.6 and ck-1.0 showed no significant ammonium oxidation over 36 days, further confirming that the ammonium was removed biologically.

Conventional Feammox oxidizes ammonium into three metabolic products, that is N₂, nitrite and nitrate [see eqn (1)–(3)], and diverse processes yielded different products. For example, A6 bacteria oxidized ammonium to nitrite instead of N₂ or nitrate,²⁹ while Feammox products were nitrite and nitrate with activated sludge as the inoculant.³⁰ For the microbial community utilized in this study, the major product of Feammox was N₂.¹⁸ However, the end products of E-Feammox turned into nitrite and a little nitrate in the MEC system. The production ratios of nitrite were 57 ± 10%, 82 ± 2% and 65 ± 1%, and the ratios of nitrate were 2 ± 0.3%, 4 ± 1.6% and 2 ± 2.4% under 0.2 V, 0.6 V and 1.0 V, respectively. The possible intermediate (NH₂OH) was not detected throughout the experiment, so that the main nitrogen loss should be attributed to microbial assimilation and N₂ production.¹⁸ A similar trend in nitrogen conversion was observed from days 36 to 66. The rate of ammonium oxidation at 0.6 V was 2.2 mg N L⁻¹ per day and was higher than that at 1.0 V [1.9 mg N L⁻¹ per day] or 0.2 V [1.2 mg N L⁻¹ per day]. The accumulation of nitrite still occurred after ammonium oxidation. On the 66th day, the levels of nitrite under above voltages were 3.4 ± 0.3 mM, 2.0 ± 0.5 mM and 1.6 ± 0.2 mM. On the 5th day of the second period, there was a significant ammonium oxidation phenomenon, indicating that the E-Feammox system can be initiated rapidly in the new period. In brief, the polarized anode altered the end products of E-Feammox to mainly nitrite, providing a possibility of combination of Feammox with anammox for autotrophic nitrogen removal.³¹

Iron cycle in the MEC system

The content of iron in the different MECs was measured since ammonium oxidation is accompanied by Fe(III) reduction in the Feammox process. Fe²⁺ and Fe³⁺ ions were not detectable in bulk solution, and all changes were carried out in solid precipitations (Fig. 2A). All iron existed in the form of Fe(II) with a content of 368 ± 3 mM in the inoculant, which was a result of complete Fe(III) reduction by Feammox. Fe(II) was slowly oxidized to Fe(III) over 36 days, corresponding to a slow increase in the ammonium oxidation rate, and finally Fe(III) contents reached 274 ± 4 mM, 245 ± 24 mM and 261 ± 4 mM at 0.2 V, 0.6 V and 1.0 V, respectively, compared to 345 ± 3 mM in ck-bio. These results suggested that the initial solid Fe(II) in E-Feammox can be oxidized to Fe(III) by the anode or iron oxidation bacteria, and the produced Fe(III) was reduced back to Fe(II) by E-Feammox bacteria, forming a complete iron oxidation–reduction cycle. The highest content of Fe(II) was generated at 0.6 V (99 ± 22 mM), followed by 1.0 V (72 ± 16 mM) and 0.2 V (53 ± 4 mM). The increase in Fe(II) was positively correlated with the amount of ammonium oxidation during the inoculant culture (Fig. S3†). The more Fe(II) that was detected in the iron cycle, the better the ammonium oxidation performance. Only a small amount of Fe(II) was generated in ck-bio without the ammonium oxidation capability. The external voltage in the MECs was necessary to maintain the iron cycle and consecutive E-Feammox reaction.

Fig. 2 Iron contents (A), X-ray diffraction spectra (B) and X-ray photoelectron spectroscopy diagram (C) of iron precipitates at the start and end of the experiment.

XRD and XPS analyses of the iron precipitates were carried out to explore the conversion of Fe(II) and Fe(III). There was a high content of vivianite in the inoculant at the beginning of the experiment (Fig. 2B). A reasonable explanation was that the iron reduction bacteria in the microbial community, such as Geobacter, reduced Fe(III) to Fe(II). Then, Fe(II) was chemically precipitated with phosphate as vivianite, which is an important form of P-containing mineral in anaerobic sludge.^32,33 The redox potential of vivianite was close to 0 V as per the Eh-pH diagram.³⁴ In the present study, the anodic potentials were all greater than 0 V, which could facilely oxidize vivianite. Thus, the process of Fe(II) oxidation in the iron cycle was achieved by the MEC. The predominant mineral in ck-bio was leucophosphite with Fe(III).³⁵ The contents of vivianite and leucophosphite in the precipitates were different under the applied voltages, confirming the presence of the iron cycle in the MECs. The chemical state of iron in mineral crystals was further investigated through XPS analysis (Fig. 2C). The mineral crystals had characteristic peaks near 711 eV and 725 eV, corresponding to Fe2p_3/2 and Fe2p_1/2, respectively.³⁶ Based on semi-quantitative fitting of iron, divalent iron accounted for 44% of the total iron in the crystals at 0.6 V, which was higher than that at 0.2 V (33%), 1.0 V (30%) or ck-bio (28%). These results supported the conversion of Fe(II) and Fe(III) in the form of vivianite and leucophosphite on the surface of the anode. The cycling of iron was successfully driven by the MEC system, which significantly enhanced the performance of E-Feammox.

Bioelectrochemical analysis

Six electrons were theoretically transferred per mol of ammonium oxidized to nitrite. According to Coulomb values shown in Table S1,† the corresponding currents of ammonium oxidation were 8.6 × 10⁻⁵ A, 16 × 10⁻⁵ A and 11 × 10⁻⁵ A under 0.2, 0.6 and 1.0 V, respectively, which were below the minimum detection limits of the data acquisition system. To verify the occurrence of an E-Feammox process on the anode surface, CV curves were scanned in fresh medium. As shown in Fig. 3, unambiguous oxidation peaks appeared at 0.8 V in all the CV curves, representing the oxidation of ammonium similar to previous reports.^24,37 While the curve of the abiotic anode was smooth without any peak at the same potential of 0.8 V. The values of I_pa under 0.2 V, 0.6 V and 1.0 V were 4.2 × 10⁻⁴ A, 10 × 10⁻⁴ A and 7.6 × 10⁻⁴ A, respectively, following the same trend as ammonium removal by E-Feammox. The oxidation peak and reduction peak at −0.05 V and −0.3 V were similar to the oxidation and reduction peaks of iron reported in previous literature.³⁸ Similar peaks also appeared in the CV curve of abiotic MEC, suggesting that iron can undergo electrochemical transformation on the anode. The differences in the CV peaks can be attributed to the different Fe(III) contents in the iron precipitates. CV results confirmed that E-Feammox performed ammonium oxidation and the iron cycle on the anode.

Fig. 3 Cyclic voltammetry curves of biofilm in the different MECs.

Microbial biofilm analysis

At the end of the experiment, CLSM was performed on the different anodes to reveal the morphology of the biofilms. No biofilm was observed on abiotic anodes (Fig. 4A). The biofilm attached to the anode was thickest at 0.6 V, with an active cell ratio of 50% (Fig. 4B). The active cells were 21% and 27% in sparse biofilms at 0.2 V and 1.0 V, showing that low and high applied voltages were detrimental to biofilm activity on the anode. Alpha diversity analysis further revealed the diversity of microbial communities at different voltages. The Simpson and Shannon indices of the bacterial community at 0.6 V were 0.95 and 6.3, which were smaller than those at other voltages (Table S2†). These smaller indices indicated a lower diversity of bacterial communities and a larger proportion of dominant bacterial communities, demonstrating that a voltage of 0.6 V was conducive to the evolution of E-Feammox bacteria.

Fig. 4 Confocal laser scanning microscopy images (A), active and inactive cells ratio (B), and cluster tree and relative abundance at the family level (C) of the different biofilms.

The colonies on the anodes under 0.2 V, 0.6 V, 1.0 V and ck-bio (Fig. 4C) were different, indicating that the applied voltages influenced the evolution of E-Feammox bacteria. The evolutionary direction of the Feammox bacteria under 0.6 V was significantly different from that at 0.2 V or1.0 V. The diverse performance of E-Feammox can be attributed to the different bacterial communities. The bacterial community in the different MECs was analyzed at the family level. The relative abundances of Geobacteraceae and Desulfobulbaceae, which are typical iron reduction bacteria, at 0.6 V were 23% and 13%.³⁹ However, Geobacteraceae and Desulfobulbaceae were not detected with no voltage control, suggesting that MEC promoted the multiplication of the iron reduction bacteria. The high abundance of the iron reduction bacteria further resulted in the highest content of Fe(II) in iron precipitates at 0.6 V. Xanthomonadaceae (1.9%), Comamonadaceae (1.9%) and Hydrogenophilaceae (0.64%) are known bacteria with ammonium oxidation ability.⁴⁰ These bacteria dominated ammonium oxidation in the E-Feammox process. Brucellaceae (0.50%), Phyllobacteriaceae (0.86%) and Hyphomicrobiaceae (1.3%) are also associated with the nitrogen cycle,⁴¹ such as denitrification, which is a potential pathway for producing N₂ due to the presence of nitrite. The high E-Feammox rate at 0.6 V should be attributed to the co-metabolism of a high abundance of the iron reduction bacteria and various nitrogen-related bacteria, in which Geobacteraceae play a crucial role. The families at 0.2 V and 1.0 V exhibited a relatively high abundance of Xanthomonadaceae, Comamonadaceae and Hydrogenophilaceae, explaining their capacity to oxidize ammonium. The presence of Geobacteraceae (10%) at 0.2 V was higher than at 1.0 V (0.92%), indicating that high potential was unfavorable for the growth of Geobacteraceae. The abundance of unclassified_Gemmatimonadetes at 1.0 V (7.6%) and no voltage (17%) was high, and they are frequently found in soil and potentially associated with Feammox.⁴²

The metabolic interactions between E-Feammox bacteria were further analyzed at the genus level. There were three main interactions among genera (Fig. 5A). First, Geobacter with the highest weight was positively correlated with unclassified_Desulfobulbaceae and unclassified_Thiobacterales, associated with Fe(III) reduction and extracellular electron transfer in the MEC. The positive correlation between unclassified_Gemmatimonadetes, Bacillus, Chryseobacterium and Dokdonella suggested interaction among them in the nitrogen cycle.^43,44 However, they exhibited a negative correlation with Geobacter. The negative interaction may be attributed to niche competition. Second, the ammonium oxidation genus Nitrosomonas formed a positive interaction with denitrification genera Thermomonas, Alicycliphilus and Pseudomonas,⁴⁵ indicating that the denitrification process was actuated when ammonium was oxidized to nitrite. The third interaction was also related to nitrogen metabolism. Ochrobactrum, Clostridium and Achromobacter were all found to have capability for ammonium oxidation,⁴⁶ but they exhibited a negative correlation with other nitrogen-related genera GOUTA19 and unclassified_Melioribacteraceae, which may be attributed to competition between them for nitrogen sources. Some denitrification bacteria, such as Thermomonas, Pseudomonas⁴⁷ and Thiobacillus,⁴⁸ were capable of oxidizing Fe(II) while reducing nitrate or nitrite, and they existed in various interaction networks. The positive interaction among iron reduction bacteria, iron oxidation bacteria, ammonium oxidation bacteria and denitrification bacteria fulfilled their critical roles in the E-Feammox process.

Fig. 5 (A) Correlation network analysis of functional bacteria at the genus level (P < 0.05, |Pearson's coefficient| ≥ 0.8); green (yellow) lines represent positive (negative) correlation; the size of nodes represent the weight of the bacteria. (B) Clusters of orthologous groups of proteins (COG) analysis of Feammox bacteria under 0.6 V and no voltage.

Based on the COG database, more functional proteins under 0.6 V compared to no voltage were analyzed to reveal the impact of the MEC on Feammox bacteria. The proportion of energy production and conversion at 0.6 V (7.6%) was higher than that for ck-bio of 6.4% (Fig. 5B), suggesting that the metabolic activity of the bacteria was more vigorous in the E-Feammox process. The proteins related to cell multiplication have a higher proportion at 0.6 V, such as cell cycle control, cell division, chromosome partitioning (1.2%), translation, ribosomal structure and biogenesis (7.5%), replication, recombination and repair (4.7%). The high multiplication capacity of the bacteria can form thicker biofilms, which also mutually corroborated the results of CLSM. The protein involved in signal transduction mechanisms should be key to the co-metabolism in the bacterial community. The proportion of 0.6 V (4.9%) was higher than that of ck-bio (4.0%), resulting in better interactions between E-Feammox bacteria under 0.6 V. The functional proteins that may be associated with electron transfer were translated more under 0.6 V, including coenzyme transport and metabolism (5.3%), cell motility (2.8%) and intracellular trafficking, and secretion and vesicular transport (3.0%). However, the pathway and mechanism of electron transfer by E-Feammox bacteria need to be further explored.

This study verified the adoption of an MEC to enhance the autotrophic Feammox performance via the iron cycle. Based on the above content, the mechanism of E-Feammox in MEC was proposed, as shown in Fig. 6. Inoculants containing Feammox bacteria and vivianite [Fe(II)] were added to the MEC system. Owing to the anodic potential being higher than the oxidation potential of vivianite and presence of iron oxidation bacteria, vivianite was gradually oxidized to leucophosphite [Fe(III)]. After the accumulation of Fe(III), the E-Feammox process was initiated to oxidize ammonium and reduce Fe(III). The regenerated Fe(II) was then oxidized again to Fe(III) by the anode and iron oxidation bacteria; thus, the cycling of iron was successfully driven in E-Feammox. Due to the positive interaction between iron-related and nitrogen-related bacteria, the iron cycle promoted the rate of the ammonium oxidation process. When ammonium was oxidized to nitrite or nitrate, some denitrification bacteria were awakened to oxidize Fe(II), which is similar to NDFO and is frequently found in the Feammox system.⁴⁹ Although the NDFO process was weak in the interaction network, it also strengthened the iron cycle. The MEC system enhanced Feammox mainly through following aspects: (1) boosting the metabolism of Feammox bacteria, including growth activity, bacterial interaction and electron transfer, to increase biomass and thick biofilm attachment to the anode; (2) promoting the evolvement of E-Feammox functional bacteria, such as Geobacteraceae, Xanthomonadaceae, Comamonadaceae, Desulfobulbaceae and Brucellaceae, to strengthen ammonium oxidation and the iron cycle; and (3) accelerating the oxidation of Fe(II) to Fe(III) to drive the cycling of iron. Therefore, MEC is a feasible technology for intensifying autotrophic Feammox in the treatment of ammonium-containing wastewater.

Fig. 6 Mechanism of E-Feammox with the iron cycle in an MEC system.

Conclusions

The ammonium oxidation rate of autotrophic Feammox was significantly enhanced in the MEC with applied voltages and was the highest at 0.6 V, followed by 1.0 V and 0.2 V. The iron cycle was successfully driven through the MEC system. The CV curves confirmed that ammonium oxidation and the iron cycle of E-Feammox occurred on the anode. The positive interactions between the functional bacteria facilitated ammonium oxidation and the iron cycle. This present study constructed an E-Feammox system to reduce the supply of iron resources and optimize the performance of autotrophic Feammox, which will be a progressive step forward for the treatment low-C/N wastewater in engineering.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was financially supported by “Pioneer” and “Leading Goose” R&D Program of Zhejiang (No. 2023C03017), the National Natural Science Foundation of China (No. 52270042, 52261145701, and 22036004), and the Ministry of Education of China (T2017002).

References

1. S. Sawayama, Possibility of anoxic ferric ammonium oxidation, J. Biosci. Bioeng., 2006, 101, 70–72,  DOI:10.1263/jbb.101.70.
2. W. H. Yang, K. A. Weber and W. L. Silver, Nitrogen loss from soil through anaerobic ammonium oxidation coupled to iron reduction, Nat. Geosci., 2012, 5, 538–541,  DOI:10.1038/ngeo1530.
3. H. Li, J. Su, X. Yang, G. Zhou, S. L. Lassen and Y. Zhu, RNA Stable Isotope Probing of Potential Feammox Population in Paddy Soil, Environ. Sci. Technol., 2019, 53, 4841–4849,  DOI:10.1021/acs.est.8b05016.
4. X. Li, L. Hou, M. Liu, Y. Zheng, G. Yin, X. Lin, L. Cheng, Y. Li and X. Hu, Evidence of Nitrogen Loss from Anaerobic Ammonium Oxidation Coupled with Ferric Iron Reduction in an Intertidal Wetland, Environ. Sci. Technol., 2015, 49, 11560–11568,  DOI:10.1021/acs.est.5b03419.
5. Z. Yao, L. Yang, N. Song, C. Wang and H. Jiang, Effect of organic matter derived from algae and macrophyte on anaerobic ammonium oxidation coupled to ferric iron reduction in the sediment of a shallow freshwater lake, Environ. Sci. Pollut. Res., 2019, 27, 25899–25907,  DOI:10.1007/s11356-019-06793-5.
6. Y. Xiong, Y. Du, Y. Deng, T. Ma and Y. Wang, Feammox in alluvial-lacustrine aquifer system: Nitrogen/iron isotopic and biogeochemical evidences, Water Res., 2022, 222, 118867,  DOI:10.1016/j.watres.2022.118867.
7. Y. Yang, Y. Zhang, Y. Li, H. Zhao and H. Peng, Nitrogen removal during anaerobic digestion of wasted activated sludge under supplementing Fe(III) compounds, Chem. Eng. J., 2018, 332, 711–716,  DOI:10.1016/j.cej.2017.09.133.
8. T. Zhu, W. Lai, Y. Zhang and Y. Liu, Feammox process driven anaerobic ammonium removal of wastewater treatment under supplementing Fe(III) compounds, Sci. Total Environ., 2022, 804, 149965,  DOI:10.1016/j.scitotenv.2021.149965.
9. L. Ding, X. An, S. Li, G. Zhang and Y. Zhu, Nitrogen Loss through Anaerobic Ammonium Oxidation Coupled to Iron Reduction from Paddy Soils in a Chronosequence, Environ. Sci. Technol., 2014, 48, 10641–10647,  DOI:10.1021/es503113s.
10. J. Cisternas, C. Rodríguez, J. Serrano and E. Leiva, Study of the key biotic and abiotic parameters influencing ammonium removal from wastewaters by Fe3+−mediated anaerobic ammonium oxidation (Feammox), Chemosphere, 2023, 339, 139463,  DOI:10.1016/j.chemosphere.2023.139463.
11. J. Zhu, T. Li, C. Liao, N. Li and X. Wang, A promising destiny for Feammox: From biogeochemical ammonium oxidation to wastewater treatment, Sci. Total Environ., 2021, 790, 148038,  DOI:10.1016/j.scitotenv.2021.148038.
12. L. Hu, X. Cheng, G. Qi, M. Zheng, Y. Dang, J. Li and K. Xu, Achieving Ammonium Removal Through Anammox-Derived Feammox With Low Demand of Fe(III), Front. Microbiol., 2022, 13, 918634,  DOI:10.3389/fmicb.2022.918634.
13. J. Li, W. Zeng, H. Liu, M. Zhan and H. Miao, Achieving deep autotrophic nitrogen removal in aerated biofilter driven by sponge iron: Performance and mechanism, Environ. Res., 2022, 213, 113653,  DOI:10.1016/j.envres.2022.113653.
14. L. Cheng, H. Liang, W. Yang, T. Yang, T. Chen and D. Gao, The biochar/Fe-modified biocarrier driven simultaneous NDFO and Feammox to remove nitrogen from eutrophic water, Water Res., 2023, 243, 120280,  DOI:10.1016/j.watres.2023.120280.
15. B. E. Logan, R. Rossi, A. A. Ragab and P. E. Saikaly, Electroactive microorganisms in bioelectrochemical systems, Nat. Rev. Microbiol., 2019, 17, 307–319,  DOI:10.1038/s41579-019-0173-x.
16. T. Zhu, Y. Zhang, Y. Liu and Z. Zhao, Electrostimulation enhanced ammonium removal during Fe(III) reduction coupled with anaerobic ammonium oxidation (Feammox) process, Sci. Total Environ., 2021, 751, 141703,  DOI:10.1016/j.scitotenv.2020.141703.
17. Z. Wang, C. Liao, Z. Zhong, S. Liu, M. Li and X. Wang, Design, optimization and application of a highly sensitive microbial electrolytic cell-based BOD biosensor, Environ. Res., 2023, 216, 114533,  DOI:10.1016/j.envres.2022.114533.
18. J. Zhu, X. Yan, L. Zhou, N. Li, C. Liao and X. Wang, Insight of bacteria and archaea in Feammox community enriched from different soils, Environ. Res., 2022, 203, 111802,  DOI:10.1016/j.envres.2021.111802.
19. D. R. Lovley and E. J. P. Phillips, Novel mode of microbial energy metabolism: organic carbon oxidation coupled to dissimilatory reduction of iron or manganese, Appl. Environ. Microbiol., 1988, 54, 1472–1480,  DOI:10.1128/aem.54.6.1472-1480.1988.
20. APHA, Standard Methods for the Examination of Water and Wastewate, American Public Health Association (APHA), Washington, DC, USA, 2005.
21. D. S. Frear and R. C. Burrell, Spectrophotometric Method for Determining Hydroxylamine Reductase Activity in Higher Plants, Anal. Chem., 1955, 27, 1664–1665,  DOI:10.1021/ac60106a054.
22. M. Ruiz-Urigüen, D. Steingart and P. R. Jaffé, Oxidation of ammonium by FeammoxAcidimicrobiaceaesp. A6 in anaerobic microbial electrolysis cells, Environ. Sci.: Water Res. Technol., 2019, 5, 1582–1592,  10.1039/c9ew00366e.
23. B. E. Logan, B. Hamelers, R. Rozendal, U. Schröder, J. Keller, S. Freguia, P. Aelterman, W. Verstraete and K. Rabaey, Microbial Fuel Cells: Methodology and Technology, Environ. Sci. Technol., 2006, 40, 5181–5192,  DOI:10.1021/es0605016.
24. A. Vilajeliu-Pons, C. Koch, M. D. Balaguer, J. Colprim, F. Harnisch and S. Puig, Microbial electricity driven anoxic ammonium removal, Water Res., 2018, 130, 168–175,  DOI:10.1016/j.watres.2017.11.059.
25. D. Zheng, W. Gu, Q. Zhou, L. Zhang, C. Wei, Q. Yang and D. Li, Ammonia oxidation and denitrification in a bio-anode single-chambered microbial electrolysis cell, Bioresour. Technol., 2020, 310, 123466,  DOI:10.1016/j.biortech.2020.123466.
26. D. R. Shaw, M. Ali, K. P. Katuri, J. A. Gralnick, J. Reimann, R. Mesman, L. van Niftrik, M. S. M. Jetten and P. E. Saikaly, Extracellular electron transfer-dependent anaerobic oxidation of ammonium by anammox bacteria, Nat. Commun., 2020, 11, 2058,  DOI:10.1038/s41467-020-16016-y.
27. R. C. Wagner, D. F. Call and B. E. Logan, Optimal Set Anode Potentials Vary in Bioelectrochemical Systems, Environ. Sci. Technol., 2010, 44, 6007–6522,  DOI:10.1021/es101013e.
28. A. Ding, Y. Yang, G. Sun and D. Wu, Impact of applied voltage on methane generation and microbial activities in an anaerobic microbial electrolysis cell (MEC), Chem. Eng. J., 2016, 283, 260–265,  DOI:10.1016/j.cej.2015.07.054.
29. M. Ruiz-Uriguen, W. Shuai and P. R. Jaffe, Electrode Colonization by the Feammox Bacterium Acidimicrobiaceae sp. Strain A6, Appl. Environ. Microbiol., 2018, 84(24), e02029-18,  DOI:10.1128/AEM.02029-18.
30. J. Wang, C. Zhu, Q. Li, D. Yan, L. Wang, T. He, Z. Cai, H. Zhou and X. Song, Anaerobic granular sludge coupled with artificial aeration or Fe-based substrate enhanced nitrogen conversion dynamic based on CW-MFCs, J. Water Process Eng., 2022, 45, 102483,  DOI:10.1016/j.jwpe.2021.102483.
31. Z. Liang, J. Shi, W. Yang, L. Dai and X. Dai, Coupling anammox and feammox via polymeric ferric sulfate: An efficient and aeration-saving way for nitrogen removal, J. Cleaner Prod., 2022, 355, 131788,  DOI:10.1016/j.jclepro.2022.131788.
32. X. Luo, L. Wen, L. Zhou and Y. Yuan, Facet Dependence of Biosynthesis of Vivianite from Iron Oxides by Geobacter sulfurreducens, Int. J. Environ. Res. Public Health, 2023, 20, 4247,  DOI:10.3390/ijerph20054247.
33. S. Wang, J. An, Y. Wan, Q. Du, X. Wang, X. Cheng and N. Li, Phosphorus Competition in Bioinduced Vivianite Recovery from Wastewater, Environ. Sci. Technol., 2018, 52, 13863–13870,  DOI:10.1021/acs.est.8b03022.
34. V. P. Lemos, M. L. D. Costa, R. L. Lemos and M. S. G. D. Faria, Vivianite and siderite in lateritic iron crust: an example of bioreduction, Quim. Nova, 2007, 30, 36–40,  DOI:10.1590/S0100-40422007000100008.
35. R. L. Frost, Y. Xi, R. Scholz, F. M. Belotti and M. C. Filho, Infrared and Raman Spectroscopic Characterization of the Phosphate Mineral Leucophosphite K(Fe3+)2(PO4)2(OH) · 2(H2O), Spectrosc. Lett., 2013, 46, 415–420,  DOI:10.1080/00387010.2012.733478.
36. Y. Liu, J. Jin, J. Li, Z. Zou, R. Lei, J. Sun and J. Xia, Enhanced Phosphorus Recovery as Vivianite from Anaerobically Digested Sewage Sludge with Magnetic Biochar Addition, Sustainability, 2022, 14(14), 8690,  DOI:10.3390/su14148690.
37. J. Tang, S. Chen, L. Huang, X. Zhong, G. Yang and S. Zhou, Acceleration of electroactive anammox (electroammox) start-up by switching acetate pre-acclimated biofilms to electroammox biofilms, Bioresour. Technol., 2017, 243, 1257–1261,  DOI:10.1016/j.biortech.2017.08.033.
38. M. Wei, J. Liu, Q. Yang, A. Xue, H. Wu, J. Ni, L. R. Winter, M. Elimelech and H. Zhao, Denitrification mechanism in oxygen-rich aquatic environments through long-distance electron transfer, npj Clean Water, 2022, 5, 61,  DOI:10.1038/s41545-022-00205-x.
39. S. I. Ishii, S. Suzuki, T. M. Norden-Krichmar, A. Wu, Y. Yamanaka, K. H. Nealson and O. Bretschger, Identifying the microbial communities and operational conditions for optimized wastewater treatment in microbial fuel cells, Water Res., 2013, 47, 7120–7130,  DOI:10.1016/j.watres.2013.07.048.
40. Z. Wang, Y. Yao, N. Steiner, H.-H. Cheng, Y.-J. Wu, S.-G. Woo and C. S. Criddle, Impacts of nitrogen-containing coagulants on the nitritation/denitrification of anaerobic digester centrate, Environ. Sci.: Water Res. Technol., 2020, 6, 3451–3459,  10.1039/d0ew00938e.
41. W. Alrashed, J. Lee, J. Park, B. E. Rittmann, Y. Tang, J. D. Neufeld and H.-S. Lee, Hypoxic methane oxidation coupled to denitrification in a membrane biofilm, Chem. Eng. J., 2018, 348, 745–753,  DOI:10.1016/j.cej.2018.04.202.
42. M. Oshiki, Y. Toyama, T. Suenaga, A. Terada, Y. Kasahara, T. Yamaguchi and N. Araki, N2O Reduction by Gemmatimonas aurantiaca and Potential Involvement of Gemmatimonadetes Bacteria in N2O Reduction in Agricultural Soils, Microbes Environ., 2022, 37(2), ME21090,  DOI:10.1264/jsme2.ME21090.
43. Z. Qiao, R. Sun, Y. Wu, S. Hu, X. Liu, J. Chan and X. Mi, Characteristics and metabolic pathway of the bacteria for heterotrophic nitrification and aerobic denitrification in aquatic ecosystems, Environ. Res., 2020, 191, 110069,  DOI:10.1016/j.envres.2020.110069.
44. J. Wang, Z. Lei, L. Wang, S. Yang, Y. Zhao, Y. Li and R. Chen, Insight into using up-flow anaerobic sludge blanket-anammox to remove nitrogen from an anaerobic membrane reactor during mainstream wastewater treatment, Bioresour. Technol., 2020, 314, 123710,  DOI:10.1016/j.biortech.2020.123710.
45. J. He, X. Xin, Z. Pei, L. Chen, Z. Chu, M. Zhao, X. Wu, B. Li, X. Tang and X. Xiao, Microbial profiles associated improving bioelectricity generation from sludge fermentation liquid via microbial fuel cells with adding fruit waste extracts, Bioresour. Technol., 2021, 337, 125452,  DOI:10.1016/j.biortech.2021.125452.
46. J. Ren, X. Bai, Y. Liu and X. Huang, Simultaneous nitrification and aerobic denitrification by a novel isolated Ochrobactrum anthropi HND19, Bioresour. Technol., 2021, 340, 125582,  DOI:10.1016/j.biortech.2021.125582.
47. W. Wang, B. Ding, Y. Hu, H. Zhang, Y. He, Y. She and Z. Li, Evidence for the occurrence of Feammox coupled with nitrate-dependent Fe(II) oxidation in natural enrichment cultures, Chemosphere, 2022, 303, 134903,  DOI:10.1016/j.chemosphere.2022.134903.
48. Y. Chen, X. Li, T. Liu, F. Li, W. Sun, L. Y. Young and W. Huang, Metagenomic analysis of Fe(II)-oxidizing bacteria for Fe(III) mineral formation and carbon assimilation under microoxic conditions in paddy soil, Sci. Total Environ., 2022, 851, 158068,  DOI:10.1016/j.scitotenv.2022.158068.
49. J. Li, W. Zeng, H. Liu, M. Zhan, H. Miao and X. Hao, Achieving deep autotrophic nitrogen removal from low strength ammonia nitrogen wastewater in aeration sponge iron biofilter: Simultaneous nitrification, Feammox, NDFO and Anammox, Chem. Eng. J., 2023, 460, 141755,  DOI:10.1016/j.cej.2023.141755.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4ew00074a

---