Introduction

The removal of chemical pollutants in water has always been an important issue in the process of production and development of human society. Among them, the method of using oxidation to destroy organic matter to reduce its harm has taken root in actual water treatment applications1. In recent years, the research on the removal of pollutants by polymerity-oriented oxidation has attracted wide attention, and this new direction seems to shake the previous mainstream views on pollutant mineralization2,3,4,5,6. Compared with the traditional mineralization pathway, the polymerization pathway cannot only reduce carbon emissions but also greatly reduce the input of chemical substances and environmental risks while ensuring the effective removal of pollutants. Furthermore, polymeric byproducts often exhibit heightened reactivity and enhanced adsorption affinity, making them more readily removed during conventional oxidative or adsorptive treatment processes. Therefore, the development of new catalytic methods to regulate the polymerization pathways for enhancing oxidant-driven pollutant removal efficiency in the traditional oxidation process is an urgent need to further promote green and resource-based water treatment.

Ferrate(VI) is widely regarded as a green water treatment agent with high efficacy in degrading phenolic contaminants6,7,8,9,10,11,12. Research has indicated that polymerization represents a key pathway in ferrate(VI)-driven pollutant oxidation. For instance, phenoxy radicals generated via single-electron transfer during phenol oxidation by ferrate(VI) undergo coupling reactions to yield polymeric products. However, the oxidative capacity of ferrate(VI) in aqueous systems can be compromised by ligand interactions with background substances. In phosphate-containing water, for example, phosphate ions form complexes with Fe(V)—a highly reactive intermediate derived from Fe(VI)—through coordination to its central iron atom, resulting in octahedral coordination structures. This process reduces the redox potential of Fe(V) and introduces steric hindrance, thereby diminishing oxidation efficiency12. Pyrophosphate, with stronger chelating ability, similarly suppresses ferrate(VI) reactivity12. Notably, ligand interactions do not universally hinder oxidation. In KMnO4-based systems (sharing oxidative similarities with ferrate(VI)), ligands such as phosphate, pyrophosphate, nitrilotriacetic acid, and humic acid significantly improve the removal of triclosan, bisphenol A (BPA), phenol, and 17β-estradiol by stabilizing in situ-generated Mn(III) species13,14,15,16,17,18. Conversely, these ligands show negligible effects on carbamazepine and trimethylsilyl oxide degradation, suggesting ligand-mediated effects are pollutant-specific18.

Harnessing beneficial ligand interactions offers a strategic avenue to optimize ferrate(VI) performance. Recent studies reveal that specific catalysts enhance direct electron transfer between ferrate(VI) and pollutants through complexation-driven mechanisms19,20,21,22. Carbon quantum dots, for instance, form surface complexes with Fe(VI), enabling electron shuttling that boosts degradation of ibuprofen, diclofenac, sulfamethoxazole, and phenols19. Similar enhancement via homogeneous catalysis is observed in ammonium-containing systems: NH3/NH4+ coordinates with transient iron species during ferrate(VI) oxidation of carbamazepine, naproxen, trimethoprim, sulfamethoxazole, and flumequine, generating reactive complexes that elevate removal rates and kinetic constants23,24. However, ligand effects may exhibit substrate specificity. Carbonate ions, for example, selectively enhance the removal of p-toluidine and sulfonamide antibiotics with aromatic amine groups by ferrate(VI) by stabilizing Fe(V), which has high reactivity toward primary amines24.

This study employed EDTA, a chelator with dual functional advantages in complexation capacity and environmental compatibility, to modulate the ferrate(VI) coordination environment, systematically investigating its pollutant removal enhancement mechanisms. Comparative performance evaluations in benchmark matrices and authentic water samples demonstrated EDTA’s catalytic synergism with ferrate(VI) for phenol degradation across multiple operational parameters. Through comparative quenching tests and electron paramagnetic resonance analysis, we mechanistically elucidated the contribution of reactive oxygen species (radicals and Fe(V)/Fe(IV)) in the optimized ferrate(VI)/EDTA system. Furthermore, spectral characterization revealed EDTA’s structural modulation of ferrate(VI). Density functional theory (DFT) calculations provided atomic-level theoretical validation, confirming the thermodynamic feasibility and electron transfer mechanisms underlying this coordination-enhanced oxidation process.

Results

Degradation of pollutants in ferrate(VI)/EDTA system

Three primary EDTA concentration gradients (no EDTA, trace, and excess) were employed to investigate their effects on the pollutant degradation by ferrate(VI). As shown in Figs. 1A and S1, with the introduction of trace EDTA, both the phenol removal efficiency by ferrate(VI) and the second-order rate constant showed significant enhancement, while EDTA alone exhibited negligible contribution to phenol removal under identical conditions. When EDTA concentrations were elevated to 50 μM and 100 μM, phenol removal rates exceeded 99%, accompanied by 3.8-fold and 4.1-fold increases in rate constants, respectively. However, the phenol removal rate decreased to 77.4% under the excess EDTA (10 mM) condition, likely limited by the competitive consumption of oxidants by EDTA. The EDTA-mediated enhancement also extended to other substrates: trace EDTA addition boosted the removal of 4-chlorophenol (4-ClP), 4-tert-butylphenol (4-tBP), acetaminophen (AAP), and bisphenol A (BPA) by up to 15%. The rate constants showed 1.8-fold, 1.7-fold, and 1.2-fold increases for 4-ClP, 4-tBP, and AAP with ferrate(VI), respectively (Fig. S2). These observations imply potential novel reaction pathways and mechanisms in the ferrate(VI)-EDTA system.

Fig. 1: Effect of EDTA on ferrate(VI) removing pollutants.
figure 1

A Effect of EDTA on ferrate(VI) removing pollutants, B phenol removal at different pH, and C, D the reaction ratio (∆[phenol]/∆[K2FeO4]). Initial concentrations of reactants: [K2FeO4]0 = 50 μM; [Phenol]0 = 5 μM in (A, B), 20 μM in (C and D); [4-ClP]0 = 10 μM; [AAP]0 = 10 μM ([Fe(VI)]0 = 20 μM); [4-tBP]0 = 20 μM; [BPA]0 = 20 μM. pH = 8.0 in (A–D). Phosphates were used for the buffer solutions of pH 5.0 and 6.0, and borate for that of other pH.

Full size image

Improved substrate removal reflects an optimized stoichiometric ratio between the substrate and oxidant. As depicted in Fig. 1C, D, the molar reaction ratio between phenol and ferrate(VI) was 0.17:1 (∆[phenol]/∆[K2FeO4]). Pre-addition of 50 μM and 100 μM EDTA increased this ratio to 0.20:1 and 0.22:1, respectively. However, further elevating EDTA concentrations to 500 μM and 1 mM did not proportionally enhance the ratio, which stabilized at 0.22:1 and 0.21:1. Notably, at excess EDTA conditions, the sharply declined ratio reveals that excessive EDTA competes with phenol for ferrate(VI) reactivity, ultimately surpassing the beneficial effects. Consequently, optimal enhancement was attained at a 1:2 molar ratio of ferrate(VI) to EDTA.

Furthermore, within the pH range typical of actual water treatment conditions (pH 5.0–9.0), the addition of EDTA significantly enhanced phenol removal efficiency by ferrate(VI) (Fig. 1B). For instance, 20 μM EDTA improved phenol removal by approximately 1.4-fold under pH 5.0–7.0. With EDTA dosage increased to 50 μM, the enhancement factor rose to ~2.3-fold at pH 5.0–6.0. In contrast, the optimization effect of EDTA under alkaline conditions (pH 8.0–9.0) was less pronounced, achieving an ~1.2-fold improvement. This discrepancy may arise from the strong pH-dependent self-decay behavior of Fe(VI). Under acidic conditions, Fe(VI) undergoes rapid self-decay25, resulting in substantially lower phenol removal compared to alkaline conditions, where Fe(VI) is more stable. However, EDTA improved the utilization efficiency of Fe(VI) by phenol, thereby leading to a more pronounced enhancement in phenol removal at lower pH values.

Polymeric products in EDTA-regulated coordination environment

The polymerization of phenoxy groups has been proved as an important pathway of Fe(VI) oxidizing phenol6,26,27. As shown in Fig. S4A, [1,1’-biphenyl]-2,4’-diol, the polymerization product of two phenoxy groups, started to generate in about 5 min during Fe(VI) oxidizing phenol. With 50 μM EDTA, [1,1’-biphenyl]-2,4’-diol generated in 30 s of the reaction, which is 4.5 min earlier than that of Fe(VI) alone. The concentration reached a maximum at 10 min and was 1.7 times that of Fe(VI) alone. As the dosage of EDTA increased to 500 μM and 1 mM, the generation of [1,1’-biphenyl]-2,4’-diol also increased, with the maximum concentrations being 4.3 and 4.9 times that without EDTA, respectively. Furthermore, EDTA promoted 2,2’-biphenol formation (Fig. S4B), whereas Fe(VI) alone failed to detect this coupling product. Higher EDTA concentrations not only induced its generation but also accelerated its appearance kinetics. Under the influence of EDTA, the dimeric product of acetaminophen also showed an increasing trend (Fig. S4C). Greater quantities and broader diversity of polymerization products caused by EDTA initially proved its direction to the polymerization pathway, which was further explained in subsequent studies.

Preliminary exploration of the reaction mechanism

To elucidate the intervention mechanisms of complexation environments on the reaction pathways between ferrate(VI) and pollutants, we first investigated the contribution characteristics of active species (e.g., free radicals, intermediate iron species) during the reactions, whose involvement has been reported in ferrate(VI)-driven systems in previous studies28,29,30,31,32,33. Given this mechanistic context, the present investigation systematically evaluated whether these reactive intermediates dominate the enhanced oxidation observed in the ferrate(VI)/EDTA system.

In ferrate(VI) oxidation systems, reactive oxygen species such as hydroxyl radicals (•OH), superoxide radicals (•O2−), and singlet oxygen (1O2) have been extensively reported as potential contributors to oxidation processes19,32,33. To assess their roles in the ferrate(VI)/EDTA system, quenching experiments were conducted using methanol (MeOH) and tert-butanol (TBA) as •OH scavengers (Fig. 2A)33. MeOH (50 mM) reduced phenol degradation by 11% in ferrate(VI)-only systems and 17% in ferrate(VI)/EDTA systems compared to controls. Notably, the EDTA-containing system retained a 10% efficiency advantage over the EDTA-free system despite inhibition. In contrast, TBA exhibited negligible influence on phenol removal. These results indicate •OH is not the dominant reactive intermediate in this system. The partial suppression observed with MeOH may arise from its competitive reductive interaction with ferrate(VI). This conclusion aligns with EPR analyses, which detected no DMPO-•OH adduct signals (generated through the reaction of 5,5-dimethyl-1-pyrroline-N-oxide with •OH) in either ferrate(VI)-only or ferrate(VI)/EDTA systems (Fig. 2B)33.

Fig. 2: Responses of different probes in the K2FeO4/EDTA system.
figure 2

A Degradation of 5 μM phenol in 50 μM K2FeO4/50 μM EDTA system, [MeOH]0 = [TBA]0 = 50 mM, [FFA]0 = 2 mM, [CCl4]0 = 10 mM; B EPR spectrum of DMPO in K2FeO4/EDTA system, [Phenol]0 = 50 μM, [K2FeO4]0 = [EDTA]0 = 100 μM, [DMPO]0 = 50 mM; C the transformation of 10 μM PMSO in 50 μM Fe(VI) and 50 μM K2FeO4/50 μM EDTA; D the effect of EDTA on the oxidation of 100 μM ABTS by K2FeO4. pH = 8.0.

Full size image

To probe the potential involvement of 1O2, 2 mM furfuryl alcohol (FFA) was used as a quencher19. As shown in Fig. 2A, FFA competitively consumed Fe(VI), reducing phenol degradation from 84% to 22.4%. However, pre-addition of EDTA partially counteracted this inhibition, restoring degradation efficiency to 47.1%. The persistence of EDTA’s enhancement effect under 1O2-quenching conditions implies that 1O2 does not constitute the critical reactive species facilitating improved oxidation. This conclusion was confirmed by EPR analysis: no TEMP-1O2 adduct signals (resulting from 2,2,6,6-tetramethyl-4-piperidinol (TEMP) trapping of 1O2) were detectable in either ferrate(VI) or ferrate(VI)/EDTA systems during phenol oxidation (Fig. S6)19,33.

Complementary experiments employing carbon tetrachloride (CCl4) as an •O2− scavenger revealed no significant alterations in phenol degradation efficiency across all tested conditions (Fig. 2A)34. Combined radical scavenger, spectroscopic, and efficiency comparison data confirm EDTA does not enhance inorganic radical pathways during ferrate(VI)-driven phenol oxidation. Instead, the observed enhancement likely arises from non-radical mechanisms requiring further study.

During self-decay or electron transfer processes, ferrate(VI) typically produces highly reactive intermediate iron species (Fe(V)/Fe(IV)) with elevated oxidative potential29,30,31. Consequently, phosphate ligands, known to complex and deactivate these transient iron species, often diminishes the contaminant removal by ferrate(VI)12. Intriguingly, in the current study, EDTA-induced enhancement persisted even under phosphate-buffered conditions at pH 5.0 and 6.0 (Fig. 1B). At pH 8.0, comparative experiments revealed analogous EDTA-mediated improvement magnitudes in both phosphate (Fig. S5) and borate buffer systems (Fig. 1B), suggesting EDTA’s mechanism may operate independently of Fe(V)/Fe(IV) modulation. To further evaluate the potential involvement of Fe(V)/Fe(IV) in ferrate(VI)/EDTA system, PMSO and ABTS were selected as probe substrates.

As a diagnostic substrate for Fe(V)/Fe(IV) detection, PMSO exhibits limited reactivity with Fe(VI) but converts to PMSO2 via oxygen atom transfer when exposed to Fe(V)/Fe(IV)35,36. Experimental data (Fig. 2C) demonstrate that 50 μM K2FeO4 fully oxidized 10 μM PMSO to 10 μM PMSO2 within 30 min, while 50 μM EDTA reduced PMSO conversion by 2.6 μM. This suggests EDTA neither promotes Fe(V)/Fe(IV) generation nor their oxidative capacity. Notably, the Fe(VI)/EDTA system still achieved full conversion of 7.4 μM PMSO to PMSO2 (100% efficiency), which ruled out radical-mediated pathways again.

Different from PMSO, ABTS functions as a single-electron transfer reagent, reacting with ferrate(VI) to produce ABTS•+ while sequentially reducing Fe(VI) to Fe(III) via Fe(V)/Fe(IV) intermediates25. Theoretically, stoichiometry predicts a 1:3 Fe(VI):ABTS•+ ratio, but experiments in 10 mM borate buffer yielded a 1:1.9 ratio (Fig. 2D), attributed to rapid self-decay of Fe(V)/Fe(IV) to Fe(III). Crucially, EDTA addition did not alter this ratio, demonstrating no increased contribution from Fe(V)/Fe(IV) species. Combined with the PMSO quenching results, these data collectively demonstrate that Fe(V)/Fe(IV) intermediates do not mediate the enhanced oxidative capacity in the ferrate(VI)/EDTA system, with Fe(VI) remaining the dominant active oxidant.

EDTA exhibits strong chelation capability toward metal ions (e.g., Mg(II) and Fe(III)) in aqueous systems, while Fe(VI) species may also interact with coexisting anions via coordination12. Therefore, it is necessary to investigate the complexation role of EDTA in this system. To this end, we designed different sequences of reactant addition and compared their effects on phenol removal. As shown in Fig. 3A, two reaction sequences were implemented: the conventional method of mixing EDTA with phenol first, followed by adding the oxidant (Fe(VI)) to initiate the reaction, and pre-mixing EDTA with Fe(VI) for 30 seconds before introducing phenol to start the reaction. Both approaches resulted in the same magnitude of improvement in phenol removal compared to using Fe(VI) alone. However, the pre-mixing of EDTA and Fe(VI) exhibited faster reaction kinetics and an earlier termination of the reaction, indicating the formation of active species during the pre-mixing stage. This suggests that the reactive species responsible for the improving process are derivatives generated from the interaction between Fe(VI) and EDTA. To further clarify the EDTA-specific effects in the K2FeO4/EDTA system, competitive metal ion experiments were conducted. As shown in Fig. 3B, incremental Mg(II) concentrations (20–200 μM) progressively reduced phenol degradation efficiency by 8%, 10%, and 19% in the Fe(VI)/EDTA system, while leaving Fe(VI)-only performance unaffected. This implies that Mg(II) occupies EDTA coordination sites, thereby diminishing its reactivity-enhancing capacity. At 200 μM Mg(II), excessive complexation formed inactive Mg-EDTA species, resulting in their negative effects. Moreover, kinetic analysis at pH 8.0 revealed that Mg(II) introduction significantly suppressed the reaction rate constant between Fe(VI) and EDTA from 5.38 M−1 s−1 to 1.2 M−1 s−1 (Fig. 3C), providing direct evidence of competitive binding at EDTA coordination sites. Analogous suppression was observed with Fe(III) substitution (50 μM Fe(III) caused 5% efficiency loss), confirming that EDTA’s enhancement originates from its Fe(VI)-specific chelation. This additionally excludes Fe(III)‘s role in system optimization, despite prior reports suggesting its potential catalytic contributions to Fe(VI) activation35.

Fig. 3: Investigation on the active complex of Fe(VI)-EDTA.
figure 3

A, B The removal of phenol by Fe(VI)/EDTA, C the rate constants of Fe(VI) reacting with EDTA, and D the in-situ Raman spectra of the Fe(VI)/EDTA system. [Fe(VI)]0 = [EDTA]0 = 50 μM in (B); [Fe(VI)]0 = 50 μM, [Mg(II)]0 = [EDTA]0 in (C); [Fe(VI)]0 = [EDTA]0 = 5 mM in (D). [Phenol]0 = 5 μM. Borate buffer solution of pH 8.0.

Full size image

Therefore, the resultant Fe(VI)-EDTA coordination complex emerges as the dominant reactive entity driving pollutant degradation, with its formation kinetics directly modulated by competing metal ions. This conclusion is corroborated by the in situ monitored Raman spectroscopy. As illustrated in Fig. 3D, a characteristic peak of EDTA was observed at 880 cm−1. This peak disappeared when EDTA coexisted with Fe(VI), and a new characteristic peak emerged at 750 cm−1, which was absent in the spectrum of Fe(VI) alone. The emergence of the new spectral peak indicates that EDTA induces the formation of new species in the Fe(VI) system, providing direct experimental evidence supporting the prior inference. Given that previous discussions have established the decisive role of EDTA complexation with Fe(VI) in this system, and other potential influencing factors have been ruled out, it is reasonable to conclude that the new peak is associated with complexation derivatives formed between the two. These findings collectively establish that the performance enhancement of EDTA stems fundamentally from its molecular-level coordination chemistry with Fe(VI), overriding potential contributions from indirect redox cycles involving radicals or intermediate iron intermediates.

Regulation of EDTA in the polymerization pathway

The oxidation of phenolic compounds by Fe(VI) predominantly proceeds through electrophilic attack on aromatic ring active sites, as established in prior studies10,26,27. For phenol, the para-carbon position exhibits maximal susceptibility to such attack due to its elevated electron density6. Mechanistic analysis reveals that Fe(VI) initially engages the phenol aromatic ring (Fig. 4), facilitating sequential electron and hydrogen transfer to generate phenoxy radicals that couple into polymeric derivatives. Accelerated intermediate generation and elevated product concentrations under EDTA supplementation (Fig. S4) indicate EDTA enhances Fe(VI)’s electrophilic reactivity at critical aromatic sites, explaining both accelerated kinetics and polymer accumulation. Moreover, uncharacterized spectral features were detected in EPR spectrum of DMPO in K2FeO4/EDTA systems (Fig. 2B), consistent with DMPO-trapped phenoxy radicals (DMPO-R)7, further validate phenoxyl radical involvement in this pathway.

Fig. 4: Products and reaction pathways.
figure 4

Reaction pathways of phenol in Fe(VI)/EDTA system.

Full size image

EDTA introduction induced negligible changes to the second-order rate constants of Fe(VI) reacting with 2,2’-biphenol and [1,1’-biphenyl]-2,4’-diol (Fig. S9), confirming EDTA selectively enhances coupling product formation without altering degradation kinetics, which results in the net accumulation of coupling products. Notably, the polymerization byproducts in this system do not exhibit sustained accumulation over prolonged reaction time, indicating that their buildup is merely a transient intermediate phase, as they are subsequently degraded by Fe(VI). In fact, the concentrations of these polymeric byproducts in the Fe(VI)/EDTA system remain extremely low. For example, the peak concentrations of 2,2’-biphenol (0.0091 μM) and [1,1’-biphenyl]-2,4’-diol (0.0221 μM) were over three orders of magnitude lower than the initial concentration of their parent compound, phenol (20 μM). Given both the trace levels of these byproducts and their enhanced removability in subsequent conventional adsorption processes, the system presents minimal secondary risks.

The inefficacy of EDTA on coupling phenols may arise from substituents occupying active sites on the aromatic ring and steric hindrance generated by bulky groups, and alterations in charge distribution may also modulate nucleophilicity. For example, the dimerization of phenol decreases the Fukui function values (f-, reflecting electronic supply capacity) at active sites by approximately 50%, substantially weakening their electron-donating capacity and consequently diminishing the attack propensity of Fe(VI)-EDTA6. Moreover, the ineffectiveness of EDTA in Fe(VI) oxidizing hydroquinone, which possesses an additional para-hydroxyl group compared to phenol, confirms the para-C atom as the primary attack site for Fe(VI)-EDTA complexes (Fig. S9C). The hydroxyl substituents likely sterically hinder Fe(VI)-EDTA’s access to the aromatic ring while simultaneously redistributing the para-C atom’s electron density, thereby potentially resulting in the ineffectiveness of EDTA in these systems6,37. Furthermore, as shown in Fig. S2, EDTA elevated the rate constant of Fe(VI) reacting with 4-tert-butylphenol and acetaminophen by 1.66 times and 1.21 times, respectively, lower than that with phenol (3.8 times). This attenuation correlates with structural features of para-substituents: the bulky tert-butyl group and acetyl amino group likely impose steric hindrance to Fe(VI) attack on aromatic reactive sites while reducing the electronegativity of the carbon atom38, which disfavors Fe(VI) interaction. In contrast, chlorine substitution in 4-chlorophenol preserves para-C atom electronegativity39, enabling a moderate 1.84-fold EDTA-induced rate constant increase, higher than 4-tert-butylphenol and acetaminophen. Nevertheless, this enhancement remains inferior to phenol due to partial occupation of reactive sites by chlorine atoms, highlighting the dual role of para-substituents in governing both electronic and steric factors during Fe(VI)-EDTA mediated oxidation processes.

Considering the addition attack pathway of Fe(VI) on the benzene ring of phenol, it can be inferred that different functional groups on the saturation, steric hindrance, and electronegativity of the active sites on the benzene ring may affect the enhancement of EDTA on phenolic compounds removing.

Thermodynamic regulation of EDTA in reaction pathways

DFT calculations were performed at the B3LYP/6-31 G level to validate the EDTA-mediated enhancement mechanism in Fe(VI)-phenol oxidation. Under pH 8.0 conditions, EDTA primarily exists in the triply deprotonated ionic state (EDTA3-) (Fig. S10), while HFeO4− and FeO42− constitute the dominant Fe(VI) species. Given the higher oxidative capacity of HFeO4− relative to FeO42−, the computational modeling specifically addressed complexation interactions between HFeO4− and EDTA3-, consistent with established mechanistic frameworks40.

The structural configuration of EDTA3-, as illustrated in Fig. 5A, features two centrally positioned nitrogen atoms (N1, N2) and four carboxyl oxygen atoms (O1–O4). Charge distribution analysis across three computational methods (CHELPG, MK, RESP) in Table S5 reveals that carboxyl oxygen atoms in EDTA3− maintain greater negative charge density compared to nitrogen atoms, with deprotonated oxygen atoms (O1-O3) exhibiting higher electronegativity than their protonated counterpart (O4). This charge gradient dictates the preferential interaction of HFeO4− with O1, the most negatively charged site, during oxidative attack. To elucidate complexation pathways, two distinct binding modes between EDTA3− and HFeO4− were computationally evaluated through Gibbs free energy (∆G) calculations. The first mode involves monodentate coordination, where EDTA3− approaches the central iron atom of HFeO4− from a single orientation, while the second mode proposes bidentate interaction through dual-directional binding. Thermodynamic calculations demonstrate spontaneity for the monodentate pathway (ΔG = -26.524 kJ/mol), confirming Fe(VI)-EDTA complex formation as the dominant process over the energetically unfavorable EDTA-Fe(VI)-EDTA configuration. These computational insights mechanistically validate the formation of Fe(VI)-EDTA complexes under experimental conditions. Moreover, this regulation induced a substantial modulation of the energy alignment between HOMO and LUMO in the molecular cluster (Fig. 5B), manifesting as a marked downshift of the HOMO energy level (from −0.302 Ha to −0.346 Ha) concurrent with an upshift of the LUMO energy level (from −0.288 Ha to −0.225 Ha). The increased energy gap (ΔE) explicitly demonstrates that the EDTA ligand synergistically suppresses electron detachment propensity through coordinated optimization of spin states and electronic energy levels. This electronic modulation drives preferential adoption of electron capture pathways during oxidation processes, thereby significantly enhancing oxidative capacity. In addition, the complexation of EDTA prolonged the bond length of Fe-O from 1.611 A0 to 1.632 A0, which might cause Fe(VI) to exhibit a stronger oxidizing ability41.

Fig. 5: DFT conclusions.
figure 5

A Gibbs free energy changes for the complex reactions of Fe(VI) and EDTA. B LUMO−HOMO energy gaps of the Fe-centered complex in the absence and presence of EDTA. C Calculated potential reaction barriers and Gibbs free energy changes for the reactions of Fe(VI) and Fe(VI)-EDTA with phenol at the B3LYP/6-31 G level.

Full size image

The para-carbon atom in phenol demonstrates heightened susceptibility to electrophilic reagent attack owing to its pronounced electronegativity, as established in prior investigations6. Regarding the phenoxylation pathway (Fig. 5C), the interaction between Fe(VI) and phenol initiates with oxygen atom coordination from Fe(VI) to the para-carbon atom, forming a transient transition state species (1TS). Concurrently, the hydrogen atom bonded to this carbon undergoes intramolecular transfer to an alternative oxygen atom within the Fe(VI) structure. This transition state process necessitates overcoming an activation energy barrier (ΔG‡) measuring 84.187 kJ/mol. Subsequent cleavage of the carbon-oxygen bond yields one mole of phenoxy radical, accompanied by electron transfer converting Fe(VI) to Fe(V) with concomitant energy release of 93.24 kJ. The overall reaction exhibits a negative Gibbs free energy change (ΔG = −9.053 kJ/mol), thermodynamically confirming spontaneous progression. Comparatively, the Fe(VI)-EDTA system demonstrates enhanced reactivity through reduced activation energy (ΔG‡ = 53.035 kJ/mol) and greater energy release (ΔG = −30.456 kJ/mol), quantitatively verifying EDTA’s catalytic promotion of phenoxylation kinetics. These computational findings align with experimental observations regarding EDTA-mediated product concentration variations detailed in Section 3.3.1.

Performance of Fe(VI)/EDTA applied to real phenol-containing water

Recent research initiatives have extensively explored ferrate(VI) application in wastewater pretreatment or tertiary treatment processes to achieve contaminant deep elimination, particularly targeting phenolic compounds42,43,44. The present investigation reveals that the addition of EDTA substantially improves phenol removal efficiency during ferrate(VI)-mediated wastewater treatment. Analytical measurements recorded phenol concentrations of 1.988 μg/L in municipal wastewater and 0.213 μg/L in secondary effluent (Fig. 6A). Implementation of 20 mg/L K2FeO4 reduced these concentrations to 0.647 μg/L (municipal wastewater) and 0.059 μg/L (secondary effluent). Remarkably, co-administration of 20 mg/L EDTA with equivalent K2FeO4 dosage achieved further reductions to 0.085 μg/L and 0.013 μg/L, respectively, representing removal efficiency improvements of 28.3% and 21.6% compared to ferrate(VI) treatment alone.

Fig. 6: Removal of phenol in municipal wastewater (MW) and secondary effluent (SE).
figure 6

Removal of phenol naturally occurring in actual water (A). Removal of high-concentration phenol in the background of secondary effluent (B) and municipal wastewater (C), [Phenol]0 = 0.5 mg/L and [EDTA]0 = 15 mg/L, reaction time = 60 min.

Full size image

The operational efficacy of the ferrate(VI)/EDTA system for high-concentration phenol removal was systematically evaluated in municipal wastewater and secondary effluent matrices. Background constituents inherent to real water systems are known to compromise ferrate(VI) performance through competitive consumption mechanisms, as documented in existing research11,45. Experimental data demonstrate this interference pattern: 10 mg/L K2FeO4 achieved 78% phenol removal (0.5 mg/L initial concentration) in secondary effluent (Fig. 6B), whereas 20 mg/L K2FeO4 attained merely 8.7% removal efficiency in municipal wastewater (Fig. 6C). The introduction of 15 mg/L EDTA markedly enhanced system performance, elevating removal rates to 85% in secondary effluent under equivalent conditions (Fig. 6B). Furthermore, complete phenol elimination in secondary effluent (below detection limits) was achieved using 15 mg/L K2FeO4 combined with 15 mg/L EDTA, outperforming standalone 30 mg/L K2FeO4 treatment. This is due to that EDTA modulates ferrate(VI) reactivity by enhancing molecular affinity toward phenol, thereby minimizing non-productive oxidant consumption by background substances. The synergistic effect proved particularly pronounced in complex municipal wastewater matrices, where 15 mg/L EDTA increased phenol removal by 20 mg/L K2FeO4 from 8.7% to 30% (Fig. 6C). Dose-response analyses revealed diminishing returns with ferrate(VI) monotherapy, necessitating >140 mg/L K2FeO4 for complete phenol removal. In contrast, the combined system effectively decoupled the non-linear dependence of trace-level pollutant removal efficiency on oxidant dosage in the presence of background impurities, which achieved complete elimination using 60 mg/L K2FeO4 with 15 mg/L EDTA, demonstrating ≥57% oxidant savings. These findings collectively validate the technical superiority of ferrate(VI)-EDTA coordination for phenolic pollutant abatement in real wastewater systems.

Discussion

This study demonstrates that EDTA significantly enhances ferrate(VI) oxidizing pollutants by mediating a complexation-driven catalytic mechanism. The formation of a metastable Fe(VI)-EDTA complex facilitates electron transfer from aromatic substrates, accelerating reaction kinetics and thermodynamically favoring phenolic group activation. Furthermore, EDTA steers the reaction toward polymerization pathway, which enhances the utilization efficiency of Fe(VI) by the substrate, and reduces Fe(VI) self-decay losses. DFT calculations corroborate the spontaneity of Fe(VI)-EDTA complexation and the regulatory role of EDTA in enhancing the thermodynamic favorability of reactions between K2FeO4 and substrates.

In the treatment of industrial and domestic wastewater, the removal of highly toxic organic pollutants is of significant importance. Ferrate(VI) is acknowledged as a green oxidant due to its strong oxidative capacity, minimal byproduct generation, and absence of residual contamination. As for EDTA, as a mature industrial raw material, it is characterized by low toxicity and high biodegradability, and is commonly used as an additive in industries such as food, healthcare, and cosmetics. Therefore, their application in wastewater treatment poses a low risk of inducing secondary environmental issues, demonstrating favorable applicability. By leveraging EDTA coordination to improve molecular recognition between Fe(VI) and phenolic substrates, this approach redirects reaction pathways, significantly reducing unproductive oxidant consumption while decoupling the non-linear dependence of trace-level pollutant removal efficiency on oxidant dosage in the presence of background impurities. This ligand-mediated approach aligns with green chemistry principles through lower chemical consumption in complex matrices, contributing to advancing the application of ferrate(VI) in water treatment. Moreover, the study provides some valuable references for developing ligand-tunable oxidation technologies for resource-efficient contaminant control.

Methods

Chemicals and reagents

All chemicals and reagents applied in this research are detailed in Text S1.

Experimental procedure

Unless otherwise specified, experiments were conducted at 25 ± 1 °C using glassware including beakers, conical flasks, TOC bottles, and other similar vessels.

Degradation studies were performed by introducing reactants and background additives into 100 mL buffered solutions. Freshly prepared ferrate(VI) solution was quantified via spectrophotometric analysis (λ = 510 nm, ε = 1150 M−1 cm−1)46. At predetermined intervals, aliquots were collected and immediately quenched with hydroxylamine hydrochloride (>20-fold molar excess relative to oxidant concentration) prior to residual pollutant quantification by high-performance liquid chromatography (HPLC). Residual ferrate(VI) was neutralized with 2,2′-azino-bis(3-ethylbenzothiazoline-6-sulfonate) (ABTS) to stoichiometrically generate ABTS•+ radicals, followed by spectrophotometric measurement (λ = 415 nm, ε = 34,000 M−1 cm−1)25. Methodology for deriving second-order rate constants is described in Text S3. The second-order rate constants for ferrate(VI)-EDTA interactions were calculated as outlined in Text S3. A stopped-flow spectrophotometer (Applied Photophysics, SX-20) operating at 510 nm was employed to quantify Fe(VI) concentrations, leveraging distinct absorption maxima between Fe(VI) and Fe(V)/Fe(IV) species46,47.

Municipal wastewater and secondary effluent samples were collected from primary and secondary sedimentation tanks, respectively, at a wastewater treatment facility in Harbin. Key water quality parameters are provided in Table S1. For phenol detection and product analysis in real water matrices, reactions were scaled up to 1 L volumes and terminated with hydroxylamine hydrochloride. Aliquots were filtered for LC-MS analysis, while remaining samples underwent solid-phase extraction (SPE) concentration before GC-MS characterization. The SPE method is described in Text S4.

Analytical methods

Pollutant concentrations were analyzed using a HITACHI HPLC system. Phenol and its derivatives were characterized with an Agilent 1260 HPLC coupled to an ABSciex QTrap 5500 MS (HPLC/ESI-QqQMS) and an Agilent 7890A-7000B GC-QQQ system. Operational parameters are specified in Text S5 and Tables S2 and S3. Radical species were identified via electron paramagnetic resonance (EPR, EMX-8/2.7, Bruker, Germany).

DFT calculations

Density functional theory (DFT) simulations were performed using Gaussian 09 software at the B3LYP/6-31 G level. These included the charge distribution analysis of EDTA, the highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO), and the geometry optimization and frequency calculations for reactants, transition states, and products. Atomic charge distributions, critical for evaluating molecular polarity, reactivity, and electrostatic interactions, were derived using electrostatic potential fitting models (CHELP, MK, RESP)41. Dispersion corrections were implemented via the Grimme DFT-D3(BJ) method. Computed atomic charges and optimized molecular coordinates are compiled in Tables S5–S8.