Influence factor of 17β-estradiol photodegradation by heterogeneous Fenton reaction

Abstract

The photocatalytic degradation of 17β-estradiol (E2), by Fenton like reaction was investigated as a function of E2 concentrations, organic co-solvents and co-existing estrogens, humic acid (HA) and other background anions. E2 degradation was effectively achieved by hydroxyl radicals that were generated in the heterogeneous photo-Fenton process. The degradation kinetics were fitted to Langmuir–Hinshelwood model with kr = 0.3140 µM/h and Kads = 2.2146L/µmol. The removal kinetics of E2 were initiated by a rapid decay and then followed by a much slower one in acetonitrile–water solutions while in methanol–water solutions they followed the first-kinetic model for the diffusion-control of hydroxyl radicals and competition between E2 and co-solvents. In addition, the lower level of co-existing substances did not significantly influence the oxidation efficiency of E2. The degradation rates of E2 were found to depend not only on the concentrations of hydrogen peroxide and iron content as reported before but also on pH, E2 concentrations and composition of co-solvents. Thus it is very important to look for the optimum conditions for the purpose of most efficiently eliminating E2 from drinking water.

---

Environmental impact

Most endocrine disrupting compounds (EDCs) are synthetic organic chemicals introduced into the environment by anthropogenic input, but they can also be naturally generated estrogenic hormones (e.g., estrone, 17β-estradiol) and therefore are ubiquitous in aquatic environments. The environmental level EDCs in the public drinking water can cause adverse effects on humans and wildlife via interactions with the endocrine system. α-FeOOH-coated resin (α-FeOOHR), was applied for the photodegradation of 17β-estradiol (E2) in this work as a function of E2 concentrations, organic co-solvents, co-existing estrogens, natural organic matter humic acid (HA) and other background anions. The results are important to evaluate the photodegradation of E2 and estrogen mimics under optimum condition. Moreover, heterogenerous photo Fenton catalysts would be of great importance for application due to their high photoactivity.

Introduction

Endocrine disrupting chemicals (EDCs) have attracted significant attention in recent years. The environmental level EDCs in the public drinking water supply can cause adverse effects on humans and wildlife via interactions with the endocrine system. For example, estrone, estradiol and 2-hydroxyestradiol have many unexpected beneficial effects in vivo. On the contrary, quinoid radicals and phenoxy radicals formed from them may also be responsible for adverse effects such as carcinogenesis. The biological effects of estrogens involve their dual role as prooxidants and effective antioxidants.^1,2 17β-estradiol (E2) is a steroid hormone produced primarily within the female ovaries and released into the environmental water from humans and domestic animals. E2 usually causes suspected adverse effects in aquatic organisms at concentrations above 0.03 nmol/L.³

The removal of EDCs becomes a challenging task to effectively remove EDCs due to its high toxicity for microorganisms and poor biodegradability. In some cases, the traditional activated sludge process is now assumed to generate EDCs in the form of free estrogens which are the result of incomplete degradation of respective parent compounds.⁴ By these reasons, advanced oxidation processes (AOPs) have emerged as available and promising alternatives for conventional water treatment systems. For example, the degradation of three EDCs, bisphenol A, ethinyl estradiol, and estradiol, was investigated via ultraviolet (UV) radiation photolysis and the UV/hydrogen peroxide advanced oxidation process (AOP). UV/H₂O₂ was more effective as compared to direct UV photolysis treatment.⁵ The TiO₂-assisted photodegradation of two natural female hormones, estrone (E1) and E2 has also been studied in two UV-photo-reactors with wavelengths of 253 nm and 238–579 nm, respectively. The high levels of hormone removal suggested the photocatalysis was an effective and rapid method.⁶ Among AOPs, Fenton reaction was also used to degrade EDCs.⁷ The acceleration for decomposition of organic compounds is believed to be a result of photolysis of iron aquacomplex, Fe(OH)²⁺, which provides an importance new source of ˙OH radicals. Furthermore, the Fe(OH)²⁺ can absorb light at wavelengths up to ca. 410 nm, while TiO₂ photocatalysis can use photons with wavelengths close to 380 nm. Therefore, the photo-Fenton process can be expected to an efficient and inexpensive method for water treatment and promotes the rate of degradation of various organic pollutants. To overcome the drawbacks of conventional Fenton reaction, various modified Fenton systems have been developed due to their advantages such as facile recovery and recycling. For example, most used heterogeneous Fenton-like systems were made using goethite and hematite,⁸ which are usually slow and need additional assistants such as UV and ultrasound. Immobilized Fenton reactions using Fe³⁺ loaded on zeolite,⁹ silica,¹⁰ clay¹¹ and alumina¹² have been developed. Recently, the heterogeneous Fenton catalyst, Fe₃O₄–poly (3,4-ethylene-dioxythiophene) (PEDOT) core-shell nanoparticles, were synthesized by acid etching-mediated chemical oxidation polymerization. The degradation rate of Reactive Black 5 by Fe₃O₄–PEDOT had an approximately 2.5 times higher catalytic activity than commercialized Fe₃O₄ nanopowder without external energy input. The low level leaching of iron (<0.6 mg/L) facilitates the catalytic reaction in the confined region of the nanoparticle surface. Polymer-encapsulated iron oxide nanoparticles as highly efficient Fenton catalysts due to the functional polymer outer layer and high effective surface area.¹³

The activity of a heterogeneous catalyst is dependent on the structure and composition of its surface—both of which can change in response to variations in the environment.¹⁴ Just like the homogeneous Fenton reaction, the mechanisms and the kinetics of heterogeneous Fenton system is very sensitive towards almost any experimental variables, i.e. co-existing solvent, ligand, co-existing substances, and the oxidation state of metal ion, which can be rapidly changed at the beginning of the experiment.¹⁵ Thus, attention must be paid to the possible competition in hydroxyl radical receiving between E2 and organic co-solvent. However, the effects of co-existing solvent on the efficiency of the Fenton or Fenton-like systems are usually ignored and limited. E2 is soluble in organic solvents such as ethanol (EtOH), methanol (MeOH), dimethylformamide (DMF), acetonitrile (ACN), acetone and dimethyl sulfoxide (DMSO) while sparingly soluble in aqueous buffer. For maximum solubility of E2 in aqueous buffer, it should be firstly dissolved in DMSO and then diluted with aqueous buffer of choice. A lot of research has reported that E2 degradation by AOPs, in which E2 working solutions were prepared by diluting E2 organic co-solvent, such as ethanol, methanol, acetonitrile or DMSO, etc. with water. The addition of organic solvent in the E2 medium was to minimize the adsorption effects by the reactor. For example, during photo-degradation of E2 in aqueous solutions by UV–Vis/Fe(III)/H₂O₂ system, acetonitrile was used as co-solvent of E2.¹⁶ In TiO₂-assisted photodegradation of two natural female hormones, estrone (E1) and E2, was investigated in two UV-photo-reactors. Its working solution containing E1 and E2 was prepared by diluting the individual stock solutions (1000 mg/L) in methanol.⁶ Japanese researchers reported estrogen and its conjugates by TiO₂ as photocatalyst under UV irradiation, with ethanol stock solution of estrogen being diluted with water.¹⁷ Basically, most of the papers reported the degradation of estrogens in aqueous solutions with organic co-solvents whether AOPs or in determination of the estrogenicity. But there are hardly any reports on the competitive effects of organic co-solvents towards E2 degradation mechanism. So it is necessary to elucidate this kind of effects so as to choose suitable organic solvents to dissolve estrogens and avoid the side effects of the organic co-solvents. L. D. Palma et al. also reported that the degradation of atrazine with Fenton's reagent in the presence of ethanol.¹⁸ While at a Fe²⁺ concentration of 3 mM atrazine practically disappeared from pure aqueous solutions within 2 h, a degradation yield of only 28.1% was observed in the presence of 4.5 vol.% ethanol even when Fe²⁺ concentration was 15 mM. Ethanol decomposed in fact simultaneously with the atrazine degradation, which also significantly retarded the atrazine degradation efficiency. Researchers also noted that the addition of 1% (v/v) methanol to the HA-modified Fenton system strongly inhibited the selectivity of all model compounds by the catalyst (NDRL/NIST Solution Kinetics Database on the Web). Not only co-solvent influence Fenton oxidation process, but iron(II) and (III) ions with hydrogen peroxide in water and acetonitrile solvents also has different mechanism of the reactivity.¹⁹ Since a co-solvent is added for avoiding adsorption effects over time, its effect in the presence of Fenton's reagent needs to be investigated as it may compete with the substrates. So this research is not only very important to the degradation of E2 in water solutions, but also to the pollutants in contaminated soil which need to be flushed with organic solvent.

In this work, the photodegradation of E2 was studied in a batch recycle reactor in the presence of ultraviolet light, hydrogen peroxide, and α-FeOOH loaded resin (α-FeOOHR). This paper intends to provide information of reaction kinetic and the factors of co-existing solvent, co-existing estrogens and the presence of humic acid (HA) in drinking water affecting the photodegradation course.

Results and discussion

The novel photocatalyst α-FeOOHR was prepared by in situ hydrolysis of Fe³⁺ ions inside the pores of photocatalytic carriers Amberlite® 200 by carbamide.²⁰ The polymer carriers make α-FeOOHR very easy to be separated from treated water systems. The crystalline of the loaded iron oxide on the resin is the typical pattern of crystalline α-FeOOH (goethite, syn). The morphology of α-FeOOHR confirmed the presence of α-FeOOH entities on the surface of resin comparing with non-coated resin. The α-FeOOH entities were mainly one dimensional bars with a cross-section of 100 × 100 nm and length of 300 to 400 nm except for a few irregular micrometre particles. The loaded α-FeOOH in the photocatalyst was 10 Fewt.%, i.e. 100 mgFe/g. A high content of iron in the Fenton catalyst will inevitably increase the photocatalytic efficiency. The factors that influenced the efficiency of E2 degradation in this Fenton reaction were investigated.²⁰

Adsorption isotherm of E2 on α-FeOOHR

It is likely that sorption of the E2 is an important step in determining photocatalytic degradation rate. Adsorption tests in the dark were carried out in order to evaluate the equilibrium constants of the adsorption of the E2 on the surface of photocatalyst α-FeOOHR. When E2 concentration was 272µg/L, E2 removal rate increased from 17.3% to 96.9% with α-FeOOHR dosage increased from 2 g/L to 50 g/L. When α-FeOOHR dosage was 2 g/L, E2 removal rate increased from 17.3% to 99.6% with E2 concentration decreased from 272 to 0.0272 µg/L. Adsorption isotherm curve showed an isotherm of Langmuir-shape. It means that there is no strong competition between the solvent and E2 to occupy the adsorbent surface sites. The experimental data are well fitted by the Langmuir adsorption model as Eq. (1) and Freudlich adsorption model as Eq. (2) to describe the adsorption of E2 on the surface of the Fenton catalyst.

Ce/Qe = Ce/Qmax + 1/(Ka × Qmax) (1)

logQe = 1/n × logCe + logK_F (2)

where Ce is the equilibrium concentration in the solution, Qe is the sorbed concentration, Ka is the adsorption equilibrium constant and Qmax is the saturated adsorption capacity, K_F the Freundlich coefficient and n is a constant representing non-linearity. The saturated adsorption amount Qmax and the adsorption equilibrium constant Ka of E2 onto α-FeOOHR were listed in Table 1. The adsorption isotherm of E2 on the α-FeOOHR can be described by both Langmuir and Freudlich model with the high regression coefficients. From the data obtained, the maximal adsorption quantity and the Langmuir adsorption constant were respectively, 2.6641 × 10⁻⁵ mol/g and 0.5879 L/µmol. It was believed that the chelating complex could form between Fe³⁺ and E2 by chemical adsorption owing to their chemical structure. This implies that the chemical adsorption should exist between E2 and iron oxides on the surface of the Fenton catalyst through hydrogen bond between hydroxyls or chelating complexes through ligand exchange reaction by hydroxyl groups. The Ka value should be attributed to the chemical adsorption in some extent.

Table 1 The adsorption equilibrium constant and photo-catalytic rate constant

Langmuir	Qmax × 10^−5 (mol/g)	Ka (L/µmol)	R^2
	2.6641	0.5879	0.990
Freundlich	n	K F
	1.1558	0.2273	0.997
Langmuir–Hinshelwood	kr (µM h^−1)	Kads (L/µmol)
	0.3140	2.2146	0.999

---

Kinetics of photo-Fenton degradation of E2

The photo-degradation of E2 by the Fenton catalysts were repeated with a range of initial concentrations of 0.1, 0.5 and 1 µmol/L shown in Fig. 1. The adsorbed E2 on the surface of the Fenton catalyst α-FeOOHR acts as an electron donor, injecting electrons from its excited state to the conduction band of the semiconductor under UV irradiation. The initial rates for each concentration were determined from the psuedo-first-order rate model and initial concentrations. The data were then fitted to Langmuir–Hinshelwood kinetic rate model (L-H), which has been used to describe the rates of photocatalytic destruction of many organic compounds in many studies. In the absence of external mass transfer limitations, kinetic data of E2 photodecomposition under UV irradiation agreed well with the L–H kinetic rate mode as Eq. (3).where k_r and K_ads are the reaction rate constant in the aqueous solution and apparent adsorption coefficient.

Fig. 1 Effect of Concentration on E2 photodegradation. [H₂O₂] = 9.7 mmol/L; pH = 7.47; T = 20 °C; [FeOOHR] = 5 g/L; [E2] = 27.2, 136 and 272 µg/L.

The rate form can be linearized for initial concentrations as Eq. (4).

where C_o is the initial E2 concentration and r_o is the initial degradation rate. The k_r and K_ads values can be obtained from the intercept and slope of the 1/r_ovs. 1/C_o plot shown in Fig. 2 and Table 1. The constant of adsorption calculated from the L-H model is different from that calculated through Langmuir adsorption isotherm. This difference may be due to the fact that adsorbed E2 on α-FeOOHR who underwent degradation under UV irradiation and then other E2 in the bulk solution readsorbed on the α-FeOOHR. So the Langmuir isotherm of adsorption regroups the adsorption on the surface of α-FeOOHR, contrary to Langmuir–Hinshelwood model which represents only the adsorption of E2 on the photo-Fenton catalyst.²¹ The larger K_ads value (K_ads = 2.2146 L/µmol), which was larger than the Langmuir adsorption constant of 0.5879 L/µmol, further indicates that catalytic degradation occurred on the surface of α-FeOOHR. The concentration of iron leaking in the liquid phase was measured during 8 h irradiation and the maximum iron leaking was below 0.2% of catalyst iron content, i.e. below 1 mg/L. Then, the contribution of the homogeneous Fenton reaction during the degradation of E2 by α-FeOOHR could be neglected, which has been proved by our paper.²⁰ The higher k_r value (k_r = 0.3140 µmol/h with R² = 0.999) indicates higher degradation rate of E2 on the surface of α-FeOOHR. The results showed that the adsorption of E2 on photocatalyst was important to photocatalyst effectively functioning. At the same time kinetic constant is also very important to evaluate the quantum efficiency of this catalytic process. In a previous report, carried out under 150 W xenon lamp the kinetic rate constant for the degradation of E2 by TiO₂ (Degussa P25) was 0.044 µM/min. The difference of photo-degradation of E2 in these two AOPs duo to the light intensity (15 W black light) of this study was far lower than the previous report. The lower light intensity leads to an energy saving in degrading environmental concentration level pollutants.

Fig. 2 Plot of 1/C versus 1/R.

Influence of co-solvent of E2 on photodegradation

This study intended to evaluate the influence of methanol and acetonitrile on the E2 degradation rate, a preliminary series of oxidation with photo-Fenton reaction tests were performed on an aqueous solution of 272 µg/L of E2 in the presence of methanol or acetonitrile of 1 or 10 vol% (results shown in Fig. 3). At an acetonitrile volume concentration of 1 or 10%, 98.4 ∼ 98.7% E2 practically disappeared from acetonitrile–water solutions within 4 h, another degradation yield from 98.4% to 99.3% was observed within later 4 h. While at a methanol volume concentration of 1 or 10%, only 63.5 ∼64.6% E2 practically disappeared from methanol–water solutions within 4 h at the same conditions, a degradation yield of 85.5 ∼ 86.4% was observed within 8 h. A comparison between E2 in methanol solutions and E2 in acetonitrile solutions revealed that E2 was significantly more effective degradation in acetonitrile water solutions. Results showed a high consumption of catalytic reagent was needed to achieve a significant E2 oxidation from methanol–water solutions as compared with that of acetonitrile–water solutions. It is interesting to note that E2 and organic co-solvents do not react identically with the hydroxyl radicals, even though ˙OH is considered to be a non selective oxidative species. These results proved that the co-solvent indeed greatly influence the photodegradation rate in water solutions. In this study, it might be due to the fact that in acetonitrile–water solutions the initially E2 concentration was relatively large and accordingly the degradation rate was mainly controlled by generation of hydroxyl radical or the concentration of H₂O₂. As the concentration of E2 was decreased to a certain level, the reaction rate is controlled by diffusion of E2 molecules from the bulk to the α-FeOOHR/water interface. Hydroxyl radicals will most probably react with sorbed species before being able to diffuse to the solution.²² Usually, the chemical reaction (˙OH) rate is relatively much quicker compare to the diffusion of residual concentration of E2 in the solutions. From the kinetic studies, it can be concluded that the degradation process of E2 is under generation of hydroxyl radical control at initial phase and at a later stage the degradation rate is controlled by mass-transfer for hydroxyl radical is generally known to react with organics at diffusion-controlled rate.²³ Another explanation of the tailing of degradation curve also suggested that competition of hydroxyl radicals between substrate and primary intermediates did exist in the systems. This was the likely reason that the lower substrate in the later stage had a much lower chance to compete for the hydroxyl radicals with primary intermediates. Therefore, in the later stage of the oxidation processes, the decay of substrate was no longer the major reaction, while the decay of primary intermediates became dominant.²⁴ All in all, the reason the tailing curve was ascribed to the functioning of many factors in synergy. While in methanol–water solutions the E2 degradation rate within 8 h was mainly controlled by the reactivity competition by E2 and methanol molecules with oxidants on the α-FeOOHR/water interface. The high reactivity of methanol towards hydroxyl radicals leads to a higher consumption of reagents and a slower E2 photo-oxidation.²⁵ Whether in acetonitrile–water solutions or in methanol–water solutions, degradation curves within 24 h irradiation of a fast first stage followed by a slow stage were both occurred (data was not shown here). This behavior was also observed in the homogeneous Fenton and can be explained by the recycling of iron (3+) and (2+) or regeneration of iron (2+) by the reaction with H₂O₂ on the catalyst surface. The oxidized iron on the solid's surface produced in the first stage could react with hydrogen peroxide to produce hydroperoxyl radicals and regenerating the catalyst on the solid's surface. As the hydroperoxyl radical is less oxidative than the hydroxyl radicals, a slow second stage occurs.²⁶ However, Fig. 3 still proves that E2 compounds are readily oxidized simultaneously with nearly the same reaction rate constants whether 1% or 10 volume% methanol–water solutions, that is, excess methanol do not influence the E2 degradation rates. It shows that reactivity of substrates toward hydroxyl radicals could be another important factor affecting the degradation rates of E2 in organic co-solvent solutions. This was proved by Puangrat Kajitvichyanukul's work, whose findings indicated that methanol was difficult to oxide when competing with formaldehyde in photo-Fenton process.²⁷ The Fenton reaction mechanism and rate constants of E2 and several organic solvents with hydroxyl radicals were listed in Table 2.^28–30 It has been reported that reaction rate constants of OH˙ with ethanol, methanol and acetonitrile are respectively 2.1 × 10⁹, 1.2 × 10⁹ and 6 × 10⁶ M⁻¹s⁻¹.³¹ In the case of E2 the reaction rate constant of undissociated form of E2 with hydroxyl radicals is (5.3 ± 0.4) × 10⁹ M⁻¹s⁻¹ which is slightly bigger than that of methanol and ethanol while far bigger than that of acetonitrile. This higher reaction rate constant of E2 should be due to an increment of the electron density on the phenyl ring because of an electron-donating group of cycloalkane. Even though the co-solvents used have lower reactivities than E2, they still can compete for ˙OH as they are present in large excess, compared to the trace E2 concentrations in the solutions.³² On the other hand, the influence of methanol on the degradation rate of E2 was higher than that of acetonitrile. The reaction rate constants of E2 with hydroxyl radicals was nearly similar with that of methanol, E2 still can be degraded fast because the reactivity was also not the only factor affecting the degradation rate of E2. Though we did not investigate the influence of ethanol on E2 degradation it can be anticipated that the extent of ethanol hindering the E2 degradation must larger that of methanol for the difference of their reaction rate constants. Thus it is very important to choose suitable organic co-solvent of E2 to eliminate the solvent effect.

Table 2 Chemical Reactions Involving H₂O₂, Iron, E2 and Scavengers

H2O2 + Fe(II) → Fe(III) + OH^− + ˙OH	(1)
H2O2 + Resin-Fe(II) → Resin-Fe(III) + ˙OH + OH^−
H2O2 + Fe(III) → Fe(II) + O^−2˙ + 2H^+	(2)
H2O2 + Resin-Fe(III) → Resin-Fe(II) + O^−2˙/HO^−2˙ + H^+
E2 + ˙OH → reaction products	(3)
˙OH + S → products of scavenging reactions (reaction rate constant k)	(4)
O^−2˙ + E2 → reaction products	(5)

---

˙OH	hydroxyl radical
O^−2˙	superoxide radical
S	concentration of individual scavengers
k	second-order rate constant of hydroxyl radical with scavengers (M^−1s^−1)
(1)	k = 76 L/mol-s of homogeneous Fenton catalysis
(2)	reaction involves soluble and solid phase iron
(3)	Rate constant of hydroxyl radical with E2 (5.3 ± 0.4) × 10^9 M^−1s^−1
(4)	rate constant for ˙OH with radical scavenging of methanol 0.97 × 10^9 M^−1s^−1
	Rate constant of hydroxyl radical with ethanol in aqueous is 1.9∼2.1 × 10^9 M^−1s^−1
	Rate constant of hydroxyl radical with acetonitrile 2.1∼3.5 × 10^6 M^−1s^−1
	Rate constant of hydroxyl radical with DMSO 6.6 × 10^9 M^−1s^−1
	Rate constant of hydroxyl radical with acetone 0.11 × 10^9 M^−1s^−1
	Rate constant of peroxyl radical with E2 1 × 10^5 M^−1s^−1
(5)	k = 1.0 × 10^5 M^−1s^−1

---

Fig. 3 The effect of organic co-solvent on E2 photodegradation in 1% or 10% methanol or acetonitrile water solutions. [E2]_o = 272 µg/L; [H₂O₂] = 9.7 mmol/L; [α-FeOOHR] = 5 g/L; T = 20 °C.

Influence of co-existing substances on E2 photodegradation

The simultaneous presence of different organic pollutants in an environmental water system is expected. Due to the low selectivity of hydroxyl radicals, the treatment efficiencies are strongly affected by the presence of other competitive substrates, including co-existing estrogens, humic substances and background anions. The photo-Fenton oxidation processes of (1) 1µmol/L of E2 in deionized water; (2) E2 in deionized water co-existing with the other four estrogens, i.e. estrone (E1), estriol (E3), ethinyl estradiol (EE2), and bisphenol A (BPA), respectively; and (3) E2 in drinking water co-existing with the other four co-existing estrogens and HA (2.5 mg/L) were compared under the same conditions. The drinking water matrice contains (F⁻ (0.45 mg/L), Cl⁻ (12.91 mg/L), NO⁻₃(4.16 mg/L), SO²⁻₄(14.57 mg/L), Na⁺ (2.19 mg/L), K⁺ (6.43 mg/L), Ca²⁺ (14.28 mg/L) and Humic acid (HA) spiked (2.5 mg/L). The photo-degradation rate of five estrogens over α-FeOOHR was simulated with the pseudo-first order reaction kinetics model shown in Table 3. All of the rate constants indicate that the hydroxyl radical based decay of these five estrogens is rapid, and the heterogeneous photo-Fenton oxidation is an effective treatment for degradation of these contaminants (Fig. 4).

Table 3 degradation rate constants of estrogens

Estrogens		E1	E2	E3	EE2	BPA
Situation (1)	Rate constants (h^−1)	0.31	0.25	0.28	0.29	0.15
	R^2	0.930	0.972	0.917	0.978	0.993
Situation (2)	Rate constants (h^−1)	0.27	0.30	0.21	0.30	0.16
	R^2	0.964	0.928	0.929	0.900	0.913
Situation (3)	Rate constants (h^−1)	0.26	0.28	0.23	0.28	0.21
	R^2	0.979	0.997	0.995	0.996	0.995

---

Fig. 4 The effect of coexisting substances on the photodegradation of E2 in three different situations. [E2]_o = 272 µg/L; [H₂O₂] = 9.7 mmol/L; [α-FeOOHR] = 5 g/L; T = 20 °C. (1) E2 in deionized water; (2) E2 in deionized water co-existing with 1µmol/L of estrone (E1), estriol (E3), ethinyl estradiol (EE2), and bisphenol A (BPA), respectively; and (3) E2 in drinking water co-existing with the other four co-existing estrogens and HA.

Maureen et al. reported the reaction rate constants of E1 with hydroxyl radicals ranged from 1.1 × 10¹⁰ M⁻¹s⁻¹ to 7.0 × 10¹⁰ M⁻¹s⁻¹ with an average value of 2.6 × 10¹⁰ M⁻¹s⁻¹.³³ Rosenfeldt reported that the reaction rate constants of EE2, BPA and E2 with hydroxyl radicals were 1.02 × 10¹⁰ M⁻¹s⁻¹, 1.41 × 10¹⁰ M⁻¹s⁻¹ and 1.08 × 10¹⁰ M⁻¹s⁻¹, respectively.⁵ Comparing these three situations in Table 3, the degradation rates of these five estrogens were basically the same in these three situations for estrogens basically have the same magnitude second order reaction rate constants with hydroxyl radicals. Thus the influence of the presence of same magnitude co-existing estrogens on E2 photo-degradation rate almost can be neglected.

The effect of HA and drinking water background substances on E2 degradation should also be studied for its prevalence and interaction with organic contaminants. Comparing situation (2) with (3), E2 degradation rates also were not influenced greatly by the presence of complex background substances in drinking water. The reaction rate constant of the average value for several aquatic humics was of (1.7 ± 0.7) × 10⁴ (mg of C/L) M⁻¹s⁻¹. Comparing the reaction rate constant with that of E2, 2.5 mg/L HA was a minor sink for hydroxyl radical, thus the HA did not obviously alter the concentration of hydroxyl radicals through scavenging.³⁴ Potentially, important effect of Humic Substances (HS) on the Fenton and Fenton-like processes still remains unclear as their relationship is complicated and the published results are conflicting. Some authors suggested that the presence of HS inhibited or had no significant effect on the Fenton processes, while others reported that HS enhanced the oxidation efficiency in such systems. Negative result, for example, HS contains many functional groups and can compete with the target compound for hydroxyl radicals but also function as iron chelators. Furthermore, binding of iron by NOM can change the formation rate of hydroxyl radical by affecting the redox cycling of iron.³⁵ Fu et al. reported that γ-HCH photo-degraded by α-Fe₂O₃ in the presence of different fulvic acid (FA) concentrations at pH 4.³⁶ It was found that FA slowed down the transformation rate of γ-HCH. They speculated the humic molecular adsorbed onto the surface of α- Fe₂O₃, acting as scavengers of valence band holes due to its electron-rich nature, thus slowing down the degradation of γ-HCH occurred on the semiconductor surface. Besides, when a preequilibrium formed between the FA and α- Fe₂O₃, it is difficult for γ-HCH to contact band hole due to the obstructiveness of the layer of the FA molecular layer. However, positive results regarding the influence of humic substances on pollutant degradation in AOPs systems has been reported. Zhang et al. reported the influence of HA (with concentration of 0.01 ∼ 10 mg/L) on the degradation of E1 and E2. Results showed that the concentration of E1 and E2 started to decrease with an increase in HA concentration. This positive influence of HA on the photodegradation of E1 and E2 is attributed to the sensitization effect of HA. Photosensitized reactions involving electronic energy transfer from triplet states of HA to organic molecules as well as photosensitized oxygenations via the singlet oxygen pathway have been widely described.⁶ Voelker and Sulzberger reported an increase in the rate of H₂O₂ degradation by the Fenton reaction was observed at pH 5 when FA was added, whereas the effect was negligible at pH 3. The authors concluded that Fe(II)–fulvate complexes formed at pH 5 are able to react more rapidly with H₂O₂ than Fe(II)–aqua complexes, leading to a higher rate of OH radicals production. They also showed that FA acts as a reductant of Fe(III), which was ascribed to a binding of Fe(III) to quinone-type structures in the FA.³⁷ Georgi et al. reported a slight deceleration of the degradation of trinitrotoluene in the presence of 40 mgC/L of HA, whereas 20 mgC/L of FA appeared to slightly promote the reaction by Fenton catalytic process in pH 3.³⁸ The influence of HA on the AOPs is more complex, the crucial is concrete analysis of specific conditions. In drinking water matrix, it also contains many other anions and cations, such as Cl⁻, SO²⁻₄, PO³⁻₄, Na⁺, Mg²⁺, etc., which play a significant role in the reaction rate of the Fenton process. The oxidation process would be usually reduced or inhibited in presence of these anions for the formation of less reactive radicals.^29,39 Laat and Le pointed out that relatively high concentrations of chloride are needed (>50–100 mM) in order to inhibit the formation of the iron(III)-peroxocomplexes and therefore to observe a significant decrease of the rate of decomposition of H₂O₂. As compared to the sulfate ion, the inhibiting effect of the chloride ion for the formation of the peroxocomplexes is much less important.¹⁵ In this study, the effect of the anions and cation in drinking water can be neglected for their low concentrations. The result was the same with that Raphael Semiat's report. He pointed out that Cl, SO₄, Ca, Na and Mg ions had no significant effect on the kinetics of phenol oxidation by iron oxide-based catalysts.⁴⁰ Consequently, considering both scavengers and formation rate, it is expected that hydroxyl radical concentrations would be relatively constant upon addition of 2.5 mg/L HA in drinking water under the current condition. Furthermore, first-order kinetic was observed for estrogens degradations, and measured rate constants for reaction with hydroxyl radicals in pure water and in drinking water were basically the same.³⁵

The effect of pH on E2 photodegradation

Under real environmental conditions, the Fenton and Fenton like systems are strongly influenced by pH, become even more complicated. The effects of pH on the percentage of E2 removal, iron leakage, and pH variations of reaction solutions were determined with a pH range of 3 to 11 and results were shown in Fig. 5. The photo-degradation process remained efficient and feasible up to pH 11.00 with the photodegradation rate of E2 remaining 63.2% comparing with 98.2% at pH 3.07 and 86.4% at pH 7.47. The initial rates as function of pH were 0.56 µM/h (R² = 0.900) at pH 3.07, 0.22 µM/h (R² = 0.992) at pH 7.47 and 0.13 µM/h (R² = 0.908) at pH 11.00. There was a general decrease in the degradation rate with increasing pH. This observation is consistent with the typical Fenton process whose optimum efficiency was carried on at pH 3.0.⁴¹ At acid pH range, Fenton catalyst is positively charged at pH < pHzpc and favors E2 adsorption. E2 would be ionized to the phenoxide and also the surface of α-FeOOH is deprotonated to the negative ions causing repulsion with substrate when pH > pHzpc, especially pH > 10 (pK_a of E2),⁴² which better explained the E2 removal rate decrease with the increase of the pH. The pH is an important parameter affecting photo-Fenton process, which affects the generation of hydroxyl radicals and the nature of iron species in solution. Zhang et al. gave the optimal pH for phenol degradation by ferromagnetic nanoparticles catalysis was restricted around 3 with the removal rate about 98%.⁴³ When the pH ranged 4 to 10, the phenol removal rate was sharply decreased below 20%.⁴³ Li and coworkers also found similar results with that of Jinbin Zhang's result. The optimum efficiency of photodegradation of bisphenol A with iron oxides and oxalate in aqueous solution also ranged 3–4.⁴⁴ While the influence of pH on degradation of endocrine disruptors by other AOPs shows different trends. Coleman et al. also investigated the effect of pH on the photocatalytic degradation of estradiol by immobilized TiO₂ under 150 W xenon lamp. Degradation rate of E2 increased with increasing pH up to 7, then with a sharp drop to pH 10 followed by a rise to pH 12.⁴⁵ The influence of pH (3 to 11) on the formation of Endocrine Disruptor intermediates in the ozonation process showed that different pH values thereby led to different reaction paths and reaction rates. This phenomena likely arisen from the different proportions of oxidizing species (principally molecular ozone and hydroxyl radicals) present in the solution at different pH.⁴⁶

Fig. 5 The effect of pH on iron leaching and pH variations during photodegradation of E2. [E2]_o = 272 µg/L; [H₂O₂] = 9.7 mmol/L; [α-FeOOHR] = 5 g/L; T = 20 °C; pH = 3.07, 7.47 and 11.00, respectively.

Additionally, the dissolved iron and pH variation were also discussed during photo-degradation of E2. The iron leaking increased with the degradation proceeding. The biggest iron leaking after 8 h irradiation were 1.5509, 0.9504 and 0.8265 mg/L at pH 3.07, 7.47 and 11.00, respectively. About the pH change during degradation can be divided into two kinds. At acid condition, the pH of effluent has slight increased from pH 3.07 to 4.08 due to the generation of the OH⁻. At neutral to basic condition, the pH of effluent has slight decreased from pH 7.47 to 7.25 and from 11.00 to 8.73, which might be caused by E2 degradation products, such as aliphatic acids and carboxylic acids, decreasing the pH and increasing the solubility of iron. During E2 degradation, the iron leaking was also observed at different pH. The dissolution of iron oxide is known as proton-promoted dissolution reaction as described by Stumm.⁴⁷ While Huang et al. pointed out the iron oxides are dissolved by intermediate degradation products, such as organic acids, via a non reductive dissolution pathway by a mechanism that is similar to that of the dissolution of iron by oxalic acid. For example, some of the intermediates of phenol that are derived from phenol degradation, such as catechol and 1,4-hydroquinone, reduce Fe(III) to Fe(II) and result in the reductive dissolution of iron.⁴⁸ In this study, whether the dissolution of iron occurs either directly or by reduction to the more soluble iron (2+), the amount of dissolved iron was quite low. Therefore, the proceeding of homogeneous Fenton reactions can be neglected, which has proved by our paper.²⁰ Similar result was also found in J. Kochany's work. In acidic media, some ferrous ions are regenerated through much slower reduction of Fe³⁺ ions. However, under nearly neutral pH conditions, ferric (Fe³⁺) ions exist mostly as hydroxy complexes and most importantly, as insoluble oxide-hydroxide phases FeOOH, Fe(OH) etc. The precipitated species do not re-dissolve readily and they do not participate in the reduction steps.⁴⁹ Combing the other studies, the E2 degradation mechanism was given in Scheme 1.^26,48,50,51

Scheme 1 The simplified reaction pathways

Experimental

Chemicals

17β-estradiol (>98%; Sigma) was used without further purification. Adsorbent/catalyst α-FeOOHR was prepared according to reference 20. All the other chemicals used were analytical grade and all the solvents used were HPLC grade. The water employed was purified by a Milli-Q system with a resistivity higher than 18 MΩ·cm⁻¹. The resin was Amberlite® 200 with matrix of styrene–divinylbenzene, Na⁺-form, strongly acidic and particle size of 20–50 mesh (0.297–0.840 mm) (Fluka). 0.1 mM stock solution was prepared by dissolving desired amount of E2 into methanol and stored in 4 °C. Working solutions were followed by diluting stock solution with Milli-Q water to desired concentrations.

Adsorption of 17β-estradiol on catalyst

The capacity of adsorption of α-FeOOHR was determined experimentally by contacting the 25 different concentrations of E2 solutions of 272, 136, 27.2, 2.72, 0.272 and 0.0272 µg/L, respectively, with α-FeOOHR dosages in the range of 2 g/L to 50 g/L without adjusting initial pHs. The flasks were shaken at 20 °C for 24 h. Afterwards, the liquid was filtered for determination of the remaining E2 concentrations using a Liquid chromatography/triple quadrupole tandem mass spectrometry, equipped with turbo ion spray interface (API2000 LC/MS/MS system, Applied Biosystems Asia Pte Ltd., USA).

Heterogeneous photo-Fenton degradation experiments

The photochemical reactor was made of open cylindrical Pyrex with diameter of 19 cm and height of 9 cm and equipped with a magnetic stirring bar. 1L of desired concentration E2 aqueous solution was prepared without adjusting pH prior to addition of required amount of α-FeOOHR (5 g/L, 0.5 gFe/L) and H₂O₂ (9.7 mM). The mixture was magnetically stirred at 200 rpm at 20 °C air cooled room. The irradiation was carried out with two 15 W black light lamps with the irradiation intensity at the center of the reactor of 0.3mW/cm² (measured by a UV radiometer, IL700, International Light, USA.) and the main emission wavelength of 365 nm. The distance between light source and the surface of the solute was 5 cm. About 25 aliquot of the suspension was collected at regular intervals and analyzed for subsequent residual E2 concentrations, pH and total dissolved iron by atomic absorption spectrum (Thermo, USA). The photo-degradation were carried out using different E2 concentrations, organic co-solvents of methanol or acetonitrile water solutions, background substances including estrogens, humic acid (HA) and water matrices were investigated through photo-Fenton experiments in batch reactor.

Conclusion

Iron oxide can react with H₂O₂ can quickly oxidize a great number of organic compounds, providing a simple and economic source of hydroxyl radicals. The main conclusions are as follows:

In this study, the photo-degradation kinetics of E2 by heterogeneous Fenton catalyst a-FeOOHR obeyed the pseudo-first order kinetic model well whether in methanol–water solution or in acetonitrile–water solutions. Due to the nonselectivity of hydroxyl radicals, methanol showed greater retardation to degradation rate of E2 compared with acetonitrile in different composition of reaction solutions. The organic solvent with low level concentration as far as possible and a lower scavenging property towards ˙OH should be recommended for this purpose to avoid E2 being adsorbed by reactor. The presence of low level estrogens hardly competes with E2 for reactive sites. Because of the coexistence of a low concentration of HA and anions in drinking water, the percentage of quenched hydroxyl radicals in side reactions is mitigated to some extent, and the treatment efficiency remains constant with a constant delivery rate of Fenton reagent within a certain range. The pH was another important factor that greatly influenced the E2 degradation efficiency and iron leaking. With an increasing pH the E2 removal efficiency decreased, signifying that there is also a decrease of hydroxyl radicals. Results showed that careful optimization of experimental conditions is required to degrade E2 with Fenton reagent.

Acknowledgements

Research support from NSF of China (20707006) and SRF for ROCS, SEM (2008) is greatly appreciated.

References

1. K. Mukai, K. Daifuku, S. Yokoyama and M. Nakano, Biochem. Biophysic. Acta., 1990, 1035, 348–352.
2. E. Niki and M. Nakano, Methods Enzymol., 1990, 186, 330–333.
3. J. E. Harries, T. Runnalls, E. Hill, C. A. Harris, S. Maddix, J. P. Sumpter and C. R. Tyler, Environ. Sci. Technol., 2000, 34, 3003–3011.
4. A. C. Johnson and J. P. Sumpter, Environ Sci Technol., 2001, 35, 4697–4703.
5. E. J. Rosenfeldt and K. G. Linden, Environ. Sci. Technol., 2004, 38, 5476–5483.
6. Y. Zhang, J. L. Zhou and B. Ning, Water Res., 2007, 41, 19–26.
7. H. Katsumata, S. Kawabe, S. Kaneco, T. Suzuki and K. Ohta, J. Photochem. Photobiol., A, 2004, 162, 297–305.
8. F. Herrera, A. Lopez, G. Mascolo, P. Albers and J. Kiwi, Appl. Catal., B, 2001, 29, 147–162.
9. H. Kušić, N. Koprivanac and I. Selanec, Chemosphere, 2004, 57, 711–719.
10. F. Martínez, G. Calleja, J. A. Melero and R. Molina, Appl. Catal., B, 2007, 70, 452–460.
11. M. A. De León, J. Castiglioni, J. Bussi and M. Sergio, Catalysis Today, 2008, 133–135, 600–605.
12. M. Inbasekaran and S. Meenakshisundaram, Sol. Energy Mater. Sol. Cells, 2008, 92, 857–863.
13. S. Shin, H. Yoon and J. Jang, Catal. Commun., 2008, 10, 178–182.
14. G. Armstrong, Nature Chemistry, 2008 Published online, 24th October.
15. J. D. Laat and T. G. Le, Appl. Catal., B, 2006, 66, 137–146.
16. X. Feng, J. Tu, S. Ding, F. Wu and N. Deng, J. Hazard. Mater., 2005, B127, 129–133.
17. K. Mitamura, H. Narukawa, T. Mizuguchi and K. Shimada, Anal. Sci., 2004, 20, 3–4.
18. L. D. Palma, P. Ferrantelli and E. Petrucci, J. Hazard. Mater., 2003, B99, 265–276.
19. P. Tomasz and S. Andrzej, J. Mol. Catal. A: Chem., 2003, 194(1–2), 1–11.
20. Y. P. Zhao and J. Y. Hu, Appl. Catal., B, 2008, 78(3–4), 250–258.
21. N. Barka, A. Assabbane, A. Nouna and Y. A. Chou, J. Hazard. Mater., 2008, 152, 1054–1059.
22. S. S. Lin and M. D. Gurol, Environ. Sci. Technol., 1998, 32, 1417–1423.
23. R. J. Watts, Remediation, 1992, 413–425.
24. C. Y. Kwan and W. Chu, Water Res., 2003, 37, 4405–4412.
25. R. Maciel, G. L. Sant'Anna Jr. and M. Dezotti, Chemosphere, 2004, 57(7), 711–719.
26. T. L. P. Dantas, V. P. Mendonc, H. J. Jośe, A. E. Rodrigues and R. F. P. M. Moreira, Chemical Engineering Journal, 2006, 118, 77–82.
27. P. Kajitvichyanukul, M.-C. Lu and A. Jamroensan, J. Environ. Manage., 2008, 86, 545–553.
28. S. G. Huling, R. G. Arnold and R. A. Sierka, United States patent, 2004, 7335246.
29. G. V. Buxton, C. L. Greenstock, W. P. Helman and A. B. Ross, J. Phys. Chem. Ref. Data, 1988, 17(2), 513–886.
30. C. Walling, Acc. Chem. Res., 1975, 8, 125–131.
31. R. H. Werner, Environ. Sci. Technol., 1992, 26(5), 1005–1013.
32. V. Flotron, C. Delteil, A. Bermond and V. Camel, Polycyclic Aromat. Compd., 2003, 23(4), 449–459.
33. N. Maureen, I. Keisuke and G. E. Mohamed, Ozone: Sci. Eng., 2008, 30(4), 249–255.
34. P. L. Brezonik and J. Fulkerson-Brekken, Environ. Sci. Technol., 1998, 32(19), 3004–3010.
35. M. E. Lindsey and M. A. Tarr, Wat. Res., 2000, 34, 2385–2389.
36. H. Fu, X. Quan and H. Zhao, J. Photochem. Photobiol., A, 2005, 173, 143–149.
37. B. M. Voelker and B. Sulzberger, Environ. Sci. Technol., 1996, 30(4), 1106–1114.
38. A. Georgi, A. Schierz, U. Trommler, C. P. Horwitz, T. J. Collins and F.-D. Kopinke, Appl. Catal., B, 2007, 72, 26–36.
39. R. Andreozzi, V. Caprio, A. Insola and R. Marotta, Catal. Today, 1999, 53, 51–59.
40. Z. Grigory and S. Raphael, Water Res., 2008, 42, 3848–3856.
41. W. Z. Tang and C. P. Huang, Environ. Technol., 1996, 17, 1371–1378.
42. A. Albert, E. P. Serjeant, 1962, Chapman and Hall, London.
43. J. Zhang, J. Zhuang, L. Gao, Y. Zhang, N. Gu, J. Feng, D. Yang, J. Zhu and X. Yan, Chemosphere, 2008, 73, 1524–1528.
44. F. B. Li, X. Z. Li, X. M. Li, T. X. Liu and J. Dong, J. Colloid Interface Sci., 2007, 311, 481–490.
45. H. M. Coleman, B. R. Eggins, J. A. Byrne, F. L. Palmer and E. King, Appl. Catal., B, 2000, 24(1), L1–L5.
46. D. Bila, A. F. Montalvao, M. Dezotti, 2005, IOA 17th World Ozone Congress, Strabourg, France.
47. W. Stumm, 1992, Wiley/interscience, New York.
48. C.-P. Huang and Y.-H. Huang, Appl. Catal., A, 2008, 346, 140–148.
49. J. Kochany and E. Lipczynska-Kochany, Environ. Technol., 2007, 28, 1007–1013.
50. M. Lu, J. Chen and H. Huang, Chemosphere, 2002, 46, 131–136.
51. J. A. Melero, G. Calleja, F. Martĺnez, R. Molina and M. I. Pariente, Chemical Engineering Journal, 2007, 131, 245–256 NDRL/NIST Solution Kinetics Database on the Web, http://www.kinetics.nist.gov/solution/index.php.

---

Footnote

† Part of a themed issue dealing with water and water related issues.

---