Zero-Valent Iron-Enhanced Nitrogen and Phosphorus Removal Performance and Mechanisms in a GAO-Dominated System

Abstract

Zero-valent iron (ZVI) improves nitrogen and phosphorus removal efficiency in municipal wastewater treatment systems. However, its effects on glycogen-accumulating organisms (GAOs) and the mechanisms of action are still not fully understood. This study investigated the influence of ZVI on GAOs by assessing its effect on the performance of a denitrification reactor dominated by Candidatus Competibacter sp. (with a relative abundance of 46.33%). After ZVI addition, the reactor harbored the iron-autotrophic denitrifying bacterium Ferruginibacter sp. (0.38%), the iron-reducing bacterium Pseudomonas sp. (0.26%), and the iron-oxidizing bacterium Comamonas sp. (0.26%). This effectively integrated ammonia-oxidizing bacteria (AOB) and nitrite-oxidizing bacteria (NOB) with heterotrophic denitrification alongside iron ammonium oxidation and nitrate-dependent ferrous oxidation. The integrated approach decreased aeration duration while improving the removal efficiency of PO₄³-P (98.01 ± 2.54%) and total inorganic nitrogen (TIN) (62.01 ± 1.45%). The combined of ZVI in systems predominantly occupied by GAOs enhanced the carbon storage capacity of the bacteria.

1. Introduction

Enhanced biological phosphorus removal (EBPR) depends on the activity of polyphosphate accumulating organisms (PAOs). In conventional EBPR systems, glycogen accumulating organisms (GAOs) are viewed as harmful competitors, depleting carbon sources during the anaerobic stage and perhaps resulting in system failure [1,2,3]. As a result, techniques to inhibit GAOs while enhancing PAOs have been rigorously investigated [4,5].

Nonetheless, recent research indicates a more complex function for GAOs in the removal of nitrogen and phosphorus, especially under a low carbon-to-nitrogen (C/N) ratio. The anaerobic carbon storage capacity of GAOs can promote endogenous denitrification, enhancing nitrogen removal efficiency [6,7,8]. Liu et al. [9] developed a combined partial nitrification, endogenous denitrification, partial denitrification, and anammox (PNEnD/A) process within an anaerobic/aerobic/anoxic (An/O/A) reactor for the treatment of low C/N wastewater, resulting in an effluent total inorganic nitrogen (TIN) concentration as low as 2.9 mg/L. Additionally, specialized denitrifying glycogen accumulating organisms (DGAOs) can collaborate with PAOs for concurrent nitrogen and phosphorus removal [10]. Research has shown that An/O/A systems executing simultaneous nitrification denitrification and phosphorus removal (SNDPR) can attain elevated removal efficiencies and in situ sludge reduction, frequently linked to a significant presence of GAOs like Candidatus Competibacter sp. [11]. These findings underscore the substantial role of GAOs in endogenous denitrification and sludge reduction.

Prior research indicates that PAOs exhibit greater competitiveness than GAOs under conditions of low dissolved oxygen (DO) levels [4,5]. This suggests that the regulation of GAOs and PAOs by DO control is a complex issue that necessitates further research to maximize the benefits of GAOs and improve the stability and efficiency of nitrogen and phosphorus removal.

Chemical phosphorus removal agents are frequently employed to enhance EBPR in municipal wastewater treatment. Nevertheless, they may impede microbial activity and compete with PAOs for phosphate. Conversely, zero-valent iron (ZVI) serves as an alternative, environmentally plentiful, and economical mineral with diverse applications in water remediation [12,13]. This transition emphasizes the potential of ZVI as an instrument for microbial community engineering and reinforces the necessity to reevaluate the functional role of GAOs in hybrid chemical–biological nitrogen and phosphorus removal systems.

This study analyzed the impact of ZVI on GAOs within a simultaneous nitrification denitrification sequencing batch reactor (SBR) enriched with GAO bacteria. The research assessed the performance of nitrogen and phosphorus removal at varying C/N ratios with the addition of ZVI and investigated the mechanisms involved in the removal processes.

2. Materials and Methods

2.1. The Composition and Operating Conditions of the An/O/A-SBR System

This study employed a laboratory-scale SBR reactor made of acrylic glass with an effective volume of 5 L (Figure 1), a drainage ratio of 2/5, and the hydraulic retention time (HRT) of 15 h. The reactor operated in An/O/A mode with four cycles per day, each lasting 6 h. Each cycle encompassed influent filling for 5 min, anaerobic stage for 55 min, aerobic stage for 120–240 min, anoxic stage for 30–150 min, settling for 20 min, and effluent withdrawal for 10 min. The aerobic stage was controlled by the programmed logic control (PLC)-DO real-time monitoring system, maintaining DO levels at 0.2 ± 0.1 mg/L.

Microscale zero-valent iron (mZVI) was added to the SBR intermittently. The particle size of the mZVI was 45 μm (analytical purity, Tianjin Damao Chemical Reagent Factory, Tianjin, China). A total of three additions were made (on days 38, 76, and 110, respectively). The specific quantitative indicators and operating conditions for ZVI addition are shown in

Table 1. Experimental operating conditions.

Phases	Days of Operation	Influent C/N Ratio	ZVI Addition Concentration 1 (g-mZVI/g-MLSS)	Add Methods
I	0–37	7	0	/
II-1	38–75	7	0.33	Intermittent dosing: dosing was performed at the initial anaerobic stage on days 38 and 76 of a specific SBR cycle
II-2	76–85	7	0.33
III-1	86–109	4	0	/
III-2	110–135	4	0.33	Intermittent dosing: dosing was performed at the initial anaerobic stage on day 110 of a specific SBR cycle
IV	136–163	4	0	/
V	164–206	6	0	/

Note: ¹ The ZVI per unit mass of sludge.

.

2.2. Sludge and Sewage

The SBR reactor with simultaneous nitrification denitrification (SND) had been stable for a year. The reactor originally possessed a biological phosphorus removal function. After one month of anaerobic terminal phosphorus deprivation and phosphate-free influent operation, it became a biological nitrogen removal reactor without phosphorus removal [14]. Mixed liquor suspended solids (MLSS) and settling velocity (SV₃₀) were kept at 4000 ± 500 mg/L and 20~30%, respectively.

The experiment used a synthetic water mixture with COD (CH₃COOH) of 420 mg/L, 240 mg/L, and 360 mg/L respectively; NH₄Cl of 60 mg-N/L; NaHCO₃ of 650 mg/L; KH₂PO₄ of 5 mg-P/L; MgSO₄·7H₂O of 15 mg/L, CaCl₂ of 11.2 mg/L, and trace elements of 1 mL/L. The trace element composition included 50 mg/L H₃BO₃, 30 mg/L CuCl₂, 50 mg/L ZnCl₂, 500 mg/L MgSO₄·7H₂O, 50 mg/L CoCl₂·6H₂O, 50 mg/L AlCl₃, 50 mg/L NiCl₂, and 1 mL/L concentrated hydrochloric acid.

2.3. Indicator Determination

The sludge water mixture was removed from the reactor. Samples were filtered through a 0.45 mm membrane filter prior to analysis for the determination of chemical oxygen demand (COD), NH₄⁺-N, NO₂⁻-N, NO₃⁻-N, PO₄³⁻-P, Fe²⁺, total iron ions, MLSS and SV₃₀ according to standard methods [15]. The Fe³⁺ concentration was calculated as the difference between total iron and Fe²⁺. The dissolved nitric oxide (NO) in the reactor was detected online by a NO-500 sensor (Unisense Company, Aarhus, Denmark).

The nitrogen reduction efficiency during the aerobic stage is denoted as SND efficiency and is calculated using Equation (1) [16].

$$\mathrm{SND}={\displaystyle \frac{{\mathrm{TIN}}_{\mathrm{a}}-{\mathrm{TIN}}_0}{{\mathrm{NH}}_{4,\mathrm{a}}^{+}-{\mathrm{NH}}_{4,0}^{+}}}\times 100\%$$

(1)

where TIN_a and ${\mathrm{NH}}_{4,\mathrm{a}}^{+}$ represent the concentrations of TIN and ammonia nitrogen at the beginning of the aerobic stage, respectively, mg/L; TIN₀ and ${\mathrm{NH}}_{4,0}^{+}$ represent the concentrations of TIN and ammonia nitrogen at the ending of the aerobic stage, respectively, mg/L.

The anaerobic organic carbon consumption (COD_consum) in the system was defined as the consumptions of influent COD during the anaerobic stage, consisting of two components. One portion represents COD consumed by exogenous denitrification of nitrates and nitrites; the other portion represents COD stored as intracellular carbon sources (referring to the activity of PAOs and GAOs). Calculated by Equations (2) and (3):

$${\mathrm{COD}}_{\mathrm{consum}} = {\mathrm{COD}}_{\mathrm{dn}} + {\mathrm{COD}}_{\mathrm{ins}}$$

(2)

$${\mathrm{COD}}_{\mathrm{dn}}=2.86{\mathrm{NO}}_{3,\mathrm{a}}^{-}+1.71{\mathrm{NO}}_{2,\mathrm{a}}^{-}$$

(3)

where COD_dn was the COD consumed by denitrification in the anaerobic stage, mg/L; COD_ins was the COD consumed by PAOs and GAOs for storage with intracellular carbon sources, mg/L. The 2.86 and 1.71 values were the COD consumption per unit mass concentration of NO₃⁻-N and NO₂⁻-N utilized by denitrifying bacteria, respectively, mg/L; ${\mathrm{NO}}_{3,\mathrm{a}}^{-}$ and ${\mathrm{NO}}_{2,\mathrm{a}}^{-}$ were the concentrations of anaerobic initial NO₃⁻-N and NO₂⁻-N, respectively, mg/L.

The COD_ins could be calculated by Equation (4):

$${\mathrm{COD}}_{\mathrm{ins}} = \Delta \mathrm{COD}-(2.86\Delta {\mathrm{NO}}_3^{-} + 1.71\Delta {\mathrm{NO}}_2^{-})$$

(4)

where ΔCOD, ΔNO₃⁻ and ΔNO₂⁻ were the concentration values of COD, NO₃⁻-N and NO₂⁻-N during the anaerobic stage, respectively, mg/L.

The COD_ins rate was defined as the percentage of COD in sewage during the anaerobic stage that was stored as an internal carbon source by PAOs and GAOs, calculated by Equation (5):

$$\mathrm{COD}\mathrm{ins}(\%) =\left(1-{\displaystyle \frac{2.86\Delta {\mathrm{NO}}_3^{-}\text{-}\mathrm{N} + 1.71\Delta {\mathrm{NO}}_2^{-}\text{-}\mathrm{N}}{\Delta \mathrm{COD}}}\right)\times 100\%$$

(5)

2.4. Microbial Community Analysis

This study employed high-throughput sequencing to analyze microbial community characteristics on days 58 and 178, with 16S rRNA high-throughput sequencing conducted by Paisenuo Biotechnology Co., Ltd. (Shanghai, China). Paired-end sequencing of sample DNA fragments was performed using the Illumina method. The resulting DNA data were compared against the 16S rDNA gene database, and gene sequences with approximately 97% similarity were grouped into operational taxonomic units (OTUs). Amplification of the V3-V4 region was performed using primers 338F and 806R, followed by taxonomic composition analysis of the samples using QIIME2 2026 software.

3. Results and Discussion

3.1. The Nitrogen and Phosphorus Removal Performance of the SBR Process

The SBR reactor was operated for 206 days to assess the impact of ZVI addition on nitrogen and phosphorus removal (Figure 2). Initial operation (Phase I, C/N = 7) showed rapid recovery of denitrification, yielding the average effluent concentrations of TIN and PO₄³⁻-P of 12.02 ± 4.03 mg/L and 4.78 ± 0.38 mg/L, respectively, but no biological phosphorus removal was observed.

A single ZVI addition (0.33 g-mZVI/g-MLSS) at the start of the anaerobic stage (Phase II-1) resulted in immediate and stable phosphate removal. During the initial period of ZVI addition, which lasted for 9 days, the effluent PO₄³⁻-P remained below 0.5 mg/L, and the effluent TIN was below 7.0 mg/L. On the 13th day of ZVI addition, the PO₄³⁻-P in the effluent gradually increased to 0.96 mg/L. The addition of ZVI was continuously monitored for 38 days, with average effluent concentrations of TIN and PO₄³⁻-P reaching 6.66 ± 2.07 mg/L and 1.33 ± 1.11 mg/L, respectively. Results demonstrated that the addition of ZVI during the initial anaerobic stage significantly reduced the effluent PO₄³⁻-P. A second identical ZVI addition confirmed this effect. During this period (Phase II-2), the average concentrations of TIN and PO₄³⁻-P in the effluent were 4.26 ± 0.77 mg/L and 0.28 ± 0.35 mg/L, respectively.

Reducing the C/N ratio from 7 to 4 (Phase III-1) caused a sharp increase in effluent nitrate, but phosphate removal remained effective. After decreasing the C/N ratio, the average concentrations of TIN and PO₄³⁻-P in the effluent were 27.96 ± 3.56 mg/L and 1.23 ± 0.85 mg/L, respectively. A subsequent third ZVI addition (Phase III-2) at this low C/N ratio immediately suppressed effluent phosphate. The average concentrations of TIN and PO₄³⁻-P in the effluent were 27.49 ± 3.40 mg/L and 0.10 ± 0.13 mg/L, respectively, confirming that ZVI’s phosphate removal efficacy was independent of the C/N ratio.

The chemical removal effect of ZVI diminished after approximately 54 days (Phase IV), leading to a rise in effluent PO₄³⁻-P (4.04 ± 1.86 mg/L). Increasing the C/N ratio from 4 to 6 in Phase V restored denitrification (the average effluent TIN of 9.63 ± 3.52 mg/L) but not biological phosphorus removal, as neither anaerobic phosphate release nor aerobic uptake was observed. This confirms that nitrogen removal was driven by GAOs, while ZVI addition effectively suppressed PAOs activity and provided chemical phosphate fixation.

At the C/N ratio of 4, the third addition of ZVI significantly improved phosphorus removal, yet the average TIN concentration in the effluent remained essentially unchanged. Performance improved as the C/N ratio was increased, implying that organic carbon limitation was the dominant factor. However, in the absence of simultaneous control experiments at a C/N ratio of 4 without ZVI, it is not possible to completely rule out the potential impact of persistent iron accumulation under low C/N conditions.

3.2. Effect of Intermittent ZVI Addition on Pollutant Removal for Typical Cyclic Systems

Cyclic performance was assessed on selected operational days: prior to ZVI addition (Day 36), and following ZVI dosing on Day 40 (3 days post-first addition), Day 46 (9 days post-first addition), Day 110 (1 day post-third addition), Day 120 (11 days post-third addition), and Day 135 (26 days post-third addition). Temporal profiles of nitrogen, phosphorus, DO, COD, NO, pH, ORP, Fe²⁺ and Fe³⁺ are shown in Figure 3 and Figure 4.

At a C/N ratio of 7, before the addition of ZVI (Day 36), the system demonstrated SND at a rate of 55.58%, and COD_dn was 0.40 mg/L. Throughout the full cycle, no accumulation of NO₂⁻-N was observed (Figure 3a). The NO concentration showed a slight increase during the late anaerobic stage and initial aerobic stage (Figure 4a). This was likely due to the accumulation of nitrite resulting from heterotrophic denitrification of nitrate during the initial aerobic stage. The free nitrous acid (FNA) formed from nitrite entered microbial cells, and the stressed response of denitrifying microorganisms to neutralize FNA toxicity led to the production and release of NO [17,18,19]. Nonetheless, no biological phosphorus removal was detected, phosphorus release during the anaerobic stage was negligible (0.43 mg/L), and the effluent PO₄³⁻-P concentration (5.35 mg/L) corresponded with the influent level, signifying a lack of PAO activity (Figure 3a). The COD_ins rate (99.66%) indicated that more COD was absorbed by GAOs and common heterotrophic bacteria in the anaerobic stage. The swift pH elevation during the anaerobic stage (Figure 4a) aligns with COD storage. As carbon sources (acetic acid) enter cells via transmembrane transport, their active transport relies on the cotransport mechanism driven by the proton gradient (ΔpH). This causes extracellular H⁺ to migrate intracellularly, thereby reducing the proton concentration in the liquid and causing system pH to increase. The process exhibits synergistic metabolic associations with efficient COD utilization and reduction power (reduced nicotinamide adenine dinucleotide, NADH) regeneration during the Poly-β-hydroxy alkanoate (PHA) synthesis stage [20].

Phosphorus elimination was promptly and consistently accomplished following the initial ZVI addition. On Days 40 and 46 (Figure 3b,c), anaerobic phosphorus uptake transpired rather than release, yielding effluent PO₄³⁻-P of 0.11 mg/L. The SND rates rose to 67.37% and 74.16%, respectively. No buildup of NO₂⁻-N was observed. The consistently elevated COD_ins rates (>98%) verified that GAOs were the principal contributors to carbon source utilization. The results indicate that the addition of ZVI facilitated steady concurrent removal of nitrogen and phosphorus in a reactor that initially lacked biological phosphorus removal capabilities.

At a reduced C/N ratio of 4, after adapting the system to a lower C/N ratio and with subsequent ZVI additions, distinct patterns emerged. On Days 110 and 120 (Figure 3d,e), anaerobic phosphorus uptake persisted, culminating in complete phosphate removal (effluent PO₄³⁻-P ≈ 0 mg/L). Notably, the efficiency of chemical phosphorus removal by ZVI appeared to enhance with repeated dosing. However, nitrogen removal dynamics changed significantly: SND rates drastically declined to 42.66% and 1.18%, respectively, accompanied by elevated effluent NO₃⁻-N and higher COD_dn values. This indicates a shift toward classical nitrification with limited concurrent denitrification under carbon-limited conditions. Correspondingly, the COD_ins rates plummeted to 0% and 14%, reflecting severely constrained internal carbon storage. On Day 135, the effluent PO₄³⁻-P concentration was 0.64 mg/L (Figure 3f). The SND rate was 3.53%, and the COD_dn was 29.61 mg/L. The NO concentration in the liquid did not exhibit a distinct peak, showing a gradual upward trend with minimal variation throughout the cycle. The COD_ins was 32.23 mg/L and the COD_ins rate was 52%.

Synoptic Analysis: A comparative analysis indicates a significant impact of the C/N ratio on system functionality following the addition of ZVI. At a C/N ratio of 7, the addition of ZVI improved SND rates while sustaining elevated COD_ins rates, signifying effective GAO-mediated carbon storage and endogenous denitrification [21]. Conversely, at a C/N ratio of 4, nitrogen removal transitioned to mostly nitrification with inadequate SND, and carbon storage was markedly inhibited. Significantly, under all conditions, intermittent ZVI addition consistently elevated the COD_ins rate within the GAO-dominated community, affirming its function in augmenting the carbon source store capacity of GAOs, despite the total nitrogen removal pathway being influenced by substrate availability.

3.3. Microbial Diversity and Community Structure Analysis

Alpha diversity and relative abundance of microbial species on days 58 and 178 of reactor operation were determined.

3.3.1. Microbial Diversity Analysis

The Chao1 and Shannon indices of the microbial community under ZVI addition were higher than those after ZVI addition was discontinued, while the Simpson index showed no significant change (

Table 2. Alpha diversity index analysis of microbial communities.

Sample Types	Chao1	Shannon	Simpson	Coverage
Addition of ZVI	1941.78	6.99	0.94	0.999115
Cessation of ZVI addition	1023.77	5.53	0.92	0.99984

). This may reflect increased species richness resulting from the addition of ZVI.

The distribution range of OTUs in samples with ZVI addition was significantly broader than that in samples where ZVI addition was discontinued (Figure 5), indicating a decline in species abundance and evenness following the cessation of ZVI addition. Comparing the abundance curves revealed that the curve for samples with ZVI addition was flatter than that for samples without ZVI addition, indicating that the OTU abundance distribution within the addition group was more balanced and exhibited higher evenness. This demonstrates that ZVI addition could enhance microbial community diversity.

3.3.2. Microbial Community Structure Analysis

The primary bacterial phyla associated with denitrification in biological sewage treatment systems include Proteobacteria, Nitrospirae, and Firmicutes [22,23]. As shown in Figure 6a, the microbial community structures of the two sludge samples are similar at the phylum level but differ in relative abundance. On Day 58, the dominant bacterial phyla were Proteobacteria (80.70%), Bacteroidetes (9.23%), Nitrospirae (4.06%), and Acidobacteria (1.94%). On Day 178, the dominant phyla were Proteobacteria (78.00%), Bacteroidetes (6.15%), Nitrospirae (5.02%), and Chloroflexi (3.91%). The dominant bacterial phylum Proteobacteria, associated with denitrification in the system, showed no significant change before and after the addition of ZVI. The Bacteroidetes phylum, associated with phosphorus removal, decreased from 9.23% during ZVI addition to 6.15% after cessation of ZVI addition, indicating that ZVI addition in the nitrogen removal-dominated reactor could enhance the relative abundance of phosphorus-removing bacterial genera to some extent.

Candidatus Competibacter sp. is common in sewage treatment systems as a GAO. Within SNDPR systems, it does not directly participate in phosphorus removal [24]. The primary bacterial genera involved in nitrification and denitrification include Nitrospira sp., Nitrosomonas sp., Azoarcus sp., Thauera sp., and Denitratisoma sp. [25,26]. Among these, Nitrosomonas sp. and Nitrospira sp. are common AOB and NOB in wastewater treatment plants [27,28]. Azoarcus sp., Thauera sp., and Denitratisoma sp. are denitrifying bacteria responsible for reducing nitrate to N₂ or nitrous oxide (N₂O) [26]. On Day 58, dominant bacterial genera included Candidatus Competibacter sp. (46.33%), Defluviicoccus sp. (7.15%), Nitrospira sp. (4.06%), Azoarcus sp. (4.88%), Terrimonas sp. (1.22%), Thauera sp. (0.8%), Nitrosomonas sp. (0.36%), and Denitratisoma sp. (0.58%) (Figure 6b). On Day 178, dominant bacterial genera included Candidatus Competibacter sp. (40.75%), Defluviicoccus sp. (6.55%), Nitrospira sp. (5.02%), Azoarcus sp. (1.11%), Thauera sp. (2.7%), Candidatus Alysiosphaera sp. (2.5%), Nitrosomonas sp. (1.03%), and Denitratisoma sp. (0.39%) (Figure 6b). Based on previous studies, common bacterial genera in activated sludge systems include Candidatus Competibacter sp., Nitrospira sp., Thauera sp., Nitrosospira sp., Defluviimonas sp., and Nitrosomonas sp., which were largely consistent with the genera detected in this study. Ferruginibacter sp. is the iron-autotrophic denitrifying bacterial genus, Comamonas sp. is an iron-oxidizing bacterium, and Pseudomonas sp. is an iron-reducing bacterium [29,30,31,32,33]. On Day 58, the relative abundances were 0.38%, 0.26%, and 0.26%, respectively. However, on day 178, none of these three iron bacteria were observed.

It is worth noting that the relative abundance of Candidatus Competibacter sp. decreased from 46.33% (On Day 58) to 40.75% (On Day 178). Despite the 5.58% decline, Candidatus Competibacter sp. retained the overwhelming dominant position, and the system’s denitrification performance recovered significantly with the increase in the C/N ratio from 4 to 6 (Phase V). The functional capacity for endogenous denitrification remained intact, and denitrification performance was determined more by carbon source limitation than by the slight decrease in the relative abundance of GAOs. Furthermore, the loss of phosphate removal capacity was associated with the depletion of ZVI’s chemical phosphate removal capacity, rather than changes in the microbial community.

3.4. The Mechanism Analysis for Nitrogen and Phosphorus Removal in the An/O/A-SBR Process

The analysis of microbial communities indicated the presence of iron-dependent nitrogen transformation pathways within the GAO-dominated system (Figure 6b). In addition to the predominant Candidatus Competibacter sp. (46.33%), iron bacteria were identified, including Ferruginibacter sp. (0.38%), Pseudomonas sp. (0.26%), and Comamonas sp. (0.26%), which might imply the possible coexistence of iron oxidation and reduction processes.

Data from the cyclic process were consistent with the putative occurrence of Feammox. Moreover, further isotope tracer experiments are required to verify this, which will form a key part of our subsequent research. Under constant aeration intensity, the reduced aeration time for complete ammonium oxidation suggests an alternative ammonia removal pathway, possibly facilitated by iron-reducing bacteria such as Pseudomonas sp., which transforms ammonia nitrogen to nitrate without the accumulation of intermediates. This was corroborated by a simultaneous decrease in Fe³⁺ and an increase in Fe²⁺ during aeration. In natural environments, the oxidation products of Feammox have been shown to be primarily nitrate, nitrite, and nitrogen gas. However, the primary product of Feammox in activated sludge for sewage treatment is nitrate [34,35]. Denitrification occurred mainly through heterotrophic pathways in GAOs and through autotrophic pathways associated with iron. The nitrate-dependent ferrous oxidation (NDFO) process was inferred to predominate during the anoxic/anaerobic stage, as suggested by the observed decrease in Fe²⁺ accompanying the reduction of NO₃⁻ to N₂ and minimal accumulation of NO₂⁻ or NH₄⁺. The persistently low levels of NO and inferred low levels of N₂O suggest complete nitrogen conversion to N₂, with biological processes predominating over abiotic reactions. Based on the above analysis, the inferred nitrogen and phosphorus removal mechanism is shown in Figure 7.

The Figure 7 indicates that during the aerobic stage, with DO maintained at 0.2 ± 0.1 mg/L, ammonia nitrogen oxidation occurs by both aerobic oxidation of ammonia nitrogen and nitrite by AOB and NOB, as well as anaerobic oxidation of ammonia nitrogen by iron-reducing bacteria Pseudomonas sp. (Feammox). The products of these processes form nitrate nitrogen. Regarding denitrification, the process involves heterotrophic denitrification with endogenous PHA as electron donor in GAOs, and autotrophic denitrification of ZVI and the NDFO process that reduces nitrate nitrogen with Fe²⁺ as electron donor, both yielding nitrogen gas as the final product. During the anaerobic and anoxic stages, heterotrophic denitrification occurs alongside autotrophic denitrification involving ZVI and Fe²⁺. Heterotrophic denitrification reduced nitrate nitrogen to nitrogen gas. Autotrophic denitrification of ZVI and Fe²⁺ oxidizes ZVI to Fe²⁺, oxidizes Fe²⁺ to Fe³⁺, and reduces nitrate to nitrogen gas.

In this study, NO concentrations remained low throughout the process, insufficient to exert inhibitory effects on microorganisms. In denitrification pathways, NO is a key intermediate that can further reduce to N₂O or N₂. The low dissolved NO concentrations observed in this study, coupled with nitrifying heterotrophic denitrification, Feammox, and NDFO for nitrogen and phosphorus removal, suggest a potential for reduced N₂O emissions in this system. However, as N₂O was not directly quantified in the present study, this remains a hypothesis that warrants direct verification in future work. The integration of nitrification–heterotrophic denitrification, Feammox, and NDFO processes improved system performance. The addition of ZVI resulted in increased nitrogen and phosphorus removal efficiencies, rising from 53.15 ± 5.85% and 75.12 ± 17.34% to 62.01 ± 1.45% and 98.01 ± 2.54%, respectively.

The biological oxidation process of ZVI is a slow process. To ensure sufficient Fe²⁺ in the solution, a sufficient amount of ZVI must be added at one time, allowing for it to be gradually oxidized in situ during the An/O/A cycle. This provides the Fe²⁺ required for nitrifying heterotrophic denitrification, and the coupled Feammox and NDFO reactions. A single adequate dose of ZVI facilitated continuous in situ generation of Fe²⁺, thereby promoting iron-mediated pathways and leading to the formation of highly insoluble iron-phosphate precipitates [36]. ZVI may be added as a single full dose followed by daily supplemental additions. The supplement quantity should be determined based on the amount discharged daily from the system as excess sludge (ensuring the requirements for chemical phosphorus removal and biological nitrogen removal under appropriate sludge retention time). The process of gradually eliminating PAOs from an An/O/A-SBR reactor capable of both nitrogen removal and phosphorus removal to cultivate a pure denitrifying An/O/A-SBR reactor revealed that within the mixed microbial community of PAOs and GAOs, GAOs likely occupied a subordinate position. Only after the deliberate elimination of PAOs did GAOs gradually transition to a dominant status. The reactors experiencing deteriorated phosphorus removal due to excessive GAOs growth may not be caused by reduced PAOs activity resulting from GAOs competition for carbon sources, but rather by the PAOs themselves failing to adapt to environmental changes. The increase in GAOs relative abundance may merely be a consequence of decreased PAOs relative abundance rather than its cause.

To rigorously validate the proposed Feammox and NDFO pathways and to quantitatively distinguish biotic from abiotic contributions to nitrogen conversion and N₂O generation, future studies should adopt advanced mechanistic validation strategies. Ma et al. [37] developed a physically integrated multiscale framework, demonstrating how physically coupled modelling can be used to quantify coupled biogeochemical processes. This approach offers a promising pathway for quantitatively disentangling the coupled contributions of heterotrophic denitrification, iron-mediated autotrophic pathways, and chemical precipitation in the GAO-dominated system. Furthermore, Wang et al. [38] explored chemically driven pathways in water treatment. By identifying and tracking key reactive intermediates, the study established a framework for distinguishing between biocatalytic transformations and chemically driven processes. For example, by tracking the dynamic variations of NO, N₂O and iron species using specific inhibitors, or by conducting isotope tracer experiments, it will be possible to clearly distinguish between biological N₂O reduction and chemical denitrification.

4. Conclusions

This study identified iron-autotrophic redox bacteria in a SBR dominated by GAOs (Candidatus Competibacter sp., relative abundance 46.33%) after ZVI enhancement. Within 26 days following a single ZVI addition, the removal efficiencies of TIN and PO₄³⁻-P reached 62.01 ± 1.45% and 98.01 ± 2.54%, respectively. The relative abundance of the iron-autotrophic denitrifying genus Ferruginibacter sp. was 0.38%, the iron-reducing bacterium Pseudomonas sp. was 0.26%, and the iron-oxidizing bacterium Comamonas sp. was 0.26%. Results hinted that autotrophic oxidation and reduction processes of iron might have occurred within the reactor.