Transformation of Nitrogen and Iron Species during NitrogenRemoval from Wastewater via Feammox by Adding FerrihydriteYafei Yang, Zhen Jin, Xie Quan, and Yaobin Zhang*
Key Laboratory of Industrial Ecology and Environmental Engineering (Ministry of Education, China), School of EnvironmentalScience and Technology, Dalian University of Technology, Dalian 116024, China
*S Supporting Information
ABSTRACT: Fe(III) reduction coupled to anaerobic ammonium oxidation,namely, Feammox, plays an important role in the Fe/N cycle in naturalenvironments. However, it has been rarely studied in wastewater treatmentsystems. To date, transformation of nitrogen and iron species of Feammoxduring anaerobic digestion remain unknown. In this study, ferrihydrite wassupplemented (50 mM) in an anaerobic digester to remove ammoniumthrough Feammox. Results showed that the ammonium removal efficiencyafter 63 days reached 69.49%, significantly higher than that of the control(35.63%). X-ray diffraction analysis showed that ferrihydrite was trans-formed into magnetite and akaganeite after the experiment. Further studyfound that Fe(II) could be oxidized by NOx
− in the inoculants taken fromthe ferrihydrite-supplemented group; i.e., possible NOx
−-dependent Fe(II)oxidation could provide the possibility for further Feammox. Microbialanalysis showed that iron-reducing bacteria and iron-oxidizing bacteria were both detected in two groups.
KEYWORDS: Feammox, NOx−-dependent Fe(II) oxidation, Anaerobic digestion, Nitrogen loss
■ INTRODUCTION
In general, the nitrification−denitrification and anaerobicammonium oxidation (anammox) processes have beenconsidered as the two main approaches to biological nitrogenremoval.1−3 The nitrification−denitrification process operatedin separate aerobic and anaerobic tanks is energy intensive andrequires amendment with extra carbon sources in many low C/N wastewaters. The anammox process, which removesammonium with nitrite as the electron acceptor under anoxicconditions (NH4
+ + NO2− → N2 + 2H2O), is sensitive to the
environmental conditions, and the proliferation of anammoxbacteria is slow.4−6
Recently, Feammox, i.e., ferric iron [Fe(III)] reductioncoupled to anaerobic ammonium oxidation, has been reportedto play an important role in the nitrogen cycles of naturalenvironments.7−10 In Feammox, Fe(III) is reduced to Fe(II),accompanied by oxidation of NH4
+ to N2, NO2−, and NO3
−
(eqs 1−3). Feammox has been estimated to metabolize 7.8−61(3.9%−31% of nitrogen fertilizer loss) kg of NH4
+/ha/year inpaddy soil.11 In our previous study, it was found that 20.1%total nitrogen removal was achieved when supplementingFe(OH)3 in a high-ammonium anaerobic sludge digester.12
Although the efficiency was not high, Feammox had showed itspotential to run as an in-site anaerobic ammonium treatmentprocess. However, little is known about the transformation andpossible interactions of nitrogen and iron species duringFeammox as well as the potential effects on the nitrogenremoval in anaerobic wastewater treatment. For example,NOx
− as a product of Feammox was reported to be capable of
oxidizing Fe(II) to Fe(III), termed as NOx−-dependent Fe(II)
oxidation (NDFO).13−15 In other words, Fe(III) is likelyregenerated by NDFO.Therefore, the aim of this study was to investigate the effects
of Feammox on nitrogen removal, including the followingcontents: (1) changes in the nitrogen and iron species duringFeammox, (2) possible interaction during ferrihydrite relatedFeammox process and their potential effects on the nitrogenremoval, and (3) bacterial communities analysis.
3Fe(OH) 5H NH 3Fe 9H O 0.5N
G 245KJ/mol3 4
22 2
r m
+ + → + +
Δ = −
+ + +
(1)
6Fe(OH) 10H NH 6Fe 16H O NO
G 164KJ/mol3 4
22 2
r m
+ + → + +
Δ = −
+ + + −
(2)
8Fe(OH) 14H NH 8Fe 21H O NO
G 207KJ/mol3 4
22 3
r m
+ + → + +
Δ = −
+ + + −
(3)
■ MATERIALS AND METHODSExperimental Procedures. Inoculant sludge with a solid content
of 10% was taken from a sludge treatment plant in Dalian (China)and stored in a refrigerator at 4 °C. The ferrihydrite used in this study
Received: June 29, 2018Revised: September 18, 2018Published: September 21, 2018
was synthesized by adding Fe(III) chloride to water that wasmaintained at a pH of 6.8−7.2.16 Morphology of the synthesizedferrihydrite characterized by a scanning electron microscope is shownin Figure S1. Prior to incubation, the sterile anoxic deionized water(autoclaved at 120 °C for 20 min) was added into the inoculantsludge at a ratio of 5:1 (v/v) and preincubated anoxically in the darkat 25 °C for 1 days statically; then, the supernatant after centrifugationwas pured out to remove indigenous NH4
+ and NOx−. The above
procedure was repeated three times. Then, the prepared sludge afterremoving the supernatant was used as the inoculant sludge.Two batch experiments were conducted in 200 mL serum vials: a
control group and a ferrihydrite group. Here, 10 mL of preparedinoculant sludge was added into each vial as the seed sludge.Ferrihydrite was supplemented in the ferrihydrite group with a dose of50 mM, while Fe(III) was not supplemented in the control. Then,NH4
+-containing medium was added to ensure that the initial volumeof each serum vial was 150 mL. The medium (pH 5.0) containsMgCl2·6H2O (0.4 g/L), CaCl2·H2O (0.1 g/L), NH4Cl (0.027 g/L),KH2PO4 (0.6 g/L), 1 mL/L of a vitamin solution, 1 mL/L of a traceelement solution, and 30 mM bicarbonate buffer.16 Afterward, theserum vials were sealed with butyl rubber septa and crimped withaluminum caps. The headspace of the vial was flushed with N2 andCO2 (80%/20%) for 30 min. Then, the vials were incubated staticallyat 25 °C in the dark for 87 days. The experiments were repeated intriplicate. The initial parameters of the mixture in the control andferrihydrite groups (before adding ferrihydrite) are shown in Table 1.
An anaerobic NDFO experiment was conducted using ahomogenized slurry collected from the ferrihydrite group after theabove-mentioned experiment. This experiment was operated underthe following three culture conditions: (i) 1 mL of slurry + 1.2 mMNO3
−, (ii) 1 mL of slurry + 1.2 mM NO2−, and (iii) 1 mL of slurry +
sterilized deionized water. These three cultures were added in three100 mL-serum vials containing 50 mL of sterilized medium. Afterflushing the headspace of the vials with N2 and CO2 (80%/20%), thevials were incubated statically at 25 °C in the dark for 30 days. Priorto sampling, the vials were thoroughly mixed to ensure samplingequally. The experiments were repeated in triplicate.Analysis Methods. The pH was determined with a dual-channel
pH-ion-conductivity-dissolved oxygen meter (X60, Fisher Scientific).The total chemical oxygen demand (TCOD) was analyzed accordingto a reference.17 The contents of protein, polysaccharide, and totalnitrogen were determined based on the method described in ourprevious report.12 The well-mixed slurry was sampled inside theanaerobic chamber and centrifuging (10,000 rpm) for 5 min. Thesupernatant was filtered through a 0.22 μm PTFE filter. The filtratewas then applied for measuring NH4
+, NO2−, and NO3
− using aDionex DX1000 ion chromatograph (IC) system. The well-mixedslurry was dried in a desiccator located inside the anoxic chamber toavoid oxidation. The dried solids were analyzed by X-ray diffraction(XRD)18,19 and X-ray photoelectron spectroscopy (XPS)20−22 toidentify the crystalline iron phases present in the sample. An XRDmeasurement was carried out via an XRD diffractometer (Empyrean,
Panalytical, Netherlands) with Cu Kα radiation at 45 kV and 200 mA,and the XRD pattern was analyzed by the software Jade. An XPSmeasurement was performed using a 250Xi system (Escalab 205Xi,Thermofisher, America), and XPSPEAK4.1 software was used forpattern analysis.
Fe(II) and Fe(III) contents in the slurry were analyzed accordingto the method described by Ding et al.11 In brief, 1 mL of slurry wasimmediately extracted with 5 mL of 0.5 M HCl for 2 h at roomtemperature, and then, the mixture was centrifuged at 10,000 rpm for10 min. The Fe(II) and total iron concentrations in the supernatantwere measured according to the standard method.23 The amount ofFe(III) was calculated as the difference between the total iron andFe(II). All extractions were conducted in an anoxic glovebox. Thedetails of Fe analysis are provided in the Supporting Information.
After the 87-day Feammox experiment, the microbial communitiesin the two groups were analyzed via high-throughput 16S rRNApyrosequencing. The detailed methods of DNA extraction, PCR, andsequencing are also provided in the Supporting Information.
■ RESULTS AND DISCUSSION
Changes in the Nitrogen Species during the Experi-ment. Figure 1a demonstrates the changes in the NH4
+-Ncontent during the experiment. In the initial 20 days, theNH4
+-N content increased drastically from 10.14 to 66.15 mg/L in the ferrihydrite group, while that in the control groupincreased slightly to 21.82 mg/L, indicating that ferrihydriteaccelerated the release of ammonium from the decompositionof proteins that was a major organic component of sludge.24 Asshown in Figure 1b, the protein of the ferrihydrite group onday 37 was undetectable, while 21.1 mg/L protein was stillobserved in the control, meaning that protein removalefficiency was higher with amending ferrihydrite. These resultsagreed well with previous reports that Fe(III) enhanced theanaerobic decomposition of protein.25 After the first 20 days,the NH4
+-N content of the ferrihydrite group exhibited anobvious downward trend to 20.05 mg/L on day 87. In contrast,the NH4
+-N content of the control increased to 35.62 mg/L onday 87; this content was higher than that of the ferrihydritegroup. Ammonium is an intermediate product of proteindegradation that might be consumed by the potentialFeammox process, which could explain the fluctuation in theammonium content observed in the two groups. In addition toammonium, organic matters could also provide electrons forthe microbial dissimilatory Fe(III) reduction. From thethermodynamic point of view, organic substrates are moresuitable to use as electron donors in the Fe(III) reduction thanammonium.26 Therefore, as the organic matters wereconsumed over time, the dissimilatory Fe(III) reduction wasmore inclined to use ammonium as the electron donor, whichcould be a main reason for the lag in ammonium removal(Figure 1a). The control group contained 49.85 mg/L ofendogenous Fe(III) (Table 1), which also resulted in nitrogenloss through Feammox, although the nitrogen loss wasobviously lower than that in the ferrihydrite group.Notably, the ammonium content of the ferrihydrite group
showed less changes during the period from day 36−47, whilethe content decreased again after sodium bicarbonate(NaHCO3) was added into each serum vial to a final contentof 1.5 mM on day 49. This change was consistent withFeammox being inorganically autotrophic27,28 and requiring asufficient amount of inorganic carbon as the carbon source. Tofurther verify that ammonium was removed in the presence offerrihydrite, on the 87th day, an extra 73.2 mg/L of NH4
+-Nwas added to the ferrihydrite group. Meanwhile, an additional
Table 1. Main Initial Characteristics before Incubationa
在200mL血清瓶中进行两次批次实验：对照组和水铁矿组。在此，将10mL制备的孕育剂污泥添加到每个小瓶中作为种子污泥。在水铁矿组中以50mM的剂量补充水铁矿，而在对照中不补充Fe（III）。然后，加入含NH 4 +的培养基以确保每个血清小瓶的初始体积为150mL。培养基（pH 5.0）含有MgCl2·6H2O（0.4 g / L），CaCl2·H2O（0.1 g / L），NH4Cl（0.027 g / L），KH2PO4（0.6 g / L），1 mL / L维生素溶液，1mL / L的微量元素溶液和30mM碳酸氢盐溶液。然后，将血清小瓶用丁基橡胶隔片密封并用铝盖卷曲。将小瓶的顶部空间用N 2和CO 2（80％/ 20％）洗涤30分钟。然后，将小瓶在25℃下在黑暗中静态孵育87天。一式三份重复实验。对照和水铁矿组中的混合物的初始参数（在加入水铁矿之前）示于表1中。
通过X射线衍射（XRD）18,19和X射线光电子能谱（XPS）20-22分析干燥的固体，以鉴定样品中存在的结晶铁相。通过XRD di ff ractometer（Empyrean，Panalytical，Netherlands）在45kV和200mA下用CuKα辐射进行XRD测量，并通过软件Jade分析XRD图谱。使用250Xi系统（Escalab 205Xi，Thermo fi sher，America）进行XPS测量，并使用XPSPEAK4.1软件进行模式分析。
bicarbonate was also added into the system as the carbonsource (final content was 1.5 mM). At this time, the content ofFe(III) in the ferrihydrite group was 1294.5 mg/L. As shownin Figure 1a, the NH4
+-N content dropped again to 54.5 mg/Lafter the following 30 days. The pH of the two systems rangedfrom 5.0 to 6.2 during the experiment, and these weakly acidicconditions prevented volatilization of the ammonium. Anabiotic experiment with no inoculant sludge was conducted inparallel and showed that the ammonium content almostremained unchanged, i.e., 7.43 ± 0.08 mg/L, when addingferrihydrite alone, further indicating that ammonium removalhere was a biotic process (Figure 1c).
In Feammox, NH4+ was directly converted to N2 or oxidized
to NOx−, which might be further denitrified into N2.
Considering that NH4+ is an intermediate in the anaerobic
decomposition of protein, the total nitrogen, including theorganic and inorganic nitrogen of a well-mixed slurry, wasdetermined to further clarify the Feammox process. On day 37,the total nitrogen in the control decreased from the initial levelof 106.48 to 96.54 mg/L, while the total nitrogen in theferrihydrite group decreased from 106.48 to 68.01 mg/L(Figure 1d). Namely, on day 37, the total nitrogen removalefficiency in the two systems was 9.33% and 36.13%,respectively. On day 63, the removal efficiency of totalnitrogen was 35.63% in the control, while that in the
Figure 1. (a) Changes in the ammonium content in the control and ferrihydrite groups during the experiment. (b) Total protein content in thedifferent groups on day 37. (c) Changes in the ammonium content under abiotic conditions. (d) Total nitrogen and removal efficiency on days 37and 63 for the two groups. Error bars represent standard deviations.
Figure 2. (a) Fe(II) contents in the different groups during the experiment. (b) Fe(III) contents in the control group during the experiment. Errorbars represent standard deviations.
ACS Sustainable Chemistry & Engineering Research Article
ferrihydrite group reached 69.49%. This difference indicatedthat ferrihydrite effectively promoted nitrogen removal.Remarkably, the nitrogen removal efficiency in the controlgroup was 9.33% on day 37 and rose to 35.63% on day 63.This was related to the presence of endogenous Fe(III) (49.85mg/L) in the sludge, which was mainly originated from thesludge dewatering process that used iron salts as coagulants.NO2
− and NO3− are also potential products of Feammox
(eqs 2 and 3); however, these species were nearly undetectedin this study (Figures S2 and S3). One reason for this resultwas that N2 was the dominant product of Feammoxthermodynamically (eqs 1−3), which lowered the productionof NOx
−. Zhou et al. also found that NOx− was nearly
undetected during the Feammox process in paddy soil.16 TheN2 directly produced from Feammox was reported to accountfrom 67% to 78% of the total N2 emission.11 On the otherhand, the NOx
− generated was easily reduced by organicmatters (namely, denitrification) and Fe(II) (E(NO2
−/N2O) =1.31 V vs E(Fe3+/Fe2+) = 0.2 V, pH 7).Changes in Iron. As a product of Feammox driven by
dissimilatory Fe(III) reduction, the Fe(II) content could beused to quantify the extent of Fe(III) reduction. As shown inFigure 2a, on day 6, the Fe(II) content in the ferrihydritegroup was 1026.78 mg/L, approximately 10-fold higher thanthat in the control group (106.98 mg/L). Besides morenitrogen removal in the ferrihydrite group (Figure 1d), TCODremoval efficiency reached 82.22%, compared with 48.52% inthe control (Figure 3). Therefore, the generation of moreFe(II) was consistent with the more nitrogen and TCODremoval in the ferrihydrite group compared with the control.
Figure 2b illustrates the changes in the Fe(III) content ofthe control. As shown in Figure 2a and b, the contents ofFe(II) and Fe(III) fluctuated. In this anoxic system, thedissimilatory Fe(III) reduction by organic matters and NH4
+
(i.e., Feammox) as electron donors was responsible for thegeneration of Fe(II). The absence of detectable NOx
− in thisstudy did not necessarily mean that no NOx
− was produced.Instead, anaerobic NDFO (eqs 4 and 5) might explain theincrease in Fe(III) and decrease in Fe(II) (Figure 2a and b).However, 8 or 6 mol Fe(II) could be produced when
producing 1 mol NO3− or NO2
− according to eqs 2 and 3,while only 5 mol (eq 4) and 2 mol (eq 5) Fe(III) could beproduced when even all NO3
− and NO2− were reduced to N2
via NDFO. The imbalance of electron transfer between Fe(III)
reduction and Fe(II) oxidation might lead to less and lessNOx
− production in each circle and then gradually weakenNDFO and Fe(III) regeneration. This meant that theproduced Fe(II) cannot be completely transformed back toFe(III) when only considering Feammox and subsequentNDFO. Regarding the fluctuant changes in the Fe(II) andFe(III) contents (Figure 2), it was assumed that other originalmetal oxides were involved in the N/Fe cycle. From EDXresults (Figure.S4), besides iron, heavy metals Cr, Pd, etc. weredetected in the sludge. These polyvalent metal oxides werereportedly capable of inducing dissimilatory metal reduc-tion,29,30 thereby likely participating in ammonium oxidationlike Feammox,31−33 which is required to be further studied. Ifthis assumption was true, the extra NOx
− might be producedfrom ammonium oxidation by these polyvalent metal oxides,further oxidizing Fe(II). Similar results of Fe(III) and Fe(II)during Feammox were also observed in sediment.16
Potential NDFO Process. NDFO was possible in thisstudy when considering that NOx
− was the product ofFeammox (eqs 2 and 3). Particularly with the consumptionof organics, denitrification gradually decreased, and NDFOlikely gradually increased. It means that Fe(III) is likelyregenerated by NDFO in this study. To further identify theiron transformation during NDFO, a NDFO experiment in theinoculants taken from the ferrihydrite-supplemented group wasconducted. As shown in Figure 4a, the NO3
−-N contentdecreased rapidly from the initial level of 17.1 to 3.09 mg/L onday 2, resulting in a decrease in the Fe(II) content from theinitial level of 51.32 to 10.98 mg/L on day 5, while the Fe(II)content of the NO3
−_free group remained nearly unchanged(Figure 4b). Similar results also occurred in the reactoramended with NO2
−; e.g., the Fe(II) content also decreasedsharply when supplementing NO2
− (Figure S5). This resultindicated that NDFO occurred in the presence of NOx
− andFe(II) in this system, which was consistent with the results ofBao and Li,34 who found that Fe(II) was oxidized to Fe(III)when NO3
− was added to an Fe(II)-containing anoxic system.Fe(II) cannot be oxidized to Fe(III) under anoxic
conditions unless stronger oxidizing agents,35 e.g., NOx−, are
present (eqs 4 and 5). The NOx− species produced from
Feammox likely served as an electron acceptor in the oxidationof Fe(II) through NDFO. In turn, the possible regeneratedFe(III) could likely participate in the Feammox process again,resulting in further nitrogen removal. The further solidevidence supporting NDFO occurrence in this system will beexplored in future studies.
Iron Speciation during Incubation. The physiochemicalnature of the Fe (hydr)oxide, including Fe(III) reduction andpossible NDFO, could influence Fe (hydr)oxide trans-formation.18 As demonstrated in Figure 5a, the XRD patternof the original sludge showed characteristic peaks at 21.2°,26.3°, 36.4°, 46.8°, and 60.19° ascribed to the (1052), (2874),(772), (706), and (662) planes, respectively, corresponding tothe standard card of iron hydroxide (Fe(OH)3) (JCPDS CardNo. 38-0032). It indicated that the endogenous Fe(III) in theoriginal sludge from the dewatering process was in the form ofiron hydroxide. The peaks located at 709.9 and 724.0 eV20 ofbinding energies shown in the XPS pattern (Figure 5b)
Figure 3. TCOD contents in the different groups on day 37. Errorbars represent standard deviations.
ACS Sustainable Chemistry & Engineering Research Article
indicated the presence of Fe(II) and Fe(III), consistent withthat Fe(II) and Fe(III) both existed in original sludge (Table1). The Fe(III) added as ferrihydrite in the ferrihydrite groupwas often inadequately described as Fe(OH)3 because thesetwo iron compounds have similar physicochemical proper-ties.36 A slight change in the XRD spectrum of the controlgroup was observed after the experiment; i.e., besides ironhydroxide (Fe(OH)3) (Figure 6a), a new iron species, irontartrate hydrate (C4H4FeO6·H2O) (Figure 6b), was formed.Specifically, the characteristic peaks at 21.2°, 26.3°, 36.4°,60.19°, and 67.92° ascribed to the (1002), (3728), (1057),(618), and (599) planes, respectively, of iron hydroxide(JCPDS Card No. 38-0032), and the characteristic peaks at
26.2°, 32.4°, 50.07° were ascribed to the (3728), (1006), and(889) planes, respectively, corresponding to the standard cardof iron tartrate hydrate (JCPDS Card No. 23-0299). While theXRD pattern of the iron minerals in the slurry of theferrihydrite group after the experiment showed that thecharacteristic peaks at 30.09°, 35.42°, and 37.05° were ascribedto the (1504), (1602), and (1602) planes, respectively,corresponding to the standard card of magnetite (JCPDSCard NO. 19-0629) (Figure 7a), and the characteristic peaks at26.72°, 35.16°, and 39.22° were ascribed to the (1438),(1326), and (1282) planes, respectively, of akaganeite (β-FeOOH) (Figure 7b) (JCPDS Card No. 34-1266); i.e.,magnetite and akaganeite were formed after experiment. It
Figure 4. Changes in the (a) NO3− and (b) Fe(II) contents with and without NO3
− under anaerobic conditions.
Figure 5. XRD (a) and Fe 2p XPS (b) spectrum patterns of the original sludge. The XRD pattern for standard rutile is shown as straight lines (a).
Figure 6. XRD pattern of the iron minerals in the sludge of the control group after the experiment. The XRD pattern for standard rutile is shown asstraight lines [iron hydroxide (a) and iron tartrate hydrate (b)].
ACS Sustainable Chemistry & Engineering Research Article
meant that the original ferrihydrite transformed into magnetiteand akaganeite. This result was consistent with those ofprevious studies,18 showing that ferrihydrite could beconverted to magnetite and other Fe(III) oxides. Besidespossible NDFO, Boland et al. indicated that with the presenceof aqueous Fe(II) ferrihydrite could be chemically transformedto more crystalline minerals.37 The observed changes in thecrystalline iron, especially in the ferrihydrite group, indicatedFe(III) reduction and regeneration.Proportion of Iron Reduction Associated with
Feammox. Under anoxic conditions, the potential ofassimilation by microbes, anammox, denitrification, andpossible NDFO could contribute to nitrogen loss (Table 2).7
However, for the assimilation, NH4+ could be converted by
microbes to nitrogenous organic matters such as proteins,which might be mostly decomposed into NH4
+ again. Thereminder still stayed in the sludge slurry with no way tobecome N2 to generate nitrogen loss during assimilation. Fordenitrification, anammox, and NDFO, these processes reliedon Feammox to generate NOx
−. Therefore, the nitrogen losscould result from the following two pathways: (i) N2 directlyproduced from Feammox. (ii) Feammox-generated NOx
−
followed by denitrification, possible NDFO, or anammox(anammox was ignored based on microbiological analysisbelow). In other words, the nitrogen loss in this study wascaused by Feammox, regardless of directly producing N2 orfirst producing NOx
− then being reduced to N2 (Figure 8).Organic materials and ammonium can both serve as electrondonors in the dissimilatory Fe(III) reduction. On the basis ofthermodynamics, the majority of Fe(III) reduction wasinvolved preferentially with the oxidation of organics ratherthan with ammonium. Therefore, the proportion of Fe(II)generated from Feammox was not high. Ding et al. found thatonly 0.81%−2.2% of the Fe(III) reduction in paddy soils wasassociated with Feammox and increasing the organic contentled to a decrease in the amount of Fe(II) generated fromFeammox.11 The organic content in the wastewater treatmentsystem in this study was much higher than that in the naturalenvironment; thus, the proportion of Fe(II) generated fromFeammox should be lower in this study than in that by Ding etal. In contrast, based on the nitrogen loss and thestoichiometric Fe/N ratio of 3 (eq 1), Fe(II) generated fromFeammox in the ferrihydrite group accounted for 30.3% and65.0% of the total Fe(II) on days 37 and 63, respectively,which were much higher than the proportions reported inprevious studies (0.81%−2.2% and 0.4%−6.1%).11,16 It waslikely because that NDFO was indispensable to reproduceFe(III) (Figure 8), consistent with an NOx
−-reducing Fe(II)oxidizer (see Microbial Community below) detected insystems, leading to the more nitrogen removal and higherproportion of Fe(III) reduction through Feammox. Notably,the endogenous Fe(III) content of 49.85 mg/L in the controlwas not sufficient for removing such a large amount of nitrogenby day 63. Specifically, only 4.15 mg/L of nitrogen couldtheoretically be removed via Feammox based on the 49.85 mg/L Fe(III) content in the control based on the stoichiometric
Figure 7. XRD pattern of the iron minerals in the sludge of the ferrihydrite group after the experiment. The XRD pattern for standard rutile isshown as straight lines [magnetite (a) and akaganeite (b)].
Table 2. Potential Pathway of Nitrogen Loss under AnoxicConditions
N loss Mechanism
Feammox eqs 1−3anammox NH4
+ + NO2− → N2 + 2H2O
assimilation by microbes Used for microbial proliferation.denitrification Organics act as electron donorsNDFOa eqs 4 and 5
aNOx−-dependent Fe(II) oxidation.
Figure 8. Diagram of nitrogen removal in this study.
ACS Sustainable Chemistry & Engineering Research Article
Fe/N ratio of 3 (eq 1). Nevertheless, 37.94 mg of nitrogen wasremoved on day 63 in the control, which was more than thetheoretical amount, implying that polyvalent metal oxidesexisting in the sludge (Figure S4) was likely involved in thenitrogen removal. Importantly, the operating time was anotherfactor that influences the proportion of Fe(III) reductionassociated with Feammox. In general, the proportion of Fe(III)reduction associated with Feammox in the ferrihydrite groupincreased over time, e.g., from 30.3% on day 37 to 65.0% onday 63. As the organic matters were depleted during digestion,ammonium gradually became the main electron donor forFe(III) reduction.Microbial Community. To date, the underlying micro-
biological process of Feammox has not been clarified. Mostresearchers have suggested that iron-reducing bacteriaparticipate in Feammox (Figure S6). For example, Zhou etal. ascribed the Feammox process in paddy soil to the functionof Geobacter, an iron-reducing bacterium.16 Huang and Jaffe ́stated that an uncultured Acidimicrobiaceae bacterium capableof reducing Fe(III) played a key role in Feammox in a forestedriparian wetland.27 In this study, the iron-reducing bacteriawere also enriched. As shown in Figure 9a, Deltaproteobacteria,which contains iron-reducing bacteria, e.g., Geobacter,accounted for 6.02% of the bacteria 16S rRNA gene sequencesat the class level in the control, while this class accounted for12.8% of the bacteria in the ferrihydrite-supplemented group.Further identification at the genus level showed that Geobactermade up 0.04% and 0.48% of the total sequences in the controland ferrihydrite groups, respectively (Figure 9b). The lowabundance of Geobacter in the sludge agreed with the results ofprevious reports,38 indicating that the inoculant sludge was notan ideal system for enriching Geobacter compared with thenatural habitat. This low abundance of Geobacter was probablyrelated to the low efficiency of nitrogen removal in this study(69.49% on day 63 in the ferrihydrite group, Figure 1d).The NO3
−-reducing Fe(II) oxidizer Acidovorax (Figure 9b)and the family Comamonadaceae (Figure 9c), belonging toiron-oxidizing bacteria, were also detected in both groups.Particularly, the abundance of Acidovorax was 1.18% and0.81% in the control and ferrihydrite groups, respectively, andthe abundance of Comamonadaceae was 1.28% and 0.44%.Putative denitrifiers, including Acinetobacter, Rhodoplanes,Bacillus, and Pseudomonas,39 were detected in the two groups,indicating that denitrification of NOx
− from Feammox was analternative to the consumption of NOx
− (Figures S2 and S3).Notably, one microbe at the genus level, Aminicenantes_ge-
nera_incertae_sedis, was abundantly enriched in the control(19.88%) and the ferrihydrite group (40.34%) (Figure 9b).Aminicenantes has been found to adapt and utilize a variety ofcomplex sugar polymers and amino acids, such as glycine,glutamate, and aspartate,40 which was consistent with theobserved polysaccharide and protein removal (Figure 1b andFigure S7). The genes for the assimilatory acquisition ofnitrogen were also identified in Aminicenantes. However, theability of Aminicenantes_genera_incertae_sedis to reduce Fe-(III), even when mediated by Feammox, is unknown.
■ IMPLICATIONAmmonium is one of mainly concerned pollutants due to therisk to cause eutrophication. Feammox has increasingly beenreported in natural systems. However, this nitrogen removalmodel has rarely been investigated in wastewater treatmentsystems. In this study, the nitrogen removal efficiency in the
ferrihydrite group was significantly higher than that of thecontrol (Figure 1d), implying that ferrihydrite inducedanaerobic NH4
+ oxidation and further nitrogen removal.Although the nitrogen removal efficiency was not high, e.g.,69.49% of nitrogen was removed during 63 days whensupplementing ferrihydrite, the study still demonstrated thepossibility of application of Feammox in anaerobic nitrogenremoval from wastewater. Although the results suggested thatFeammox was related to nitrogen removal, the direct evidencestill lacked. 15N-labeled isotopic is an ideal tool to characterizenitrogen loss in Feammox, which will be investigated in ournext study.
Figure 9. Relative abundances of bacteria at the class level (a), genuslevel (b), and family level (c) in the different groups on day 87.
ACS Sustainable Chemistry & Engineering Research Article
■ CONCLUSIONSIn this study, ferrihydrite was supplemented in an anaerobicwastewater treatment reactor to investigate its effects onnitrogen removal. The results showed that nitrogen removalefficiency increased by 33.86% after 63 days compared to thecontrol. As part products of Feammox, Fe(II) and NOx
− likelyreacted together in this study, resulting in NDFO to generateFe(III). XRD results showed that new species of Fe(III)compounds, magnetite and akaganeite, were detected duringFe (Hydr)oxide transformation. Microbial community analysisdemonstrated that iron-reducing bacteria and iron-oxidizingbacteria were enriched in the two groups, indicating that thecooperation of two bacteria was essential for the ammoniumremoval in Fe(III)-containing wastewater treatment systems.
■ ASSOCIATED CONTENT*S Supporting InformationThe Supporting Information is available free of charge on theACS Publications website at DOI: 10.1021/acssusche-meng.8b03083.
Morphology of synthesized ferrihydrite via a scanningelectron microscope. Specific method of Fe(II) andFe(III) determination. DNA extraction, PCR amplifica-tion, and high-throughput 16S rRNA pyrosequencing.NO3
− and NO2− contents during the experiment. EDX
of sludge slurry taken from a sludge treatment plant.Changes in Fe(II) contents with and without NO2
−
under anoxic conditions. Mechanism model of Feam-mox. Total polysaccharide content in different groups onday 37. (PDF)
■ ACKNOWLEDGMENTSThe authors acknowledge the financial support from theNational Natural Scientific Foundation of China (21777016).
■ REFERENCES(1) Cavigelli, M. A.; Robertson, G. P. The functional significance ofdenitrifier community composition in a terrestrial ecosystem. Ecology2000, 81 (5), 1402−1414.(2) Spott, O.; Russow, R.; Stange, C. F. Formation of hybrid N2Oand hybrid N2 due to codenitrification: First review of a barelyconsidered process of microbially mediated N-nitrosation. Soil Biol.Biochem. 2011, 43 (10), 1995−2011.(3) Wu, L.; Liang, D.; Xu, Y.; Liu, T.; Peng, Y.; Zhang, J. A robustand cost-effective integrated process for nitrogen and bio-refractoryorganics removal from landfill leachate via short-cut nitrification,anaerobic ammonium oxidation in tandem with electrochemicaloxidation. Bioresour. Technol. 2016, 212, 296−301.(4) Dalsgaard, T.; Thamdrup, B.; Canfield, D. E. Anaerobicammonium oxidation (anammox) in the marine environment. Res.Microbiol. 2005, 156 (4), 457−464.
(5) Xing, B.; Guo, Q.; Zhang, Z.; Zhang, J.; Wang, H.; Jin, R.Optimization of process performance in a granule-based anaerobicammonium oxidation (anammox) upflow anaerobic sludge blanket(UASB) reactor. Bioresour. Technol. 2014, 170, 404−412.(6) Dale, O. R.; Tobias, C. R.; Song, B. Biogeographical distributionof diverse anaerobic ammonium oxidizing (anammox) bacteria inCape Fear River Estuary. Environ. Microbiol. 2009, 11 (5), 1194−1207.(7) Yang, W. H.; Weber, K. A.; Silver, W. L. Nitrogen loss from soilthrough anaerobic ammonium oxidation coupled to iron reduction.Nat. Geosci. 2012, 5 (8), 538−541.(8) Clement, J.; Shrestha, J.; Ehrenfeld, J.; Jaffe, P. Ammoniumoxidation coupled to dissimilatory reduction of iron under anaerobicconditions in wetland soils. Soil Biol. Biochem. 2005, 37 (12), 2323−2328.(9) Shrestha, J.; Rich, J. J.; Ehrenfeld, J. G.; Jaffe, P. R. Oxidation ofAmmonium to Nitrite Under Iron-Reducing Conditions in WetlandSoils Laboratory, Field Demonstrations, and Push-Pull RateDetermination. Soil Sci. 2009, 174 (3), 156−164.(10) Xu, H.; Wang, X.; Li, H.; Yao, H.; Su, J.; Zhu, Y. BiocharImpacts Soil Microbial Community Composition and NitrogenCycling in an Acidic Soil Planted with Rape. Environ. Sci. Technol.2014, 48 (16), 9391−9399.(11) Ding, L. J.; An, X. L.; Li, S.; Zhang, G. L.; Zhu, Y. G. Nitrogenloss through anaerobic ammonium oxidation coupled to ironreduction from paddy soils in a chronosequence. Environ. Sci. Technol.2014, 48 (18), 10641−7.(12) Yang, Y.; Zhang, Y.; Li, Y.; Zhao, H.; Peng, H. Nitrogenremoval during anaerobic digestion of wasted activated sludge undersupplementing Fe(III) compounds. Chem. Eng. J. 2018, 332, 711−716.(13) Picardal, F. Abiotic and Microbial Interactions duringAnaerobic Transformations of Fe(II) and NOX
−. Front. Microbiol.2012, 3 (3), 112.(14) Klueglein, N.; Kappler, A. Abiotic oxidation of Fe(II) byreactive nitrogen species in cultures of the nitrate-reducing Fe(II)oxidizer Acidovorax sp. BoFeN1 - questioning the existence ofenzymatic Fe(II) oxidation. Geobiology 2013, 11 (2), 180−190.(15) Oshiki, M.; Ishii, S.; Yoshida, K.; Fujii, N.; Ishiguro, M.; Satoh,H.; Okabe, S. Nitrate-dependent ferrous iron oxidation by anaerobicammonium oxidation (anammox) bacteria. Appl. Environ. Microbiol.2013, 79 (13), 4087−4093.(16) Zhou, G.; Yang, X.; Li, H.; Marshall, C. W.; Zheng, B.; Yan, Y.;Su, J.; Zhu, Y. Electron Shuttles Enhance Anaerobic AmmoniumOxidation Coupled to Iron(III) Reduction. Environ. Sci. Technol.2016, 50 (17), 9298−9307.(17) Zhang, J.; Loh, K.; Li, W.; Lim, J. W.; Dai, Y.; Tong, Y. W.Three-stage anaerobic digester for food waste. Appl. Energy 2017, 194,287−295.(18) Mejia, J.; Roden, E. E.; Ginder-Vogel, M. Influence of Oxygenand Nitrate on Fe (Hydr)oxide Mineral Transformation and SoilMicrobial Communities during Redox Cycling. Environ. Sci. Technol.2016, 50 (7), 3580−3588.(19) Brostoff, L. B.; Centeno, S. A.; Ropret, P.; Bythrow, P.; Pottier,F. Combined X-ray Diffraction and Raman Identification of SyntheticOrganic Pigments in Works of Art: From Powder Samples to Artists’Paints. Anal. Chem. 2009, 81 (15), 6096−6106.(20) Figueira, M. M.; Volesky, B.; Mathieu, H. J. InstrumentalAnalysis Study of Iron Species Biosorption by Sargassum Biomass.Environ. Sci. Technol. 1999, 33 (11), 1840−1846.(21) Wang, D.; Li, Y.; Wang, Q.; Wang, T. Nanostructured Fe2O3−graphene composite as a novel electrode material for supercapacitors.J. Solid State Electrochem. 2012, 16 (6), 2095−2102.(22) Yamashita, T.; Hayes, P. Analysis of XPS spectra of Fe2+ andFe3+ ions in oxide materials. Appl. Surf. Sci. 2008, 254 (8), 2441−2449.(23) Hinman, J. J., Jr. Standard Methods for the Examination of Water,Sewage, and Industrial Wastes, 10th ed.; 1955; Vol. 45, pp 821−821,DOI: 10.2105/AJPH.45.6.821-a.
ACS Sustainable Chemistry & Engineering Research Article
(24) Feng, Y.; Zhang, Y.; Quan, X.; Chen, S. Enhanced anaerobicdigestion of waste activated sludge digestion by the addition of zerovalent iron. Water Res. 2014, 52, 242−250.(25) Zhao, Z.; Li, Y.; Quan, X.; Zhang, Y. Towards engineeringapplication: Potential mechanism for enhancing anaerobic digestionof complex organic waste with different types of conductive materials.Water Res. 2017, 115, 266.(26) Lovely, D. R.; Phillips, E. Organic-Matter Mineralization withreduction of Ferric Iron in Anaerobic Sediments. Appl. Environ.Microb. 1986, 51 (4), 683−689.(27) Huang, S.; Jaffe,́ P. R. Characterization of incubationexperiments and development of an enrichment culture capable ofammonium oxidation under iron-reducing conditions. Biogeosciences2015, 12 (3), 769−779.(28) Huang, S.; Chen, C.; Peng, X.; Jaffe,́ P. R. Environmentalfactors affecting the presence of Acidimicrobiaceae and ammoniumremoval under iron-reducing conditions in soil environments. SoilBiol. Biochem. 2016, 98, 148−158.(29) Diao, Z.; Du, J.; Jiang, D.; Kong, L.; Huo, W.; Liu, C.; Wu, Q.;Xu, X. Insights into the simultaneous removal of Cr6+ and Pb2+ by anovel sewage sludge-derived biochar immobilized nanoscale zerovalent iron: Coexistence effect and mechanism. Sci. Total Environ.2018, 642, 505−515.(30) Gong, Y.; Werth, C. J.; He, Y.; Su, Y.; Zhang, Y.; Zhou, X.Intracellular versus extracellular accumulation of Hexavalent chro-mium reduction products by Geobacter sulfurreducens PCA. Environ.Pollut. 2018, 240, 485−492.(31) Luther, G. W.; Sundby, B. R.; Lewis, B. L.; Brendel, P. J.;Silverberg, N. Interactions of manganese with the nitrogen cycle:Alternative pathways to dinitrogen. Geochim. Cosmochim. Acta 1997,61 (19), 4043−4052.(32) Weng, T.; Liu, C.; Kao, Y.; Hsiao, S. S. Isotopic evidence ofnitrogen sources and nitrogen transformation in arsenic-contaminatedgroundwater. Sci. Total Environ. 2017, 578, 167−185.(33) Gilson, E. R.; Huang, S.; Jaffe,́ P. R. Biological reduction ofuranium coupled with oxidation of ammonium by Acidimicrobiaceaebacterium A6 under iron reducing conditions. Biodegradation 2015,26 (6), 475−482.(34) Bao, P.; Li, G. Sulfur-Driven Iron Reduction Coupled toAnaerobic Ammonium Oxidation. Environ. Sci. Technol. 2017, 51(12), 6691−6698.(35) Melton, E. D.; Swanner, E. D.; Behrens, S.; Schmidt, C.;Kappler, A. The interplay of microbially mediated and abioticreactions in the biogeochemical Fe cycle. Nat. Rev. Microbiol. 2014, 12(12), 797−808.(36) Kappler, A. Geomicrobiological Cycling of Iron. Rev. Mineral.Geochem. 2005, 59 (1), 85−108.(37) Boland, D. D.; Collins, R. N.; Miller, C. J.; Glover, C. J.; Waite,T. D. Effect of Solution and Solid-Phase Conditions on the Fe(II)-Accelerated Transformation of Ferrihydrite to Lepidocrocite andGoethite. Environ. Sci. Technol. 2014, 48 (10), 5477−5485.(38) Yang, Y.; Zhang, Y.; Li, Z.; Zhao, Z.; Quan, X.; Zhao, Z. Addinggranular activated carbon into anaerobic sludge digestion to promotemethane production and sludge decomposition. J. Cleaner Prod. 2017,149, 1101.(39) Han, M.; Li, Z.; Zhang, F. The Ammonia Oxidizing andDenitrifying Prokaryotes Associated with Sponges from Different SeaAreas. Microb. Ecol. 2013, 66 (2), 427−436.(40) Robbins, S. J.; Evans, P. N.; Parks, D. H.; Golding, S. D.;Tyson, G. W. Genome-Centric Analysis of Microbial PopulationsEnriched by Hydraulic Fracture Fluid Additives in a Coal BedMethane Production Well. Front. Microbiol. 2016, 7, 731DOI: 10.3389/fmicb.2016.00731.
ACS Sustainable Chemistry & Engineering Research Article
