Mechanistic insights into As(III) removal from groundwater using iron net-based zero-valent iron: reaction pathways and inhibitory effects of phosphorus and silicon

Abstract

Arsenic (As) contamination of groundwater poses a serious public health risk in many regions worldwide. Zero-valent iron (ZVI)–based systems are widely used for As removal. Nevertheless, their performance is strongly influenced by coexisting groundwater constituents such as phosphorus (P) and silicon (Si). This study systematically investigates the removal mechanisms of arsenite (As(III)) from groundwater using an iron net-based ZVI (IN-ZVI) system. Batch experiments were conducted to evaluate As(III) removal kinetics, iron dissolution behavior, and adsorption isotherms under varying P and Si concentrations. Spectroscopic analyses using X-ray Absorption Fine Structure (XAFS) and X-ray Photoelectron Spectroscopy (XPS) were employed to elucidate As(III) oxidation pathways and surface transformations of the iron net. Results show that As(III) removal followed first-order kinetics, whereas Fe(II) dissolution exhibited zero-order behavior. Increasing P and Si concentrations significantly suppressed both As(III) removal and iron dissolution, with P exerting a stronger inhibitory effect than Si. Langmuir and Freundlich isotherm analyses revealed a progressive decrease in adsorption capacity with increasing P and Si, attributed to competitive site occupation and surface passivation. XAFS results confirmed complete oxidation of As(III) to As(V) in the IN-ZVI system, while XPS analysis demonstrated extensive formation of Fe(III) oxides and deposition of P- and Si-bearing surface scales. These findings indicate that As(III) removal proceeds via corrosion-induced Fe redox cycling, reactive oxygen species-mediated oxidation, and subsequent adsorption of As(V) onto hydrous ferric oxides. The study highlights the critical role of groundwater chemistry in controlling ZVI reactivity and provides mechanistic insights for optimizing iron-based arsenic removal technologies in P and Si rich groundwater.

Introduction

Arsenic (As) is a recognized toxicant and carcinogen, and use of water containing it poses significant health risks¹. In many developing nations, groundwater pollution is the primary pathway for human exposure to arsenic. The World Health Organization (WHO) establishes the recommended value for arsenic in drinking water at 10 μg/L for public health purposes². Nevertheless, several nations with severely contaminated aquifers continue to use a permissible limit of 50 μg/L³. Natural arsenic pollution is prevalent in shallow aquifers located in arid and semi-arid regions. It originates from geological rocks that include arsenic⁴. Recent research in Saudi Arabia indicates that some aquifer systems have arsenic concentrations above 10 μg/L⁵. In central Saudi Arabia, groundwater serves as the primary supply of potable water⁶. This indicates that a significant number of individuals may be subjected to exposure via both oral and dermal contact.

Arsenic in groundwater originates from the weathering of alluvial sediments containing arsenic-bearing pyrite minerals. The main parameters influencing its mobility across aquifers are the salt content, pH, and redox potential⁴. Arsenic typically exists in two oxidation states: arsenite (As(III), represented as H₃AsO₃ or HAsO₂) and arsenate (As(V), denoted as H₂AsO₄⁻ or HAsO₄²⁻). Under reducing circumstances, As(III) is the predominant form. It is also more hazardous and more difficult to eliminate using conventional adsorption and co-precipitation techniques. Pre-oxidation to As(V) facilitates the removal of As(III), although it complicates treatment operations and increases expenses^7–9. An alternative, prevalent technique involves the oxidation of Fe(II) followed by the coprecipitation of As with the resultant Fe(III) hydroxides^10,11. As(III) is concurrently oxidized to As(V) during the oxidation of Fe(II), facilitating its removal¹². Fe(II) is more reactive than Fe(III), and naturally available dissolved Fe(II) in groundwater may fulfill the need for Fe(II) in treatment. Zerovalent iron (ZVI) [Fe(0)] is a cost-effective and straightforward method for obtaining Fe (II), and it is more prevalent in contemporary treatment systems^13,14.

The composition of the groundwater significantly influences the efficacy of Fe-based coprecipitation in arsenic removal^15,16. Prevalent anions such as carbonate, silicate, sulfate, phosphate, and natural organic materials might complicate the removal of both As(V) and As(III)^17,18. Phosphate (P) and silicate (Si) significantly hinder the adsorption process. They alter the configuration of ferric hydroxide, impeding As absorption and decelerating the oxidation rate of Fe^19,20.

Phosphate (P) and silicate (Si) ions are particularly elevated in Bangladesh’s groundwater¹⁴, complicating the removal of arsenic. Phosphate has a significant affinity for iron oxides and oxyhydroxides; hence, it competes with arsenate for the same binding sites. This diminishes its efficacy in adsorbing As(V). Previous investigations reported substantial reductions in As(V) adsorption at all pH levels with just 0.3 mg/L of PO₄-P present²¹. Silicate complicates the removal of As when ferric hydroxides serve as the adsorbent²². Liu et al.²³ demonstrated that silicate adsorption and surface complexation reduce the ζ-potential of ferric hydroxides, particularly at elevated pH levels²³. This process intensifies the repulsion between particles, hence inhibiting the formation of flocs. The resultant smaller flocs possess less efficient sites for arsenate sorption, hence diminishing the total removal efficacy. Therefore, elevated concentrations of phosphate and silicate in groundwater remain a significant challenge for arsenic removal technology. In addition to competing with As for adsorption sites, P and Si act as inhibitors of iron dissolution in the zero-valent iron (ZVI) system²⁴.

Several zero-valent iron (ZVI)–based arsenic removal technologies, such as the SONO filter, Kanchan arsenic filter (KAF), and arsenic removal ceramic filter, have been successfully implemented in arsenic-affected regions^14,25,26. These technologies typically employ ZVI as the iron source, where Fe(II) released during ZVI corrosion is subsequently oxidized to form iron (hydr)oxides [e.g., Fe(OH)₃], which adsorb As from groundwater. Such filters are commonly fabricated using corroding metallic iron materials, including iron nets, iron nails, or iron filings. However, it has been observed that As removal efficiency decreases significantly with prolonged filtration and, in many cases, fails to meet drinking water standards. Therefore, elucidating the mechanisms of As(III) removal by ZVI is critical for the effective and sustainable implementation of As removal technologies. To develop an efficient and economical water filtration system, a more profound understanding of the competitive interactions between As and coexisting groundwater constituents is essential.

Despite extensive research on ZVI-based As removal, the mechanistic influence of coexisting groundwater constituents such as P and Si on iron-net-based ZVI (IN-ZVI) systems remains insufficiently understood. Previous studies have shown that phosphate and silicate can strongly interfere with As adsorption by competing for reactive surface sites and altering iron corrosion products^3,23,24. However, their combined influence on Fe dissolution kinetics, Fe(II)/Fe(III) transformation pathways, reactive oxygen species (ROS) generation, and surface passivation mechanisms under iron-net configurations relevant to decentralized groundwater treatment has not been comprehensively investigated. In particular, the interactions between P/Si and hydrous ferric oxide (HFO) surfaces may significantly regulate As(III) oxidation and subsequent As(V) immobilization. Therefore, understanding the coupled influence of P and Si is essential for improving the long-term performance and reliability of ZVI-based As treatment systems in complex groundwater environments.

The present study aims to mechanistically evaluate As(III) removal using an IN-ZVI system in the presence of P and Si. The specific objectives were to: (i) investigate As(III) removal and Fe dissolution kinetics of the iron net under varying P and Si concentrations in groundwater; (ii) evaluate the adsorption behavior of As(III) through isotherm modeling in the presence of different P and Si concentrations; (iii) elucidate the oxidation mechanism of As(III) using X-ray Absorption Fine Structure (XAFS) spectroscopy; and (iv) examine changes in the surface oxidation states of the iron net using X-ray photoelectron spectroscopy (XPS). Accordingly, a locally available iron net wire was collected, and systematic batch experiments were conducted.

Materials and methods

Materials

A commercially available iron net wire (0.75 mm diameter) was purchased from the local market and used as the source of zero-valent iron (ZVI). Elemental analysis of the iron net confirmed that Fe was the dominant element on the surface, accounting for the majority of the detected composition (99–99.5%). Minor amounts of oxygen were also detected, indicating the presence of a thin native oxide layer formed due to exposure to air. Trace levels of other elements, such as carbon and silicon, were observed, which may be attributed to surface contamination or manufacturing-related impurities. These results confirm that the iron net mainly consisted of elemental iron with slight surface oxidation prior to reaction. The iron net was cut into pieces smaller than 2 mm, thoroughly washed with pure water, dried in an oven at 100 °C, and used directly in this study without any additional pretreatment. The same pretreatment procedure was consistently applied to all experiments to ensure reproducibility. All reagent-grade chemicals were obtained from Wako Company (Wako, Japan) or Nacalai Tesque Inc. (Kyoto, Japan). Laboratory-grade NaAsO₂ and Na₂HAsO₂ powder reagents were used as the sources of As(III) and As(V), respectively. Reagent-grade FeSO₄·7H₂O and FeCl₃ powders served as the sources of Fe(II) and Fe(III). Likewise, K₂HPO₄ and Na₂SiO₃·5H₂O were used as the phosphorus (P) and silicon (Si) sources. Stock solutions of all reagents were prepared using ultrapure water and stored at 4 °C. These stock solutions were subsequently diluted to obtain the required concentrations of As(III), As(V), Fe(II), Fe(III), P, and Si for the adsorption experiments. Synthetic groundwater simulating the geochemical characteristics of Bangladesh groundwater was used in this study. Selected chemical reagents were dissolved in ultrapure water to achieve the desired ionic composition of the synthetic groundwater. The concentrations of the synthetic groundwater constituents are presented in Table 1. The concentrations of Fe, P, and Si in the synthetic groundwater were adjusted according to the specific experimental conditions.

Table 1.

Composition of synthetic groundwater used in this study.

Parameters	Unit	Concentrations
pH	–	7.0–7.5
As(III)/As(V)	µg/L	500
Fe	mg/L	***
SiO2	mg/L	***
Mn	mg/L	0.45
Ca2+	mg/L	100
Mg2+	mg/L	40
K +	mg/L	25
Na+	mg/L	140
PO43-	mg/L	***
SO42-	mg/L	200
HCO3-	mg/L	350
Cl-	mg/L	210

***Concentrations were adjusted according to the specific experimental conditions

Batch adsorption experiments

Three types of batch adsorption experiments, adsorption kinetics, adsorption isotherms, and comparative adsorption, were conducted to investigate the mechanisms of As(III) removal by the iron net-based- ZVI (IN-ZVI) system. In the batch kinetics experiments, the As(III) adsorption rate was evaluated under varying concentrations of P and Si in the synthetic groundwater. For each experiment, 1000 mg of iron net was mixed at 40 rpm for 120 h in 1 L of synthetic groundwater containing 500 µg/L As(III), using a jar-test apparatus. A total of six kinetic experiments were performed: three at different P concentrations (0, 1, and 2 mg/L) and three at different Si concentrations (0, 30, and 50 mg/L). The selected P (1 and 2 mg/L) and Si concentrations (30 and 50 mg/L) were chosen based on reported concentration ranges commonly observed in arsenic-affected groundwater, particularly in Bangladesh and other South Asian aquifers where phosphate and silicate coexist at elevated levels¹⁴.

Water samples were collected at predetermined intervals (0, 5, 24, 48, 72, and 120 h) and analyzed for total As, dissolved Fe(II), total Fe (T-Fe), P, and Si. For the analysis of total As, dissolved Fe(II), P, and Si, water samples were filtered through 0.45 µm membrane filters (Advantec, Japan) prior to analysis. In contrast, samples for total iron (T-Fe) analysis were collected without filtration and digested using aqua regia on a hot plate. After digestion, the samples were cooled, diluted with deionized water, and subsequently filtered through 0.45 µm membrane filters to remove residual particulate matter before ICP analysis. Accordingly, T-Fe represented the combined concentration of Fe(II) and Fe(III) released from Fe(0) corrosion into the system. In contrast, the iron concentration measured after filtration without digestion was considered to represent the dissolved Fe(II) fraction.

For the batch isotherm experiments, 100 mL of synthetic groundwater containing 500 µg/L As(III) and varying doses of iron net (500, 250, 100, 50, and 20 mg/L) was added to Teflon bottles and shaken horizontally at 30 rpm for 120 h. Five sets of isotherm experiments were conducted to evaluate the adsorption characteristics and adsorption capacity of the IN-ZVI system under different P and Si conditions. The first set was carried out without any added P or Si, while two additional sets were performed with P concentrations of 1 and 2 mg/L, respectively. The remaining two sets were conducted using synthetic groundwater containing 30 and 50 mg/L of Si. Samples were collected after 120 h, filtered through a 0.45 µm membrane filter (Advantec, Japan), and analyzed for total As.

Comparative adsorption experiments were carried out to evaluate the efficiency and elucidate the mechanism of As(III) removal by ZVI (Fe(0)) relative to other iron species, namely Fe(II) and Fe(III). A total of six batch experiments were conducted using Fe(0), Fe(II), and Fe(III) as adsorbents. In each experiment, 5 mg/L of Fe(0), Fe(II), or Fe(III) was added separately to 1 L of synthetic groundwater containing 500 µg/L As(III) or As(V). The suspensions were mixed at 40 rpm in a jar-test apparatus for 120 h. After the reaction, samples were collected, filtered through a 0.45 µm membrane filter (Advantec, Japan), and analyzed for total As. All experiments were conducted under ambient aerobic conditions at pH 7.0–7.5, dissolved oxygen (DO) concentrations of 6.0–6.5 mg/L, and oxidation–reduction potential (ORP) values of 100–150 mV. The suspensions were continuously mixed using a jar test apparatus to maintain homogeneous conditions and facilitate oxygen transfer throughout the reaction period. pH, DO, and ORP were periodically monitored using calibrated portable probes during the experiments. No external chemical buffer was added. Each experiment was performed in triplicate to ensure reproducibility, and the results are presented as mean values with corresponding standard deviations/error bars.

Analytical procedure

The water samples were analyzed according to the standard methods described by the American Public Health Association²⁷. pH and oxidation–reduction potential (ORP) were measured using a standard pH probe connected to a HORIBA D-54S meter, following the manufacturer’s operating manual. Dissolved oxygen (DO) was determined using a standard DO probe (HACH HQ30d) in accordance with the HACH user manual. Total arsenic (As) was quantified using inductively coupled plasma mass spectrometry (ICP–MS; Yokogawa HP 4500) following APHA Method 3125. Iron (Fe) and silicon (Si) concentrations were measured using inductively coupled plasma atomic emission spectrometry (ICP–AES; Seiko SPS 4000) in accordance with APHA Method 3120. Phosphorus (P), expressed as PO₄–P, was analyzed using a UV–visible spectrophotometer (Shimadzu, Japan) following APHA Method 4500-P.

X-ray photoelectron spectroscopy (XPS) and X-ray absorption fine structure (XAFS) analysis

X-ray photoelectron spectroscopy (XPS) was conducted to examine the oxidation states of the iron net surface after the batch experiments. Following each experiment, the residual iron nets were collected, air-dried, and directly analyzed by XPS without further pretreatment. XPS measurements were performed using a PHI Quantera SXM system equipped with Al Kα radiation (1486.6 eV). Standard reference materials, including Fe(0) powder, FeSO₄ for Fe(II), and FeCl₃ for Fe(III), were used to support peak identification and interpretation of the Fe oxidation states.

X-ray Absorption Fine Structure (XAFS) spectroscopy was employed to determine the oxidation state of arsenic in the batch experiments. The As L₃-edge spectra were obtained for hydrous ferric oxide (HFO) flocs collected from the reactors, which were air-dried before analysis. Measurements were conducted using dispersive crystals at the BL-10 beamline. Two arsenic standards, NaAsO₂ as As(III) and Na₂HAsO₄·7H₂O as As(V), were used as reference materials. The spectrum peaks of the HFO samples were compared with those of the standard reference compounds to distinguish between As(III) and As(V) species.

Results and discussion

As(III) removal kinetics of iron net-ZVI

The As(III) removal kinetics of iron net–ZVI at varying phosphorus (P) and silicon (Si) concentrations are shown in Fig. 1. A gradual decrease in aqueous As(III) concentration with increasing contact time was observed under all conditions. However, the removal rate markedly decreased with increasing P concentration, indicating that phosphate exerts a negative effect on As(III) removal by iron net–ZVI. Phosphate ions strongly compete with arsenic species for sorption sites on the iron (hydr)oxide layers formed during ZVI corrosion. Owing to their ability to form stable inner-sphere bidentate or binuclear complexes with Fe(III)–OH surface groups, phosphate species (PO₄–P) exhibit a higher binding affinity than predominantly neutral As(III) species under circumneutral pH conditions, resulting in preferential occupation of reactive iron sites²⁸. Consequently, the presence of phosphate significantly suppresses As(III) removal by ZVI, primarily through blockage of sorption sites on Fe(III) corrosion products such as ferrihydrite, lepidocrocite, and goethite²⁹. This competitive inhibition leads to reduced arsenic uptake, diminished oxidation of As(III), and prolonged persistence of dissolved arsenic in solution³⁰. Recent studies further indicate that phosphate induces a pronounced “passivating effect” on ZVI under aerobic conditions by limiting iron oxidation and dissolution³¹. The concurrent formation of iron–phosphate or related surface precipitates contributes to surface passivation, thereby reducing the generation of reactive Fe(II)/Fe(III) species required for effective arsenic removal³². Overall, phosphate acts as a strong inhibitory species that diminishes the effectiveness of ZVI-based arsenic remediation through a combination of competitive site blocking and surface passivation mechanisms. The temporary increase in the measured As concentration at P = 2 mg/L after 5 h was likely caused by experimental fluctuations associated with sample heterogeneity and analytical uncertainty during the early reaction period.

Fig. 1.

Kinetics of As(III) removal by iron-net-based zero-valent iron (ZVI) under different concentrations of (a) phosphorus (P) and (b) silicon (Si).

The impact of Si on the As(III) removal rate is presented in Fig. 1b. Results showed that relatively lower Si (30 mg/L) did not influence the As(III) removal rate. However, at higher Si concentration in groundwater (50 mg/L) the As(III) removal rate was significantly reduced. Silicon is well known to reduce Fe dissolution by promoting the formation of stable, Si-enriched passive films on iron surfaces. The formation of silica-stabilized ferrihydrite contributes to corrosion inhibition, as Si-bearing ferrihydrite exhibits markedly lower solubility and a more resistant surface structure³³. In zero-valent iron (ZVI) systems, these interactions result in decreased Fe (II) release and slower corrosion. Collectively, these studies demonstrate that dissolved silicate consistently acts as a passivating agent that suppresses iron dissolution by stabilizing the iron oxide corrosion layer. As a result, the As(III) removal rate decreases with increasing Si in the groundwater.

As (III) removal by IN-ZVI followed the first-order adsorption kinetics and was significantly influenced by the presence of P and Si. As shown in Table 2, the rate constant for P = 0, Si = 0 is 0.041 h^-1, indicating the highest adsorption rate in the absence of P and Si in the groundwater. Increasing P concentrations to 1 mg/L and 2 mg/L progressively decreases the rate constant to 0.021 h^-1 and 0.012 h^-1, respectively, demonstrating that phosphorus strongly inhibits As(III) adsorption, likely due to competitive binding at reactive Fe(III) sites and partial passivation of the iron surface. Similarly, the addition of dissolved silicon (Si = 30 mg/L and 50 mg/L) reduces the adsorption rate constants to 0.032 h^-1 and 0.018 h^-1, respectively, indicating that silicate also slows As(III) removal, presumably through surface modification and formation of Si-enriched passive layers that limit access to reactive Fe sites. Comparing the adsorption rate constant for P and Si demonstrated that P exhibits a stronger suppressive effect (lower rate constant) on As(III) removal than Si (higher rate constant) at the tested concentrations. The high correlation coefficients (R² = 0.935–0.998) confirm that first-order kinetics adequately describe As(III) adsorption under all tested conditions. The apparent first-order kinetics of As(III) removal suggest that the removal rate depends on the availability of reactive iron corrosion products and adsorption sites generated during Fe(0) oxidation. Thus, As(III) removal is governed by coupled oxidation–adsorption processes occurring at the iron/HFO interface rather than by simple equilibrium adsorption alone.

Table 2:

1^st order adsorption rate of As(III) and Zero-order Fe dissolution rate by Iron net ZVI.

Experiments	As adsorption rate		Fe dissolution rate
	1st order rate constant (h-1)	R2	Zero-order rate constant (mg/L/h)	R2
P = 0, Si = 0	0.041	0.987	0.267	0.997
P = 1	0.021	0.955	0.229	0.935
P = 2	0.012	0.935	0.117	0.919
Si = 30	0.032	0.998	0.203	0.972
Si = 50	0.018	0.997	0.108	0.918

Fe(II) dissolution and oxidation rate during the As(III) removal by iron net- ZVI

Fe(II) dissolution and subsequent oxidation during As(III) removal were evaluated in the batch reactor, and the results are presented in Fig. 2. Total dissolved iron (T-Fe) gradually increased with reaction time across all experimental conditions, reflecting the continuous corrosion and dissolution of iron from the iron net. However, the presence of phosphate (P) and silica (Si) had a pronounced inhibitory effect on iron release. Higher concentrations of both P and Si led to substantially lower T-Fe levels, indicating that these anions form surface complexes or protective layers on iron surfaces, thereby suppressing Fe dissolution and ultimately slowing the As(III) removal process. In the absence of P, T-Fe reached approximately 35 mg/L after 120 h, whereas increasing P to 1 and 2 mg/L reduced T-Fe to 27 mg/L and 14 mg/L, respectively. A similar trend was observed in groundwater containing elevated Si (30 and 50 mg/L), confirming that Si also plays a major role in inhibiting Fe release.

Fig. 2.

Fe(II) dissolution and oxidation rate during the As(III) removal by Iron net- ZVI at various P and Si concentrations.

Dissolved Fe(II) concentrations remained extremely low (< 0.5 mg/L) in all test conditions, with the only exception being a slight increase observed at 5 and 24 h in the Si-50 treatment (Fig. 2). This behavior indicates that although Fe(II) is continuously produced at the iron surface through corrosion, it is rapidly converted to Fe(III) hydroxides once released into the aqueous phase. Thus, Fe dissolution and Fe(II) oxidation proceed concurrently in the reactor.

The rapid depletion of Fe(II) is consistent with the physicochemical conditions of the experiments. The system operated at pH 7.0–7.5, DO 6.0–6.5 mg/L, ORP 100–150 mV, and a temperature of 25–26 °C, conditions known to strongly promote the abiotic oxidation of Fe(II) by dissolved oxygen. Classical Fe(II) oxidation kinetics³⁴ show that oxidation rates increase dramatically near circumneutral pH, with half-lives decreasing to the scale of seconds to a few minutes. Under these conditions, Fe(II) rarely accumulates in solution because it is rapidly transformed into Fe(III) precipitates or colloids. Therefore, dissolved Fe(II) remained negligible throughout the tests despite continuous Fe corrosion from the iron net. Overall, the combined evidence shows that Fe(II) is steadily produced but almost immediately oxidized to Fe(III), whereas Fe dissolution is inhibited by P and Si, and finally, the availability of dissolved Fe species directly controls As(III) oxidation and removal efficiency. The simultaneous occurrence of Fe(II) release and its rapid oxidation is therefore a key mechanistic feature of the iron-net ZVI system.

Based on the Fe(II) dissolution and total iron (T-Fe) results presented in Fig. 2, it was assumed that the Fe(II) dissolution rate closely reflected the T-Fe concentrations in the reactor over time, since dissolution and oxidation occurred simultaneously. Therefore, the Fe(II) dissolution rate from the iron net was estimated from the T-Fe graph. The analysis showed that the dissolution followed zero-order kinetics, and the dissolution rate decreased with increasing concentrations of phosphorus (P) and silicon (Si) (Table 2). The highest zero-order Fe(II) dissolution rate constant was 0.267 mg/L/h in the absence of P and Si. This rate decreased progressively to 0.229 mg/L/h and 0.117 mg/L/h as P increased to 1 and 2 mg/L, respectively. Similarly, the rate decreased to 0.208 mg/L/h and 0.108 mg/L/h as Si increased to 30 and 50 mg/L. The observed zero-order kinetics of Fe release indicate that iron dissolution from the Fe(0) surface was predominantly governed by surface corrosion reactions, in which the dissolution rate depended mainly on the availability of reactive surface sites rather than the dissolved Fe concentration. Consequently, a relatively constant Fe release rate was maintained during the initial stage of the reaction.

P and Si removal

Removal of P and Si during the As(III) removal was investigated, and the results are presented in Fig. 3. P removal was achieved 100% for both the P1 and P2 batches after 120 h. Si removal was lower and achieved 48% and 37%, respectively, for Si 30 and Si 50. The results indicate that P and Si removal occurred simultaneously with As removal, and P was adsorbed more effectively onto oxidized Fe(III) phases. Both P and Si adsorb onto hydrous ferric oxides (HFO) through different mechanisms and with different adsorption strengths. Phosphorus, typically present as phosphate (PO₄^3-), strongly binds to Fe(OH)₃ surfaces through inner-sphere ligand-exchange reactions, in which phosphate replaces surface hydroxyl groups and forms stable mono- or bidentate surface complexes²⁴. Moreover, phosphate may promote the formation of Fe–P surface precipitates and strongly suppress Fe dissolution through surface passivation.

Fig. 3.

P and Si removal %. P1 = P removal at P = 1 mg/L, P2 = P removal at P = 2 mg/L, P1 = Si removal at Si = 30 mg/L and Si removal at Si = 50 mg/L.

In contrast, dissolved silicon is predominantly present as monomeric silicic acid (H₄SiO₄) under circumneutral conditions. Silicic acid generally forms hydrogen-bonded outer-sphere complexes or weak inner-sphere complexes and can progressively polymerize to form a silicate layer coating the Fe(OH)₃ surfaces³⁵. Although both Si and P accumulate on Fe(OH)₃ surfaces, phosphate exhibits much stronger, and more competitive adsorption, whereas silicate adsorption is more gradual and mainly contributes to surface passivation rather than strong site-specific binding. Therefore, both P and Si not only inhibit iron dissolution but also reduce As(III) adsorption through occupying or blocking reactive Fe surface sites. These processes reduce Fe(II) availability and suppress ROS generation, thereby limiting As(III) oxidation. Thus, the inhibitory effects of P and Si involve not only competitive adsorption but also the regulation of corrosion-driven oxidation pathways and HFO surface transformations.

As(III) adsorption isotherm modelling

The adsorption behavior of the iron net -ZVI system under varying phosphorus (P) and silicon (Si) concentrations was evaluated using both the Langmuir and the Freundlich isotherm models. Table 3 summarizes the model parameters, while Fig. 4 presents the visual model fits. The Langmuir model showed reasonably good to excellent agreement with the experimental results under all tested conditions (R² = 0.904–0.999), indicating predominantly monolayer adsorption on relatively homogeneous sites. The maximum adsorption capacity (Q_max) was 2000 µg/g at P = 0, Si = 0, progressively decreasing to 1250 µg/g at P = 2 and 1000 µg/g at Si = 50, suggesting that higher P and Si concentrations reduce the number of accessible adsorption sites, potentially due to site blockage or surface modification. The Langmuir constant (K_L) was relatively low (0.004–0.006 L/µg) in the presence of P but increased in the presence of Si, reaching 0.036 L/µg at Si = 30 and 0.028 L/µg at Si = 50, indicating enhanced binding affinity at the remaining available active sites. Furthermore, the separation factor (R_L) values ranged from 0.728 to 0.961, confirming that adsorption remained favorable under all tested conditions³⁶.

Table 3.

Langmuir and Freundlich isotherm model parameter for As (III) removal by iron net.

Model parameters	P = 0, Si = 0	P = 1	P = 2	Si = 30	Si = 50
Langmuir model
R2	0.998	0.970	0.999	0.904	0.994
Qmax (µg/g)	2000	1667	1250	1429	1000
KL (L/µg)	0.006	0.004	0.004	0.036	0.028
RL	0.939	0.961	0.957	0.728	0.774
Freundlich model
R2	0.997	0.967	0.957	0.926	0.911
Kf	103.59	64.68	92.73	199.53	234.42
n	2.49	2.33	3.01	3.30	4.90

Fig. 4.

Langmuir and Freundlich isotherm fits for As(III) adsorption onto iron-net ZVI under different phosphorus and silicon concentrations.: (a) P = 0, Si = 0; (b) P = 1 mg/L; (c) P = 2 mg/L; (d) Si = 30 mg/L; and (e) Si = 50 mg/L.

The Freundlich model also described the adsorption behavior well, with R² values ranging from 0.911 to 0.997, highlighting the contribution of surface heterogeneity to the overall adsorption process. Notably, both K_f and n values were lower in the presence of phosphorus (K_f = 64.68–92.73; n = 2.33–3.01), indicating that P reduced adsorption intensity and limited access to high-energy adsorption sites. This is consistent with the well-documented tendency of PO₄-P to form strong inner-sphere complexes with iron surfaces, which can reduce surface reactivity and narrow the spectrum of available sorption sites. In contrast, the presence of Si markedly increased both K_f (199.53–234.42 µg/g) and n (3.30–4.90), suggesting stronger adsorption affinity and a greater degree of surface heterogeneity under Si-rich conditions. These increases likely reflect Si-induced surface restructuring, such as the formation of Fe–Si complexes or polymerized silica layers. These surface modifications may alter the surface charge distribution and create a broader variety of adsorption microenvironments³⁷.

The concurrent applicability of both the Langmuir and the Freundlich models suggests that the adsorption process involves a mixed mechanism of monolayer adsorption and surface heterogeneity under varying P and Si conditions. Phosphorus primarily reduces the adsorption capacity and intensity through site occupation and competitive interactions. Under certain Si conditions, the apparent increase in adsorption affinity may be attributed to enhanced surface heterogeneity and the formation of localized high-affinity adsorption domains induced by Si-related surface restructuring. Therefore, understanding these mechanisms is essential for improving the efficiency of iron-based adsorbents in groundwater, where P and Si frequently coexist.

Comparative study and XAFS analysis for As(III) oxidation and removal pathways

Figure 5 shows the comparative removal efficiencies of As(III) and As(V) in different forms of Fe. Removal of As(III) by Fe(III) was approximately 26%, likely due to adsorption and/or co-precipitation with ferric hydroxides. However, As (III) removal was higher (60%) with Fe(II), probably due to oxidation of As(III) to As(V) followed by Fe(III) precipitation. Zero-valent iron (ZVI) achieves the highest removal (96%) of As(III), reflecting the superior efficiency of this material with enhanced oxidation and co-precipitation mechanisms. Figure 5 also shows removal of As(V) by different iron forms. As(V)-Fe(III) achieved efficient removal (48%) of As(V) mainly via adsorption to Fe(III) hydroxides. Whereas Fe(II) was more efficient (94%) than Fe(III) for As(V) removal. Fe(0) showed the highest efficiency (96%) in removing As(V). Under aerobic conditions, dissolved oxygen oxidized Fe(II) to Fe(III) during Fe(0) corrosion, leading to the formation of reactive iron hydroxides that effectively adsorbed As(V).

Fig. 5.

Comparison of As (III) and As(V) removal by Fe (III), Fe (II), and Fe (0). Initial As (III) or As (V) = 500 ug/L, Fe (III) or Fe (II) or Fe (0) = 5 mg/L, pH = 7.0–7.5, DO = 6.0–6.5 mg/L and ORP = 100–150 mV.

In a comparison of As(III) and As(V) removal by different forms of Fe, it was observed that As(V) removal was higher than that of As(III) for all forms of Fe (Fig. 5), indicating that As(III) is less efficient and needs to be oxidized to As(V) for efficient removal. Fe(III) primarily facilitates As removal through adsorption and co-precipitation mechanisms. It is more effective for As(V) than As(III), as As(III) is less strongly adsorbed onto Fe(III) hydroxides³⁸. Whereas Fe(II) can partially oxidize As(III) to As(V) in the presence of dissolved oxygen but is less efficient than Fe(0). As shown in Fig. 5, the Fe(0) system exhibited high removal efficiencies for both arsenic species, achieving approximately 89% for As(III) and 96% for As(V). Therefore, As removal by Fe(0) can be considered one of the most efficient alternatives, owing to the combined effects of As(III) oxidation to As(V) via reactive oxygen species generated during Fe(0) corrosion (Fe (0) → Fe (II) → Fe(III)), followed by adsorption and/or co-precipitation with ferric hydroxides^39,40. Among all tested iron forms, ZVI exhibited the highest As(III) removal efficiency through an integrated mechanism involving: (i) corrosion-induced Fe(II)/Fe(III) cycling, (ii) in situ generation of reactive oxygen species, and (iii) the strong affinity of As(V) for iron (hydr)oxide surfaces.

For further mechanistic understanding of As removal by the iron-net ZVI system, hydrous ferric oxide (HFO) sludges generated from selected batch experiments were characterized using XAFS L₃-edge spectroscopy. Reagent-grade NaAsO₂ and Na₂HAsO₄ were used as reference standards representing As(III) and As(V), respectively. As shown in Fig. 6, the standard As(III) spectrum displayed a single peak at 1326 eV, whereas As(V) exhibited two distinct peaks at approximately 1327.5 eV and 1331 eV. These well-defined reference spectra were used to interpret the oxidation state of arsenic associated with the HFO samples. In the As(III)–Fe(III) batch system (Fig. 6a), the spectra did not show clear As(III) or As(V) peaks, indicating that As(III) remained largely unreactive and weakly adsorbed in the presence of Fe(III). Conversely, in the As(V)–Fe(III) batch, the two characteristic peaks at 1327.5 and 1331 eV confirm that As(V) was strongly adsorbed onto HFO flocs generated from the direct addition of Fe(III). These observations are consistent with the known behavior of As(V), which readily complexes with freshly precipitated ferric (hydr)oxides.

Fig. 6.

XAFS L₃-edge spectra of iron hydroxide flocs: (a) As(III)–Fe(III) and As(V)–Fe(III), (b) As(III)–Fe(II) and As(V)–Fe(II), and (c) As(III)–iron-net ZVI at different phosphorus and silicon concentrations, with As(III) and As(V) standards shown as references.

The As(III)-Fe(II) spectra (Fig. 6b) revealed a distinct shoulder/doublet at approximately 1327 and 1331 eV, which is characteristic of As(V). This observation indicates that a substantial fraction of As(III) was oxidized to As(V) during the transformation of Fe(II) to Fe(II) hydroxides. No discernible spectral peak was observed at ~ 1326 eV, corresponding to the As(III) standard, suggesting negligible adsorption of As(III) species onto the HFO. Similar dual peaks were also detected in the As(V)-Fe(II) system, confirming the efficient adsorption of As(V) by HFO formed during Fe(II) oxidation. These results demonstrate that Fe(II) serves as a precursor for the in situ generation of reactive Fe(III) hydroxides that both promote As(III) oxidation and provide strong adsorption sites for As(V).

The As(III)–iron-net systems (Fig. 6c) exhibited two well-defined peaks at approximately 1327 and 1331 eV, confirming that As(V) was the dominant arsenic species present on the resulting HFO surfaces. The absence of As(III)-related peaks suggests complete oxidation of As(III) before or during adsorption. These findings support the proposed mechanism whereby Fe(0) corrosion proceeds through sequential Fe(0) → Fe (II) → Fe (III) transformations, generating reactive surface sites and associated reactive oxygen species (ROS) that drive As(III) oxidation and subsequent immobilization as As(V).

The oxidation of As(III) in the Fe(0)–oxygen system is driven by a sequence of ROS-mediated reactions. Dissolution of Fe(0) releases Fe(II) and electrons, which reduce dissolved oxygen to superoxide radicals (O₂•⁻)⁴¹. These radicals rapidly undergo protonation and disproportionation to form hydrogen peroxide (H₂O₂), which then reacts with Fe(II) in classical Fenton pathways to generate hydroxyl radicals (•OH)^42,43. Hydroxyl radicals are extremely strong oxidants, capable of rapidly converting As(III) to As(V), which subsequently adsorbs onto freshly formed HFO surfaces. This combination of Fe(0) dissolution, ROS generation, and efficient As(V) adsorption explains the superior removal performance of Fe(0) relative to Fe(II) or Fe(0) systems (Fig. 4). Overall, the XAFS results clearly demonstrate that arsenic removal in the iron-net ZVI system occurs predominantly through As(III) oxidation to As(V) followed by strong adsorption of As(V) onto HFO flocs generated during iron corrosion. Although the XAFS results confirmed that As associated with the Fe corrosion products was predominantly present as As(V), the temporal evolution of dissolved and surface-bound As(III) and As(V) species was not quantitatively monitored in this study. Therefore, the relative contributions of direct As(III) adsorption and oxidative transformation followed by As(V) adsorption could not be clearly distinguished. Future investigations involving time-resolved As speciation analysis in both the aqueous phase and on Fe surfaces, particularly under varying phosphate and silicate concentrations, are recommended to further elucidate the coupled oxidation–adsorption mechanisms governing As removal in the IN-ZVI system. Moreover, the proposed ROS-mediated oxidation pathway is supported by XAFS observations and established Fe(0)-oxygen reaction mechanisms reported in the literature^42,43. However, direct experimental verification of ROS generation was not performed in this study. Future investigations involving radical scavenging experiments and electron paramagnetic resonance (EPR) analysis are recommended to further confirm the contribution of hydroxyl and superoxide radicals during As(III) oxidation. A detailed reaction pathway illustrating these coupled redox and adsorption processes, derived from the XAFS results, is presented in Fig. 7.

Fig. 7.

Reaction diagram for As(III) oxidation and removal process in ZVI system.

XPS characterization of surface transformations on the iron net

To investigate the surface transformation of the iron net after the adsorption experiments, X-ray photoelectron spectroscopy (XPS) analysis was performed, and the results are presented in Fig. 8. The Fe 2p spectra (Fig. 8a) indicates that the iron-net surface was predominantly covered with Fe(III) species, with no detectable Fe(II) or Fe(0) peaks in either sample. This observation confirms that extensive oxidation of Fe occurred during the adsorption process. The absence of Fe(II) and Fe(0) peaks suggests that the iron-net surface was fully converted into corroded Fe(III) oxides, which likely inhibited further Fe(II) dissolution as the reaction progressed. As shown in Fig. 8b, additional peaks corresponding to Si 2 s, P 2p, and Fe 3 s were observed, confirming the deposition of silicon (Si) and phosphorus (P) on the iron-net surface. The presence of P and Si indicates the formation of scale layers on the metal surface, which can further suppress Fe(II) release during the reaction²⁴. Although detailed peak deconvolution and quantitative surface speciation analysis were beyond the scope of the present work, the XPS results clearly support the formation of Fe(III)-dominated passive layers associated with phosphate and silicate accumulation. Overall, the XPS results demonstrate that the iron-net surface gradually becomes less reactive due to the accumulation of Fe(III) corrosion products and the deposition of P- and Si-containing scales, even under favorable pH, dissolved oxygen, and ORP conditions. Consequently, periodic surface regeneration through acid or alkali washing is necessary to restore the reactivity of the iron net. Further research is required to optimize regeneration frequency, cleaning agents, and operational protocols to ensure practical applicability under field conditions.

Fig. 8.

XPS spectra of (a) Fe2p, and (b) Si 2 s, P 2p, Si 2p, and Fe 3 s of the iron net collected after batch experiments.

Practical implication

The results of this study provide valuable insights for the design, operation, and long-term sustainability of IN-ZVI As removal systems, particularly in regions where groundwater contains elevated concentrations of phosphorus (P) and silicon (Si) with arsenic. The mechanistic insights gained from kinetic analyses, adsorption modeling, and advanced spectroscopic techniques provide a scientific basis for improving the performance and reliability of iron net–based ZVI filtration systems in field applications. The demonstrated inhibitory effects of P and Si on Fe dissolution and As(III) removal highlight the need to explicitly consider groundwater chemistry during system design and site selection. In many arsenic-affected regions, such as Bangladesh, India, Nepal, Pakistan, and parts of the Middle East, groundwater commonly contains elevated phosphate and silicate concentrations. Under these conditions, ZVI filters may experience rapid performance decline due to surface passivation and competitive adsorption. Therefore, routine characterization of P and Si in source water should be incorporated into pre-installation assessments, enabling informed decisions on filter sizing, iron loading, and expected service life.

The results show that iron dissolution is a key rate-limiting step controlling As(III) oxidation and removal. The observed zero-order Fe(II) dissolution behavior indicates that maintaining an active iron surface is critical for sustained treatment efficiency. The accumulation of Fe(III) oxides and P/Si-bearing scales on the iron net surface, as confirmed by XPS, directly suppresses Fe release over time. From an operational perspective, this finding highlights the necessity of periodic regeneration or replacement of iron media. Simple chemical regeneration methods, such as mild acid or alkaline washing, could be implemented at the household or community scale to remove passivating layers and restore iron reactivity. Optimizing regeneration frequency based on water chemistry and usage patterns can significantly extend filter lifespan and reduce operational costs. The strong competitive adsorption of phosphate relative to As suggests that staged treatment configurations may be advantageous in P-rich waters. Double-layered filter designs, where sacrificial iron media are placed upstream to preferentially bind phosphate, may mitigate passivation effects on the primary arsenic removal unit. These design strategies are particularly relevant for household-based filters, where simplicity and robustness are essential. The confirmed oxidation of As(III) to As(V) via corrosion-induced Fe redox cycling and reactive oxygen species (ROS) generation has important practical significance. Unlike conventional treatment systems that require separate pre-oxidation steps, the iron net–based ZVI system enables in situ As(III) oxidation under ambient conditions. This reduces system complexity, chemical requirements, and operational costs. However, the efficiency of this oxidative pathway is directly linked to iron corrosion rates, which are sensitive to groundwater composition. Maintaining adequate dissolved oxygen levels and avoiding excessive surface passivation are therefore critical for ensuring effective As(III) oxidation in field systems.

The comparative performance of Fe (0), Fe(II), and Fe(III) observed in this study reinforces the superiority of ZVI-based systems for treating mixed As(III)/As(V) groundwater. While Fe(III) relies primarily on adsorption and Fe(II) provides partial oxidation capacity, ZVI uniquely combines sustained iron release, As(III) oxidation, and strong As(V) adsorption. This multifunctional property makes IN-ZVI particularly suitable for decentralized water treatment in rural and resource-limited settings, where operational simplicity and chemical-free treatment are critical. From a policy and implementation perspective, the use of locally available iron nets offers a scalable and economically viable solution for arsenic mitigation. The affordability and widespread availability of iron nets make them suitable for household and community-level deployment. However, this study demonstrates that long-term performance cannot be assumed without proper management of surface passivation and groundwater chemistry. Integrating scientific guidelines on regeneration, monitoring of coexisting anions, and system maintenance into public health programs will be essential for ensuring sustained compliance with drinking water standards. Overall, this study guides the translation of laboratory-scale insights into practical As removal solutions. By accounting for groundwater composition, managing iron surface passivation, and optimizing system design and maintenance, iron net–based ZVI filters can be effectively adapted for long-term use in arsenic-affected regions worldwide.

From the operational perspective of the ZVI system, excessive Fe dissolution may create operational and environmental concerns if residual iron remains in the treated water. Under the aerobic conditions employed in this study, dissolved Fe(II) was rapidly oxidized and precipitated as Fe(III) hydroxides, resulting in negligible dissolved iron concentrations in the aqueous phase. However, long-term operation may still require careful monitoring of residual iron concentrations, sludge generation, and the appropriate handling and disposal of Fe–As-rich precipitates to minimize potential secondary environmental impacts.

Conclusion

This study presents a mechanistic assessment of As(III) removal from groundwater using an iron net–based zero-valent iron (IN-ZVI) system, with particular emphasis on the inhibitory effects of phosphorus (P) and silicon (Si). As(III) removal involved iron corrosion, Fe(II)/Fe(III) redox cycling, oxidation of As(III) to As(V), and subsequent adsorption or co-precipitation with hydrous ferric oxides (HFO). As(III) removal follows first-order kinetics, while Fe(II) dissolution exhibits zero-order kinetics, indicating that arsenic removal is governed by the availability of reactive iron species generated at the iron surface. The presence of P and Si substantially decreased Fe dissolution and arsenic removal performance. Phosphorus exerted a stronger inhibitory effect by competing for adsorption onto Fe(III) surface sites and blocking reactive hydroxyl groups. Silicon reduced system reactivity by promoting the formation of Si-enriched passive layers, although its impact was comparatively weaker. Adsorption isotherm analysis confirmed that increasing concentrations of P and Si reduced the adsorption capacity of the IN-ZVI system. The Langmuir model suggested dominant monolayer adsorption, while the Freundlich model indicated increased surface heterogeneity under Si-rich conditions. XAFS results confirmed the complete oxidation of As(III) to As(V), highlighting the role of corrosion-induced reactive oxygen species in arsenic oxidation. XPS analysis further revealed extensive surface passivation of the iron net by Fe(III) oxides and P- and Si-bearing scales, explaining the observed decline in long-term reactivity. Overall, P and Si impair arsenic removal through competitive adsorption, surface site blockage, and inhibition of iron dissolution. Despite these challenges, the IN-ZVI system remains a cost-effective option for arsenic mitigation, particularly in resource-limited settings. Periodic surface regeneration of the iron net is recommended to sustain long-term performance. Future studies should incorporate combined P–Si experiments together with detailed Fe speciation and time-resolved As(III)/As(V) speciation analyses in both aqueous and solid phases under varying phosphate and silicate conditions. Additional characterization techniques, including SEM, ATR-FTIR, and advanced XPS deconvolution analyses, are also recommended to further elucidate the coupled oxidation, adsorption, and surface passivation mechanisms governing iron-based groundwater treatment systems.

Author contributions

Md. Shafiquzzaman: Conceptualization, Methodology, Investigation, Formal analysis, Supervision, Writing—original draft. Mohammad Shahedur Rahman: Formal analysis, visualization, Writing—original draft; Writing—review & editing. Amimul Ahsan: Writing—review & editing, Abdelkader T. Ahmed: Data curation, Writing—review & editing.

Funding

This work was supported and funded by the Deanship of Scientific Research at Imam Mohammad Ibn Saud Islamic University (IMSIU) (grant number IMSIU-DDRSP2602).

Data availability

The datasets used and/or analysed during the current study available from the corresponding author on reasonable request.

Declarations

Competing interests

The authors declare no conflict of interest.

Ethical approval

All authors have read, understood, and have complied as applicable with the statement on "Ethical responsibilities of Authors" as found in the Instructions for Authors.