Adsorption of Fluoroquinolones from water by TiO2-modified zeolites and photodegradation under solar light

Journal of Environmental Chemical Engineering 2 (2014) 2170–2176

TiO2-modified zeolites for fluoroquinolones removal from wastewatersand reuse after solar light regeneration

Federica Maraschi a, Michela Sturini a, Andrea Speltini a, Luca Pretali b,Antonella Profumo a,*, Anna Pastorello a, Vimal Kumar a, Maurizio Ferretti c,d,Valentina Caratto c,d

aDepartment of Chemistry, University of Pavia, Via Taramelli 12, Pavia 27100, Italyb Parco Tecnologico Padano, Via Albert Einstein, Lodi 26900, ItalycDepartment of Chemistry and Industrial Chemistry, University of Genoa, Via Dodecaneso 31, Genova 16146, ItalydCNR-Spin Genoa, Corso Perrone 24, Genova 16156, Italy

A R T I C L E I N F O

Article history:Received 26 June 2014Received in revised form 30 July 2014Accepted 8 August 2014

Keywords:FluoroquinolonesEmerging pollutantsTitanium dioxideZeoliteWastewater treatment

A B S T R A C T

Adsorption and photocatalytic removal from water of marbofloxacin (MAR) and enrofloxacin (ENR), twofluoroquinolone (FQ) antibiotics widely present in surface waters, were investigated on zeolite Yderivatized with three different TiO2 catalysts (P25 Degussa and TiO2 obtained by optimization of a sol–gel method). The prepared materials were characterized by X-rays powder diffraction (XRPD), scanningelectron microscopy (SEM), BET analysis and diffuse reflectance spectroscopy (DRS). TiO2-derivatizationimproved the adsorption capacities of zeolite Y for FQs and promoted the sunlight-induced degradationof FQs adsorbed (solid-state photodegradation), not investigated in literature yet. A large FQs removal(3 mg g�1) occurred in about 2 h (residual adsorbed FQs < 0.03%). The degradation curves of FQs and theevolution profiles of the photoproducts were monitored, after microwave-assisted extraction, by high-performance liquid chromatography coupled to UV detection (HPLC-UV), thus proving the effectivenessof solar radiation in the photocatalytic removal of the drugs adsorbed on TiO2-modified zeolites, that canbe recycled for further adsorption/photodegradation cycles, with no loss of sorption capability andphotocatalytic activity. A complete removal of FQs from the outlet of a wastewater treatment plantsamples, spiked with MAR and ENR at the micrograms per litre levels, was obtained with the bestperforming modified zeolite.

ã 2014 Elsevier Ltd. All rights reserved.

Contents lists available at ScienceDirect

Journal of Environmental Chemical Engineering

journal homepage: www.elsevier .com/ locate / je ce

Introduction

Enrofloxacin (ENR) and marbofloxacin (MAR) are two syntheticantibiotics largely employed in cattle and swine farms in NorthernItaly [1]. They belong to the class of fluoroquinolones (FQs), whichare among the most important antibacterial agents used both inhuman and in veterinary medicine. After administration, FQsexcreted unchanged for a large portion [2] and are only partiallyremoved by wastewater/sewage treatment plants [3,4]. As a

Abbreviations: AOP, advanced oxidation process; ACN, acetonitrile; BET,Brunauer–Emmett–Teller analysis; DRS, diffuse reflectance spectroscopy; ENR,enrofloxacin; FD, fluorescence detector; FQs, fluoroquinolones; HPLC, highperformance liquid chromatography; IDL, instrumental detection limit; IQL,instrumental quantification limit; MAR, marbofloxacin; MDL, method detectionlimit; MQL, method quantification limit; SEM, scanning electron microscopy; XRPD,X-rays powder diffraction.* Corresponding author. Tel.: +39 0382 987581; fax: +39 0382 528544.E-mail address: [email protected] (A. Profumo).

http://dx.doi.org/10.1016/j.jece.2014.08.0092213-3437/ã 2014 Elsevier Ltd. All rights reserved.

consequence, these compounds have been detected up to few partsper billion in the environment, in particular in surface/wastewaterand soil [5–7]. FQs have been defined as emerging micropollutantsbecause they may lead to antibiotic resistance in microorganismsalso due to the formation of various photoproducts, and thus as athreat to the ecosystem and human health [8–11].

Efficient, sustainable and cost-effective methods for theremediation of such recalcitrant pollutants from contaminatedwaters are expected. In this context, new composite materialsconsisting of immobilized or N-doped TiO2 have been synthesizedas a promising photocatalysts [12–16]. In particular, TiO2-modifiedzeolites have been tested for the photocatalytic degradation ofpollutants in aqueous solution [15,16], but presently no reportshave been focused on the remediation of FQs-contaminated waterscombining TiO2-zeolite adsorption/solar light regeneration. Zeo-lites were chosen for the high adsorption capability, low cost andlarge availability [17], while TiO2 was meant to photodegrade theadsorbed drugs under solar light. Differently from the literaturecited above [15,16], TiO2-modified zeolites were used to adsorb FQs

F. Maraschi et al. / Journal of Environmental Chemical Engineering 2 (2014) 2170–2176 2171

from water and successively exposed to natural solar light forquantitatively removing the adsorbed drugs, too. Solid-statephotodegradation of adsorbed FQs allowed the reuse of zeolites.

Solar light degradation of adsorbed MAR and ENR as function ofirradiation time was investigated on zeolite Y derivatized withthree different TiO2 catalysts, previously characterized by X-rayspowder diffraction (XRPD), scanning electron microscopy (SEM),diffuse reflectance spectroscopy (DRS) and Brunauer–Emmett–Teller (BET) analysis. The formation/decomposition of FQs photo-products was monitored, after microwave-assisted extraction(MAE), by high-performance liquid chromatography coupled toUV detection (HPLC-UV).

The obtained materials were compared both in terms ofadsorption capacity and photocatalytic efficiency. Preliminaryapplications of the most performing material to real situationswere undertaken, at laboratory scale, on water samples from theoutlet of a wastewater treatment plant, fortified with realisticamounts of both FQs.

Experimental

Materials and reagents

All the chemicals employed were reagent grade or higher inquality and were used without any further purification. ENR andMAR, in the injectable form, were purchased from Bayer (Baytril25 mg mL�1) and Vétoquinol (Marbocyl 20 mg mL�1) respectively.Degussa P25 titanium dioxide (80% anatase, 20% rutile, averageparticle size 30 nm, surface area 50 � 15 m2g�1) from Degussa AG,Frankfurt, Germany. HCOOH (>95% w/w) and hexahydrate Mg(NO3)2 (97% w/w) were supplied by Fluka (Sigma–Aldrich), HPLCgradient grade acetonitrile (ACN) by VWR, H3PO4 (85% w/w) andammonia solution (30% w/w) by Carlo Erba. Zeolite Y in sodiumform, titanium (IV) tetraisopropoxide (97% v/v) and 2-propanol(99.5% v/v) were delivered by Sigma–Aldrich.

Ultrapure water (resistivity 18.2 MV cm�1 at 25 �C) wasproduced in laboratory by means of a Millipore Milli-Q system.Aqueous FQs solutions were prepared at different concentrationsand stored in dark environment/room at 4 �C before use.

Synthesis of TiO2-modified zeolite

Commercial nanopowders P25 Degussa and nanopowders TiO2

synthesized by sol–gel method [18,19] were used. Titaniumisopropoxide, 2-propanol and water (1:2:10) were vigorouslystirred for 4 h at room temperature, then the mixture wassubmitted to a thermal treatment at 100 �C for 12 h (20 �C min�1,atmospheric pressure) in order to eliminate residual water andorganic compounds, and to densify the amorphous aerogel. Thematerial was finally calcined to obtain crystalline nano-scaledpowder. Temperature tuning allowed to obtain different crystalstructures. Specifically, the treatment at 350 �C for 1 h wassufficient to achieve the complete conversion of the TiO2

amorphous phase into anatase. The average grain size was19 nm, with a standard deviation of 2 nm [19]. Experiments wereperformed on three differently derivatized zeolites Y:

� ZT1, synthesized by mixing 1 g of P25 Degussa with 10 mL waterand 0.5 g of H+ zeolite and magnetically stirring the resultantsuspension for 20 min. Powders were filtered and washed, andfinally the samples were dried in an oven at 100 �C overnight, atatmospheric pressure.

� ZT2, obtained by stirring magnetically for 20 min the resultingsuspension of 1 g of amorphous TiO2 mixed with 0.5 g H+ zeolitein 10 mL of water. Powder was filtered, washed, and finally dried

in an oven at 100 �C overnight at atmospheric pressure. Thesample was then treated at 350 �C for 1 h to convert amorphousTiO2 into anatase.

� ZT3, obtained mixing 1 g of anatase TiO2 with 10 mL water and0.5 g of H+ zeolite and magnetically stirring the resultantsuspension for 20 min. Powder was filtered, washed and finallydried in oven at 100 �C overnight at atmospheric pressure.

Characterization of TiO2-modified zeolites

Samples were characterized by XRPD (Philips PW 1830generator), SEM (Cambridge S360 microscope), BET surface areaanalysis and DRS.

Before SEM analysis, powders were coated with gold in lowvacuum (2 � 10�2mbar) to have a conductive material (Au layerthickness 10 nm).

Samples (about 1 g) submitted to BET analysis were overnightpre-heated at 200 �C under vacuum to remove adsorbed water; thevolumetric adsorption measurements were performed on Micro-meritics ASAP 2000 and the report was the specific surface area.

DRS was performed using a JASCO V-570 UV–vis-NIR spectro-photometer. Samples were prepared in tablets by pressing 400 mgof powder at 4 atm, and every scan was performed on indepen-dently prepared tablets in order to obtain accurate results. Theenergy gap (Egap) of each sample was derived according to therelation: hcl�1 = hn = Egap, by considering the wavelength corre-sponding to the edge of the experimental spectrum [20].

Sorption experiments

FQ sorption was investigated using a batch equilibrationmethod in tap water (pH 7.9 � 0.1, conductivity 271 mS cm�1,DOC 0.43 mg L�1) at FQs concentrations (C0) ranging from 20 to600 mg L�1 for MAR, and from 20 to 500 mg L�1 for ENR. 0.25 gzeolite was weighted into 50 mL PP centrifuge tubes wrapped withaluminium foils to prevent FQ light-induced decomposition, mixedwith 25 mL of FQ solutions and shaken on a rotating plate(150 rpm) at room temperature for 24 h to ensure equilibration.MAR and ENR equilibrium concentrations (Ce) were measured by aUVmini-1240 UV–vis spectrophotometer (Shimadzu Corporation)at 290 and 270 nm, respectively, on the centrifuged (4500 rpm,15 min) and filtered (a 0.22 mm membrane) suspensions. Linearregression on four standards (0 and 11 mg L�1, R2 > 0.999) gavedetection and quantification limits of 0.1 and 0.37 mg L�1 for MARand 0.2 and 0.47 mg L�1 for ENR. FQs adsorbed amounts (Cs) werecalculated from the difference between C0 and Ce. All theexperiments were performed in triplicate (n = 3).

Irradiation experiments

Zeolite samples (0.5 g) were fortified with MAR and ENR(3 mg g�1) as described in Section Sorption experiments. After 24 h,the solids were separated from the aqueous solutions by filtration(0.22 mm), air-dried and homogeneously dispersed on glass Petricapsulae (Ø 8.5 cm, depth 2 cm) to obtain monolayers (thicknessbelow 1–2 mm) to be exposed to natural sunlight (9:00 am–5:00pm) during the summer (June–September), at temperaturesranging from 25 to 35 �C. The solar power ranged from 170 to470 W m�2 (in the visible range) and from 8 to 30 W m�2 (in theUV), respectively. The flux was measured by means of an HD 9221(Delta OHM) (450–950 nm) and of Multimeter (CO.FO.ME.GRA)(295–400 nm) radiometers. At regular intervals, each one of theirradiated samples (n = 3) was extracted to determine the FQscontent by the procedure described below.

2172 F. Maraschi et al. / Journal of Environmental Chemical Engineering 2 (2014) 2170–2176

Microwave-assisted extraction and chromatographic analyses

A high performance instrument equipped with an infraredtemperature control system, stirring and cooling options (DiscoverSP1 microwave system, CEM S.r.l., Cologno al Serio, Italy) wasemployed for MAE from zeolite samples enriched with FQs (seeSection Irradiation experiments), before and after irradiation.Briefly, 16 mL of an aqueous solution 50% (w/v) Mg(NO3)2�6H2Oand 4% (v/v) NH3 were added to 0.5 g of the adsorbed zeolite in aPyrex1 tube of 35 mL and introduced into the microwave cavity.After 2 min stirring, microwave irradiation (200 W) was performedto reach a temperature of 110 �C, which was kept constant for5 min, then heating to 115 �C for 15 min. After cooling at roomtemperature, the extract was centrifuged for 10 min at 4500 rpm,filtered (0.22 mm) and acidified with H3PO4 (1:5) before HPLC-UVanalysis. HPLC-UV and HPLC–ESI-MS/MS analyses were performedfollowing the procedures previously reported [21].

Regeneration of TiO2-modified zeolite

TiO2-modified zeolite (0.5 g) was fortified with MAR and ENR(3 mg g�1) at pH 7.5 (rotating plate,150 rpm, 24 h), filtered through a0.22 mm membrane, air-dried and homogeneously dispersed onglass Petri capsulae (Ø 8.5 cm, depth 2 cm) to obtain a monolayer(thickness below 1–2 mm), exposed to sunlight outdoor (9:00 am–

5:00 pm) during the summer (June–September) for 5 h, at temper-atures ranging from 25 to 35 �C. Then, the sample was extracted andanalysed as described in Section Microwave-assisted extraction andchromatographic analyses. Then, the sorbent was air-dried andsubmitted to a further adsorption/photodegradation cycle.

Preparation of samples from the outlet of a treatment plant

Water samples were collected at the outlet of a wastewatertreatment plant located in Pavia, Northern Italy (pH 7.5 � 0.1, COD16.5 mg L�1, BOD5 < 5 mg L�1).

Aliquots of 250 mL of sample were analysed before (WW) andafter equilibration with TiO2-zeolite modified (WW + ZT2), using asolid phase extraction (SPE) procedure previously reported [1].Wastewater samples, spiked with MAR and ENR at concentrationlevels of 5 mg L�1 and 1 mg L�1 were contacted with 0.25 g of ZT2and shaken on a rotating plate (150 rpm) at room temperature, for24 h. Afterwards, samples were centrifuged and the supernatantwas analysed after 0.22 mm filtration.

Results and discussion

Catalysts characterization

The preparation of the samples ZT2 and ZT3 was simpler andreproducible compared to that of the ZT1 sample because thenanopowders obtained with the sol–gel technique were smallerand more homogeneous than the commercial P25 Degussa. This isdue to the smaller size, round shape and regular grains of anatasewhich form a spongy aggregate. In particular, in the case of ZT2 thenanoparticles were already dispersed in water as a colloidalsuspension with dimensions of few nanometers. This greatlyfacilitated the homogeneous coating of zeolite. On the contrary, theP25 particles form aggregates that were hardly disaggregated inwater. SEM analysis proved that samples ZT2 and ZT3 are morehomogenous than ZT1 (Fig. 1a).

The XRPD spectra show the anatase crystalline structure of ZT2and ZT3 compared with Anatase Pearson’s Crystal Data; no othercrystalline phase of TiO2 is present (see Fig.1b). The high baseline isdue to zeolite. The XRPD spectrum of ZT1 (not reported) is wellknown in literature [22,23].

The dimension,the surfacearea(thatdependsonthe morphologyand the quality of dispersion), the crystal structure and the Egap ofsamples, expected to exert a key role in the photocatalytic activity,were determined for each sample to ascertain the relation betweenthe synthetic procedure and the properties of the obtained TiO2

samples. In order to maximize charge separation on the photo-catalyst surface, the absorbed light should be at a wavelengthshorter than Egap. Table 1 shows the shift of the absorption edge ofZT2 and ZT3 towards the visible region in comparison with ZT1. Thisis due to the decrease of Egap. However, the photocatalytic efficiencyis highly dependent on the surface area, too. In this context, amongthe prepared materials, the ZT1 sample exhibits the lowest surfacearea, and an Egap value not significantly different from those of ZT2and ZT3 (uncertainty 0.1 eV).

Adsorption of marbofloxacin and enrofloxacin on TiO2-modifiedzeolites

The adsorption isotherms of the two FQs, performed in tapwater because of its invariant composition [9], were determinedprimarily to define the adsorption capacity of the modifiedzeolites.

The adsorption profiles of MAR and ENR, reported in Fig. S1(Supplementary data), indicate that sorption capacities are higheron ZT2 and ZT3 than on ZY and ZT1. These results are ascribable totheir larger surface area and different morphology with respect toZY and ZT1, as discussed in Section Catalysts characterization.

The experimental data were fitted by using the Langmuir(Eq. (1)) and Freundlich models (Eq. (2)):

qe ¼KLCe

1 þ KLCe

� �qm (1)

where qm and KL are the Langmuir constants related to themonolayer adsorption capacity and the energy of adsorption,respectively.

qe ¼ KfCne (2)

where qe is the amount of the adsorbed antibiotic at theequilibrium, Kf the Freundlich adsorption coefficient and n is theFreundlich constant.

The isotherm parameters were obtained by a dedicatedsoftware (Origin1). The correlation coefficients (R2) showed thatthe Langmuir model (Eq. (1)) fitted the experimental points betterthan the Freundlich model (Eq. (2)) in all the cases (Table 2),indicating that adsorbed FQs form a monolayer coverage on thezeolite surface [24,25].

Photodegradation experiments

It has recently been demonstrated that FQs are removed byphotodegradation both in aqueous solution and in solid matrices[6,26–28]. In the present work, TiO2-modified zeolites have beenapplied for FQs adsorption from wastewaters, degradation of theadsorbed pollutants and regeneration of the material, undernatural solar light.

The photodegradation profiles of the adsorbed FQs wereinvestigated on ZT2 and ZT3, and for comparison on ZY, toevaluate the effective contribution of the TiO2 coating both interms of adsorption capability and photodegradation of thetrapped pollutants. Fig. 2 shows the photodegradation profilesof the three samples (0.5 g) fortified with 3 mg g�1 of eachantibiotic. The degradation rate was greatly enhanced on theTiO2-modified zeolites ZT2 and ZT3; specifically, 91% of the FQsinitial amount was removed in about 2 h irradiation while morethan 20 h were required to obtain a comparable degradation onuntreated zeolite (ZY).

Fig. 1. TiO2-modified zeolites characterization: (a) SEM images (dimension: 10 mm, magnification: 300 K, EHT: 20.0 kV); top: cluster of ZT1; middle: cluster of ZT2; bottom:cluster of ZT3. (b) XRPD spectra of ZT2 (red) and ZT3 (blue) samples compared with anatase Pearson’s Crystal data (green).

Table 2Parameters obtained by fitting experimental data to a Langmuir equation model (inparentheses standard deviation, n = 3).

F. Maraschi et al. / Journal of Environmental Chemical Engineering 2 (2014) 2170–2176 2173

It is apparent that the TiO2 coating markedly increases thephotodegradation rate of the adsorbed pollutants, while on ZY thedegradation mechanism is purely photochemical (see Fig. 3, path i)and ascribable to the solid state photoreactivity of FQs [26,29].

As a matter of fact, on ZT2 and ZT3, the photocatalyticdegradation (about 70% degraded FQs) happened at the first20 min of irradiation. Although ZT1 had a similar efficiency for bothFQs (data not shown), it was not considered for remediationpurpose due to its low sorption capacity and non-homogeneousTiO2 coating.

Table 1Absorption wavelength, energy gap and surface area values determined for ZY, ZT1,ZT2 and ZT3 samples.

Sample name Absorption wavelength(nm)

Energy gap(eV)

Surface area (m2 g�1)

ZY 188 3.7 175ZT1 377 3.3 135ZT2 383 3.2 228ZT3 384 3.2 215

The observed decrease in FQs concentration was due tophotodegradation only, as no degradation was observed on zeolitesamples spiked with 3 mg g�1 FQs and stored in the dark at roomtemperature for 50 h; as a proof, the MAE recovery tests gavequantitative extraction for both drugs (95%, RSD 5%, n = 3), thus

Sample KL (L mg�1) qm (mg g�1) R2 R2a

MAR ZY 0.07(2) 7.5(4) 0.9228 0.9278ZT1 0.018(2) 7.4(2) 0.9929 0.9474ZT2 0.020(2) 23.1(7) 0.9919 0.9581ZT3 0.020(3) 19.5(7) 0.9851 0.9606

ENR ZY 0.35(3) 3.60(5) 0.9930 0.8338ZT1 0.035(5) 15.1(7) 0.9852 0.8949ZT2 0.09(1) 21.3(5) 0.9912 0.9218ZT3 0.20(3) 24.1(6) 0.9887 0.9242

a R2 obtained from Freundlich model fitting.

0

10

20

30

40

50

60

70

80

90

100

0 2 4 6 8 10 12 14 16 18 20 22 240

10

20

30

40

50

60

70

80

90

100

0 2 4 6 8 10 12 14 16 18 20 22 24

a b

Irradiation time (h) Irradiati on time (h)

% %

Fig. 2. Photodegradation profiles of MAR (a) and ENR (b) on different zeolites: (^) ZY, (�) ZT2, (&) ZT3 (RSD <10%, n = 3).

2174 F. Maraschi et al. / Journal of Environmental Chemical Engineering 2 (2014) 2170–2176

demonstrating that the FQ residuals after irradiation are not due toa poor recovery, but rather due to photodegradation.

The relative percentage distribution of FQs photoproducts wasmonitored by HPLC-UV (chromatograms not shown). The maxi-mum amount of intermediates, not exceeding 10% of MAR and ENRinitial concentrations, was observed after 20 min irradiation forZT2 and ZT3. Both FQs and their photoproducts were almostcompletely decomposed in about 2 h, as the drug residuals resulted<0.2% with respect to the starting amount. Main photoproductsstructures have been identified by means of HPLC–ESI-MS/MS (seeSupplementary data). As previously reported by our group [29],three main photochemical pathways are possible after lightabsorption by 6-FQs, viz. (i) amine side-chain degradativeoxidation, (ii) photosubstitution and (iii) reductive dehalogenation(see Fig. 3). This is a general process except for MAR, which has acompeting outcome, N—N bond cleavage (path (iv), Fig. 3) [30]. It isknown that in heterogeneous photocatalytic processes mediatedby TiO2 most of the light is absorbed by the semiconductorparticles [31,32], that then interact by hydrogen or electrontransfer with the adsorbed drug. In the experimental conditionsexplored in this work, side-chain photodegradation (i) is the mainactive path and all the characterized products both from MAR andENR arise from this route. This result is in accordance with ourprevious observations about photocatalytic degradation of FQs[21] and FQs photodegradation on solid matrix [26,29], furtherconfirming that all the photoproducts are ascribable to thephotoreaction of MAR and ENR adsorbed onto the surface of thephotoreactive material.

In more detail, ENR main photoproducts (Fig. 4a) arise frompiperazine moiety degradation. Most of them (B, m/z = 334; Q, m/z = 332; E, m/z = 374; E5, m/z = 306) have already been identified inprevious works and the structures were assigned in comparisonwith literature MS/MS data [21]. The structure of product E12,having a m/z ratio of 280, was assigned on the basis of the MS/MSfragmentation spectrum (see Supplementary data) characterized

FQ 1FQ* 3FQ*

amine sid e-chaindegrada tiv e oxida tion

photosubstitution

reductiv e dehalogena tio n

i

ii

iii

N-N bon d clevea ge

hν

H2O

ISC

iv

Fig. 3. General scheme of the photodegradation pathways of the FQs.

by three main losses ascribable to piperazine and carboxylicmoiety (�33 amu, —C2H5N; �18 amu, —H2O; �61 amu, —CH3NO2)and was proposed to arise from product E5 by partial degradationof the cyclopropane ring.

MAR photoproducts distribution (Fig. 4b) is a little morecomplex and 9 products have been characterized. Two of them(Ma1, Ma2) have already been found in previous work [21] and thestructures were assigned by comparison of the MS/MS spectra, theother structures were proposed on the basis of the MS/MSfragmentations (see Supplementary data). All the identifiedcompounds arise from step by step piperazine side-chain oxidativedegradation; the proposed structures for Ma11 (m/z = 373), Ma13(m/z = 391), and Ma8 (m/z = 377) are characterized by differentoxidation levels of piperazine ring and correspond to thehydroxylated (Ma11) mono-oxo (Ma8) and di-oxo (Ma13) deriv-atives, respectively. All the other products can be justified to arisefrom these three first generation products by stepwise oxidation.Ma10 differs from the parent drug MAR for 2 amu and thisdifference was attributed to C—C bond break on the piperazinering. On the contrary, Ma2, which differs from MAR for 14 amu, wasproposed to come from demethylation of the piperazine. ProductsMa9 (m/z = 351, �14 amu), Ma1 (m/z = 337, �28 amu), Ma4 (m/z = 323, �42 amu) differ from each other for the number ofmethylene groups lost with respect to Ma10 (m/z = 365), from oneto three, respectively. The last structure, Ma12 (m/z = 308), wasassigned to the most degraded observed product where almost allthe piperazine is lost, remaining only a dimethylamino residue.

Regeneration of TiO2-modified zeolite

Further experiments were carried out on adsorbed ZT2 toinvestigate the potential reuse of the material after irradiation interms of both sorption capacity and photocatalytic activity. After5 h irradiation under solar light, the adsorbed MAR and ENRcontent was lower than 1 mg g�1 (0.03% of FQs initial concentra-tion). Then, two further adsorption–irradiation cycles were carriedout on the same ZT2 sample observing no loss of efficiency in termsof both adsorption capacity (23.0 mg g�1 for MAR and 19.9 mg g�1

for ENR) and photocatalytic activity (FQs <0.03%, 5 h radiation).

Actual wastewater samples

FQs removal by ZT2, which resulted the most performingamong the three TiO2-modified zeolites, was investigated atlaboratory scale on actual wastewater samples spiked at realisticconcentrations with MAR and ENR. The samples collected at theoutlet of a wastewater treatment plant were analysed before and

Fig. 4. Products distribution and photodegradation path of ENR (a) and MAR (b).

Table 3Percentage removal of MAR and ENR from wastewater samples.

FQs WW (mg L�1 added) WW + ZT2 (mg L�1) % removal

MAR 5–1000 <0.2 96ENR 5–1000 <0.07 98

F. Maraschi et al. / Journal of Environmental Chemical Engineering 2 (2014) 2170–2176 2175

after contact with ZT2: the drastic decrease in FQs concentrationafter adsorption on ZT2 is clearly evident in Table 3, andcorresponds to a removal rate >95%.

After solar light exposure, the reusability of ZT2 was proved by afurther treatment, obtaining as well a quantitative removal (>95%)of the two pollutants from wastewater.

Conclusions

In the present study three different TiO2-modified zeolites havebeen tested as sorbent/photoreactive materials for FQs removalfrom wastewaters. Among them, ZT2 proved to be the mostefficient for removing trace levels of such emerging pollutantsfrom complex aqueous matrices. As expected, TiO2 coatingincreased the photodegradation rate of adsorbed pollutants withrespect to the untreated zeolite. The drugs were completelyremoved by adsorption from water and quantitatively

photodegraded under natural solar light, once TiO2-modifiedzeolite was removed from the aqueous solution. Furthermore, thiscombined procedure allowed this new material to be reused forfurther adsorption/photodegradation cycles with no loss ofadsorption capacity and photocatalytic activity. FQs photoreactiv-ity has to be ascribed to drug molecules absorbed onto TiO2-modified zeolite surface, as evidenced by FQs photoproducts,arisen from amine side-chain oxidative degradation, that wereefficiently decomposed in few hours.

The first encouraging results obtained on actual wastewatersamples suggest the possible use of TiO2-modified zeolites for theend-of-pipe treatment of wastewater containing recalcitrantpharmaceuticals not completely removed by conventional treat-ment plants.

The results presented here contribute to a better understandingof the behaviour of antibiotics under conventional sewagetreatment and advanced treatment processes.

Acknowledgements

This work was supported by European Union, Regioneautonoma Valle d’ Aosta and Ministero del Lavoro e delle PoliticheSociali and partially by the Italian MIUR through the FIRB ProjectRBAP115AYN “oxides at the nanoscale: multifunctionality andapplications”.

2176 F. Maraschi et al. / Journal of Environmental Chemical Engineering 2 (2014) 2170–2176

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at http://dx.doi.org/10.1016/j.jece.2014.08.009.

References

[1] M. Sturini, A. Speltini, L. Pretali, E. Fasani, A. Profumo, Solid-phase extractionand HPLC determination of fluoroquinolones in surface waters, J. Sep. Sci. 32(2009) 3020–3028, doi:http://dx.doi.org/10.1002/jssc.200900135. 19662641.

[2] S. Kusari, D. Prabhakaran, M. Lamshöft, M. Spiteller, In vitro residual anti-bacterial activity of difloxacin, sarafloxacin and their photoproducts afterphotolysis in water, Environ. Pollut. (Barking, Essex: 1987) 157 (2009) 2722–2730, doi:http://dx.doi.org/10.1016/j.envpol.2009.04.033. 19473738.

[3] E.M. Golet, A.C. Alder, W. Giger, Environmental exposure and risk assessmentof fluoroquinolone antibacterial agents in wastewater and river water of theGlatt valley watershed, Switzerland, Environ. Sci. Technol. 36 (2002) 3645–3651, doi:http://dx.doi.org/10.1021/es0256212. 12322733.

[4] R.H. Lindberg, K. Björklund, P. Rendahl, M.I. Johansson, M. Tysklind, B.A.V.Andersson, Environmental risk assessment of antibiotics in the Swedishenvironment with emphasis on sewage treatment plants, Water Res. 41 (2007)613–619, doi:http://dx.doi.org/10.1016/j.watres.2006.11.014. 17187841.

[5] K. Kümmerer, Antibiotics in the aquatic environment – a review – part I,Chemosphere 75 (2009) 417–434, doi:http://dx.doi.org/10.1016/j.chemo-sphere.2008.11.086. 19185900.

[6] A. Speltini, M. Sturini, F. Maraschi, A. Profumo, Fluoroquinolone antibiotics inenvironmental waters: sample preparation and determination, J. Sep. Sci. 33(2010) 1115–1131, doi:http://dx.doi.org/10.1002/jssc.200900753. 20187033.

[7] A. Speltini, M. Sturini, F. Maraschi, A. Profumo, A. Albini, Analytical methods forthe determination of fluoroquinolones in solid environmental matrices, TrACTrends Anal. Chem. 30 (2011) 1337–1350, doi:http://dx.doi.org/10.1016/j.trac.2011.04.011.

[8] K. Kümmerer, Antibiotics in the aquatic environment – a review – part II,Chemosphere 75 (2009) 435–441, doi:http://dx.doi.org/10.1016/j.chemo-sphere.2008.12.006. 19178931.

[9] M. Sturini, A. Speltini, F. Maraschi, A. Profumo, L. Pretali, E. Fasani, A. Albini,Photochemical degradation of marbofloxacin and enrofloxacin in naturalwaters, Environ. Sci. Technol. 44 (2010) 4564–4569, doi:http://dx.doi.org/10.1021/es100278n. 20481547.

[10] M. Sturini, A. Speltini, F. Maraschi, L. Pretali, A. Profumo, E. Fasani, A. Albini, R.Migliavacca, E. Nucleo, Photodegradation of fluoroquinolones in surface waterand antimicrobial activity of the photoproducts, Water Res. 46 (2012) 5575–5582, doi:http://dx.doi.org/10.1016/j.watres.2012.07.043. 22901305.

[11] A. Speltini, M. Sturini, F. Maraschi, A. Profumo, A. Albini, Microwave-assistedextraction and determination of enrofloxacin and danofloxacin photo-transformation products in soil, Anal. Bioanal. Chem. 404 (2012) 1565–1569, doi:http://dx.doi.org/10.1007/s00216-012-6249-3. 22814972.

[12] Y.C. Lee, H.U. Lee, K. Lee, B. Kim, S.Y. Lee, M.H. Choi, W. Farooq, J.S. Choi, J.Y. Park,J. Lee, Y.K. Oh, Y.S. Huh, Aminoclay-conjugated TiO2 synthesis for simultaneousharvesting and wet-disruption of oleaginous Chlorella sp, Chem. Eng. J. 245(2014) 143–149, doi:http://dx.doi.org/10.1016/j.cej.2014.02.009.

[13] H.U. Lee, G. Lee, J.C. Park, Y.C. Lee, S.M. Lee, B. Son, S.Y. Park, C. Kim, S.G. Lee, S.C.Lee, B. Nam, J.W. Lee, D.R. Bae, J.S. Yoon, J. Lee, Efficient visible-light responsiveTiO2 nanoparticles incorporated magnetic carbon photocatalysts, Chem. Eng. J.240 (2014) 91–98, doi:http://dx.doi.org/10.1016/j.cej.2013.11.054.

[14] H.U. Lee, Y.C. Lee, S.C. Lee, S.Y. Park, B. Son, J.W. Lee, C.H. Lim, C.J. Choi, M.H.Choi, S.Y. Lee, Y.K. Oh, J. Lee, Visible-light-responsive bicrystalline (anatase/brookite) nanoporous nitrogen-doped TiO2 photocatalysts by plasma treat-ment, Chem. Eng. J. 254 (2014) 268–275, doi:http://dx.doi.org/10.1016/j.cej.2014.06.011.

[15] D.I. Petkowicz, S.B.C. Pergher, C.D.S. da Silva, Z.N. da Rocha, J.H.Z. dos Santos,Catalytic photodegradation of dyes by in situ zeolite-supported titania, Chem.Eng. J. 158 (2010) 505–512, doi:http://dx.doi.org/10.1016/j.cej.2010.01.039.

[16] C. Zhao, H. Deng, Y. Li, Z. Liu, Photodegradation of oxytetracycline in aqueousby 5A and 13X loaded with TiO2 under UV irradiation, J. Hazard. Mater. 176(2010) 884–892, doi:http://dx.doi.org/10.1016/j.jhazmat.2009.11.119.20034739.

[17] S. Wang, Y. Peng, Natural zeolites as effective adsorbents in water andwastewater treatment, Chem. Eng. J. 156 (2010) 11–24, doi:http://dx.doi.org/10.1016/j.cej.2009.10.029.

[18] L.L. Hench, J.K. West, The sol–gel process, Chem. Rev. 90 (1990) 33–72, doi:http://dx.doi.org/10.1021/cr00099a003.

[19] V. Caratto, L. Setti, S. Campodonico, M.M. Carnasciali, R. Botter, M. Ferretti,Synthesis and characterization of nitrogen-doped TiO2 nanoparticles preparedby sol–gel method, J. Sol-Gel Sci. Technol. 63 (2012) 16–22.

[20] L. Kavan, T. Stoto, M. Gratzel, D. Fitzmaurice, V. Shklover, Quantum size effectsin nanocrystalline semiconducting titania layers prepared by anodic oxidativehydrolysis of titanium trichloride, J. Phys. Chem. 97 (1993) 9493–9498.

[21] M. Sturini, A. Speltini, F. Maraschi, A. Profumo, L. Pretali, E.A. Irastorza, E.Fasani, A. Albini, Photolytic and photocatalytic degradation of fluoroquino-lones in untreated river water under natural sunlight, Appl. Catal. Part B:Environ. 119–120 (2012) 32–39.

[22] H.R. Jafry, M.V. Liga, Q. Li, A.R. Barron, Single walled carbon nanotubes (SWNTs)as templates for the growth of TiO2: the effect of silicon in coverage and thepositive and negative synergies for the photocatalytic degradation of Congored dye, New J. Chem. 35 (2011) 400–406, doi:http://dx.doi.org/10.1039/c0nj00604a.

[23] X.T. Zhou, H.B. Ji, X.J. Huang, Photocatalytic degradation of methyl orange overmetalloporphyrins supported on TiO2 Degussa P25, Molecules (Basel,Switzerland) 17 (2012) 1149–1158, doi:http://dx.doi.org/10.3390/mole-cules17021149. 22277995.

[24] H.M. Otker, I. Akmehmet-Balcio�glu, Adsorption and degradation of enroflox-acin, a veterinary antibiotic on natural zeolite, J. Hazard. Mater. 122 (2005)251–258, doi:http://dx.doi.org/10.1016/j.jhazmat.2005.03.005. 15967281.

[25] A. Alberti, A. Martucci, L. Pasti, N. Marchetti, A. Cavazzini, F. Dondi, Adsorptionof pharmaceuticals from aqueous solutions on synthetic zeolites, MicroporousMesoporous Mater. 148 (2012) 174–183, doi:http://dx.doi.org/10.1016/j.micro-meso.2011.07.009.

[26] M. Sturini, A. Speltini, F. Maraschi, A. Profumo, L. Pretali, E. Fasani, A. Albini,Sunlight-induced degradation of soil-adsorbed veterinary antimicrobials,Chemosphere 86 (2012) 130–137, doi:http://dx.doi.org/10.1016/j.chemo-sphere.2011.09.053. 22051342.

[27] Y. Picó, V. Andreu, Fluoroquinolones in soil—risks and challenges, Anal.Bioanal. Chem. 387 (2007) 1287–1299, doi:http://dx.doi.org/10.1007/s00216-006-0843-1. 17082879.

[28] W. Chenxi, A.L. Spongberg, J.D. Witter, Determination of the persistence ofpharmaceuticals in biosolids using liquid-chromatography tandem massspectrometry, Chemosphere 73 (2008) 511–518, doi:http://dx.doi.org/10.1016/j.chemosphere.2008.06.026. 18674794.

[29] M. Sturini, A. Speltini, F. Maraschi, L. Pretali, A. Profumo, E. Fasani, A. Albini,Environmental photochemistry of fluoroquinolones in soil and in aqueous soilsuspensions under solar light, Environ. Sci. Pollut. Res. Int. (2013) , doi:http://dx.doi.org/10.1007/s11356-013-2124-9. 24026207.

[30] L. Pretali, E. Fasani, D. Dondi, M. Mella, A. Albini, The unexpectedphotochemistry of marbofloxacin, Tetrahedron Lett. 51 (2010) 4696–4698,doi:http://dx.doi.org/10.1016/j.tetlet.2010.06.137.

[31] T. Paul, P.L. Miller, T.J. Strathmann, Visible-light-mediated TiO2 photocatalysisof fluoroquinolone antibacterial agents, Environ. Sci. Technol. 41 (2007) 4720–4727, doi:http://dx.doi.org/10.1021/es070097q. 17695920.

[32] X. Van Doorslaer, K. Demeestere, P.M. Heynderickx, H. Van Langenhove, J.Dewulf, UV-A and UV-C induced photolytic and photocatalytic degradation ofaqueous ciprofloxacin and moxifloxacin: reaction kinetics and role ofadsorption, Appl. Catal. B: Environ. 101 (2011) 540–547, doi:http://dx.doi.org/10.1016/j.apcatb.2010.10.027.