Heterogeneous Fenton oxidation of paracetamol using iron oxide(nano)particles

OATAO is an open access repository that collects the work of Toulouse

researchers and makes it freely available over the web where possible

Any correspondence concerning this service should be sent

to the repository administrator:

tech-oatao@listes-diff.inp-toulouse.fr

This is an author’s version published in: http://oatao.univ-toulouse.fr/20826

To cite this version:

Velichkova, Filipa Aleksandrova

and Julcour-Lebigue, Carine

and Koumanova, Bogdana and Delmas, Henri

Heterogeneous

Fenton oxidation of paracetamol using iron oxide(nano)particles.

(2013) Journal of Environmental Chemical Engineering, 1 (4).

1214-1222. ISSN 2213-3437

Heterogeneous Fenton oxidation of paracetamol using iron oxide (nano)particles

F. Velichkova

a,b

, C. Julcour-Lebigue

b,

*

, B. Koumanova

a

, H. Delmas

b

a_{Department of Chemical Engineering, University of Chemical Technology and Metallurgy, 8 Kliment Ohridsky Blvd., 1756 Sofia, Bulgaria} b_{Universite´ de Toulouse, Laboratoire de Ge´nie Chimique, 4 alle´e Emile Monso, 31432 Toulouse, France}

Introduction

For more than a decade, the presence of pharmaceutical compounds in water resources has been considered as a worrying environmental issue. Due to their increasing input and persistence in the environment, they have been detected in surface water[1], groundwater [2]and even in drinking water[3]. Many of these substances are suspected to have adverse environmental and health effects, although still not clearly defined [3]. Beside, available data clearly indicate that some pharmaceutical com-pounds develop multi-resistant strains in microorganisms [4], exert toxic effects on algae and invertebrates[5] and affect the endocrine system of fishes[6].

The pharmaceutical compounds are designed to have a complex structure and as a consequence they are not completely removed by conventional wastewater treatment techniques. On the other hand, Advanced Oxidation Processes (AOPs) have proved to be efficient for the elimination of pharmaceuticals from water, either as an alternative or a complement to the conventional treatments

[7,8]. With AOPs used as a pre-treatment step, the pollutants can be oxidized to by-products that are easily biodegradable and less toxic, preventing the death of microorganisms in the subsequent biological treatments[8].

Among these processes, Fenton and Fenton-like oxidations have retained attention of many researchers, as they use easy to handle reagents (Fe2+ _{or Fe}3+ _{and H}

2O2), require very little energy

compared to other oxidation technologies and are able to destroy a large number of organic compounds. They involve numerous parallel and consecutive reactions, but are generally described by the following dominant reactions[9]:

Fe2þ_þH

2O2!Fe3þþOH þ OHÿ (1)

Fe3þ_þH

2O2$Fe2þþOOH þ Hþ (2)

in which reaction(2)is several orders of magnitude slower than reaction(1).

The generated hydroxyl radicals (

OH) are powerful oxidant species, which exhibit a very high oxidation potential of 2.80 V, only being surpassed by fluorine[10].

However the homogeneous Fenton oxidation suffers from a few drawbacks[7,11]: (i) the recovery of dissolved iron ions from the treated solution requires additional treatments, (ii) Fenton catalytic cycle is restrained by the formation of stable iron complexes with phosphates anions or generated carboxylic acids and (iii) the process needs to be operated in a narrow pH range, between 2 and 4, to avoid the formation and subsequent precipitation of iron oxyhydroxides. To overcome these drawbacks an adequate solution seems to be the use of heterogeneous catalysts in the so-called heterogeneous Fenton-like oxidation, which also might avoid the initial acidification.

A R T I C L E I N F O Keywords: Water treatment Advanced oxidation Paracetamol Iron oxide Magnetite Maghemite A B S T R A C T

100 mg Lÿ1_{paracetamol aqueous solutions were treated by heterogeneous Fenton oxidation at acidic pH} (2.6). Three types of iron oxides – nano- and submicro-structured magnetite, nanostructured maghemite – were tested as catalysts for that purpose. For each system, the paracetamol conversion and mineralization yield (Total Organic Carbon removal) were evaluated, as well as the catalyst stability upon recycling.

The influence of reaction parameters such as temperature, iron amount, and hydrogen peroxide dosage was also investigated. Paracetamol mineralization was improved by high temperature and low oxidant dosage due to radical scavenging effects. In best conditions (two times the stoichiometric amount of H2O2, a temperature of 60 8C, a catalyst concentration of 6 g Lÿ1), paracetamol was fully degraded after 5 h, but total mineralization was not yet achieved: TOC removal reached about 50% when magnetite powders were used as catalysts. All iron oxides exhibited low iron leaching (<1%) and stable catalytic activity upon first recycling.

* Corresponding author. Tel.: +33 534323709; fax: +33 534323700.

E-mail address:carine.julcour@ensiacet.fr(C. Julcour-Lebigue).

Zero-valent iron nanoparticles[12,13], iron containing clays

[14,15], iron oxide immobilized on zeolites[11,16]and activated carbons[17,18], iron oxide minerals[19–21]have been investi-gated as heterogeneous Fenton catalysts. The latter are especially interesting, because they are ubiquitous in the environment and can be easily applied in in-situ remediation processes[20].

Paracetamol or acetaminophen is a mild analgesic and anti-inflammatory, commonly used for humans and animals. Concen-trations up to 6

m

g Lÿ1have been reported in European sewage

treatment plant effluents[22], up to 10

m

g Lÿ1in natural waters in

USA[23], and up to 61

m

g Lÿ1in a WWTP located in East Lansing,

Michigan[24]. Moreover it is transformed into toxic compounds by the chlorination process used in wastewater treatment plants[25]. Almost all past research pertaining to the removal of paracetamol by the Fenton process used soluble Fe2+as catalyst

[26–30]. A recent successful example can be found in the work of Almeida et al. [29] on solar photoelectro-Fenton process: the authors obtained a complete degradation of paracetamol and 75% reduction of TOC after 120 min. Other heterogeneous advanced oxidation processes have been also investigated for its remedia-tion: photocatalysis using TiO2 [31–33] and oxidative process

using zero valent aluminum-acid system[34].

The aim of the present work is to investigate the degradation of paracetamol by heterogeneous Fenton oxidation using three types of iron oxides as catalysts: nano- and submicro-structured magnetite powders (MGN1 and MGN2 respectively) and nano-structured maghemite powder (MGM). The influence of catalyst amount, H2O2dosage, as well as temperature is also considered.

Materials and methods

Chemicals

Paracetamol (N-(4-hydroxyphenyl)acetamide) was supplied by BioXtra with a purity 99.0%.

Hydrogen peroxide (Ph Eur; 30% w/w) and magnetite or maghemite were used as Fenton’s reagents. These minerals have the same spinel structure and are ferromagnetic. This property makes them particularly interesting as they could be magnetically separated from the treated solution. Their main difference relies in the oxidation state of iron. Three catalyst powders supplied by Sigma-Aldrich were investigated here: MGN1 (Fe3O4

pow-der < 50 nm, purity  98%), MGN2 (Fe3O4powder < 5

m

m, purity

of 95%) and MGM (

g

-Fe2O3powder < 50 nm).

During the Fenton experiments, samples of the reactant mixture were withdrawn at different reaction times and mixed with a quenching solution containing KI (0.1 mol Lÿ1_{), Na}

2SO3

(0.1 mol Lÿ1_{) and a phosphate buffer (mixture of KH} 2PO4,

0.05 mol Lÿ1 _{and Na}

2HPO4
2H2O, 0.05 mol Lÿ1). KI and Na2SO3

reduce H2O2, while the phosphate buffer precipitates leached iron

[35]. All these reagents were analytical grade and supplied by Sigma-Aldrich.

Experimental setup

The experimental setup used for Fenton oxidation consisted in a 650 mL stirred Pyrex reactor, equipped with a jacket to maintain the temperature of the solution. For all the experiments, 500 mL of paracetamol solution with an initial concentration of 100 mg Lÿ1 were used. Agitation speed was set to 350 rpm. The initial pH of the solution was adjusted to 2.6 with 10% H2SO4, after which the

catalyst powder was added. The suspension was stirred during 90 min to ensure preliminary adsorption equilibrium, as confirmed by the chromatography analysis of samples withdrawn at different times (0, 30, 60 and 90 min). Then H2O2was added to initiate the

Fenton oxidation.

During the 5 h oxidation, 7 mL aliquots were withdrawn at selected time intervals (after 5, 10, 60, 180 and 300 min of oxidation process), treated with the quenching solution, filtered through 0.2

m

m nylon membrane syringe filters and analyzed.

Paracetamol conversion was monitored by liquid chromatography and the extent of mineralization by Total Organic Carbon (TOC) analyzer. The experimental runs were performed at different temperatures (30 8C and 60 8C). Concentrations of catalyst (1, 3 and 6 g Lÿ1_{) and H}

2O2(28 and 153 mmol Lÿ1, corresponding to 2 and

11 times the stoichiometric amount of H2O2required for complete

mineralization, respectively) were also varied for the optimization of the degradation process. The stoichiometric amount of H2O2

required for the total oxidation of paracetamol was calculated based on the following equation:

C8H9NO2þ21H2O2¼ 8CO2þHNO3þ25H2O (3) Analytical methods

Analysis of aqueous solution samples

Prior to TOC analysis, 5 mL of the reacting solution were treated with 2 mL of the quenching solution (KI, 0.1 mol Lÿ1_{; Na}

2SO3,

0.1 mol Lÿ1_{; KH}

2PO4, 0.05 mol Lÿ1; Na2HPO4
2H2O, 0.05 mol Lÿ1) to

prevent any further oxidation by reducing remaining oxidant and precipitating dissolved iron. The sample was filtered and diluted with ultrapure water to 24 mL. The TOC concentration was obtained from the difference between Total Carbon and Inorganic Carbon concentrations measured with a Shimadzu TOC-VCSNanalyzer.

Paracetamol concentration was measured by a high perfor-mance liquid phase chromatograph (HPLC) with UV detection (UV6000 diode array detector, Thermo Finnigan). The separation was achieved on a C18 reverse phase column (ProntoSIL C18 AQ) using an isocratic mobile phase (10/90 mixture of acetonitrile and deionised water acidified at pH 2.0 with H3PO4) fed at 1 mL minÿ1.

The wavelength was set to 254 nm for paracetamol detection. Quantification was made from a calibration curve periodically updated with fresh standard solutions. In this case, only the phosphate buffer (0.5 mL) was added to the reacting solution (1 mL) in order to precipitate Fe2+_{and avoid any interference of KI}

and Na2SO3. Samples were injected in the chromatograph

immediately after the addition of the buffer solution.

At the end of the experiments, residual H2O2was titrated with a

0.02 M permanganate solution (standardized with sodium oxalate prior to use).

To evaluate Fe leaching, the concentration of dissolved iron was measured in the final solution, after filtration through a 0.2

m

m

cellulose acetate membrane, by Inductively Coupled Plasma Atomic Emission Spectroscopy (Ultima 2, Horiba Jobin-Yvon).

Characterization of the iron oxide catalysts

X-ray diffraction (XRD) was used to check for the presence of crystalline impurities and calculate crystallite size from Scherrer method. XRD patterns were collected on a theta/theta powder X-ray diffraction MPD Pro system (PANalytical Company) equipped with a fast X’celerator detector. The diffractograms were recorded in 5–708 2

Q

range with a 0.0178 step size.

Brunauer–Emmett–Teller (BET) surface area of the iron oxide catalysts was measured using nitrogen adsorption on a Micro-meritics ASAP 2010 instrument. The solids were degassed 2 h at 60 8C prior to the measurement.

Scanning Electron Microscopy (SEM) was used to determine the morphology of the catalysts. The SEM images were obtained on a LEO 435 VP apparatus.

The particle size distribution of the iron oxide agglomerates in solution was analyzed by laser diffraction on a Mastersizer 2000 (Malvern Co) instrument.

The values of pHpzc, pH at the point of zero charge, were

measured by mass titration method[36]. Results and discussion

Characterization of iron oxide catalysts

The diffractograms of all iron oxides are shown inFig. 1. The diffraction pattern of nano- and submicro-structured Fe3O4could

be associated to magnetite (ICDD cards 00-001-1111 and 01-085-1436, respectively) and that of nanostructured Fe2O3 to

maghe-mite (ICDD card 00-039-1346). No peaks attributable to other phases were observed. Note that magnetite and maghemite cannot be easily differentiated from each other by XRD analysis.

Diffraction peaks of MGN1 are broader and lower in intensity as compared to MGN2, which results from smaller crystallites. Their average size (dc) was calculated from the full width at

half-maximum of the strong intensity peaks using Scherrer equation, giving a ratio of about 2 between the two magnetite powders.

The elemental particle diameter (dep) could be also calculated

from BET surface area (SBET) assuming spherical shape: dep= 6/ SBET

r

where

r

is the density of the oxide. All these properties are

listed in Table 1. For MGN1 and MGM, DRX and BET size measurements were in good agreement, confirming that their elemental particles consist in nanosized single crystallites. Conversely it could be concluded that MGN2 is composed of submicronic polycrystalline grains.

This was confirmed by SEM analysis which shows that MGN2 is a homogeneous powder formed by small rounded grains of about 200 nm in size (Fig. 2B). MGN1 and MGM (Fig. 2A and C) are polydispersed powders consisting in porous agglomerates up to 100 and 50

m

m, respectively.

Suspension of MGN1 in osmosed water showed a broad monomodal distribution of particle sizes (2–300

m

m), while

MGN2 and MGM exhibited bimodal distributions, most of the agglomerates ranging between 1 and 10

m

m. Their apparent

volume mean diameter (d43) is also given inTable 1. Selection of iron oxides as catalysts in Fenton oxidation Decomposition of hydrogen peroxide over iron oxides

In order to investigate the capacity of the different iron oxides to produce hydroxyl radicals, decomposition of H2O2 (without

pollutant) was first investigated. All the experiments were conducted at 60 8C with 6 g Lÿ1_{of catalyst and an initial pH set}

to 2.6. The corresponding results are presented inFig. 3. No H2O2

consumption was observed in absence of catalyst. In the heterogeneous Fenton type process, the reaction between ferrous or ferric ions and hydrogen peroxide takes place at the surface of the solid catalyst and then depends on the specific surface area (SSA) of the catalyst[14]. As expected MGN2 which exhibits the lowest SSA also yielded the lowest H2O2 decomposition rate.

Surprisingly MGN1 and MGM nanopowders showed very similar activities despite their different oxidation states. One reason could be the fast conversion of magnetite into maghemite by oxidation

with H2O2[37]. The catalytic reaction obeyed first-order kinetics

with respect to H2O2 concentration and Table 2 gives the

corresponding kinetic constants both on a weight basis and on a surface basis. When referring to the constant per surface area unit MGN2 appears in fact more active than MGN1. It might be explained by the presence of a surfactant on the surface of the nanostructured oxide inhibiting its activity (carbon being detected by EDX analysis), or by some diffusional limitations as MGN1 agglomerates are much larger than those of MGN2 (Table 1).

Effect of iron oxide type and concentration on the degradation of paracetamol

Before oxidation was started by addition of H2O2, adsorption

equilibrium was achieved by 90 min of equilibration. No signifi-cant adsorption of paracetamol was observed for any of the catalysts (with less than 10% paracetamol uptake from the aqueous phase for a catalyst concentration of 6 g Lÿ1_).

The effect of iron catalyst concentration on the evolution of paracetamol and TOC concentrations is depicted inFigs. 4 and 5, respectively. Other conditions were: temperature of 60 8C, H2O2

concentration of 153 mmol Lÿ1_{, and initial pH of 2.6.}

The experiments with 6 g Lÿ1_{of each catalyst were duplicated,}

showing mean deviations of less than 10% for paracetamol and TOC concentrations as a function of time. Without iron oxide, paracetamol was not significantly oxidized by H2O2 (less than

1% conversion after 5 h). For all the catalysts, a complete degradation of paracetamol was achieved after 5 h of oxidation with 6 g Lÿ1_{of iron oxide, while mineralization yield reached 43%,}

34% and 39% with MGN1, MGN2 and MGM, respectively. The conversion of TOC leveled off after long reaction times as the result of surface passivation or formation of refractory oxidation intermediates.

Table 1

Physicochemical properties of iron oxide catalysts.

Label Chemical formula dc[nm]a SBET[m2/g] dep[nm]b d43[mm]c _{%Fe [%wt]}d _pH pzc

MGN1 Fe3O4 29 39.6 (0.3) 29 90.0 67.5 6.2 (0.1)

MGN2 Fe3O4 67 5.84 (0.07) 208 12.9 69.7 7.4 (0.1)

MGM Fe2O3 48 35.2 (0.3) 35 4.8 69.2 4.6 (0.1)

a _{Mean diameter of crystallites calculated from XRD pattern using Scherrer’s equation.} b _{Mean diameter of elemental particles calculated from S}

BETassuming spherical shape. c _{Volume mean diameter of agglomerates in solution from laser diffraction measurement.} d _{As given by the supplier (from titration with Na}

2S2O3) – theoretical values for magnetite and maghemite: 72.4% and 69.9%, respectively.

[(Fig._1)TD$FIG]

As previously found for H2O2decomposition, MGN1 exhibited a

better performance than MGN2 at the same concentration. However comparing paracetamol conversion obtained with MGN1 at 1 g Lÿ1 and MGN2 at 6 g Lÿ1 (see small chart in

Fig. 4B), MGN1 seemed again less active on surface area basis. A higher concentration of catalyst insures a higher number of active sites which decompose H2O2, but also a higher

concentra-tion of leached iron, promoting homogeneous Fenton reacconcentra-tion. However above 3 g Lÿ1 _{the effect of catalyst content tended to}

vanish for MGN1 and MGM nanopowders (cf.Figs. 4A, 5A and 4C, 5C, respectively). Such behavior or even a reaction inhibition at high iron oxide concentration has been already reported[19,37]

and attributed to radical scavenging by the oxide surface.

Leaching tests

Iron leaching can occur due to two main mechanisms: ‘‘standard’’ dissolution by protons at low pH and promoted dissolution with the help of complexing agents, such as oxalic acid formed as oxidation intermediate. It is a surface mechanism in which the oxalate ion binds to one or two iron atoms, which weakens the Fe–O bond more efficiently than water alone, thereby increasing dissolution kinetics[38,39]. Whether it is free or as a complex, dissolved iron will play a different role in the oxidation process efficiency.

At the end of the experiments, the total amount of iron in solution was measured by ICP/AES. Moreover homogeneous Fenton oxidation was performed using the final solutions obtained after the heterogeneous reactions, as believed to be more representative than synthetic Fe2+_{or Fe}3+_{solutions with the same}

iron content (see also Section ‘‘Comparison to the homogeneous system’’). To account for the effect of dissolved iron only, the oxides

were filtered off and the solution passed through a membrane with 0.2

m

m pore size. In all cases, the pH of the filtrate was found

almost equal to 2.6. Paracetamol and H2O2concentrations were

titrated and readjusted to match previous initial conditions.

Fig. 6compares the conversion of paracetamol observed with 1 g Lÿ1 _{of oxides and corresponding filtrates (containing 2.9, 1.7}

and 5.5 mg Lÿ1 _{of dissolved iron for MGN1, MGN2 and MGM,}

respectively).

Iron species leached from magnetite powders showed a negligible activity, either as the result of their too small concentration to allow the reaction to proceed within a reasonable period of time (a minimal threshold concentration of 3–15 mg Lÿ1

is usually reported[40]) or because they formed stable complexes with oxidation by-products.

Conversely, a significant conversion of paracetamol was observed when using the filtrate from MGM, whereas the concentration of iron was in the same order of magnitude. Therefore the contribution of the homogeneous Fenton reaction could not be neglected in this case and might also explain the apparent high activity of this oxide as compared to magnetite.

Catalyst stability

Recycling tests were performed in order to evaluate the stability of the iron oxides during successive Fenton reactions and the possibility of their use in a continuous membrane coupling process. The catalysts from 6 g Lÿ1_{experiments were filtered off,}

washed with distilled water several times and dried at 70 8C for 24 h. Then aliquots of these spent catalysts (equivalent to 3 g Lÿ1₎

were recycled.Fig. 7compares the activity of recycled catalysts to that of fresh ones. All the catalysts show a good stability upon recycling. Yet MGM was found slightly less active in the second run (with a final TOC conversion of 34% against 43% in the first run,

Fig. 7B). This could be surprising as this oxide should be more chemically stable. As shown in Table 3, a slightly lower concentration of leached iron was found in the second solution. This might result in a lower contribution of the homogeneous Fenton reaction that proved to be significant (cf. Section ‘‘Leaching tests’’). It can be also concluded that MGN1 and MGN2 did not exhibit significant surface passivation as shown by an unchanged – or even improved – initial rate of paracetamol oxidation on recycled magnetites (Fig. 7A).

[(Fig._2)TD$FIG]

Fig. 2. SEM images of (A) MGN1, (B) MGN2 and (C) MGM.

[(Fig._3)TD$FIG]

Fig. 3. Decomposition of hydrogen peroxide (without pollutant). Operating conditions: metal oxide concentration, [MO] = 6 g Lÿ1_{; initial concentration of}

hydrogen peroxide, [H2O2]0= 153 mmol Lÿ1; initial pH, pH0= 2.6; temperature, T = 60 8C.

Table 2

First-order kinetic constants of H2O2decomposition over the three oxides (same

operating conditions as forFig. 3).

Label k (L gÿ1_minÿ1_{) k}0_{(L m}ÿ2_minÿ1₎

MGN1 2.2  10ÿ4 _{5.7  10}ÿ6

MGN2 7.2  10ÿ5 _{1.2  10}ÿ5

Analysis of reaction parameters for paracetamol degradation using nano- and submicro-structured magnetite powders as catalysts

According to previous results, maghemite was not selected for further investigation of reaction parameter due to the significant contribution of the homogeneous reaction and the slight loss of activity upon recycling.

Effect of temperature

The influence of reaction temperature on the degradation of paracetamol and its mineralization is shown inFig. 8 for both magnetite powders. Two sets of data are displayed for each catalyst at a given temperature, corresponding to almost the same ratio of catalyst to H2O2concentration. They correspond to two different

initial concentrations of H2O2, 28 and 153 mmol Lÿ1, equivalent to

2 and 11 times the stoichiometric amount for complete minerali-zation, respectively. Previous works on homogeneous Fenton reaction showed that this ratio is a critical parameter for

[(Fig._4)TD$FIG]

Fig. 4. Effect of metal oxide (MO) amount on the evolution of paracetamol concentration for (A) MGN1, (B) MGN2 and (C) MGM. The small chart enclosed in (B) compares the performance of MGN1 and MGN2. Operating conditions: initial concentration of paracetamol, [P]0= 100 mg Lÿ1; initial concentration of hydrogen

peroxide, [H2O2]0= 153 mmol Lÿ1; initial pH, pH0= 2.6; temperature, T = 60 8C.

[(Fig._5)TD$FIG]

Fig. 5. Effect of metal oxide (MO) amount on the evolution of TOC concentration for (A) MGN1, (B) MGN2 and (C) MGM (same operating conditions as forFig. 4). The small chart enclosed in (B) compares the results of MGN1 and MGN2.

optimizing the process efficiency when the Fenton’s reagent is in sufficient amount[41,42].

A higher temperature increases the rate of

OH formation as per Arrhenius law, but it also promotes the decomposition of hydrogen peroxide into oxygen and water, reducing the efficiency of its utilization[43]. The increase of temperature from 30 8C to 60 8C showed here a beneficial effect for all the investigated conditions.

[(Fig._6)TD$FIG]

Fig. 6. Effect of leached iron on the degradation of paracetamol (using filtrates from heterogeneous Fenton reaction with 1 g Lÿ1 _{MO). Operating conditions:}

[P]0= 100 mg Lÿ1, [H2O2]0= 153 mmol Lÿ1, pH0= 2.6, T = 60 8C. Time zero

corresponds to the addition of H2O2.

[(Fig._7)TD$FIG]

Fig. 7. MO stability upon recycling: evolution of (A) paracetamol concentration and (B) TOC in solution. Operating conditions: [P]0= 100 mg Lÿ1, [MO] = 3 g Lÿ1,

[H2O2]0= 153 mmol Lÿ1, pH0= 2.6, T = 60 8C.

Table 3

Concentration of leached iron measured at the end of the Fenton reaction with fresh and recycled catalysts. The values in brackets represent the percentage of leached iron with respect to the initial iron content in oxide.

Iron oxide Total concentration of iron in solution [mg Lÿ1_]

(and iron fraction in aqueous phase)

Fresh catalyst Recycled catalyst MGN1 7.4 (0.37%) 6.8 (0.34%) MGN2 3.8 (0.18%) 5.1 (0.24%) MGM 7.6 (0.37%) 5.5 (0.26%) Operating conditions: [P]0= 100 mg Lÿ1, [MO] = 3 g Lÿ1, [H2O2]0= 153 mmol Lÿ1,

pH0= 2.6, T = 60 8C.

[(Fig._8)TD$FIG]

Fig. 8. Effect of temperature on the degradation of paracetamol: evolution of paracetamol concentration with (A) MGN1 and (B) MGN2, (C) TOC removal after 5 h of oxidation. Operating conditions: [P]0= 100 mg Lÿ1, [MO]/[H2O2]01.1 g/g,

Effect of initial H2O2concentration

Fig. 8C confirms that, at a given temperature and pH, the ratio [MO]/[H2O2]0 is the key parameter that controls the

mineralization yield, no matter of the H2O2 excess amount.

Nevertheless it seems from Fig. 9A and B that the initial degradation rate of paracetamol is improved by a low amount of both reagents.

Fig. 9 illustrates the effect of H2O2 concentration (28 or

153 mmol Lÿ1_{) at a given magnetite concentration (1 or 6 g L}ÿ1_).

In all cases, using 153 mmol Lÿ1_{of H}

2O2resulted in a decrease

of both paracetamol degradation rate (Fig. 9A and B) and final TOC conversion (Fig. 9C). It was especially observed at low iron oxide concentration. This behavior may be explained by the parallel and undesired reaction between H2O2 and hydroxyl

radicals[44]:

_{OH þ H}

2O2!H2O þOOH (4)

The produced 

OOH radicals have an oxidation potential significantly lower than that of 

OH. As a consequence of this scavenging effect an excess of H2O2 led to a less efficient

degradation. Eq. (4) has a reference rate constant of 2.7  107_{L mol}ÿ1_sÿ1 _{[45], while the reaction between hydroxyl}

radicals and paracetamol exhibits a rate constant of 1.9  109_{L mol}ÿ1_sÿ1_{[46]; therefore at the highest H}

2O2

concen-tration the scavenging of

OH by H2O2becomes predominant to the

detriment of pollutant oxidation.

As expected, the consumption of H2O2when in large excess

(153 mmol Lÿ1_{) was similar with or without pollutant (cf.Figs. 3}

and 9D). Logically, it increased upon decreasing initial H2O2

concentration or increasing magnetite concentration.

Comparison to the homogeneous system

A few homogeneous experiments were conducted to assess the efficiency of the heterogeneous process. For nanosized magnetite such as MGN1 the main surface plane is (1 1 1) face associated with an iron density of 9.8 Fe atom/nm2_{[47]. Therefore 1 g L}ÿ1_{of iron}

oxide should correspond to 0.6 mmol Lÿ1of accessible iron on the magnetite surface.Fig. 10compares the degradation of paraceta-mol for both systems at two temperatures: 45 8C and 60 8C. The homogeneous tests were performed using either FeSO4
7H2O

(Fe2+_{) or FeCl}

3
6H2O (Fe3+) with an initial pH of 2.6. No iron

precipitate was observed in these conditions. As already reported, H2O2 being in large excess with respect of iron species, similar

rates were obtained in both cases no matter the oxidation state of iron ion in solution[48]. It is striking that, despite MGN1 yielded much lower initial rates of paracetamol and TOC degradation, similar TOC concentrations were reached after 5 h of oxidation. MGN1 seems to exhibit an induction period that may be caused by the organic stabilizer coating the particle surface. This organic matter may need to be (partially) oxidized before surface iron can be available for paracetamol degradation.

Finally two experiments were performed using 2 and 10 mg Lÿ1

of Fe3+_{(0.035 mmol L}ÿ1_{and 0.18 mmol L}ÿ1_{, respectively), which}

correspond to the lowest and highest bound of leached iron measured with magnetites. At these concentrations, iron ion proved to be already active, resulting in paracetamol depletion within about 10 min, a much lower time than observed with MGN1. Yet the final mineralization yield was lower than in case of MGN1. Therefore it confirms that Fenton oxidation over MGN1 should mainly proceed through heterogeneous mechanism.

[(Fig._9)TD$FIG]

Fig. 9. Effect of initial concentration of hydrogen peroxide on the degradation of paracetamol: evolution of paracetamol concentration with (A) MGN1 and (B) MGN2, (C) TOC removal and (D) consumption of hydrogen peroxide after 5 h of oxidation. Operating conditions: [P]0= 100 mg Lÿ1, pH0= 2.6, T = 60 8C.

Conclusions

Catalytic wet peroxide oxidation of paracetamol could be successively performed using three types of iron oxides, leading to a complete conversion of paracetamol within 5 h and a TOC removal of about 50% with nano- and submicro-structured magnetite. This process was positively influenced by a rise in temperature from 30 8C to 60 8C. A small excess of oxidant should be preferred because of the occurrence of scavenging effect. The homogeneous contribution proved to be significant with maghe-mite only.

Low cation leaching and good catalytic stability are essential parameters for the process economy and its environmental impact: magnetite powders showed promising results as their activity was maintained upon first recycling while exhibiting iron loss of much less than 1%.

Acknowledgements

We are grateful to the French embassy in Sofia for the thesis scholarship of F. Velichkova and to the Working Community of the Pyrenees for financial support (CTP project no. 11051939). The authors also thank M.L. Pern, S. Desclaux, G. Raimbeaux, M.L. de Solan Bethmale, C. Rey-Rouch (LGC), L. Vendier (LCC) for their help on the characterization techniques, and J.L. Labat for the implementation of the Fenton experimental setup.

References

[1]B. Kasprzyk-Hordern, R.M. Dinsdale, A.J. Guwy, The occurrence of pharmaceuti-cals, personal care products, endocrine disruptors and illicit drugs in surface water in South Wales, UK, Water Res. 42 (2008) 3498–3518.

[2]D.J. Lapworth, N. Baran, M.E. Stuart, R.S. Ward, Emerging organic contaminants in groundwater: a review of sources, fate and occurrence, Environ. Pollut. 163 (2012) 287–303.

[3]O. Jones, J.N. Lester, N. Voulvoulis, Pharmaceuticals: a threat to drinking water? Trends Biotechnol. 23 (2005) 163–167.

[4]M. Naviner, L. Gordon, E. Giraud, M. Denis, C. Mangion, H. Le Bris, J.P. Ganie`re, Antimicrobial resistance of Aeromonas spp. isolated from the growth pond to the commercial product in a rainbow trout farm following a flumequine treatment, Aquaculture 315 (2011) 236–241.

[5]M. Crane, C. Watts, T. Boucard, Chronic aquatic environmental risks from expo-sure to human pharmaceuticals, Sci. Total Environ. 367 (2006) 23–41.

[6]W. Sanchez, W. Sremski, B. Piccini, O. Palluel, E. Maillot-Mare´chal, S. Betoulle, A. Jaffal, S. Aı¨t-Aı¨ssa, F. Brion, E. Thybaud, N. Hinfray, J.M. Porcher, Adverse effects in wild fish living downstream from pharmaceutical manufacture discharges, En-viron. Int. 37 (2011) 1342–1348.

[7]M. Klavarioti, D. Mantzavinos, D. Kassinos, Removal of residual pharmaceuticals from aqueous systems by advanced oxidation processes: a review, Environ. Int. 35 (2009) 402–417.

[8]I. Oller, S. Malato, J.A. Sa´nchez-Pe´rez, Combination of advanced oxidation pro-cesses and biological treatments for wastewater decontamination – a review, Sci. Total Environ. 409 (2011) 4141–4166.

[9]J.J. Pignatello, Dark and photo-assisted Fe3+_{-catalyzed degradation of}

chlorophe-noxy herbicides by hydrogen peroxide, Environ. Sci. Technol. 26 (1992) 944–951.

[10]L.M. Nieto, G. Hodaifa, S. Rodrı´guez, J.A. Gime´nez, J. Ochando, Degradation of organic matter in olive-oil mill wastewater through homogeneous Fenton-like reaction, Chem. Eng. J. 173 (2011) 503–510.

[11]E.V. Kuznetsova, E.N. Savinov, L.A. Vostrikova, V.N. Parmon, Heterogeneous catalysis in the Fenton-type system FeZSM-5/H2O2, Appl. Catal. B 51 (2004)

165–170.

[12]L. Xu, J. Wang, A heterogeneous Fenton-like system with nanoparticulate zero-valent iron for removal of 4-chloro-3-methyl phenol, J. Hazard. Mater. 186 (2011) 256–264.

[13]I. Grcˇic´, S. Papic´, K. Zˇizˇek, N. Koprivanac, Zero-valent iron (ZVI) Fenton oxidation of reactive dye wastewater under UV-C and solar irradiation, Chem. Eng. J. 195/196 (2012) 77–90.

[14]E.G. Garrido-Ramı´rez, B.K.G. Theng, M.L. Mora, Clays and oxide minerals as catalysts and nanocatalysts in Fenton-like reactions – a review, Appl. Clay Sci. 47 (2010) 182–192.

[15]H. Hassan, B.H. Hameed, Fe–clay as effective heterogeneous Fenton catalyst for the decolorization of Reactive Blue 4, Chem. Eng. J. 171 (2011) 912–918.

[16]M.B. Kasiri, H. Aleboyeh, A. Aleboyeh, Degradation of Acid Blue 74 using Fe-ZSM5 zeolite as a heterogeneous photo-Fenton catalyst, Appl. Catal. B 84 (2008) 9–15.

[17]T.L.P. Dantas, V.P. Mendonc¸a, H.J. Jose´, A.E. Rodrigues, R.F.P.M. Moreira, Treatment of textile wastewater by heterogeneous Fenton process using a new composite Fe2O3/carbon, Chem. Eng. J. 118 (2006) 77–82.

[18]M.A. Fontecha-Ca´mara, M.A. A´lvarez-Merino, F. Carrasco-Marı´n, M.V. Lo´pez-Ramo´n, C. Moreno-Castilla, Heterogeneous and homogeneous Fenton processes using activated carbon for the removal of the herbicide amitrole from water, Appl. Catal. B 101 (2011) 425–430.

[19]X. Xue, K. Hanna, N. Deng, Fenton-like oxidation of Rhodamine B in the presence of two types of iron (II, III) oxide, J. Hazard. Mater. 166 (2009) 407–414.

[20]S.P. Sun, A.T. Lemley, p-Nitrophenol degradation by a heterogeneous Fenton-like reaction on nano-magnetite: process optimization, kinetics, and degradation pathways, J. Mol. Catal. A: Chem. 349 (2011) 71–79.

[21]M.C. Pereira, L.C.A. Oliveira, E. Murad, Iron oxide catalysts: Fenton and Fenton-like reactions – a review, Clay Miner. 47 (2012) 285–302.

[22]T.A. Ternes, Occurrence of drugs in German sewage treatment plants and rivers, Water Res. 32 (1998) 3245–3260.

[23]D.W. Kolpin, E.T. Furlong, M.T. Meyer, E.M. Thurman, S.D. Zaugg, L.B. Barber, H.T. Buxton, Pharmaceuticals, hormones and other organic wastewater contaminants

[(Fig._10)TD$FIG]

Fig. 10. Comparison of Fenton oxidation over MGN1 and in homogeneous system: evolution of (A) paracetamol concentration at 60 8C, (B) TOC concentration at 60 8C and (C) TOC concentration at 45 8C. Other operating conditions: [P]0= 100 mg Lÿ1,

in US streams, 1999–2000: a national reconnaissance, Environ. Sci. Technol. 36 (2002) 1202–1211.

[24]P. Gao, Y. Ding, H. Li, I. Xagoraraki, Occurrence of pharmaceuticals in a municipal wastewater treatment plant: mass balance and removal processes, Chemosphere 88 (2012) 17–24.

[25]M. Bedner, W. Maccrehan, Transformation of acetaminophen by chlorination produces the toxicants 1,4-benzoquinone and N-acetyl-p-benzoquinone imine, Environ. Sci. Technol. 40 (2006) 516–522.

[26]A.G. Trovo´, S.A.S Melo, R.F.P Nogueira, Photodegradation of the pharmaceuticals amoxicillin, bezafibrate and paracetamol by the photo-Fenton process – applica-tion to sewage treatment plant effluent, J. Photochem. Photobiol. A 198 (2008) 215–220.

[27]A. Dura´n, J.M. Monteagudo, A. Carnicer, M. Ruiz-Murillo, Photo-Fenton minerali-zation of synthetic municipal wastewater effluent containing acetaminophen in a pilot plant, Desalination 270 (2011) 124–129.

[28]L. Santos-Juanes Jorda´, M.M. Ballesteros Martı´n, E. Ortega Go´mez, A. Cabrera Reina, I.M. Roma´n Sa´nchez, J.L. Casas Lo´pez, J.A. Sa´nchez Pe´rez, Economic evalu-ation of the photo-Fenton process. Mineralizevalu-ation level and reaction time: the keys for increasing plant efficiency, J. Hazard. Mater. 186 (2011) 1924–1929.

[29]L.C. Almeida, S.G. Segura, N. Bocchi, E. Brillas, Solar photoelectro-Fenton degra-dation of paracetamol using a flow plant with a Pt/air-diffusion cell coupled with a compound parabolic collector: process optimization by response surface meth-odology, Appl. Catal. B 103 (2011) 21–30.

[30]A. Bernabeu, S. Palacios, R. Vicente, R.F. Vercher, S. Malato, A. Arques, A.M. Amat, Solar photo-Fenton at mild conditions to treat a mixture of six emerging pollu-tants, Chem. Eng. J. 198/199 (2012) 65–72.

[31]X. Zhang, F. Wu, X. Wu, P. Chen, N. Deng, Photodegradation of acetaminophen in TiO2suspended solution, J. Hazard. Mater. 157 (2008) 300–307.

[32]L. Yang, L.E. Yu, M.B. Ray, Degradation of paracetamol in aqueous solutions by TiO2photocatalysis, Water Res. 42 (2008) 3480–3488.

[33]J. Radjenovic´, C. Sirtori, M. Petrovic´, D. Barcelo´, S. Malato, Solar photocatalytic degradation of persistent pharmaceuticals at pilot-scale: kinetics and characteri-zation of major intermediate products, Appl. Catal. B 89 (2009) 255–264.

[34]H. Zhang, B. Cao, W. Liu, K. Lin, J. Feng, Oxidative removal of acetaminophen using zero valent aluminum-acid system: efficacy, influencing factors, and reaction mechanism, J. Environ. Sci. 24 (2) (2012) 314–319.

[35] M. Haddou, De´gradation de de´rive´s de l’acide benzoı¨que par les proce´de´s d’oxy-dation avance´e en phase homoge`ne et he´te´roge`ne: proce´de´ Fenton, photo-Fenton

et photocatalyse, Universite´ Toulouse III – Paul Sabatier, France, 2010(PhD thesis, accessed 01.06.13)http://thesesups.ups-tlse.fr/871/.

[36]J.S. Noh, J.A. Schwarz, Estimation of the point of zero charge of simple oxides by mass titration, J. Colloid. Interface Sci. 130 (1989) 157–164.

[37]K. Rusevova, F.D. Kopinke, A. Georgi, Nano-sized magnetic iron oxides as catalysts for heterogeneous Fenton-like reactions – influence of Fe(II)/Fe(III) ratio on catalytic performance, J. Hazard. Mater. 241/242 (2012) 433–440.

[38]M.A. Blesa, H.A. Marinovich, E.C. Baumgartner, A.J.G. Maroto, Mechanism of dissolution of magnetite by oxalic acid-ferrous ion solutions, Inorg. Chem. 26 (1987) 3713–3717.

[39]W. Stumm, Reactivity at the mineral-water interface: dissolution and inhibition, Colloids Surf. A: Physicochem. Eng. Aspects 120 (1997) 143–166.

[40]N. Koprivanac, H. Kusˇic´, AOP as an effective tool for the minimization of hazardous organic pollutants in colored wastewater; chemical and photochemical process-es, in: K. Wada, Lewinsky (Eds.), Hazardous Materials and Wastewater – Treat-ment, Removal and Analysis, Nova Science Publishers, New York, 2007 (Chapter 5).

[41]I. Gulkaya, G.A. Surucu, F.B. Dilek, Importance of H2O2/Fe2+ratio in Fenton’s

treatment of a carpet dyeing wastewater, J. Hazard. Mater. B 136 (2006) 763–769.

[42]A. Bach, H. Shemer, R. Semiat, Kinetics of phenol mineralization by Fenton-like oxidation, Desalinisation 264 (2010) 188–192.

[43]M. Neamtu, C. Catrinescu, A. Kettrup, Effect of dealumination of iron(III) – exchanged Y zeolites on oxidation of Reactive Yellow 84 azo dye in the presence of hydrogen peroxide, Appl. Catal. B 51 (2004) 149–157.

[44]W.P. Kwan, B.M. Voelker, Rates of hydroxyl radical generation and organic compound oxidation in mineral-catalyzed Fenton-like systems, Environ. Sci. Technol. 37 (2003) 1150–1158.

[45]G.V. Buxton, C.L. Greenstock, W.P. Helman, A.B. Ross, Critical review of rate constants for reactions of hydrated electrons, hydrogen atoms and hydroxyl radicals (

OH/

O) in aqueous solution, J. Phys. Chem. Ref. Data 17 (1988) 513–886.

[46]L. Yang, L.E. Yu, M.B. Ray, Photocatalytic oxidation of paracetamol: dominant reactants, intermediates, and reaction mechanisms, Environ. Sci. Technol. 43 (2009) 460–465.

[47]T.J. Daou, S. Begin-Colin, J.M. Grenche, F. Thomas, A. Derory, P. Bernhardt, P. Legar, G. Pourroy, Phosphate adsorption properties of magnetite-based nanoparticles, Chem. Mater. 19 (2007) 4494–4505.

[48]R. Munter, Advanced oxidation processes – current status and prospects, Proc. Estonian Acad. Sci. Chem. 50 (2001) 59–80.