Low-Temperature Anammox Supported by Zero-Valent Iron (ZVI): Microbial and Physicochemical Changes during Treatment of Synthetic and Municipal Wastewater

Abstract

The anaerobic ammonium oxidation (anammox) process offers a sustainable and energy-efficient alternative for nitrogen removal in wastewater treatment, but its performance at low temperatures remains a major challenge. This study investigated the role of zero-valent iron (ZVI) in enhancing anammox activity across a temperature range of 10–30 °C using both synthetic and municipal wastewater (MWW). Short-term batch tests demonstrated that low-dose ZVI (5–10 mg/L) stimulated specific anammox activity (SAA) particularly at 13–20 °C, while ZVI increasing concentration (1–10 mg/L) enhanced the enzymatic activity of HDH and decreased NIR activity, as well as modulated oxidative stress (ROS and GSH balance). In contrast, the long-term operation of the anammox process in sequencing batch reactors (SBR) showed that while ZVI (5 mg/L) improved SAA and microbial stability under synthetic conditions at 13 °C in compared to control (without ZVI), these benefits diminished once real municipal wastewater was introduced, most likely due to biomass stress and organic load. Metataxonomic analysis confirmed that ZVI selectively promoted genera such as Candidatus Brocadia, Denitratisoma, Micavibrionales_unclassified, while reducing overall microbial diversity. These results indicate that low-dose ZVI can temporarily enhance anammox resilience at suboptimal temperatures. However, its long-term application in MWW requires further optimization to mitigate potential inhibitory effects and iron passivation.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00248-025-02615-z.

Introduction

Given the high operating costs associated with conventional biological wastewater treatment, the development of alternative technologies with greater efficiency and reduced energy demand has become essential. Over the past three decades, anaerobic ammonium oxidation (anammox) has been established as a cost-effective and sustainable pathway, enabling the direct conversion of ammonium and nitrite into dinitrogen gas [10, 70]. However, the growth rate of anammox bacteria (AnAOB) is considerably slower than that of other nitrogen-cycle microorganisms, and their optimal growth temperature (30 ± 5 °C) is higher than the average temperature of municipal wastewater entering treatment plants (8 ± 20 °C) [3, 23]. AnAOB cells are characterized by three membrane systems: (i) an outer proteinaceous cell wall with a thin peptidoglycan layer, (ii) a cytoplasmic membrane, and (iii) an additional membrane enclosing a specialized organelle called the anammoxosome [15]. The anammoxosome is the site where all catabolic nitrogen transformations occur, and its internal membrane harbors the key enzymes involved in these reactions [46]. Trace metals play an important role in microbial metabolism as cofactors of enzymes, stimulating key biochemical pathways [31]. Among them, iron is particularly critical, functioning in numerous metabolic processes including redox reactions and electron transfer [15]. Within AnAOB, the anammoxosome contains most of the cellular iron, primarily in the form of heme cytochromes (heme c) and iron–sulfur proteins. This suggests that the anammoxosome not only serves as the metabolic hub of the cell but also as a storage site for iron and heme-containing enzymes, thereby supporting AnAOB activity and growth.

Zero-valent iron (ZVI) nanoparticles have been extensively studied for the remediation of wastewater, drinking water, and groundwater. They possess a large specific surface area, a core–shell structure, and magnetic properties, which together contribute to their strong reductive potential (E° =  − 0.44 V) [70]. In aquatic environments, ZVI can undergo biological or chemical transformation into Fe(II) and Fe(III) [47]. In the context of the anammox process, previous studies demonstrated that ZVI could enhance the anammox process by promoting the secretion of extracellular polymeric substances (EPS), thereby facilitating the granulation of AnAOB [49]. Yan et al. [68] further confirmed that ZVI addition enables the recovery of anammox activity after severe dissolved oxygen (DO) inhibition. Coexisting of ZVI with Fe(II) may increase nitrogen removal efficiency by anammox consortia [5], and an increase of the Fe(II) from 0.03 to 0.12 mM gives positive results in increasing growth rate acceleration for anammox biomass [40]. Both Fe(II) and Fe(III) stimulate the process as essential cofactors of heme-containing enzymes, including cytochromes, hydrazine synthase, and hydrazine dehydrogenase [10]. Iron, therefore, plays a vital role in electron transfer and enhances AnAOB activity [46]. Although a recent study confirmed that ZVI can support anammox at optimal for anammox bacteria temperatures [36], the overall effect of ZVI on the process under such conditions remains insufficiently understood. Liu et al. [36] reported that the addition of 5 g/L ZVI enhanced the activity of nitrogen-metabolizing enzymes and improved Fe(II) utilization in AnAOB, leading to a 77.5% nitrogen removal efficiency at 17 ± 1 °C. On the other hand, [36] showed that operating partial nitritation–anammox (PN/A) below 15 °C may cause nitrate accumulation, and therefore proposed an integrated partial nitritation–anammox and iron-based denitrification (PNAID) process. Using PNAID, a nitrogen removal efficiency of 82.2% was achieved at 12 °C with ZVI dosed at 0.08 g/g TS. Similarly, Chen et al. [9] combined ZVI with biochar (nZVI@BC) at a Fe:C mass ratio of 1:30. The addition of 30 g/L nZVI@BC maintained anammox performance at about 80% efficiency at 15 °C. These findings indicate that ZVI can enhance anammox under low-temperature conditions, which may be crucial for its future application in mainstream wastewater treatment.

Nevertheless, key gaps remain. Most existing studies employed ZVI at concentrations in the gram-per-liter range, which Semerád and Cajthaml [54] suggested may pose ecotoxicological risks. Furthermore, these studies were largely conducted with synthetic wastewaters optimized for anammox growth, without fully addressing the complexity and inhibitory potential of real municipal wastewater (MWW). Therefore, the aim of this study is to determine whether ZVI at concentrations below 10 mg/L can support anammox bacteria at low temperatures. The research specifically evaluates the impact of ZVI on anammox biomass by assessing specific anammox activity (SAA), oxidative stress, enzymatic activity, and microbial community structure. Both short-term effects (batch tests) and long-term effects (treatment of synthetic and municipal wastewater) were investigated.

Materials and Methods

Anammox Biomass

Suspended anammox biomass originated from the laboratory sequencing batch reactor (SBR) described in the previous studies [4, 77]. Before the current work, SBR was operated at stable conditions at a temperature of 31 ± 2 °C, pH 7.6 ± 0.2 and a total nitrogen loading rate of 0.8 ± 0.05 g N/L·d. It was fed with a mineral medium (van [62]) containing: 1.2 g NH₄Cl/L, 2.1 g NaNO₂/L, 0.048 g KHCO₃/L, 0.041 g KH₂PO₄/L, 0.228 g MgSO₄· 7 H₂O/L, 0.007 g FeSO₄· 7 H₂O/L, 0.004 g EDTA/L.

Zero Valent Iron (ZVI)

Commercially available ZVI nanopowder (US Research Nanomaterials, Inc., USA) used in all assays was characterized by particles size 65–75 nm, specific surface area: 8 m²/g, elemental analysis (percentage by weight): Fe > 99.5%, Mn < 0.13%, Ni < 0.12%, Al < 0.091%, Ca < 0.044%, Mo < 0.024%, Si < 0.02%, Cr < 0.015%, Sn < 0.008%, Cu < 0.005%, Ti < 0.005%, Mg < 0.004%, Pb < 0.004%.

Short-Term effects

The central composite design (CCD) followed by the response surface methodology (RSM) was used to obtain a mathematical relationship for the independent and simultaneous influence of the temperature and the ZVI concentration on the specific anammox activity (SAA), which was then approximated by a polynomial quadratic equation.

Set-up for a 2 factors analysis (k = 2) consists of 9 experimental points. The center point and k² factorial points allow to estimate the first-order regression coefficients, while 2·k axial points are required to calculate the second-order model. The distance from the center point to the factorial and axial points is the same and is α = √k, which results in a round, rotatable design. Finally, the full CCD experimental set-up, including 4 center point repetitions, are presented in Table 1. Based on the literature review [40, 48], the ZVI concentration range was set between 4 and 11 mg/L. The temperature range was set between 10 and 30 °C.

Table 1.

Central composite design experimental plan. T—temperature, ZVI – zero valent iron concentration, α ≈ 1.41

Experiment number	Coded values (T, ZVI)	Natural values
		T (°C)	ZVI (mg/L)
1	0, 0	20	7.5
2	−α, 0	10	7.5
3	−1, −1	13	5
4	0, 0	20	7.5
5	+ 1, + 1	27	10
6	0, 0	20	7.5
7	0, + α	20	11
8	0, 0	20	7.5
9	0, −α	20	4
10	+ 1, −1	27	5
11	+ α, 0	30	7.5
12	−1,  + 1	13	10

Specific anammox activity was determined in batch tests according to the methodology described in the previous study by Tomaszewski et al. [59]. The batch tests were performed in batch anaerobic reactors with a working volume of 125 mL. The test medium contains a phosphate buffer (final concentration: 0.14 g KH₂PO₄/L and 0.75 g K₂HPO₄/L). pH was adjusted at 7.5 using 10% NaOH or 10% HCl and a pH meter (WTW pH 330i). Desired ZVI doses were dispersed in the medium and reactors were incubated for 20 h in a thermostatic cabinet to acclimatize, reach the desired temperature, and remove residual ammonium and nitrite nitrogen. Test reactors were purged with dinitrogen gas before the assay to remove any dissolved oxygen. Despite the temperature and ZVI concentration batch tests were performed under the following conditions: an average biomass concentration of 1.0 ± 0.2 g VSS/L, mixing at 250 rpm, 95.5 mg/L of NH₄Cl and 147.9 mg/L of NaNO₂ (resulted in 25 mg NH₄-N/L and 30 mg NO₂-N/L). Samples from the reactors were periodically collected to determine the concentrations of N-NH₄ and N-NO₂, while the stoichiometric ratio was used to calculate NO₃-N [56]. The duration of the test was adapted to the process activity, depending on temperature, in the range from 2 to 8 h. Measurements of ammonium and nitrite nitrogen were determined by the photometric tests (MERCK Millipore) using a photometer (MERCK Spectroquant® NOVA60). All tests were carried out in triplicate. Moreover, an abiotic test without biomass addition indicated that ZVI has no influence on the nitrogen concentration.

Specific anammox activity (SAA) (g N/g VSS·d) was calculated based on the nitrogen decrease in the linear range of the removal of the substrate. The ZVI influence was presented as a percentage of SAA in relation to the control test without the nanomaterial addition. For the statistical analysis of the CCD experiment, STATISTICA StatSoft® software was used. Coefficients of the second-order polynomial equation were calculated based on the regression analysis. Finally, the analysis of variance (ANOVA) was used to test the mathematical model.

Oxidative Stress

The reactive oxygen species (ROS) production was evaluated using 2’,7’-dichlorofluorescin diacetate (DCFDA), which is a kind of ROS-detecting agent [7]. DCFDA is deacetylated to 2’,7’-dichlorofluorescin (DCFH) by cellular esterases and then oxidized by ROS into 2’,7’- dichlorofluorescein (DCF), which is highly fluorescent. The level of intracellular ROS is indicated by the increasing fluorescence intensity of DCF. In detail, bacteria suspension was washed with 1 × phosphate-buffered saline (PBS) (0.01 M, pH 7.4) and resuspended in the same buffer. Then, 1.25 mL of the bacteria suspension was mixed with the ZVI dispersions (0.25 mL) to achieve five final nanomaterial concentrations: 0, 1, 2.5, 5, 7.5 and 10 mg/L. Distilled water was used to prepare the negative control. All samples were prepared in triplicate and incubated for 2 h at 20 °C. After this time, samples were washed three times with 1 × PBS and resuspended in 1.5 mL of 1 × PBS. Samples were mixed with a drop of DCFDA (final concentration of 10 μM) and incubated in the dark for the next 1 h. Then, the bacteria were washed with 1 × PBS to remove residual DCFDA. The emission generated by the DCF was measured at 520 nm (excitation at 485 nm) on a microplate reader Varioskan LUX (Thermo Scientific). The level of intracellular ROS was calculated as a percentage of the negative control.

The thiol group (-SH) of glutathione (GSH) is a very important antioxidant. They bind to the un-shared electrons of the free radicals, which reduces their number and the level of oxidative stress [7]. The bacteria suspension was washed with 1 × PBS (0.01 M pH 7.4) and resuspended in the same buffer. Then, 1.25 mL of the bacteria suspension was mixed with the ZVI dispersions (0.25 mL) to achieve five final nanomaterial concentrations: 0 (control), 1, 2.5, 5, 7.5, and 10 mg/L. All samples were prepared in triplicate and incubated for 2 h at 20 °C. After this time, samples were washed three times with distilled water, centrifuged and the supernatant was removed. Then, the bacteria were lysed by a mechanical method using 300 mg of bead-beating glass balls (Carl ROTH) and 1 mL of extraction buffer (100 mM TRIS–HCl,100 mM EDTA; 1.5 M NaCl; pH = 8.0). The samples were vortexed and incubated horizontally with shaking (1400 rpm) for 20 min. Next, 200 μL of 10% SDS (Sigma-Aldrich) was added and the samples were incubated vertically with shaking (1400 rpm) for 30 min at 65 °C. After incubation, the samples were centrifuged (15,000 rpm, 4 °C, 15 min) and the supernatant was used for the assay. The loss of thiol groups was quantified using a Fluorometric Thiol Quantification Kit MAK151 (Sigma-Aldrich), according to manufacturer’s protocol on the 96-well plate. The fluorescence intensity was measured at 535 nm (excitation at 490 nm) in relation to GSH standard (included in the kit) on the microplate reader Varioskan LUX (Thermo Scientific). The loss of thiol groups was calculated as a percentage of negative control.

Enzymes Activity

For enzymes activity experiments cell extract was prepared according to the modified method of Schüler et al. [50]. Five grams of the bacteria suspension (wet weight) was washed with 1 × PBS (0.02 M, pH 7.4) three times and resuspended in 5 mL of the same buffer. Cells were lysed by means of mechanical dispersion using an Ultra-Turrax Tube Drive homogenizer (IKA Works GmbH & Co. KG) at 6000 rpm during three cycles of 60 s, followed by freezing (−80 °C) and thawing (room temperature). Cell mass was centrifuged (15,000 rpm, 4 °C, 15 min) and the supernatant was used as a cell extract for the next steps. Protein concentration was measured using Pierce™ BCA Protein Assay Kit (Thermo Scientific) according to standard protocol.

The activity of hydrazine dehydrogenase (HDH) was determined according to the method described by Shimamura et al. [52]. The 1 mL of reaction mixture consisted of cell extract (final protein concentration 982 ± 25 μg/mL), 100 mM PBS (pH 8.0), 50 μM horse heart cytochrome c (oxidized form), 25 μM hydrazine, and five nanomaterial concentrations: 0, 1, 2.5, 5, 7.5, and 10 mg/L. All samples were prepared in triplicate and 1 × PBS was used to negative control. Tests were performed in 1 mL cuvettes purged with dinitrogen gas to remove any dissolved oxygen and sealed with stoppers to keep under anoxic conditions. The increase in the absorbance of the cytochrome c at 550 nm in time of 2 h at 20 °C, against the blank sample without cell extract was measured using UV–Vis spectrophotometer (UV5600, Shanghai Metash Instruments Co., Ltd.). The reaction rate was expressed as a percentage of negative control.

Nitrite reductase (NIR) activity was assayed on the basis of the method used by Hira et al. [25]. The 1 mL of reaction mixture consisted of cell extract (final protein concentration 981 ± 13 μg/mL), 20 mM PBS (pH 7.4), 0.2 mM benzyl viologen, 0.24 mM sodium dithionite, 500 μM sodium nitrite and five nanomaterial concentrations: 0, 1, 2.5, 5, 7.5, and 10 mg/L. All samples were prepared in triplicate and 1 × PBS was used to negative control. Test were performed in 1 mL cuvettes purged with dinitrogen gas to remove any dissolved oxygen and sealed with stoppers to keep under anoxic conditions. The oxidation of reduced benzyl viologen results in the decrease of the absorbance at 550 nm. It was monitored in time of 2 h at 20 °C, against the blank sample without cell extract using UV–Vis spectrophotometer (UV5600, Shanghai Metash Instruments Co., Ltd.). The oxidation rate was expressed as a percentage of the negative control.

Long-Term Effects and Real Wastewater Treatment

Two sequencing batch reactors (SBRs) with a working volume of 5 L had been inoculated with the same anammox-activated sludge and then were operated for 7 weeks prior to this experiment. The first reactor (C-reactor) was used as a control. The doses of ZVI (5 mg/L) were selected based short-term test and added to the second reactor (Z-reactor) at days 0, 77, 119, 143, and 203 of the experiment, which lasted a total of 230 days. ZVI was added to the reactors when the temperature decreased to 15 °C, 14 °C, and 13 °C, as well as during an increase in MWW concentration. These ZVI concentrations were chosen based on the short-term experiments. During the study, SBRs were operated under the following conditions: hydraulic retention time (HRT) = 1 d, pH = 7.6 ± 0.3, dissolved oxygen (DO) < 0.1 mg/L. Temperature in both reactors was gradually decreased from 30 to 10 °C.

In days from 0 to 189 and from 214 to 230 they were fed with a mineral medium based on van de Graaf et al. [62], characterized by the total nitrogen concentration, designed to correspond with the low-loaded wastewater streams (115–355 mg N/L). It was regulated by the addition of NH₄Cl and NaNO₂. Other ingredients were dosed in the following amounts: 0.048 g/L KHCO₃, 0.041 g/L KH₂PO₄, 0.228 g/L MgSO₄· 7 H₂O, 0.007 g/L FeSO₄· 7 H₂O, 0.004 g/L EDTA. The overall course of the experiment is shown in the Fig. 1. In days from 190 to 213 reactors were fed with real wastewater after mechanical treatment from the municipal WWTP. Municipal wastewater was firstly diluted in a ratio of 1:1 with tap water (days 190–203), 3:1 (days 204–211) and without dilution (days 212–213). Prepared influent was characterized by the total nitrogen (TN) concentration in range 38–138 mg N/L.

Fig. 1.

Overall course of the experiment divided into six phases. Arrows indicate days of ZVI addition and sampling for MGS (metaxonomic sequencing) while red bracket shows period of real MWW (municipal wastewater) treatment

Measurement and Analysis

Ammonium (NH₄⁺-N), nitrite (NO₂⁻-N), and nitrate (NO₃⁻-N) nitrogen concentrations were measured using fast photometric tests (MERCK Millipore) with a photometer (MERCK Spectroquant® NOVA60). Temperature and pH were measured by JUMO tecLine HD & pH combination electrode. DO concentration was monitored using an ELMETRON Conductivity/Oxygen Meter CCO-505 with an ELMETRON COG-1 oxygen sensor. The VSS concentration was determined according to the APHA standard [1].

Nitrogen loading rate (NLR) was calculated as Eq. 1:

$$NLR\left(k,g,N,\bullet ,m^{-3},{\bullet d}^{-1}\right)=\frac{{\mathit{TN}}_{\mathit{influent}}}{\mathit{HRT}}$$ 1

Nitrogen removal rate (NRR) was calculated as Eq. 2:

$$NRR\left(k,g,N,\bullet ,m^{-3},{\bullet d}^{-1}\right)=\frac{({\mathit{TN}}_{\mathit{influent}}-{\mathit{TN}}_{\mathit{effluent}})}{\mathit{HRT}}$$ 2

Specific anammox activity (SAA) was calculated as Eq. 3:

$$SAA\left(k,g,N,\bullet ,{\mathit{kgVSS}}^{-1},{\bullet d}^{-1}\right)=\frac{\mathit{NRR}}{\mathit{VSS}}$$ 3

16S rRNA Coding Gene Metataxonomicsequencing

The total genomic DNA was isolated using GeneMATRIX Soil DNA Purification Kits (EurX®, Poland) according to the manufacturer’s instruction, then qualitative evaluation of the isolated DNA was performed with Qubit® Fluorometer (Invitrogen, USA) and was stored at − 45 °C. The high-throughput sequencing based o 16S rRNA coding gene was made according to the methodology described by Gamoń et al. [19]. Based on the obtained results the changes in community diversity were evaluated with Shannon and Simpson and Chao1 indices.

Results and Discussion

Short-Term Effects

The combined impact of temperature and ZVI on the anammox process was tested over a temperature range of 10 to 30 °C, which is below the optimal range for the anammox process [61]. The effect of ZVI on anammox biomass consortia was assessed by measuring the percentage of SAA in relation to a control without ZVI addition. The influence of temperature and ZVI on activity was mathematically modeled using a quadratic polynomial formula (R² = 0.90), as shown in Eq. 4:

$$SAA(\%)=177.00-4.06T-12.67ZVI+0.06T2+0.48ZVI2+0.33TZVI$$ 4

This model was validated using ANOVA (Table 2), including a lack-of-fit test, which indicated no significant lack of fit (p < 0.05), suggesting the model is appropriate for describing the relationship between the studied parameters [55]. The statistical analysis confirmed that ZVI significantly affects bacterial activity, as evidenced by a high Fisher (F) value and p < 0.05 (Table 2). The empirical research was performed as batch experiments according to regression model and the results are shown in Fig. 2. Those demonstrated an enhancement of the anammox process activity by 30.5% (27 °C) and 17.7% (27 °C) caused by addition of 10 mg ZVI/L and 5 mg ZVI/L, respectively. Additionally, the increase in SAA value in compared to control was also observed in at temperature 20 °C and ZVI by 6.7% (7.5 mg ZVI/L), 0.2% (7.5 mg ZVI/L) as well as lower temperature of 13 °C by 7.7% (5 mg ZVI/L). Therefore, the maximum activity reached 30.5% at 27 °C (10 mg ZVI/L), which reflected to the regression model (22%) stimulation obtained for CCD experiment (Fig. 2). Li and Tabussum [32] examined the impact of four ZVI concentrations (50, 100, 500, and 1000 mg/L) across a temperature range of 30 °C to 15 °C on the removal efficiency of NH₄⁺ and NO₂⁻ in the anammox process during 56-h batch experiments. Their results showed that, under gradually decreasing temperatures, a ZVI concentration of 100 mg/L achieved NH₄⁺ and NO₂⁻ removal efficiencies of 93.93% and 96.69%, respectively. Similarly, Gao et al. [21] confirm that the addition of ZVI (1.5 g/L) increases anammox activity by 6.1% within 10 days (35 °C). However, both these studies were carried out with a ZVI concentration significantly higher than that used in the present study. On the other hand, although Yan et al. [68] investigated ZVI concentrations comparable to those used in the present study and they reported a decline in anammox activity during 300-min batch tests relative to the control (without ZVI). At concentration of 5 mg ZVI/L, anammox activity decreased by 18%, and the inhibitory effect intensified with increasing ZVI concentration—resulting in reductions of 32% at 25 mg/L and 61% at 75 mg/L. Several studies have highlighted that the effectiveness of ZVI is influenced by a range of factors, including the applied dosage and the biomass concentration [35, 66, 68]. Moreover, the interaction between ZVI and nitrite (NO₂⁻) can also impact the performance of the anammox process. Notably, NO₂⁻ may be abiotically reduced to either ammonium (NH₄⁺) or nitrogen gas (N₂) in the presence of ZVI [68], as represented by Eqs. 5 and 6:

Table 2.

ANOVA for quadratic model (F – Fisher’s F value; p – probability value). * indicates the statistically significant values (p < 0.05)

Source	F	p
ZVI linear	39.03284	0.022*
ZVI square	35.74394	0.043*
Temperature linear	3.038843	0.123
Temperature square	0.9473829	0.532
Interaction	4.949938	0.104
Lack of fit	1.294832	0.343

Fig. 2.

Influence on the specific anammox activity (% in relation to the control without the ZVI addition) obtained in the batch experiments and calculated based on the regression model, bars represent standard error

$$3F{e}^0+N{O}_2^{-}+8H^{+}\to N{H}_4^{+}+2H_{2}O+3Fe(II)$$ 5

$$3F{e}^0+2N{O}_2^{-}+8H^{+}\to N_2+4H_{2}O+3Fe(II)$$ 6

The results clearly indicate that temperature is a key factor influencing AnAOB activity; higher activity is generally observed at temperatures closer to the process optimum. Nevertheless, there is substantial evidence demonstrating that the anammox process can still function effectively at moderately low temperatures. For instance, SAA values of 0.466 g N/g VSS·d at 16 °C [44] and 0.138 g N/g VSS·d at 15 °C [2] have been reported. However, as shown by Ma et al. [45], process efficiency declines significantly at temperatures below 15 °C. Despite this, several studies have demonstrated the feasibility of anammox-based treatment under psychrophilic conditions, albeit with considerably reduced activity. Reported SAA values include 0.022 g N/g VSS·d at 13 °C (Sánchez Guillén et al., 2016), 0.092 g N/g VSS·d at 12.5 °C [29]. In the present study, comparable SAA values ranging from 0.035 to 0.025 g N/g VSS·d were obtained at 13 °C (described in ” section), further confirming the temperature sensitivity of the anammox process, particularly under low-load wastewater operational conditions.

The findings demonstrate that the combined application of low-dose ZVI and suboptimal temperatures can positively influence anammox activity, with the extent of stimulation being dependent on both parameters. While elevated temperatures remain, the dominant factor enhancing bacterial performance, the addition of small amounts of ZVI particularly under low-temperature conditions can provide a measurable benefit. However, due to the complex interactions between ZVI, nitrite, and microbial consortia, further research is needed to optimize operational parameters and better understand the underlying mechanisms.

Oxidative Stress

The ZVI was previously described as an effective bio-augmenter for improving the anammox process activity due to its reducibility, ecological sustainability, and stable release of Fe(II) [11, 24]. On the other hand, it is also suspected that ZVI can inhibit AnAOB through oxidative damage [68], but so far these phenomena have not been thoroughly studied. Semerád et al. [54] reported that the toxicity of ZVI is mainly associated with the release of Fe(II) and Fe(III) ions during the oxidation of ZVI, according to the Eqs. 7–12. Inside the cell, Fe(II) reacts with intracellularly produced H₂O₂ to generate ROS via Fenton chemistry (Seveu et al., 2011). Chen et al. [8] compared the toxicity of ZVI with that of Fe(II) and Fe(III) across various bacterial species from both Gram-positive and Gram-negative groups. The results indicated that it is the presence of Fe(II) and Fe(III) that exhibits greater toxicity than ZVI itself.

$${\mathit{Fe}}^0+2H2O\to {\mathit{Fe}}^{2+}+H_2+2{\mathit{OH}}^{-}$$ 7

$${2Fe}^0+2H2O+O_2\to {2Fe}^{2+}+4{\mathit{OH}}^{-}$$ 8

$${4Fe}^{2+}+H^{+}+O_2\to {4Fe}^{3+}+H2O$$ 9

$${\mathit{Fe}}^0+O_2+{2H}^{+}\to {\mathit{Fe}}^{2+}+H_2{O}_2$$ 10

$${\mathit{Fe}}^0+H_2{O}_2\to {\mathit{Fe}}^{2+}+2{\mathit{OH}}^{-}$$ 11

$${\mathit{Fe}}^{2+}+H_2{O}_2\to {\mathit{Fe}}^{3+}+{OH\bullet +OH}^{-}$$ 12

In 2012, a novel mode of programmed cell death called ferroptosis was found [13]. The mechanism of ferroptosis is related to the accumulation of lipid peroxidation products and reactive oxygen species (ROS) delivered from Fe-related metabolism [11]. The generation of ROS leads to the destruction of intracellular membranes and cellular structures resulting in cell death [6]. Therefore, the generation of ROS in the AnAOB might be involving ferroptosis. Moreover, GSH as a very important antioxidant binds to the un-shared electrons of the free radicals resulting in a reduced level of oxidative stress [7]. Therefore, increases and decreases in GSH concentration should be associated with the ROS content.

To investigate whether ZVI induces ferroptosis in anammox cells and affects the oxidation and antioxidation process, the ROS generation presented as DFC fluorescence (Fig. 3a) and GSH concentration (Fig. 3b) were studied. It can be seen that both DFC content and relative GSH concentration have the lowest value at 5 mg/L of ZVI (103.6% and −13%, respectively), while the highest levels at 7.5 mg/L of ZVI (115% and 21%, respectively) while at 10 mg/L slightly decreased to 112.5% and 1%, respectively. Negative values of GSH mean that its concentration in the test with ZVI addition was lower than that of the control (without ZVI addition). The results presented by Jiang et al. [28] showed that the concentration of Fe(II) at 5 mg/L does not affect oxidative stress and only concentrations above 25 mg/L can cause cellular damage caused by the formation of reactive oxygen species. As previously claimed generation of ROS by ZVI is related to the presence of oxygen despite its release of Fe(II) [69]. In AnAOB cells, the oxygen cannot be internally generated through the metabolic pathway, thus the oxidative damage are limited. This thesis can be supported by the results presented by [73, 75], where even a concentration of 200 mg/L did not stimulate ROS generation after 24 h. Although the lowest ROS level was obtained at a ZVI concentration of 5 mg/L, the results indicate that at both lower and higher ZVI concentrations ROS was generated intensively. This situation may result from intracellular regulation of iron concentration. Chen et al. [8] demonstrated that at low concentrations of Fe(II) and Fe(III), which correspond to ZVI concentration, bacteria can regulate the presence of iron inside the cell, which may influence ROS generation to varying degrees.

Fig. 3.

a Reactive oxygen species (ROS) generation expressed by DFC fluorescence and b thiol oxidation (GSH) after bacteria incubation with zero-valent iron (ZVI) for 2 h at 20 °C. The sample without nanomaterial was used as the control, bars represent standard error. Asterisks indicate significance (****p < 0.0001, ***p < 0.001, **p < 0.01, *p < 0.05) according to Student’s t-test statistical analysis

Brief, ZVI can enhance anammox activity but may also trigger oxidative stress and ferroptosis due to ROS generation. This study found that ROS and GSH levels in anammox cells vary with ZVI concentration, peaking at 7.5 mg/L. However, due to limited oxygen in anammox metabolism, oxidative damage remains minimal, with significant effects only at higher Fe(II) levels, which highly depend on ZVI concertation. Moreover, this is confirmed by statistically significant equations between DFC and GSH levels in the ZVI concentrations tested (Fig. 3).

Enzymes Activity

The conversion of nitrogen in the anammox process is related to the bacteria energy metabolic process associated with a series of enzymatic reactions [28]. As demonstrated previously, ZVI can stimulate the activity of genes involved in the nitrogen metabolism of the anammox process by releasing Fe(II), which is required for the synthesis of heme c—an important co-enzyme that participates in the synthesis of nitrogen-functional genes such as (NIR, HZS, HAO, and HDH). In this study, the activity of HDH and NIR under different concentrations of ZVI was evaluated. The results are presented in Fig. 4 (a, relative HDH activity; b, relative NIR activity). In the anammox metabolic pathway, NIR is responsible for ${\text{NO}}_2^{-}$ conversion into NO, while HDH participates in the final step of the nitrogen cycle pathway in the anammox process where N₂O₂ is converted into N₂ (Straus et al., 2006). As obtained in this study, the HDH activity gradually increased from 101% at 1 mg ZVI/L to 121% at 10 mg ZVI/L in comparison to the control. The relationship between ZVI concentration and HDH activity was indicated by regression analysis (Fig. 4c), where the R² fit was 0.9431. Liu et al. [35–37] demonstrated that the presence of ZVI can lead to increased HDH gene activity during the interaction with the anammox process involving Cu(II). This results from the enhanced regulation of genes related to ferric iron reductase (fhuF) and enzymes involved in iron uptake (feoA, feoB, fur, etc.), leading to increased activity of iron-dependent genes [5, 67]. On the other hand, the activity of NIR was increased from 88% at 1 mg ZVI/L to 114% at 2.5 mg ZVI/L in relation to control, then gradually decreased to 71% at 10 mg ZVI/L. [28] suggested that a high concentration of Fe(II) might inhibit metabolic activity by restricting the processes involved in NIR activity, which can explain the decline in relative NIR activity under ZVI at a concentration > 2.5 mg/L. In the case of NIR (Fig. 4d), regression analysis showed a low degree of fit (R² = 0.1068). However, the negative slope of the curve (a = –1.929) supports the findings of Jiang et al. [28], indicating that increasing ZVI concentrations may inhibit NIR activity.

Fig. 4.

Effects of different ZVI concentrations on a hydrazine dehydrogenase (HDH) and b nitrite reductase (NIR) activity, bars represent standard error. c Linear regression between HDH activity results and d NIR activity results

Long-Term Effects and Real Wastewater Treatment

The nitrogen removal performance of the sequencing batch reactors (SBRs) was assessed under a temperature range of 32–13 °C in six operational phases (Fig. 5). Both, C-reactor and Z-reactor were operated, and the nitrogen loading rate (NLR), nitrogen removal rate (NRR), and specific anammox activity (SAA) were assessed as the key performance indicators (Table 3).

Fig. 5.

Performance of the sequencing batch reactors under various temperatures during the experiment. NLR, nitrogen loading rate; Z-NRR, nitrogen removal rate of ZVI reactor; C-NRR, nitrogen removal rate of control reactor; I-VI, phases of the experiment

Table 3.

Performance of the sequencing batch reactors under various temperatures. C-reactor control reactor, Z-reactor ZVI reactor, NLR nitrogen loading rate, NRR nitrogen removal rate, SAA specific anammox activity

Phase number		I	II	III	IV	V	VI
Experiment days		0–77	78–119	120–151	152–189	190–213	214–230
Temperature (°C)		32–17.5	15	14–13	13	13	13
NLR [kg N/m3·d]		0.223 ± 0.016	0.224 ± 0.026	0.239 ± 0.008	0.250 ± 0.015	0.090 ± 0.044	0.172 ± 0.070
C-reactor	NRR [kg N/m3·d]	0.151 ± 0.024	0.153 ± 0.052	0.147 ± 0.027	0.119* ± 0.020	0.048 ± 0.036	0.087 ± 0.050
	SAA [kg N/kg VSS·d]	0.109 ± 0.017	0.111 ± 0.038	0.106 ± 0.020	0.086* ± 0.015	0.035 ± 0.026	0.063 ± 0.036
Z-reactor	NRR [kg N/m3·d]	0.143 ± 0.026	0.141 ± 0.033	0.124 ± 0.029	0.142* ± 0.020	0.045 ± 0.038	0.048 ± 0.015
	SAA [kg N/kg VSS·d]	0.136 ± 0.025	0.134 ± 0.031	0.118 ± 0.027	0.135* ± 0.019	0.043 ± 0.036	0.046 ± 0.015

Throughout the experiment, a clear trend of decreasing reactor performance with lowering temperatures was observed. During Phase I (32–15 °C), both the C- and Z-reactor maintained stable operation, with NRR values of 0.151 ± 0.024 kg N/m³·d for the C-reactor and 0.143 ± 0.026 kg N/m³·d for the Z-reactor. Likewise, Lv et al. [43] ran the anammox process with decreasing temperatures from 25 to 15 °C and achieved a stable NRR of 0.3 kg N/m³·d throughout, indicating that anammox sludge can adapt to low temperatures. On the other hand, presented results show negligible effect of ZVI on the anammox performance. This contrasts with Erdim et al. [15], who reported improved nitrogen removal in the presence of ZVI at 25 °C. Liu et al. [38] reported similar findings, demonstrating that the application of ZVI at a concentration of 27 mg/L-wastewater (0.08 g/gTS) enhanced nitrogen removal in the PN/A system (28 °C) compared to the system without ZVI supplementation. The difference likely stems from sludge history, as Erdim et al. [15] biomass was already acclimated to that temperature, whereas in this study no such adaptation occurred.

In Phase II (15 °C), reactor performance remained relatively stable. NRR values remained similar (C-reactor: 0.153 ± 0.052; Z-reactor: 0.141 ± 0.033 kg N/m³·d), while SAA in the Z-reactor remained consistently higher (0.134 ± 0.031) than in the C-reactor (0.111 ± 0.038), suggesting a possible stimulatory role of ZVI even at a low concentration (5 mg/L). Liu et al., [35] confirmed a 77.5% increase in anammox NRE at 17 °C with ZVI at a concentration of 5 g/L. This concentration is much higher than that used in the present study (5 mg/L), suggesting that the process efficiency may depend on ZVI concentration, which likely needs to be much higher than what was used here. Nonetheless, the presence of ZVI does influence Fe(II) and Fe(III) levels, which have also been shown to affect the anammox process even at much lower concentrations. Li et al. [33] reported increased activity of marine AnAOB with Fe(III) bioaugmentation at concentrations between 4 and 250 mg/L at 15 °C.

A clearer distinction between the reactors appeared at lower temperatures (Phases III–IV, 14–13 °C). While both systems experienced slight declines, under stable 13 °C conditions the Z-reactor significantly outperformed the control, showing higher NRR and SAA (p < 0.05). A similar observation was reported by Liu et al. [36], who conducted the anammox process at 15 °C and 12 °C and achieved stable nitrogen removal of 80% at both temperatures using ZVI at 0.08 g/gTS. This finding suggests that ZVI supports anammox metabolism at low temperatures, most likely by improving electron transfer and redox stability. ZVI contributes to the stabilization of redox conditions within the reactor. Its standard reduction potential (E⁰ = –0.44 V) facilitates the promoting a strongly anaerobic environment favorable for anammox metabolism [18, 35]. Previous studies demonstrated that ZVI facilitates anaerobic conditions and maintains favorable levels of reactive oxygen species (ROS), which can promote resilience on lower temperatures [66, 68]. Moreover, the results confirm the notion that 15 °C represents a metabolic threshold for anammox bacteria [27, 41, 61], with distinct changes in process kinetics occurring below this temperature.

When municipal wastewater (MWW) was introduced in Phase V (13 °C), reactor performance deteriorated sharply. During this period, both reactors exhibited marked reductions in NRR and SAA, with the Z-reactor showing similar NRR performance (0.045 ± 0.038 kg N/m³·d) compared to the C-reactor (0.048 ± 0.036 kg N/m³·d), although the Z-reactor maintained a higher SAA. This suggests that prolonged exposure to low temperatures and treatment of MWW may diminish the initial benefits of ZVI, potentially due to biomass stress or the gradual passivation of the iron particles. Previous studies have shown that chloride ions (Cl⁻) can disrupt the formation of a passivation layer on the surface of ZVI by initiating corrosion and damaging the existing passive oxide film [53]. As presented in Table S1 (see Supplementary Materials) the concentration of Cl⁻ in raw MWW was 312 mg/L, which can contribute to the passivation of ZVI. The presence of MWW influences the restructuring of the bacterial community (described in the ” section), promoting the development of denitrifying bacteria. This is also observed through a decrease in the N–NO₃⁻ produced/N–NH₄⁺ removed ratio (Figure S1, see Supplementary Materials) during Phase V, falling below the theoretical value of 0.26 reported by Strous et al. [56, 57]. This observation is consistent with the findings of Zhou et al. [76], which suggested that in a partial denitrification–anammox system, the presence of ZVI may preferentially support denitrifiers, as ZVI (Fe⁰) can biologically react with NO₃⁻–N according to Eqs. 13–15:

$$2{\mathit{NO}}_3^{-}+2{\mathit{Fe}}^0+8H^{+}\to 2{\mathit{Fe}}^{3+}+N_2+{4H}_2{O}^{-}$$ 13

$$2{\mathit{NO}}_3^{-}+4{\mathit{Fe}}^0+4H^{+}\to 4{\mathit{Fe}}^{2+}+{\mathit{NH}}_4^{+}+{3H}_{2}O$$ 14

$$2{\mathit{NO}}_3^{-}+5{\mathit{Fe}}^0+12H^{+}\to 5{\mathit{Fe}}^{2+}+N_2+{6H}_{2}O$$ 15

In the final Phase VI, switching back to synthetic wastewater led to partial recovery of anammox activity. Nevertheless, overall performance remained below that of the initial phases, and the C-reactor recovered slightly better than the Z-reactor. This suggests that the prolonged exposure to ZVI and MWW may have caused lasting inhibition. One possible mechanism involves overproduction of ROS such as superoxide (O₂⁻), which in moderate amounts may promote stress tolerance but, at elevated levels, can suppress AnAOB activity [68]. Initially, O₂⁻ may assist in restoring cellular activity by enhancing the organism’s protective capacity—for example, by inducing the expression of protective genes [42, 68]. However, excessive production of O₂⁻ can ultimately suppress AnAOB activity [68]. The incomplete recovery observed here therefore indicates that while ZVI has short-term benefits under stable cold conditions, its long-term application in MWW may lead to biomass inhibition rather than stimulation. Taken together, these findings demonstrate that ZVI (15 mg/L) can enhance anammox activity under stable low temperatures (15–13 °C) by improving redox conditions, but its effectiveness is limited by dosage, sludge adaptation, and wastewater characteristics. The results emphasize the importance of operational conditions—especially temperature thresholds and influent composition—in determining the net impact of ZVI on anammox performance.

Microbial Community Structure Changes

The influence of nanomaterials on microbial community within activated sludge systems has been the subject of numerous studies [65, 70], with particular attention given to the effects of ZVI on the anammox process [21], [20]. These investigations have provided valuable insights into how ZVI can enhance or inhibit microbial activity under controlled conditions. Nevertheless, a significant knowledge gap still exists regarding the specific role of ZVI in shaping the structure and function of anammox bacterial communities under more challenging conditions—specifically at low temperatures and in the presence of MWW.

In this study, the microbial community composition at the phylum level revealed distinct differences between the C-reactor (Fig. 6a) and Z-reactor (Fig. 6b) over the experimental period. Seven major bacterial phyla were identified across both treatment conditions: Acidobacteriota, Bacteroidota, Chloroflexi, Nitrospirota, Planctomycetota, Proteobacteria, and Verrucomicrobiota. Similar community was obtained in the anammox process supported by rGO under low-temperature performance [60]. In both reactors, Proteobacteria consistently represented the dominant phylum; however, their relative abundance was substantially higher in the C-reactor, ranging from 32.98% to 42.43%, compared to 24.84% to 51.32% in the ZVI reactor. Notably, ZVI addition appeared to stimulate Proteobacteria proliferation, peaking at over 50% on day 230, suggesting enhanced activity or selection under iron-amended conditions. Moreover, switching to MWW led to slight decreases in the abundance of Proteobacteria in both reactors. Subsequently, a renewed increase was observed in the Z-reactor following the reintroduction of synthetic wastewater (Phase VI). Planctomycetota also displayed higher proportions in the Z-reactor (up to 28.14% on day 128), in contrast to the control where values fluctuated around 10–13%. This trend indicates potential responsiveness of this phylum to ZVI-associated biochemical changes, especially stimulate enzyme production, because Fe(II) and Fe(III) are components of heme-containing enzymes [10]. Nevertheless, despite the growth of Planctomycetes abundance under ZVI-support, it still ranked second after Proteobacteria. Similar trend was observed by Fu et al., [18] in iron-assisted anammox process, but under optimal temperature condition for anammox process. Bacteroidota, known for their role in organic matter degradation, maintained moderate abundances in both conditions but showed more pronounced fluctuations in the control reactor (up to 26.43% on day 167). The ZVI system displayed slightly lower and more stable levels (ranging 13.53–18.42%), possibly reflecting altered substrate availability or competitive dynamics. On the other hand, Bacteroidetes has an iron-reducing capabilities and experimentally confirmed its involvement in the Feammox reaction [12]. Other dominant phyla: Chloroflexi, Nitrospirota, Acidobacteriota, and Verrucomicrobiota, remained minor constituents in both reactors. Overall, the introduction of ZVI led to a notable restructuring of the microbial community. The most marked changes were observed in the increased dominance of Proteobacteria and Planctomycetota, suggesting these groups may play key roles in the biogeochemical processes stimulated by ZVI, such as metal cycling, denitrification, or degradation of complex organics.

Fig. 6.

Analysis of community structures of anammox process at phylum level in C-Reactor a and Z-Reactor b; at genus level in C-Reactor c and Z-Reactor d; PCA analysis between dominant genus and type of reactor e; biodiversity Shannon f, Simpson g and Chao1 h indexed in both reactors

At the genus level, the microbial communities in both the control (Fig. 6c) and ZVI-amended (Fig. 6d) reactors exhibited complex temporal dynamics and notable compositional differences. Across both systems, several dominant genera were consistently detected, including Micavibrionales_unc, Nitrosomonas, Saprospiraceae_unc, Candidatus Brocadia, OLB8, and members of Comamonadaceae and Anaerolineaceae.

Micavibrionales_unclassified was the dominant genus in both reactors, though its dynamics differed between systems. In the C-reactor, abundance peaked at 10.7% on day 48 before declining slightly, while in the Z-reactor it raised steadily from 2.6% to over 10% by day 167, suggesting stimulation by ZVI. After switching to MWW, Micavibrionales_unclassified increased in the C-reactor (10.4% on day 209) but dropped in the Z-reactor before partially recovering in Phase VI. A similar pattern was observed for Nitrosomonas, an ammonia-oxidizing bacterium (AOB), indicating a possible correlation. Indeed, in the Z-reactor a strong positive correlation (r = 0.8, p < 0.05) was found between Micavibrionales_unclassified and Nitrosomona (Figure S2, see Supplementary Materials), consistent with the report of Shi et al. [51] linking Micavibrionales with the amoA gene encoding ammonia monooxygenase (AMO). As AMO requires iron as a cofactor [22], ZVI may have enhanced Nitrosomonas activity [65], indirectly supporting Micavibrionales_unclassified growth. The abundance of Nitrosomonas, similar to Micavibrionales_unclassified, decreased after after switching to MWW (day 209). This may be related to unfavorable conditions and competition for iron with iron-associated heterotrophs, which are promoted by the presence of organic compounds in MWW [38, 39].

Candidatus Brocadia, a key anammox genus, exhibited higher abundance in the Z-reactor, reaching 14.015% on day 48, compared to 5.208% in the C-reactor. However, both reactors were operated at 25 °C, which is close to the optimal temperature for AnAOB [60]. When the temperature dropped to 14 °C, the abundance of Candidatus Brocadia in the Z-reactor decreased to 7.545% on day 128, while in the C-reactor slightly increased to 5.243%. A similar trend was observed for the second dominant anammox genus, Candidatus Jettenia, whose abundance in the Z-reactor decreased from 3.122% (day 48) to 1.551% (day 128), while in the C-reactor it increased from 0.450% to 1.074%. Afterward, both Candidatus Brocadia and Candidatus Jettenia maintained relatively stable levels in the Z-reactor, with only minor fluctuations. Similarly, Liu et al. [38] reported that Candidatus Brocadia was the dominant anammox bacterium in a system operated at 28 °C with the addition of ZVI (0.08 g/gTS). Under similar ZVI and temperature conditions, but with varying HRT, the dominance of Candidatus Brocadia was also confirmed by Liu et al. [39]. Liu et al. [35–37] investigated the anammox bacterial community at low temperature (12 °C) in the presence of iron and revealed that ZVI further enriched the community, with Candidatus Brocadia identified as the predominant genus. During MWW feeding, the abundance of Candidatus Brocadia in the Z-reactor raised to 8.946% on day 209, followed by a slight decline to 6.704% on day 230 (after switching to mineral medium). In the C-reactor, Candidatus Brocadia remained relatively stable even during MWW feeding, while Candidatus Jettenia showed a gradual decline, reaching 0.381% by day 230. Meanwhile, the organic compounds present in municipal wastewater should typically inhibit the growth of AnAOB, yet a slight increase in their abundance was observed in both reactors, particularly in Candidatus Brocadia. As previously demonstrated, Candidatus Brocadia possesses key enzymes involved in the heterotrophic nitrate reduction to ammonium (DNRA) pathway. Moreover, the DNRA pathway within the anammox process relies on a carbon source as an electron donor [14, 76]. This suggests that under unfavorable temperature conditions (13 °C) present in the reactor during the MWW feeding period, and potentially supported by the presence of ZVI, Candidatus Brocadia may have activated the DNRA pathway to sustain cellular energy production. Despite the overall higher abundance of AnAOB in the Z-reactor, a comparison of microbial community data with nitrogen removal rate (NRR) results (see Table 3) indicates that, following the change from MWW to synthetic wastewater, the increase in NRR was lower in the Z-reactor than in the C-reactor. This trend is also reflected in the observed changes in microbial community structure. On day 230, the abundance of Candidatus Brocadia was 0.790% higher in the C-reactor than in the Z-reactor. This discrepancy may be attributed to the presence of organic compounds in the MWW, combined with the influence of ZVI, which likely stimulated the growth of denitrifiers. These denitrifiers may have continued to perform endogenous denitrification even after the switch to synthetic wastewater. This hypothesis is supported by the findings of Zhang et al. [74], who demonstrated that the development of denitrifiers in their study was linked to the utilization of organic compounds derived from extracellular polymeric substances (EPSs) and dead cells to support metabolic processes in wastewater-fed systems without the addition of external organic carbon. Additionally, Feng et al. [17] suggest that the addition of ZVI promotes the production of EPSs, which enhance hydrolytic acidification processes. This, in turn, could inhibit AnAOB activity while favoring the proliferation of denitrifiers. Under these conditions—and at a relatively low operational temperature of 13 °C—the development of AnAOB may have remained suppressed even after transitioning to synthetic wastewater.

Other genera, such as Ferruginibacter, Comamonadaceae_unclassified, and Anaerolineaceae (unclassified), were also consistently but modestly represented across both treatments. Anaerolineaceae_unclassified exhibited higher abundance in the Z-reactor compared to the C-reactor up to day 167. However, following the switch to MWW, the distribution shifted, with the C-reactor showing a higher abundance, reaching 2.027% on day 209, while the Z-reactor dropped to 0.803%. Hu et al. [26] reported that Anaerolineaceae are capable of microbial extracellular electron transfer (EET) to external electron acceptors such as biochar or Fe(III) [63, 64]. This ability may explain their initially higher abundance in the Z-reactor up to day 167. Conversely, Li et al. [34] demonstrated that Anaerolineaceae can degrade organic compounds into intermediates that are readily utilized by denitrifying bacteria. Therefore, after switching to real wastewater, a cooperative interaction with denitrifiers may have occurred, contributing to the increased abundance observed in the C-reactor on days 209 and 230, while ZVI may have disrupted this cooperation. Comamonadaceae_unclassified are heterotrophic bacteria common in anammox systems, where they degrade detritus and peptides from AnAOB and recycle nitrate to nitrite [30]. They can also adapt to temperatures below 15 °C [43], which may have supported their growth under low-temperature conditions. In this research, their abundance was consistently higher in the Z-reactor, peaking at 5.8% during MWW feeding (day 209) while in the C-reactor remaining below 1.5% throughout the entire experimental period. The presence of MWW clearly promoted the growth of Comamonadaceae_unclassified in both reactors, likely due to the availability of organic compounds. Moreover, Zhang et al. [70] reported that Comamonadaceae can interact with iron-reducing microorganisms as well as possess the ability to ferrous oxidation as part of the denitrification process. This capability may explain the significantly greater abundance of Comamonadaceae_unclassified in the Z-reactor. Ferruginibacter responded most clearly to ZVI, increasing from negligible levels to over 3%. As a Feammox bacterium, it can use Fe(III) to oxidize N₂, NO₂⁻-N, or NO₃⁻-N [58], explaining its selective enrichment in the Z-reactor. Interestingly, Denitratisoma, known for denitrification, showed a gradual increase in both systems, but reached higher levels in the Z-reactor. When comparing both reactors, Denitratisoma exhibited the highest abundance on day 209 in Z-reactor, reaching 8.586%, while in the C-reactor during the same period, its abundance was 2.115%. Its growth during MWW feeding likely reflected the combined effects of iron and elevated organic concentration, which supports denitrification pathways. A similar conclusion was presented by Zhou et al. [76], who observed an increase in the abundance of Denitratisoma from 0.12% to 0.33% during a two-stage partial denitrification-anammox process with the addition of ZVI (0.625 mg ZVI/L). Jiang et al. [28] also found that the relative abundance of Denitratisoma increased with the introduction of Fe(II).

In summary, ZVI addition significantly influenced genus-level community structure, favoring genera like Micavibrionales_ uncultured, Candidatus Brocadia, and Ferruginibacter, while potentially suppressing others like Saprospiraceae_ uncultured and Nitrosomonas. These shifts point to a reorganization of nitrogen and carbon cycling functions in response to iron amendment, with implications for system performance and biogeochemical stability.

The principal component analysis (PCA) revealed a clear relationship between specific bacterial species and the type of reactor used (Fig. 6e). Notably, the species Candidatus Jettenia, Denitratisoma, and Comamonadaceae_unknown formed a distinct cluster in the PCA plot, indicating that their abundance and distribution are strongly influenced by the presence of ZVI in the reactor. This suggests that the addition of ZVI creates specific environmental conditions that selectively promote the growth or activity of these microbial groups. A similar pattern was observed for Candidatus Brocadia, which also exhibited a preferential association with the ZVI-amended reactor. These findings imply that ZVI may play a significant role in shaping the microbial community structure, potentially enhancing the performance of the anammox process under low-temperature conditions. In contrast, a significant association with the C-reactor was observed for the species Nitrosomonas, Lacunisphaera, OM190, and Anaerolineaceae_uncultured. These microorganisms also formed a distinct cluster in the PCA analysis, suggesting that lack of ZVI favor the growth or metabolic activity of these particular taxa. Their clustering indicates a shared ecological niche or similar response to the physicochemical conditions present in the control setup.

Changes in the composition of bacterial communities, as indicated by the Shannon (Fig. 6f), Simpson (Fig. 6g), and Chao1 (Fig. 6h) diversity indices, show that the Z-reactor exhibited lower overall biodiversity throughout the experiment, but greater specialization toward specific bacterial groups such as AnAOB, which indicates higher abundance in the reactor supplemented with ZVI. In contrast, the C-reactor maintained a more balanced and stable microbial community, which may enhance the long-term resilience of the microbiome to environmental disturbances.

Possible Mechanisms and Theoretical Implications

The application of ZVI in anammox systems has garnered significant interest due to its multifaceted role in promoting microbial activity, system resilience, and nitrogen removal efficiency. The beneficial effects of ZVI are particularly evident under stress conditions such as low temperature or low substrate availability, where traditional anammox systems often exhibit reduced performance.

Upon contact with aqueous environments, ZVI undergoes corrosion, releasing ferrous ions (Fe(II)) and ferric ions (Fe(III)) (Fig. 7). Fe(II) is particularly important for AnAOB, serving as a cofactor in key enzymatic systems including hydrazine synthase (HZS), hydrazine oxidoreductase (HZO), and nitrite reductase (NIR). These enzymes are essential for the core anammox pathway converting NH₄⁺ and NO₂⁻ to N₂ gas. Additionally, Fe(II) supports the biosynthesis of cytochrome c (Cyt C), a critical component in electron transport chains within AnAOB [20, 71]. Liu et al. [35] also described that the exogenous addition of ZVI at low temperatures increased the abundance of the bfr gene, enhancing the synthesis of Bfr-proteins by AnAOB for the storage of excess Fe(II). This stored Fe(II) could later be mobilized to support AnAOB metabolic processes when environmental iron became limited. Moreover, ZVI increase abundance of iron-sulfur protein genes which enhance the efficiency of electron transfer between anammox function enzymes [17]. Recent studies have demonstrated that the long-term addition of ZVI (5 g/L) at 17 °C leads to a significant increase in enzymatic activity (e.g., + 94.12% for NIR, + 36.36% for HZS), thereby enhancing metabolic rates and nitrogen removal efficiency under low-temperature conditions [36]. Beyond metabolic contributions, ZVI promotes the production of extracellular polymeric substances (EPSs) and quorum sensing molecules such as C6-HSL and C8-HSL. These molecules facilitate microbial communication and aggregation, resulting in denser biofilm and granule structures [36]. EPSs production not only enhances biomass retention but also provides protection against toxic shocks and shear stress, particularly in continuous-flow reactors [9]. ZVI can also function as a chemical reductant in autotrophic denitrification pathways, reducing nitrate (NO₃⁻) to nitrite (NO₂⁻), the latter being a key electron acceptor in the anammox process. This reaction provides a sustainable source of NO₂⁻, particularly in mainstream applications where its accumulation via partial nitrification or partial denitrification is often unstable [72, 76]. The involvement of ZVI in processes such as Feammox (Fe(III)-mediated ammonium oxidation) and nitrate-dependent ferrous oxidation (NAFO) further underscores its role in integrated nitrogen removal mechanisms as can be seen in Fig. 7 [71].

Fig. 7.

Potential mechanism of ZVI on anammox process and nitrogen removal

In systems with significant dissolved organic nitrogen (DON), ZVI facilitates the breakdown of high-molecular-weight and highly oxygenated compounds through decarboxylation and nitro group reduction. This not only reduces the toxicity of effluents but also alters the molecular profile of DON, making it more amenable to microbial degradation [16]). These chemical transformations occur via secondary reactions with ZVI and contribute to improved total nitrogen removal and effluent quality.

Particularly relevant for temperate and cold climate applications, ZVI has been shown to enhance system stability at temperatures as low as 15 °C. Systems supplemented with nano-ZVI modified biochar (nZVI@BC) demonstrated higher heme c content (+ 36.6–91.45%), improved granulation, and upregulation of key functional genes (e.g., hzsA/B/C, hzo, hdh) compared to controls [9]. These molecular-level changes enable AnAOB to maintain high activity under thermal stress, ensuring consistent nitrogen removal.

Conclusion

This study comprehensively evaluated the short- and long-term effects of zero-valent iron (ZVI) on the anammox process under suboptimal temperature conditions, including synthetic and municipal wastewater treatment. The findings demonstrated that low-dose ZVI (5 mg/L) supplementation can stimulate specific anammox activity (SAA), particularly under moderate to low temperature conditions (15–13 °C), by enhancing enzymatic activities, modulating oxidative stress responses, and influencing microbial community dynamics. The stimulatory effect was most pronounced in short-term assays and under stable low-temperature operation (13 °C). However, under long-term operation and during exposure to real municipal wastewater, the benefits of ZVI supplementation diminished. Factors such as organic load, biomass adaptation, and potential iron passivation likely played a role in this performance decline. Community structure analysis revealed that ZVI fosters the growth of specific microbial taxa, including Candidatus Brocadia and Denitratisoma, while also shifting overall microbial diversity towards specialized communities. These shifts have implications for reactor resilience and nitrogen removal efficiency. In summary, while ZVI shows promise as an effective additive for enhancing anammox activity, particularly at moderately low temperatures, its long-term utility—especially in complex and variable wastewater matrices—requires further optimization. Future work should focus on refining dosing strategies, understanding microbial interactions at the functional gene level, and integrating ZVI-based enhancement into full-scale wastewater treatment scenarios.

Supplementary Information

Below is the link to the electronic supplementary material.

Author Contributions

Initially, the draft was prepared by FG, MĆW, AZB, GC. This draft was reviewed and edited by AZB, MMH, SM. Further, the manuscript was revised as per reviewer comments by GC, AZB, MT. All authors have read and agreed to the published version of the manuscript.

Funding

The study was financed by the National Science Centre, Poland (UMO-2017/25/N/NZ9/01159 and UMO-2013/09/D/NZ9/02438), and support by “Nauka dla Społeczeństwa II” program at Gdansk University of Technology, Faculty of Civil and Environmental Engineering. Moreover, this research was financially supported by the SUT with the project no. 08/070/BKM24/0033 (BKM-715/RIE7/2024). A. Ziembińska-Buczyńska is financed by SUT with the project no. 08/070/BK_25/0038 (BK-283/RIE7/2025).

Data Availability

No datasets were generated or analysed during the current study.

Declarations

Conflict of interest

The authors declare no competing interests.