Surface modification of nano-Fe₃O₄ with EDTA and its use in H₂O₂ activation for removing organic pollutants

Abstract

The effect of EDTA on the H₂O₂ activation ability of Fe₃O₄ nanoparticles was investigated for removing organic pollutants. Regular Fe₃O₄ nanoparticles were observed to have moderate catalytic activity, which was not suitable for the degradation of various organic pollutants. The addition of EDTA enhanced the activation of H₂O₂ on the surface of Fe₃O₄ nanoparticles, thereby accelerating the formation of reactive oxygen species and increasing the degradation rates of pentachlorophenol, sulfamonomethoxine, and Rhodamine B by 84.4, 48.3, and 17.5 times, respectively, at pH 5.0 and 40 °C. Based on spectroscopic and density functional theory studies, adsorption mechanisms for H₂O₂ and EDTA on the surface of Fe₃O₄ nanoparticles were proposed. It was clarified that the enhancing effect of EDTA was attributed to an appreciable improvement of Fe³⁺/Fe²⁺ recycling on the surface of Fe₃O₄ nanoparticles, and to the simultaneous degradation of EDTA and target pollutants.

---

Introduction

The activation of hydrogen peroxide (H₂O₂) over heterogeneous catalysts such as Fe-bearing minerals on silica, alumina, and zeolites has become increasingly popular for treating organic pollutants, where the treatment can be conducted at a neutral pH, overcoming the drawbacks of the required acidification and Fe sludge disposal in homogeneous Fenton and Fenton-like processes.^1–4 Recently, it was reported that Fe₃O₄ magnetic nanoparticles (MNPs) could induce the degradation of phenolic compounds and aniline in the presence of H₂O₂.^5,6 However, this system was also found to produce only slow degradation rates for refractory organics such as pentachlorophenol (PCP).^5–7 This motivated us to develop an effective means of improving the catalytic ability of Fe₃O₄ MNPs. We found that the use of ultrasonic irradiation enhanced the degradation of Rhodamine B (RhB) in the H₂O₂–Fe₃O₄ system by 6 times.⁸ We also developed a way of preparing Fe₃O₄ MNPs by using ultrasonic irradiation, and the obtained Fe₃O₄ MNPs were found to increase the degradation rate of RhB by ca. 12 times in comparison with those prepared by conventional reverse co-precipitation.⁹ It was also reported that the appropriate doping of metallic elements (Mn²⁺, Co²⁺, Cr³⁺, V⁵⁺, Ti⁴⁺, Bi³⁺) into Fe₃O₄ MNPs promoted the degradation of organic compounds in the presence of H₂O₂.^10–14

Another possible approach for improving the catalytic ability of Fe₃O₄ MNPs may be the addition of an organic agent with a strong chelating ability. In the preliminary research, we assessed a series of organic chelating compounds and found that EDTA was one of the best candidates. EDTA is often used as a chelating agent in industrial processes such as electroplating processes. However, it may increase the chemical oxygen demands of the associated wastewater. In addition, EDTA may mobilize toxic heavy metals, thereby increasing the environmental risk of heavy metals leaching into groundwater. This has a negative influence on ecosystems and human health.¹⁵ Some efforts have been made to remove EDTA by using advanced oxidation processes.^16,17 On the other hand, EDTA has a low toxicity to aquatic organisms and is very unlikely to accumulate in the food chain.¹⁸ Therefore, studies focused on its usage in improving the rates of reactions in green chemical schemes are increasing. It has been reported that the addition of EDTA is favorable for the degradation of organic pollutants including 4-chlorophenol, PCP, phenol, and malathion over zero valent iron (ZVI).^19–21 Keenan et al. found that EDTA increased the generation of reactive oxidant in O₂-containing solutions of ZVI nanoparticles.²² We also observed that some ferrous chelates such as Fe–EDTA were capable of raising the activation of O₂.²³ In the present work, we studied the role of EDTA in enhancing the catalytic activity of Fe₃O₄ MNPs for H₂O₂ activation by using PCP, sulfamonomethoxine (SMM), and RhB as model pollutants for chlorophenols, antibiotics and dyes, respectively. Spectrographic results and theoretical analysis offered new insights into the interactions between EDTA, H₂O₂ and Fe₃O₄, and the mechanism of the Fe₃O₄-involved heterogeneous catalytic reactions.

Experimental section

Chemical and materials

SMM was obtained from Acros (Belgium). N,N-diethyl-p-phenylenediamine sulfate (DPD) and 5,5-dimethyl-1-pyrroline-N-oxide (DMPO) were purchased from Aldrich. Horseradish peroxidase (HRP, specific activity of 290–330 units mg⁻¹, RZ ≥ 3.1) was obtained from Shanghai Xueman Biologic Engineering Corp. (China). Other reagents were provided by Sinopharm Chemical Reagent Co., Ltd (China). All the chemical reagents were of analytical grade and were used as purchased. EDTA solution was prepared by dissolving Na₂H₂EDTA in distilled water. Fe³⁺ solution (10 mmol L⁻¹) was prepared by dissolving Fe(NO₃)₃·9H₂O in 10 mmol L⁻¹HNO₃. Solution pH was adjusted with dilute aqueous HCl and NaOH solutions (0.1 mol L⁻¹).

Degradation of organic pollutants

Fe₃O₄ MNPs (specific surface area 82.5 m² g⁻¹) were synthesized with ultrasonic-assisted reverse co-precipitation as described previously⁹ and stock dispersions of Fe₃O₄ MNPs (60 mmol L⁻¹) were prepared by dispersing Fe₃O₄ MNPs in distilled water. The Fe₃O₄ MNPs stock dispersion (1.0 mL) was transferred into 50 mL EDTA solution (0.5 mmol L⁻¹) in the presence of the pollutant (0.020 mmol L⁻¹) at pH 5.0, followed by shaking at 150 rpm at 40 °C for 15 min on a constant-temperature shaker (SHZ-82A, Guohuayiqi, China). After the adsorption–desorption equilibrium was attained for 20 min, the concentration of pollutant was measured and taken as the initial concentration (c₀). The pollutant degradation was then initiated by rapidly adding H₂O₂ (5.0 mmol L⁻¹) into the solution. Aliquots (2.0 mL) were sampled at specified time intervals and filtered immediately through a 0.22 μm filter, followed by the analysis of the organic pollutants in the filtrate. As a control experiment, the degradation was also conducted in the homogeneous Fenton system, the Fe₃O₄ MNPs being replaced with a certain amount of Fe³⁺ or Fe²⁺ ions. Each experiment was carried out at least in duplicate.

Analytical methods

The concentration of RhB was measured by measuring its absorbance at 553 nm with a Cary 50 UV-visible spectrophotometer (Varian, USA). By slightly modifying the analytical procedures reported previously,^24–26 the concentrations of PCP, SMM and EDTA were monitored by high performance liquid chromatography on a PU-2089 HPLC (JASCO, Japan), equipped with a C18 ODS column and an UV detector (JASCO UV-2075) (See S1a of ESI† for analysis details). The concentration of dissolved iron was monitored by atomic absorption spectroscopy (AAS, Analyst 300, P.E. Inc.) and the 1,10-phenanthroline method using a Cary 50 UV-visible spectrophotometer (Varian, USA) at 510 nm.⁹ The site density or the concentration of the replaceable surface groups of Fe₃O₄ MNPs was studied by measuring the adsorption of fluoride on the surface of the catalyst according to the described method^7,27 (See S1b of ESI† for analysis details). The concentration of H₂O₂ was estimated by a DPD spectrophotometric method in the presence of HRP to form the colored product DPD˙⁺ with a maximum absorption wavelength at 551 nm.⁸Reactive oxygen species (ROS) were identified by electronic spin resonance (ESR) measurement.⁹ Because of its instability in water solutions, the ESR signal of O₂⁻˙/HO₂˙ was measured in dimethyl sulfoxide.

ATR-FTIR analysis was carried out with a VERTEX 70 Micro Fourier Transform Infrared/Raman spectroscope equipped with a high sensitivity DLATGS detector and a horizontal ATR-FTIR attachment (Brüker, Germany). Raman spectra were obtained with a Thermo Scientific DXR Raman microscope with excitation at 532 nm and a laser power of 0.6 mW. The samples were vacuum dried for 4 h at room temperature before the FTIR and Raman measurements were carried out. The Density Functional Theory (DFT) method was employed to investigate the adsorption process of H₂O₂ on the Fe₃O₄ (111) surface using the Material Studio 4.4 software package.^28,29 The configuration optimization was implemented by the Vosko-Wilk-Nusair (VWN) functional of the local density approximation (LDA) method with double numerical basis sets plus polarization function (DNP).²⁹ More details of the DFT calculations can be found in section S2 of the ESI†.

Results and discussion

Increased catalytic ability of Fe₃O₄ MNPs in the presence of EDTA

Fig. 1 illustrates the H₂O₂ activation ability of Fe₃O₄ MNPs for removing organic pollutants. It reveals that PCP, SMM and RhB can be degraded in the H₂O₂–Fe₃O₄ system. However, the weak H₂O₂-activation ability of Fe₃O₄ MNPs only results in a very limited degradation of the tested pollutants: less than 20% of the added organic pollutant was removed within 120 min (curve 5 in Fig. 1a, and curves 1 and 2 in Fig. 1b). If EDTA (0.5 mmol L⁻¹) is added, the degradation of the organic pollutants is significantly accelerated, and the removal of all the tested organics is more than 90% within 120 min (curve 6 in Fig. 1a, and curves 3 and 4 in Fig. 1b).

Fig. 1 Degradation of organic pollutants (0.02 mmol L⁻¹) at pH 5.0 and 40 °C. (a) RhB in solutions of (1) H₂O₂ + EDTA, (2) EDTA + H₂O₂ + Fe³⁺, (3) EDTA + H₂O₂ + Fe²⁺, (4) Fe₃O₄ + EDTA, (5) H₂O₂ + Fe₃O₄, (6) EDTA + H₂O₂ + Fe₃O₄; (b) SMM (1, 3) and PCP (2, 4) in solutions of (1, 2) H₂O₂ + Fe₃O₄, and (3, 4) EDTA + H₂O₂ + Fe₃O₄; (c) EDTA in solutions of (1) Fe₃O₄, (2) H₂O₂, (3) RhB + H₂O₂ + Fe₃O₄, and (4) H₂O₂ + Fe₃O₄. Other conditions were 1.2 mmol L⁻¹Fe₃O₄, 5 mmol L⁻¹H₂O₂, 0.5 mmol L⁻¹EDTA, and 0.20 mmol L⁻¹Fe²⁺ or Fe³⁺.

The significantly enhanced pollutant degradation by EDTA was accompanied by the degradation of EDTA itself (curves 3 and 4 in Fig. 1c). When only EDTA was added to the Fe₃O₄ MNPs and H₂O₂ system, it was almost completely degraded within 80 min (curve 4 in Fig. 1c), indicating that Fe₃O₄ MNPs have a high catalytic activity for the oxidation of EDTA in the presence of H₂O₂. Because the control experiments confirmed that the degradation of target pollutants (PCP, SMM, and RhB) and EDTA was comparably negligible in the systems of EDTA–H₂O₂ and EDTA–Fe₃O₄, we concluded that EDTA promotes the H₂O₂-activation ability of Fe₃O₄ MNPs for the degradation of organics, and the very fast degradation of the organics is attributed to the intricate interaction among EDTA, H₂O₂ and nano-Fe₃O₄.

The degradation of each pollutant within 120 min was observed to follow approximately a pseudo-first-order reaction in kinetics. The apparent rate constant k (min⁻¹) can be estimated by the slope of the ln(c_t/c₀) against time curve. All the target pollutants showed a small k value (3.2 × 10⁻⁴–1.2 × 10⁻³ min⁻¹) in the H₂O₂–Fe₃O₄ system. The addition of EDTA greatly raisedthe value of k to 0.027, 0.029, and 0.021 min⁻¹, for PCP, SMM, and RhB, respectively, being ca. 84.4, 48.3, and 17.5 times that of the corresponding H₂O₂–Fe₃O₄ system. Meanwhile, due to the competition between the target pollutants and EDTA for consuming ROS, the degradation of EDTA itself was depressed by the target pollutants. For example, k_EDTA decreased from 0.028 min⁻¹ in the absence of RhB to 0.011 min⁻¹ in the presence of RhB, although the added EDTA could be completely removed in the latter by prolonging the reaction time to 180 min (Fig. S2 in ESI†). The significant enhancement of the catalytic ability of Fe₃O₄ MNPs by EDTA can provide a novel highly-efficient way of removing refractory toxic organics.

The possible contribution of dissolved iron is minor

Metal-chelating agent may increase the dissolution of iron oxides, inducing homogeneous Fenton reactions. For example, Xue et al. reported that EDTA and oxalate increased the dissolution of iron oxide, and the resultant homogeneous Fenton reaction was mainly responsible for the improved degradation of PCP by micro-sized Fe₃O₄ and H₂O₂.³⁰ And Liet al. reported that oxalate improved the catalytic ability of the photo-Fenton system with the porous iron catalysts (FeOOH and Fe₂O₃), and that the Fe-oxalate complex both in the solution and at the catalyst surface made contributions to the degradation of the organic compounds.³¹ However, Ferraz et al., according to the DFT calculations and empirical evidence, pointed out that the enhancement effect of the oxidative power of the Fenton-like system by HCOOH was mainly dependant on the interaction between HCOOH and H₂O₂ over the iron oxide surface to form a stable H-bond linked system and to obtain the H₂O₂//HCOOH pair, which could accept a single electron leading to the formation of ˙OH more strongly than H₂O₂ alone.³² Thus, in the present work it was necessary to discuss the dissolved iron and the surface effect of Fe₃O₄ MNPs on the degradation of the organic compounds. Here, the RhB–H₂O₂–EDTA–Fe₃O₄ system was chosen as the object of study. The total dissolved iron amount in this degradation system gradually increased with time (Fig. S3 in ESI†), it reached 0.20 mmol L⁻¹ after a 2 h reaction period, corresponding to the dissolution of 5.6% of the added Fe₃O₄, and the Fe²⁺ ions were not detectable due to the presence of H₂O₂. When either Fe³⁺ or Fe²⁺ ions at such a concentration (0.20 mmol L⁻¹) were added to the reaction solution instead of Fe₃O₄ MNPs, then the rate of RhB degradation was much slower (curves 2 and 3 in Fig. 1a). This indicates that the leached iron ions make, if any, a minor contribution to the enhancing effect of EDTA on H₂O₂ activation by Fe₃O₄ MNPs, and the activation of H₂O₂ is induced predominately by heterogeneous catalysis on the Fe₃O₄ surface modified in situ with EDTA.

The mechanism for the enhancing effect of EDTA in the nano-Fe₃O₄ system is different from that in the micro-sized Fe₃O₄ system. In the heterogeneous catalyst system, the decomposition of H₂O₂ is significantly dependent on the number of surface reaction sites (catalytic sites) on the catalyst surface.^33,34 We measured the site density or the concentration of the replaceable surface groups of Fe₃O₄ by fluoride adsorption according to the method of Xue et al.⁷ As shown in Table 1, the available site density at the catalyst surface was 5.2 and 3.6 μmol m⁻² for the nano-Fe₃O₄ in the present work and the micro-Fe₃O₄ in the work of Xue et al. However, because of the much different specific surface areas, the total site density of nano-Fe₃O₄ was 49.7 times that of the micro-sized Fe₃O₄. This indicated that the larger surface area of Fe₃O₄ MNPs is more beneficial for H₂O₂ decomposition than that of micro-sized Fe₃O₄. The micro-sized Fe₃O₄ has limited active sites so that its surface could not make a great contribution to the catalytic activation of H₂O₂, and hence the enhancing effect of EDTA is attributed to the homogeneous actions of the dissolved iron as reported by Xue et al. In contrast, H₂O₂ activation with Fe₃O₄ MNPs is predominately by heterogeneous catalysis on the Fe₃O₄ surface modified with EDTA, but not by the dissolved iron as demonstrated in Fig. 1a.

Table 1 Physicochemical properties of nano- and micro-Fe₃O₄

Catalyst	Specific surface area (m^2 g^−1)	Site density (μmol m^−2)	Total site density (μmol g^−1)	R ^b
a The data are cited from Ref. 30 b R is the ratio of T1/T2.
Nano-Fe3O4	82.5	5.2	429 (T1)	49.7
Micro-Fe3O4	2.4^a	3.6^a	8.64 (T2)

---

Effects of EDTA, pH and temperature

The adsorption of EDTA on the surface of the catalyst is a prerequisite for its enhancing effect. When the initial concentration of EDTA increased from 0 to 0.5 mmol L⁻¹, the amount of EDTA (q_e) adsorbed on the surface of Fe₃O₄ MNPs rapidly increased from 0 to 21.5 mmol mol⁻¹ (Fig. 2a), indicating surface modification of Fe₃O₄ by EDTA. Because of the surface modification, an increase in the initial concentration of EDTA from 0 to 0.5 mmol L⁻¹ results in a 17.5 fold increase in k_RhB (from 0.0012 to 0.021 min⁻¹). A further increase in EDTA concentration from 0.5 to 1.0 mmol L⁻¹ results in little increase in the value of q_e, demonstrating that the adsorption of EDTA on the surface of the Fe₃O₄ MNPs roughly achieved saturation at 0.5 mmol L⁻¹EDTA. At the same time, a further increase in the initial concentration of EDTA to 1.0 mmol L⁻¹ resulted in a drop in k_RhB to 0.010 min⁻¹ (Fig. 2a). There are two reasons for the EDTA dependence of k_RhB. Firstly, EDTA favors H₂O₂ activation and the generation of ROS, promoting RhB degradation. Secondly, EDTA is also degraded by ROS, leading to a partial inhibition of RhB degradation. The amount of H₂O₂ consumed in each case was roughly the same (ca. 0.08 mmol) within 2 h of reaction, but the amount of EDTA consumed in the case of an initial EDTA concentration of 1.0 mmol L⁻¹ was measured to be 0.032 mmol, which was much greater than that (0.019 mmol) obtained with an initial EDTA concentration of 0.5 mmol L⁻¹ (Fig. S4 in ESI†). Therefore, the introduction of an excessive amount of EDTA results in a decrease in k_RhB.

Fig. 2 Variation of the degradation rate constant of RhB (0.02 mmol L⁻¹), k_RhB and adsorbed amount of EDTA (q_e, mmol mol⁻¹) over Fe₃O₄ MNPs (2.4 mmol L⁻¹) as a function of initial EDTA concentration (a), pH (b) and temperature (c). (d) Plots of ln k_RhB against T⁻¹ for the systems of (1) RhB–H₂O₂–Fe₃O₄–EDTA and (2) RhB–H₂O₂–Fe₃O₄. The pH, temperature, and the concentrations of H₂O₂, Fe₃O₄ MNPs and EDTA were 5.0, 40 °C, 5 mmol L⁻¹, 1.2 mmol L⁻¹, and 0.5 mmol L⁻¹, if not specified.

Fig. 2b shows the effects of EDTA on the pH dependence of k_RhB. In all the tested pH range (pH 4.0 to 10.0), the addition of EDTA is capable of enhancing appreciably the catalytic activity of nano-Fe₃O₄, and significantly accelerating the degradation of RhB. This enhancement effect of EDTA is almost uninfluenced by pH, implying that pre-adjusting of solution pH is unnecessary, a highly desirable nature in practical wastewater treatments. Moreover, the pH dependence of the q_e value closely resembles that of k_RhB, supporting again that the enhancing effect of EDTA primarily originates from the in situ surface modification of Fe₃O₄ MNPs by the adsorbed EDTA molecules.

Fig. 2c illustrates the effect of reaction temperature on the EDTA adsorption and the catalytic activity of Fe₃O₄ MNPs. When the temperature was increased from 25 to 60 °C, the q_e value increased from 20.8 to 32.5 mmol mol⁻¹, suggesting that EDTA adsorption on the Fe₃O₄ surface was definitely endothermic in nature and that a higher temperature favors the adsorption of EDTA .³⁵ Correspondingly, the k_RhB value in the H₂O₂–EDTA–Fe₃O₄ system significantly increased from 0.0018 to 0.098 min⁻¹ with a factor of 54.4, whereas in the absence of EDTA the k_RhB value increased slightly from 0.00093 to 0.004 min⁻¹, i.e., 4.3 times . The EDTA-induced iron leaching was also increased up to 0.29 mmol L⁻¹ at 60 °C, but the EDTA-driven homogeneous Fenton reaction in the presence of iron ions at this level was negligible at 60 °C in comparison with the fast RhB degradation in the similar H₂O₂–EDTA–Fe₃O₄ system (Fig. S5 in ESI†). Furthermore, based on the Arrhenius equation of ln k = lnA − (E_a/RT), where R is the universal gas constant (8.314 J mol⁻¹ K⁻¹), the apparent activation energy (E_a ) and the pre-exponential factor (A) of the reaction were evaluated by plotting ln k_RhB against T⁻¹ (Fig. 2d). From the slope of the ln k_RhB-T⁻¹ plots, the E_a value was found to be 36.7 kJ mol⁻¹ for the H₂O₂–Fe₃O₄ system, and 94.5 kJ mol⁻¹ for the H₂O₂–EDTA–Fe₃O₄ system. The increased E_a in the presence of EDTA may be due to the fact that the endothermic adsorption behavior of EDTA required some amount of activation.³⁵ However, the value of the apparent pre-exponential factor A in the Arrhenius equation for the H₂O₂–EDTA–Fe₃O₄ system is 8.1 × 10¹³, which is 3.9 × 10¹⁰ times greater than that (2.1 × 10³) of the H₂O₂–Fe₃O₄ system. This implies that the presence of EDTA would greatly increase the effective collision proportionality between H₂O₂ and Fe₃O₄, which is evidenced by the FTIR analysis and DFT calculations in the following section. Thus, the RhB degradation is much faster in the H₂O₂–EDTA–Fe₃O₄ system than in the H₂O₂–Fe₃O₄ system under similar experimental conditions.

FTIR characteristics of surface modified Fe₃O₄

The interactions among EDTA, Fe₃O₄, and H₂O₂ were examined by ATR-FTIR. In the spectra of bare Fe₃O₄ MNPs (curve 1 in Fig. 3a and b), the 1626 cm⁻¹ peak and the broad 3185 cm⁻¹ band (not shown here) were characteristic peaks of the O–H bending vibration (δ_OH) and stretching vibration (v_OH), respectively, indicating the presence of hydroxyl groups on the surface of the Fe₃O₄ MNPs.³⁶ After the Fe₃O₄ MNPs were immersed in H₂O₂ solutions for 1 h, δ_OH was slightly widened (curve 2 in Fig. 3a and b), suggesting that the H₂O₂ molecule can be adsorbed onto the Fe₃O₄ surface via direct binding to the surface Fe atom or by hydrogen bonding (a and b in Scheme 1).

Fig. 3 (a,b) ATR-FTIR spectra of (1) bare Fe₃O₄ MNPs, (2) H₂O₂-adsorbed Fe₃O₄, (3) EDTA-adsorbed Fe₃O₄, and (4) Fe₃O₄ simultaneously adsorbed with H₂O₂ and EDTA. The bare and adsorbed Fe₃O₄ samples were prepared by pre-immersing Fe₃O₄ MNPs in solutions of distilled water, H₂O₂ (5 mmol L⁻¹), EDTA (0.5 mmol L⁻¹), and the H₂O₂–EDTA mixture at pH 5.0 (a) and pH 10.0 (b) for 1 h at room temperature. Curve 5 is that for pure Na₂H₂EDTA solid. (c) ATR-FTIR spectra of EDTA-adsorbed Fe₃O₄ prepared at various pH values.

Scheme 1 Surface interactions of Fe₃O₄ with H₂O₂ (a, b), EDTA (c, d, e), and H₂O₂–EDTA (f).

The adsorption of EDTA on the Fe₃O₄ surface widened the 1626 cm⁻¹ band and produced a new peak at 1383 cm⁻¹ (curve 3 in Fig. 3a and b). In comparison with pure EDTA (curve 5 in Fig. 3a and b), the EDTA-adsorbed Fe₃O₄ involved the asymmetric vibration of the carboxyl group (v_as(COO⁻)) of EDTA at 1600 cm⁻¹, being partially overlapped with the δ_OH (1626 cm⁻¹) of Fe₃O₄, while the new peak at 1383 cm⁻¹ could be assigned to the symmetric vibration of the carboxyl group (v_s(COO⁻)) of EDTA.³⁷ In the spectrum of EDTA-adsorbed Fe₃O₄, both v_as(COO⁻) and v_s(COO⁻) bands were shifted to lower wavenumbers by ca. 10 cm⁻¹ compared to those of pure EDTA v_as(COO⁻), 1611 cm⁻¹; v_s(COO⁻), 1393 cm⁻¹), suggesting a strong interaction between EDTA molecules and Fe₃O₄ surface. It was reported that the separation between the carboxylate stretch bands (Δν = v_as(COO⁻) − v_s(COO⁻)) can be used to distinguish between the coordination modes of carboxylate and metal: Δν is generally larger than 200 cm⁻¹ for monodentate binding, in the range from 180 to 150 cm⁻¹ for bridging bidentate (binuclear), and smaller than 100 cm⁻¹ for chelating binding (mononuclear bidentate) (Scheme S1 in ESI†).³⁸ The Δν value of 217 cm⁻¹ in the present EDTA–Fe₃O₄ system indicates that EDTA is surface coordinated with Fe₃O₄ by forming a monodentate inner-sphere (directly coordinated to surface metal ions, c in Scheme 1) or outer-sphere complexes involving ester linkages through one carboxylate group of EDTA (d and e in Scheme 1). Guan et al. pointed out that the direct bonding between the O atom of the carboxyl group of EDTA and the Fe atom on the Fe₃O₄ surface will shorten the C=O bond and increase v_as(COO⁻).³⁹ Because the opposite trends are observed in our study, we assumed that the adsorption of EDTA onto the surface of Fe₃O₄ is dominated by the formation of outer-sphere complexes viahydrogen bonding and/or an H₂O-bridge (d and e in Scheme 1), with similar structures proposed for the interaction of carboxylate groups at the hematite–water interface.⁴⁰ This can be verified by the strong pH-dependence of the interactions between EDTA and Fe₃O₄, being reflected by the strong pH-dependence of the intensity of v_as(COO⁻) and v_s(COO⁻) on the EDTA-adsorbed Fe₃O₄ samples (Fig. 3c).

Furthermore, the characteristic bands of v_as(COO⁻) and v_s(COO⁻) in the spectra of EDTA-adsorbed Fe₃O₄ were distinctly broadened by the present H₂O₂, especially at pH 10.0, implying that the EDTA-modified Fe₃O₄ surface can bind to the H₂O₂ molecule. It was also noted that the co-existence of H₂O₂ made the peaks of v_as(COO⁻) and v_s(COO⁻) on the EDTA-adsorbed Fe₃O₄ shift to lower values, suggesting that there was hydrogen bonding between H₂O₂ and EDTA. The adsorbed H₂O₂ is considered to be a proton donor for attracting deprotonated carboxyl groups or amino groups, thereby changing the asymmetric environment around the ligand. However, these interactions can easily be masked by the hydrogen bonding between EDTA and Fe₃O₄ in the pH range of 4.0–7.0. Accordingly, the difference between the EDTA–Fe₃O₄ and H₂O₂–EDTA–Fe₃O₄ systems at pH 10.0 is more obvious than that at pH 5.0. Moreover, the δ_OH band near 1626 cm⁻¹ in the H₂O₂–EDTA–Fe₃O₄ system was much wider and stronger than that in the H₂O₂–Fe₃O₄ and EDTA–Fe₃O₄ systems, indicating that the EDTA-modified Fe₃O₄ surface has a stronger affinity for H₂O₂ molecules than the corresponding unmodified surface.

The adsorption of H₂O₂ on Fe₃O₄ and the interaction between EDTA and Fe₃O₄ were investigated by DFT calculations. H₂O₂ prefers to adsorb on the (111) surface of Fe₃O₄ through hollow sites, and the O atoms of H₂O₂ directly interact with the four-coordinate iron ion, shortening the O–O bond of H₂O₂ from 1.464 to 1.300 Å (Fig. S1b in ESI†). If both EDTA and H₂O₂ are present, then EDTA is adsorbed on the Fe₃O₄ surface to form monodentate outer-sphere complexes and H₂O₂ is hydrogen bonded to the O atom on the top sites of Fe₃O₄ through one H atom of the former. The hydrogen bonding between the other H atom of H₂O₂ and the N atom and COO⁻ of EDTA made the H₂O₂ adsorption configuration changes to a more stable bridge mode (Fig. S1c in ESI†). Compared with the case where EDTA is absent, the interaction between H₂O₂ and EDTA in the present case not only favors the dissociation of H from H₂O₂ but also stretches the O–O bond of H₂O₂ to 1.380 Å (see more details in section S2 of ESI†). Therefore, the EDTA-modified Fe₃O₄ MNPs are capable of catalyzing efficiently the breakage of H₂O₂ to generate ˙OH and O₂⁻˙/HO₂˙ radicals, as was confirmed by the following ESR measurements.

Identification of reactive oxygen species

The DMPO spin-trapping ESR spectra of ˙OH radicals and O₂⁻˙ radicals are shown in Fig. 4. Two types of ROS are produced in the H₂O₂–Fe₃O₄ systems, but the ESR signals were negligible in the absence of Fe₃O₄ MNPs. The four-fold characteristic peak with an intensity ratio of 1 : 2 : 2 : 1 is attributed to the DMPO-OH adduct,⁹ whereas the sextet characteristic peak is attributed to the DMPO-O₂⁻˙/HO₂˙ adduct.⁴¹ The peak intensities of both the ˙OH and O₂⁻˙/HO₂˙ radicals in the H₂O₂–EDTA–Fe₃O₄ system are much higher than those in the H₂O₂–Fe₃O₄ system, which is consistent with the above-mentioned theoretical prediction. Recently, Keenan and Sedlak proposed that the ferryl ion (Fe⁴⁺) would generate in a system composed of ZVI, oxygen and EDTA under circumneutral pH.²² To assess the essential role of the ROS including Fe⁴⁺, ˙OH and O₂⁻˙/HO₂˙, the RhB degradation was conducted in the H₂O₂–EDTA–Fe₃O₄ system with the addition of different scavengers (2.0 mmol L⁻¹) at various pH values (Fig. 4c), where methanol (MeOH) was used to scavenge both ˙OH radicals (k = 8 × 10⁸ M⁻¹ s⁻¹) and Fe⁴⁺,²²t-butyl alcohol to scavenge only ˙OH radicals (k = 5.2 × 10⁸ M⁻¹ s⁻¹),⁴² and benzoquinone was used as the quencher of O₂⁻˙ radicals (k = 1 × 10⁹ M⁻¹ s⁻¹).⁴³ The RhB degradation was found to be inhibited by the three scavengers, and the inhibitory effect is increased in the order of methanol < t-butyl alcohol < benzoquinone in the pH range of 4–10. This implied that O₂⁻˙/HO₂˙ radicals made a greater contribution than the other two oxidants (˙OH and Fe⁴⁺) to the RhB degradation in the H₂O₂–EDTA–Fe₃O₄ system, similar to the H₂O₂–Fe₃O₄ system in our previous report.⁹ Compared to t-butyl alcohol, the weaker inhibitory effect of methnol suggested that the ferryl ion might play a negligible role in the RhB degradation, possibly due to the fact that the ferryl ion was unable to react with the aromatic contaminant.²² Thus, the RhB degradation was mainly due to ˙OH and O₂⁻˙/HO₂˙ radicals in the catalytic system of H₂O₂–EDTA–Fe₃O₄, and the O₂⁻˙/HO₂˙ radicals perfomed a more important function than ˙OH radicals.

Fig. 4 DMPO spin-trapping ESR spectra of ˙OH radicals pH 5.0 (a) and O₂⁻ radicals (b) and degradation rate constant (k_RhB) of RhB (0.02 mmol L⁻¹) as a function of pH with different scavengers of ˙OH or O₂⁻˙ radicals in the EDTA–H₂O₂–Fe₃O₄ system (c). The temperature, H₂O₂, Fe₃O₄ MNPs and EDTA concentration were 40 °C, 5 mmol L⁻¹, 1.2 mmol L⁻¹, and 0.5 mmol L⁻¹, if not specified.

Proposed catalytic mechanism

As discussed above, the activation of H₂O₂ by Fe₃O₄ MNPs originates from the surface Fenton like reactions. The catalytic mechanism of EDTA-modified Fe₃O₄ MNPs can be summarized by eqn (1)–(9).

≡Fe²⁺ + (H₂O₂)_s → ≡Fe³⁺ + (˙OH)_s + OH⁻ (1)

≡Fe³⁺ + (H₂O₂)_s → ≡Fe²⁺ + (HO₂˙)_s + H⁺ (2)

≡Fe²⁺ + (˙OH)_s → ≡Fe³⁺ + OH⁻ (3)

≡Fe³⁺ + (HO₂˙)_s → ≡Fe²⁺ + O₂ + H⁺ (4)

(ROS)_s → ROS (5)

HO₂˙ → O₂⁻˙ + H⁺ (6)

≡Fe²⁺⋯EDTA + (H₂O₂)_s → ≡Fe³⁺⋯EDTA + (˙OH)_s + OH⁻ (7)

≡Fe³⁺⋯EDTA + (H₂O₂)_s → ≡Fe²⁺⋯EDTA⋯H⁺ + (HO₂˙)_s (8)

(ROS)_s/ROS + organics → degradation products (9)

The adsorbed H₂O₂ is catalyzed by the bound Fe²⁺ and Fe³⁺ on the Fe₃O₄ surface to produce ˙OH and O₂⁻˙/HO₂˙ radicals, respectively, being accompanied by the initiation of a series of chain reactions expressed in eqn (1)–(4). Among these reactions, eqn (2) has the slowest rate, yielding a poor Fe³⁺/Fe²⁺ recycling, which is unfavorable to the chain propagation of reactive radicals, thereby limiting the catalytic ability of Fe₃O₄ MNPs. As suggested by eqn (2) and (4), a sufficient amount of H₂O₂ or O₂⁻˙/HO₂˙ is required for the reduction of Fe³⁺ to Fe²⁺ in order to accelerate the chain reactions. Although O₂⁻˙/HO₂˙ species are more effective than H₂O₂ in the reduction of Fe³⁺, the former will consume the latter. This is why a very high concentration of H₂O₂ (40–2580 mmol L⁻¹) is required in an H₂O₂–Fe₃O₄ MNPs system, being 2–4 orders of magnitude higher than that of the target pollutants (0.02–3 mmol L⁻¹), as reported previously.^5–9 However, EDTA modifies the Fe₃O₄ surface, and the formed surface of the Fe–EDTA complex reduces the redox potential of Fe³⁺/Fe²⁺ from the conventional level of 0.77 V to 0.17 V,³⁰ making the Fe³⁺-involved surface Fenton-like reaction more thermodynamically favorable (eqn (8)). In addition, the adsorbed EDTA affects the adsorption mode and the decomposition of H₂O₂ on the Fe₃O₄ surface as evidenced by ATR-FTIR analysis and DFT calculations. The hydrogen bonding between EDTA and H₂O₂ can drive the dissociation of H from H₂O₂, enhancing the breakage of H₂O₂ by surface Fe³⁺ to generate HO₂˙ radicals (eqn (8)). Raman measurements revealed that the presence of EDTA improved the Fe³⁺/Fe²⁺ recycling on the Fe₃O₄ surface (Fig. S6 in ESI†). The enhanced reduction of Fe³⁺ by H₂O₂ can reduce the relative contribution of O₂⁻˙/HO₂˙ to Fe²⁺ generation (eqn (4)), leading to the high usage of ROS for attacking organic pollutants. Furthermore, DFT calculations demonstrated that the presence of EDTA stretches the O–O bond of H₂O₂, leading to an enhancement in the homolysis of H₂O₂ by surface Fe²⁺ to generate ˙OH radicals (eqn (7)). Both of the above mentioned effects of EDTA are favorable to improving the iron cycle on Fe₃O₄ surface and accelerating the continuous production of ROS for target pollutant degradation.

Conclusions

The surface physical and chemical environments of the Fe₃O₄ MNPs were affected by the added EDTA according to the strong interaction between Fe₃O₄ MNPs and EDTA, which lead to the efficient H₂O₂-activation and then to the efficient removal of organics. ATR-FTIR analysis and DFT calculations demonstrated that there were surface interactions between EDTA, Fe₃O₄ and H₂O₂, and revealed that the outer-sphere complexes were formed on the Fe₃O₄ MNPs surface viahydrogen bonding and/or H₂O bridging after the EDTA adsorption. The EDTA-modified Fe₃O₄ surface had a stronger affinity for H₂O₂ molecule and was beneficial for the breakage to produce ˙OH and O₂⁻˙/HO₂˙ in comparison to the unmodified surface. ESR measurements and the ROS scavengers experiments confirmed the much promoted generation of ˙OH and O₂⁻˙/HO₂˙ in the Fe₃O₄–H₂O₂ system with EDTA in comparison to the system without EDTA. In addition, Raman analysis showed that the added EDTA was favorable to the Fe³⁺/Fe²⁺ cycle on the Fe₃O₄ surface, which is beneficial for the Fenton-like reaction and then to accelerating the continuous production of ROS for target pollutant degradation.

Acknowledgements

This work was supported by the National Science Foundation of China (Grant Nos. 20877031, 21077037 and 21177044), the Brainstorm Project of Science and Technology of Wuhan City, China (Grant No. 201060623258), and the Fundamental Research Funds for the Central Universities of China (Grant No. 2011TS121 and CZZ11008). The Analytical and Testing Center of Huazhong University of Science and Technology is thanked for its help in the characterization of catalysts.

References

1. P. Shukla, S. Wang, H. Sun, H. Ang and M. Tadé, Chem. Eng. J., 2010, 164, 255–260.
2. A. L. Pham, C. Lee, F. M. Doyle and D. L. Sedlak, Environ. Sci. Technol., 2009, 43, 8930–8935.
3. C. Huang, Y. Huang, H. Cheng and Y. Huang, Catal. Commun., 2009, 10, 561–566.
4. M. B. Kasiri, H. Aleboyeh and A. Aleboeh, Appl. Catal., B, 2008, 84, 9–15.
5. J. B. Zhang, J. Zhuang, L. Z. Gao, Y. Zhang, N. Gu, J. Feng, D. L. Yang, J. D. Zhu and X. Y. Yan, Chemosphere, 2008, 73, 1524–1528.
6. S. X. Zhang, X. L. Zhao, H. Y. Niu, Y. L. Shi, Y. Q. Cai and G. B. Jiang, J. Hazard. Mater., 2009, 167, 560–566.
7. X. F. Xue, K. Hanna, M. Abdelmoula and N. S. Deng, Appl. Catal., B, 2009, 89, 432–440.
8. N. Wang, L. H. Zhu, M. Q. Wang, D. L. Wang and H. Q. Tang, Ultrason. Sonochem., 2010, 17, 78–83.
9. N. Wang, L. H. Zhu, D. L. Wang, M. Q. Wang, Z. F. Lin and H. Q. Tang, Ultrason. Sonochem., 2010, 17, 526–535.
10. R. C. C. Costa, M. F. F. Lelis, L. C. A. Oliveira, J. D. Fabris, J. D. Ardisson, R. R. V. A. Rios, C. N. Silva and R. M. Lago, Catal. Commun., 2003, 4, 525–529.
11. F. Magalhães, M. C. Pereira, S. E. C. Botrel, J. D. Fabris, W. A. Macedo, R. Mendonça, R. M. Lago and L. C. A. Oliveira, Appl. Catal., A, 2007, 332, 115–123.
12. X. L. Liang, S. Y. Zhu, Y. H. Zhong, J. X. Zhu, P. Yuan, H. P. He and J. Zhang, Appl. Catal., B, 2010, 97, 151–159.
13. S. J. Yang, H. P. He, D. Q. Wu, D. Chen, X. L. Liang, Z. H. Qin, M. D. Fan, J. X. Zhu and P. Yuan, Appl. Catal., B, 2009, 89, 527–535.
14. W. Luo, L. H. Zhu, N. Wang, H. Q. Tang, M. J. Cao and Y. B. She, Environ. Sci. Technol., 2010, 44, 1786–1791.
15. L. T. Chen, T. Liu and C. A. Ma, J. Phys. Chem. A, 2010, 114, 443–454.
16. J. D. Englehardt, D. E. Meeroff, L. Echegoyen, Y. Deng, F. M. Raymo and T. Shibata, Environ. Sci. Technol., 2007, 41, 270–276.
17. M. Sillanpää, K. Pirkanniemi and A. Sorokin, Environ. Technol., 2009, 30, 1593–1600.
18. C. K. Schmidt and H. J. Brauch, Environ. Toxicol., 2004, 19, 620–637.
19. C. Noradoun, M. D. Engelmann, M. McLaughlin, R. Hutcheson, K. Breen, A. Paszczynski and I. F. Cheng, Ind. Eng. Chem. Res., 2003, 42, 5024–5030.
20. I. Sanchez, F. Stüber, J. Font, A. Fortuny, A. Fabregat and C. Bengoa, Chemosphere, 2007, 68, 338–344.
21. C. E. Noradoun, C. S. Mekmaysy, R. M. Hutcheson and I. F. Cheng, Green Chem., 2005, 7, 426–430.
22. C. R. Keenan and D. L. Sedlak, Environ. Sci. Technol., 2008, 42, 6936–6941.
23. W. Luo, M. E. Abbas, L. Zhu, W. Zhou, K. Li, H. Tang, S. Liu and W. Li, Anal. Chim. Acta, 2009, 640, 63–67.
24. X. Shen, L. Zhu, G. Liu, H. Tang, S. Liu and W. Li, New J. Chem., 2009, 33, 2278–2285.
25. A. F. Forti and G. Scortichini, Anal. Chim. Acta, 2009, 637, 214–219.
26. B. Nowack, F. G. Kari, S. U. Hilger and L. Sigg, Anal. Chem., 1996, 68, 561–566.
27. L. Sigg and W. Stumm, Colloids Surf., 1981, 2, 101–117.
28. A. K. Kandalam, B. Chatteriee, S. N. Khanna, B. K. Rao, P. Jena and B. V. Reddy, Surf. Sci., 2007, 601, 4873–4880.
29. H. Munakata, Y. Oumi and A. Miyamoto, J. Phys. Chem. B, 2001, 105, 3493–3501.
30. X. F. Xue, K. Hanna, C. Despas, F. Wu and N. S. Deng, J. Mol. Catal. A: Chem., 2009, 311, 29–35.
31. F. B. Li, J. J. Chen, C. S. Liu, J. Dong and T. X. Liu, Biol. Fertil. Soils, 2006, 42, 409–417.
32. W. Ferraz, L. C. A. Oliveira, R. Dallago and L. Conceicao, Catal. Commun., 2007, 8, 131–134.
33. M. Hermanek, R. Zboril, I. Medrik, J. Pechousek and C. Gregor, J. Am. Chem. Soc., 2007, 129, 10929–10936.
34. A. B. Hart, J. Mcfadyen and R. A. Ross, Trans. Faraday Soc., 1963, 59, 1458–1469.
35. K. G. Bhattacharyya and A. Sharma, Dyes Pigm., 2005, 65, 51–59.
36. K. Yang, H. B. Peng, Y. H. Wen and N. Li, Appl. Surf. Sci., 2010, 256, 3093–3097.
37. C. Lanigan and K. Pidsosny, Vib. Spectrosc., 2007, 45, 2–9.
38. X. Gao, D. W. Metge, C. Ray, R. W. Harvey and J. Chorover, Environ. Sci. Technol., 2009, 43, 7423–7429.
39. X. H. Guan, G. H. Chen and C. Shang, J. Environ. Sci., 2007, 19, 438–443.
40. H. C. Chang, T. W. Healy and E. Matijevic, J. Colloid Interface Sci., 1983, 92, 469–478.
41. M. M. Cheng, W. H. Ma, J. Li, Y. P. Huang and J. C. Zhao, Environ. Sci. Technol., 2004, 38, 1569–1575.
42. J. D. Rush and W. H. Koppenol, J. Biol. Chem., 1986, 261, 6730–6733.
43. K. B. Patel and R. L. Willson, J. Chem. Soc. Faraday Trans., 1973, 1, 814–825.

---

Footnote

† Electronic supplementary information (ESI) available: Analytical methods of PCP, SMM, and EDTA with HPLC (S1a), sorption experiment of fluoride (S1b), methods and model of the DFT calculations (S2), enhanced Fe³⁺/Fe²⁺ recycle on the surface of EDTA-modified Fe₃O₄ nanoparticles (S3), Fig. S1–S6, and Scheme S1. See DOI: 10.1039/c1cy00260k

---