Sunlight photodegradation of marbofloxacin and enrofloxacin adsorbed on clay minerals

Journal of Photochemistry and Photobiology A: Chemistry 299 (2015) 103–109

Sunlight photodegradation of marbofloxacin and enrofloxacinadsorbed on clay minerals

Michela Sturini a,*, Andrea Speltini a, Federica Maraschi a, Elisa Rivagli a, Luca Pretali b,Lorenzo Malavasi a, Antonella Profumo a, Elisa Fasani a, Angelo Albini a

aDepartment of Chemistry, University of Pavia, via Taramelli 12, Pavia 27100, Italyb Parco Tecnologico Padano, via Albert Einstein, Lodi 26900, Italy

A R T I C L E I N F O

Article history:Received 24 September 2014Received in revised form 14 November 2014Accepted 15 November 2014Available online 18 November 2014

Keywords:FluoroquinolonesClaysPhotodegradationPhotoproductsX-rays diffraction

A B S T R A C T

Fluoroquinolone antibiotics (FQs) are important “emerging” micropollutants, and their environmentaldiffusion is an issue of great concern. In this study, the photochemical degradation of marbofloxacin(MAR) and enrofloxacin (ENR) adsorbed on montmorillonite (MMT) and kaolinite (KAO) clays wasinvestigated. Being FQs photosensitive molecules, the sunlight-induced degradation of clay-adsorbedFQs and of their photoproducts was monitored as function of irradiation time. The photoproducts wereidentified by high performance liquid chromatography electrospray tandem mass spectrometry(HPLC-ESI-MS/MS) and the photochemical pathways have been elucidated. X-ray diffraction (XRD)has been employed to deeply study the solid-state photodegradation process of MAR and ENR on MMT.Interestingly, the XRD results clearly evidenced a significative variation of the interlayer spacing ofMMT not only as function of FQs adsorbed amounts, but also in relationship with their photodegradation,not investigated by means of XRD as yet. The results confirmed that FQs adsorption is a bulk phenomenonthat occurs in the interlayer spacing of the MMT structure; moreover, it was proved that sunlight largelydegraded the adsorbed FQs, both on the external surface and in the interlayer spacing.

ã 2014 Elsevier B.V. All rights reserved.

Contents lists available at ScienceDirect

Journal of Photochemistry and Photobiology A:Chemistry

journal homepa ge: www.elsev ier .com/ locate / jphotochem

1. Introduction

The widespread of pharmaceuticals in the environment andtheir fate, effects and threats associated to their occurrence arematter of great concern. Among these drugs, fluoroquinolones(FQs) are an important class of “emerging” micropollutants [1–5].

In view of their broad activity spectrum against Gram bacteriaand their good oral intake, FQs are the fourth largest class of humanantibiotics and they are widely administrated also to animals fortherapeutic purposes, and as feed additives to support growth inlivestock [6]. Once administrated, FQs are metabolized to a minorextent and a large part of the unmodified initial dose reaches urbansewage treatment plants (STPs) that are not able to quantitativelyremove such complex molecules [7,8]. This causes a continuousrelease of FQs in surface water overcoming their transformationand removal rates. These drugs rapidly move from water bodies tothe soil compartment. Additionally, veterinary antibacterials candirectly reach soil and groundwater through the common practiceof recycling manure from animal husbandry and/or sewage sludge

* Corresponding author. Tel.: +39 0382 987347; fax: +39 0382 528544.E-mail address: [email protected] (M. Sturini).

http://dx.doi.org/10.1016/j.jphotochem.2014.11.0151010-6030/ã 2014 Elsevier B.V. All rights reserved.

from STPs as fertilizers. Indeed FQs have been detected up to fewmicrograms per kilogram also in soil and sediment [9] due to theirstrong binding to natural organic matter and minerals, with thehigh distribution coefficients observed in loamy matrices [10].Recent studies have attributed to FQs ecotoxicity and genotoxicityagainst Vibrio Fischeri,Daphnia Magna and Pseudomonas Putida[11–13]. Furthermore, FQ residuals are able to induce environ-mental bacterial resistance [14,15].

Although photochemistry represents a natural removal path-way for these antimicrobials from the environment [1–3,5], on theother hand the photochemical processes involve the release ofsignificant amounts of photoproducts that contribute to the overallenvironmental impact of the parent compounds [16,17]. Indeed, itwas proved that FQ photoproducts maintain significant antibioticactivity [4].

Montmorillonite (MMT) and kaolinite (KAO) clays are naturalconstituents of soils and, due to their high surface area, havebeen recently proposed for adsorption of 4-chlorophenol and2,4,6-trichloroaniline [18], tetracycline antibiotics [19], diphenhy-dramine [20], polychlorinated biphenyl and other organicpollutants [21,22]. With regard to FQs, recent studies haveinvestigated the mechanistic interactions of ciprofloxacin ontoMMT [23] and KAO [24], and those of ENR on smectite clays [25].

104 M. Sturini et al. / Journal of Photochemistry and Photobiology A: Chemistry 299 (2015) 103–109

In this paper we report the solid-state photochemicaldegradation of two largely employed FQs, marbofloxacin (MAR)and enrofloxacin (ENR), adsorbed on MMT and KAO clays. MAR andENR were chosen as probe molecules because largely used in cattleand swine farms in the South Lombardy plain (Italy) and as a resultdirectly released in ditches and rivers by farms wastewaters [26].Moreover, they present different photochemical behavior both inwater and soil [2,3]. Batch sorption experiments were carried outin water at the milligrams per litre levels, and the adsorptioncapacities were calculated by sorption isotherms, described by theLangmuir model. Then clay samples were fortified with the twodrugs and irradiated under natural sunlight. The decay profilesover irradiation time were traced and the main photodegradationpathways were proposed on the basis of the FQs photoproductsidentified by HPLC-ESI-MS/MS. The solar light-induced degrada-tion of MMT-adsorbed FQs was also studied by X-ray diffraction(XRD). The variation of the MMT interlayer spacing wasinvestigated before and after irradiation.

2. Experimental

2.1. Reagents and materials

Ca-MMT STx-1 was acquired from the Clay Mineral Society andKAO was distributed as an international standard by the CeramicSociety of Slovakia (Zettlitz). These materials were used asreceived, with no further purification, so their physical–chemicalproperties are the ones reported in literature [27–29].

All the chemicals employed were reagent grade or higher inquality. MAR and ENR, in the injectable form, were purchased fromBayer (Baytril 25 mg mL�1) and Vétoquinol (Marbocyl 20 mgmL�1), respectively. HPLC gradient grade acetonitrile (ACN) wasfrom VWR. H3PO4 (85%, w/w) and NH3 solution (30%, w/w) werepurchased from Carlo Erba Reagents; HCOOH (98–100%, w/w) wasfrom Merck. Hexahydrate Mg(NO3)2 (97%, w/w) was fromSigma–Aldrich.

Ultra-pure water (resistivity 18.2 MV cm�1 at 25 �C) wasproduced by a Millipore (Milan, Italy) Milli-Q system. FQs stocksolutions were prepared and stored in the dark at 4 �C for amaximum of three months. Working solutions were renewedweekly.

2.2. Adsorption experiments

FQs adsorption experiments, performed in triplicate by a batchequilibration method [30], were carried out in distilled water withinitial FQs concentrations ranging from 25 to 2000 mg L�1.

2.3. Sample preparation

For irradiation experiments, 0.5 g of clays were fortified withMAR and ENR at concentrations in the range 15–77 m g g�1 of eachdrug, according to the batch procedure described in Ref. [30]. Afterequilibration and filtration, the remaining pastes were dried at50 �C and homogeneously dispersed on glass Petri capsulae(; 8.5 cm, depth 2 cm) to obtain a monolayer (thickness below1–2 mm). Afterwards, samples were exposed to natural sunlight(9.00 am–5.00 pm) during the summer (June–September), attemperatures ranging from 25 to 35 �C. The solar power rangedfrom 170 to 470 W m�2 (in the visible range) and from 8 to30 W m�2 (in the UV), respectively. The flux was measured bymeans of a HD 9221 (Delta OHM) (450–950 nm) and of Multimeter(CO.FO.ME.GRA) (295–400 nm) radiometers. At regular intervals,each one of the irradiated samples was extracted by microwave-assisted extraction (MAE) according to a green method [31]selective for FQs. All experiments were performed in triplicate. The

extracts were analyzed by HPLC–UV for tracing degradationdecays, and by HPLC-ESI-MS/MS for photoproducts identification.

For XRD analyses, MMT samples (0.3 g) were enriched atdifferent concentrations of MAR and ENR ranging from 15 to 100%of their maximum adsorption capacity (see Table S1). After 24 hequilibration, the suspensions were filtered and the remainingpastes were dried at 50 �C. All the samples were measured bymeans of high-resolution XRD, before and after irradiation.

2.4. Analytical determinations

The HPLC-ESI-MS/MS analyses were performed by using anAgilent 1100 HPLC with a Luna C18 (150 � 4.6 mm, 5 mm) column,maintained at 30 �C. The mobile phase was HCOOH 0.5% v/v inultrapure water-ACN (85:15). The flow rate was 1.2 mL min�1 andthe injection volume was 5 mL. The MS/MS-system consisted of alinear trap Thermo LXQ. ESI experiments were carried out inpositive-ion mode under the following constant instrumentalconditions: source voltage 4.5 kV, capillary voltage 20 V, capillarytemperature 275 �C and normalized-collision energy 35.

The HPLC–UV system consisted of a PU-1580 pump (JASCO)equipped with a programmable UV-1575 UV–vis detector (JASCO).The detection wavelength was 275 nm. 20 mL of each sample wereinjected into a 150 � 4.6 mm, 5 mm Symmetry Column (Waters)coupled with a similar guard-column. The mobile phase was water(pH adjusted to 2.5 with 37% HCl)-ACN (90:10), at a flow rate of1.2 mL min�1.

3. Results and discussion

3.1. Adsorption experiments

The adsorption isotherms were determined both to confirm theadsorption capacity of the two clays and to better evaluate FQsconcentrations to be considered for the irradiation experiments.

The adsorption profiles of MAR and ENR on MMT and KAO arereported in Fig. 1. It can be seen that MAR and ENR adsorptioncurves show a similar trend, but the FQs observed adsorptioncapacities on MMT are higher than those found on KAO, due to thehigher surface area of the former (84 m2g�1 for MMT and 10 m2g�1

for KAO) and due to its interlayer spacing. These results arereported in Table S1 and are comparable with those found in aprevious investigation [30].

Langmuir and Freundlich models were used to fit experimentaldata, according to the Eqs. (1) and (2), respectively:

qe ¼KLCe

1 þ KLCe

� �qm (1)

qe ¼ KfCne (2)

qe is the amount of the adsorbed antibiotic at the equilibrium andCe is the equilibrium antibiotic concentration. KL is the Langmuirconstant and qm represents the total number of sorption sites(Eq. (1)); Kf is the Freundlich adsorption constant and n is theFreundlich exponent (Eq. (2)).

The isotherm parameters were obtained by a dedicatedsoftware (Origin1). The calculated Langmuir and Freundlichadsorption constants are listed in Table S1.

The correlation coefficients (R2) and chi-square (x2) showedthat the Langmuir model (Eq. (1)) better fitted the experimentalpoints than the Freundlich model (Eq. (2)) in all cases. The highervalues of correlation coefficient (R2), as well as the lower standarddeviations of Langmuir parameters compared to those obtained forthe Freundlich model, indicated that the adsorption of the twodrugs occurs through a monolayer coverage [30,32].

Fig. 1. Adsorption profiles of MAR (a) and ENR (b) on MMT (~) and KAO (*).

M. Sturini et al. / Journal of Photochemistry and Photobiology A: Chemistry 299 (2015) 103–109 105

3.2. Photodegradation experiments

It has been recently assumed [33], and also demonstrated [3,5],that photodegradation of FQs is less efficient in soil than in aqueoussolution, unless modifying soil constituents (e.g. clay minerals) toincrease their adsorption ability and to make them efficientphotocatalysts for removing the adsorbed organic pollutants[34,35]. Nevertheless, in the present investigation sunlightirradiation has been proved to contribute substantially to thephotodegradation of FQs adsorbed on clay minerals. In Fig. 2 thephotodecomposition profiles of MAR and ENR on MMT and KAO areshown.

The drugs showed a similar behavior, as they were largelydegraded up to 15% in about 120 h when adsorbed on KAO, whileon MMT their concentration decreased more slowly and about30–40% of the initial amount was conserved after 180 h ofexposure. Experimental data are quite consistent with those

Fig. 2. Degradation profiles of MAR on KAO (a) and on

previously obtained on soil, where the incomplete degradation upto 150 h was not due to soil particles not exposed to solar light, butrather to part of analyte differently located/complexed [3].Particularly, this effect is more evident in the case of MMT, asthe adsorption of MAR an ENR occurs both onto the externalsurface and in the interlayer spacing [30].

The decrease of FQs concentration as function of irradiationtime was due to photochemical processes only, as no degradationwas observed in the control samples fortified (15–77 m g g�1) andstored in the dark at room temperature for 1 month, and thensubmitted to MAE extraction [31]. These findings excluded thepotential biodegradation route that, in fact, is inhibited by the highfixation rates of FQs to external and interlayer sites of the matrix, inaccordance with previous suggestions [36].

Recovery tests from not irradiated clays samples enriched with15–77 m g g�1 FQs and extracted gave good results for both drugs(71–75%, RSD 10%, n = 3), thus confirming that the FQs residual

MMT (c), and ENR on KAO (b) and on MMT (d).

106 M. Sturini et al. / Journal of Photochemistry and Photobiology A: Chemistry 299 (2015) 103–109

after irradiation is not ascribable to the failure in the recovery rate,but it is really due to photodegradation.

As a matter of fact, the solid-state photodecomposition of theadsorbed pollutants was accompanied by the formation of variousphotoproducts. After HPLC-ESI-MS/MS identification, their relativepercentage distribution was monitored by HPLC–UV. The maxi-mum amounts of photoproducts were reached after 90 and 110 h ofirradiation (natural solar light) for KAO and MMT, respectively, andgenerally were in the range 5–8% with respect to the initialamounts of the parent compounds.

To summarize, these experiments suggest that part of the drugs(15% for KAO and ca. 30–40% for MMT) interacts more stronglywith clays or is located in the interlayer spacing. Despite this,photodegradation represents a significant degradation path forthese recalcitrant molecules strongly adsorbed onto clays.

3.3. Photochemical paths

Photochemical pathways have been traced for both FQs, asreported in the following. As well established in literature (see [37]and references therein) and widely stated in our previous works[2–5,38], FQ photochemistry firmly relies on a triplet statereactivity. Upon absorption of a photon of a proper wavelength,a singlet excited state is primarily formed and quickly populates asecond state with a triplet multiplicity via intersystem crossing(ISC). This is indeed the reactive state from which the mainphotoproducts are formed, mostly through three different reactivepaths (Fig. 3), viz. (i) amine side-chain degradative oxidation, (ii)photosubstitution, and (iii) reductive dehalogenation. In the caseof MAR a fourth and predominant route is opened through N—Nbond cleavage [39]. When FQ molecules are adsorbed on solidmatrix, such as soils, path (ii) and (iii) are almost completelyhampered [3,5]. Indeed, this occurs also when ENR and MAR areadsorbed onto clays.

For MMT-adsorbed MAR four main products (F, Ma3, Ma6, Ma7;Fig. 4a) were formed. Product F is one of the two mainphotoproducts of MAR photoreaction via path (iv) and it wasalready found and characterized both in aqueous and solidmatrices [2,3,38,39]. Product Ma3 is thought to arise from path(i) and its structure was attributed basing on its MS/MS spectra (seeSupplementary Data). Ma3 ([M + 1]+ = 336 a.m.u.) differs from MAR([M + 1]+ = 363 a.m.u.) by 27 a.m.u. and it is characterized by threemain losses of 18 a.m.u. (H2O), 28 a.m.u. (C2H4) and 46 a.m.u.(C2H6O). Ma6 and Ma7 came from compound F, from which differby 26 and 42 a.m.u., respectively. Ma6 structure ([M + 1]+ = 296 a.m.u.) is attributed on the basis of the MS fragmentation pattern(see Supplementary Data) and Ma7 ([M + 1]+ = 280 a.m.u.) differsfrom Ma6 by 16 a.m.u. This is ascribable to OH/F substitutionaccompanied by a further loss of a methyl group from thepiperazine side-chain. The MS/MS fragmentation spectra ofMa7 show 4 main losses (18, 43, 61 and 87 a.m.u., seeSupplementary Data).

Seven main photoproducts were generated by irradiation ofMMT-adsorbed ENR (Fig. 4b), arising from amine side-chainoxidative degradation (path (i)). Four structures (B, E, Q, E5) havebeen already identified and characterized [2,3,38]. CompoundsE9 and E10 resulted from E and Q via further degradation. In more

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

details, E10 derived from Q via piperazine ring oxidation; the massdifference between Q and E10 (4 a.m.u) cannot be easily explainedby a single reaction step and, instead, it is supposed to involvemultiple radical initiated reaction steps, as supported by massfragmentation analyses (see Supplementary Data). CompoundE9 resulted from either E10 or E5 through a stepwise degradationthat also affects the cyclopropane ring. Finally, product E11 wasproposed to arise directly from ENR via path (i), in competitionwith the formation of E. E11 is intermediate between ENR and thealready identified product B, from which differs by 14 a.m.u.(CH2 fragment).

In the case of KAO a closely related scenario was found, with8 and 5 products identified for MAR and ENR, respectively (Fig. 5aand b). In the case of MAR, two already known products wereidentified (F and Ma1) in previous works [2,3,38]. Ma3 was thesame compound previously identified in the MAR–MMT experi-ment. Compound Ma4 ([M + 1]+ = 323 a.m.u.), which differs by 14 a.m.u. (loss of a methylene group) from Ma1 ([M + 1]+ = 337 a.m.u.),came from the step-wise degradation of amine side-chain of Ma1,as confirmed by MS/MS analysis. The last identified product Ma5 isa primary product and differs by 16 a.m.u. from MAR. It is proposedto arise by addition of a hydroxyl group on the piperazine ring.

For ENR, 3 products (Q, B, E5) were also identified in MMTexperiments, while the structures of A and D have been assigned bycomparison of HPLC–MS and MS/MS fragmentation with literature[2,3,38]. The remaining structures (E6, E7, E8) were proposedbasing on the HPLC–MS/MS analysis. Product E6 ([M + 1]+ = 348 a.m.u.) together with Q ([M + 1]+ = 332 a.m.u.) directly arose fromENR via path (i). E6 differs from ENR by 1 carbon unit and it ischaracterized by 4 main losses, affecting mostly the easiestcleavable amine side-chain (see Supplementary Data). Product E7([M + 1]+ = 320 a.m.u.) differs from its parent compound B([M + 1]+ = 334 a.m.u.) and from E5 ([M + 1]+ = 306 a.m.u.) by 14 a.m.u. (CH2 residue) and its structure is consequently proposed asthe intermediate between the other two and confirmed by theMS/MS fragmentation (see Supplementary Data). Finally, productE8 ([M + 1]+ = 372 a.m.u.) was supposed to arise from D([M + 1]+ = 358 a.m.u.) after a second photochemical step thatinitiates the oxidative degradation of the piperazine ring (path (i)).

As expected, the reactivity of clay-adsorbed FQs mainlyproceeds through path (i) while path (ii) and (iii) play a minorrole in MAR and ENR photodegradation. Due to its unique chemicalstructure, for MAR a fourth path is possible through N—N bondcleavage and it efficiently competes with path (i), as well noticedespecially in the case of MMT. Probably due to the long irradiationtimes required for the solid-state photoreaction (up to 165 h),multi-photon processes were possible and some identifiedbyproducts arose from subsequent photo-initiated reactions. Thisconfirms that also the primary products are still photoreactive andthat light is an efficient route to remove both parent drugs andtheir degradation products from the environment.

3.4. XRD studies

In order to understand the effect on the MMT crystal structure ofthe MAR and ENR adsorption, we investigated the variation of theMMT lattice parameters upon FQs adsorption for a range ofconcentrations (38–236 m g g�1). The results (see Fig. S1 andTable S2) show that at adsorption levels up to about 150 m g g�1

the d0 0 1 value decreases with respect to the MAR “free” MMT, whileat the highest MAR concentration we investigated (236 m g g�1)there is an expansion of the c-axes. The reduction of the (0 0 1)interplanar spacing observed was correlated to the partial removalof water molecules during the exchange between interlayer cationsand MAR molecules [30]. An analogous XRD investigation wascarried out for the MMT samples enriched with ENR.

Fig. 4. Degradation paths and photoproducts of MAR (a) and ENR (b) on MMT.

Fig. 5. Degradation paths and photoproducts of MAR (a) and ENR (b) on KAO.

M. Sturini et al. / Journal of Photochemistry and Photobiology A: Chemistry 299 (2015) 103–109 107

Table 1Interplanar spacing (d001) and Full Width at Half Maximum (FWHM) values for MARand ENR adsorbed on the MMT samples before and after irradiation.

FQ Sample d0 0 1 (Å) FWHM 0 0 1

MAR(77 m g g�1)

Not irradiated 14.35589(2) 0.04560(9)Irradiated 110 h 14.2686(2) 0.06649(7)Irradiated 165 h 14.2232(1) 0.06969(8)

ENR(55 m g g�1)

Not irradiated 14.30636(2) 0.05396(4)Irradiated 110 h 14.1725(2) 0.05910(5)Irradiated 165 h 14.0487(2) 0.06279(5)

108 M. Sturini et al. / Journal of Photochemistry and Photobiology A: Chemistry 299 (2015) 103–109

The trend of lattice parameters variation upon ENR concentra-tion has been determined and reported in Fig. S2 and Table S3. As inthe case of MAR adsorption, the d0 0 1 value changes upon ENRadsorption. In particular, up to about 90 m g g�1 of ENR adsorptionthe c-axes shrinks, while, at the highest concentration probed inthe present work (151 m g g�1), there is a significant expansion ofthe MMT. The in plane parameters a and b seem to be insensitive tothe enrichment with both FQs.

The role of irradiation on MMT crystal structure was furtherinvestigated by means of XRD. Fig. 6a reports the XRD patterns ofMMT samples loaded with 77 m g g�1 of MAR before (black line)and after 110 h (red line) and 165 h (blue line) irradiation,respectively.

The XRD patterns clearly show that the irradiation leads to acontraction of the MMT c-axes while the in-plane parameters arenot affected by irradiation. Values for the d-spacings, as deter-mined from the fit of the (0 0 1) reflection for the three samplesreported in Fig. 6a, are listed in Table 1. The lattice contractionobserved after irradiation correlates to the photodegradation ofthe intercalated MAR molecules, thus further confirming that thisis a bulk phenomenon. In addition, the XRD patterns of the sampleafter irradiation show a significant broadening of the (0 0 1)reflection – the FWHM (full width at half maximum) increasesfrom 0.045 to ca. 0.070. This indicates that there is a widerdistribution of d0 0 1-spacings in the irradiated samples due to thephotodegradation of the adsorbed MAR within the MMT matrixand a possible loss of order along the c-axes. Note that the (0 2 0)

Fig. 6. XRD patterns for MMT enriched with (a) MAR (77 m g g�1) and (b) ENR(55 m g g�1) before and after irradiation. (For interpretation of the references tocolor in the text, the reader is referred to the web version of this article.)

reflection does not broaden after irradiation. Finally, it is possibleto note that the XRD patterns for the samples irradiated for 110 hand 165 h are very similar, thus indicating that the MAR photo-degradation under these conditions slows down after 100 h. Thesefindings are in good agreement with MAR photodegradationprofile.

Finally, the role of irradiation has been further probed onsamples loaded with 55 m g g�1 of ENR which underwent lightexposure for 110 and 165 h in an analogous way as for MARsamples. The d-spacing values obtained, as determined from the fitof the (0 0 1) reflection for the three samples reported in Fig. 6b, arelisted in Table 1.

As for the MAR case, the XRD patterns clearly show that theirradiation leads to a contraction of the MMT c-axes. The d0 0 1

values of the samples exposed to sunlight decrease as aconsequence of the actual photodegradation of ENR. A prolongedexposure to sunlight up to 165 h leads to a further contraction ofthe c-axes suggesting that, possibly, some further photodecompo-sition is promoted by a prolonged irradiation, as confirmed by theENR decay profile. The reflections containing contributions fromthe a and b lattice parameters seem to be insensitive to the ENRdegradation, as for MAR case.

Summing up, XRD results confirm again that the FQ adsorptionis a bulk phenomenon that occurs in the interlayer spacing of theMMT structure [23,30] and indicates that sunlight largely degradedthe adsorbed FQs both on the external surface and in the interlayerspacing.

To further support these findings, additional experiments werecarried out on MMT samples irradiated for 165 h under solar light.These were re-fortified with 77 and 55 m g g�1 of MAR and ENR,respectively, and the measured FQ uptake resulted to be 78 m g g�1

for MAR and 57 m g g�1 for ENR, confirming no loss of efficiency interms of sorption capacity for both antibiotics. This result wellevidences that photodegradation is an efficient path for FQsremoval from clays.

4. Conclusions

In this paper, the solid-state photochemistry of two largelyemployed FQs adsorbed on clay minerals has been investigated.Natural solar light proved to degrade efficiently such drugs. Forboth antibiotics degradation rates were slower onto MMT than onKAO, reasonably due to interactions that occur both onto theexternal surface and in the interlayer spacing. As demonstrated byXRD analyses on MMT samples enriched with MAR and ENR andthen irradiated, the interplanar spacing of the samples exposed tosunlight decreased with increasing the irradiation time, togetherwith a significant loss of order along the c-axes, effectively due tothe photodegradation of the adsorbed FQs within the MMT matrix,and to the formation of by-products. These have been identified,thus showing that amine side-chain oxidation (path (i)) is the mostimportant one, particularly for ENR, while for MAR the cleavage ofthe weak N—N bond occurs.

M. Sturini et al. / Journal of Photochemistry and Photobiology A: Chemistry 299 (2015) 103–109 109

Summing up, the results obtained show that the FQ adsorptionis a bulk phenomenon that occurs also in the interlayer spacing ofthe MMT structure and demonstrate that the sunlight largelydegraded the adsorbed FQs bonded both to the external surfaceand in the interlayer spacing.

Acknowledgment

The authors are indebted to Dr. Alessandro Granata (LabAnal-ysis S.r.l, Casanova Lonati, Pavia, Italy) for the HPLC-ESI-MS/MSmeasurements.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at http://dx.doi.org/10.1016/j.jphoto-chem.2014.11.015.

References

[1] 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.

[2] 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.

[3] M. Sturini, A. Speltini, F. Maraschi, A. Profumo, L. Pretali, E. Fasani, A. Albini,Sunlight-induced degradation of soil-adsorbed veterinary antimicrobialsmarbofloxacin and enrofloxacin, Chemosphere 86 (2012) 130–137.

[4] 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.

[5] 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. (2013), doi:http://dx.doi.org/10.1007/s11356-013-2124-9.

[6] S.K. Khetan, T.J. Collins, Human Pharmaceuticals in the aquatic environment: achallenge to green chemistry, Chem. Rev. 107 (2007) 2319–2364.

[7] R.H. Lindberg, P. Wenneberg, M.I. Johansson, M. Tysklind, B.A.V. Andersson,Screening of human antibiotic substances and determination of weekly massflows in five sewage treatment plants in Sweden, Environ. Sci. Technol. 39(2005) 3421–3429.

[8] E. Zuccato, S. Castiglioni, R. Bagnati, M. Melis, R. Fanelli, Source, occurrence andfate of antibiotics in the Italian aquatic environment, J. Hazard. Mater. 179(2010) 1042–1048.

[9] A. Speltini, M. Sturini, F. Maraschi, A. Profumo, A. Albini, Analytical methods forthe determination of fluoroquinolones in solid environmental matrices,Trends Anal. Chem. 30 (2011) 1337–1350.

[10] J. Tolls, Sorption of veterinary pharmaceuticals in soils: a review, Environ. Sci.Technol. 35 (2001) 3397–3406.

[11] Y. Li, J. Niu, W. Wang, Photolysis of enrofloxacin in aqueous systems undersimulated sunlight irradiation: kinetics, mechanism and toxicity of photolysisproducts, Chemosphere 85 (2011) 892–897.

[12] C. Sirtori, A. Zapata, W. Gernjak, S. Malato, A. Agüera, Photolysis of flumequine:identification of the major phototransformation products and toxicitymeasures, Chemosphere 88 (2012) 627–634.

[13] M.I. Vasquez, M. Garcia-Käufer, E. Hapeshi, J. Menz, K. Kostarelos, D. Fatta-Kassinos, K. Kümmerer, Chronic ecotoxic effects to Pseudomonas putida andVibrio fischeri, and cytostatic and genotoxic effects to the hepatoma cell line(HepG2) of ofloxacin photo(cata)lytically treated solutions, Sci. Total Environ.450–451 (450) (2013) 356––.

[14] 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. 157 (2009) 2722–2730.

[15] L. Rizzo, C. Manaia, C. Merlin, T. Schwartz, C. Dagot, M.C. Ploy, I. Michael, D.Fatta-Kassinos, Urban wastewater treatment plants as hotspots for antibioticresistant bacteria and genes spread into the environment: a review, Sci. TotalEnviron. 447 (2013) 345–360.

[16] P. Sukul, S. Lamshöft, M. Kusari, S. Zühlke, M. Spiteller, Metabolism andexcretion kinetics of 14C-labeled and non-labeled difloxacin in pigs after oraladministration, and antimicrobial activity of manure containing difloxacin andits metabolites, Environ. Res. 109 (2009) 225–231.

[17] D. Dolar, K. Košuti�c, M. Periša, S. Babi�c, Photolysis of enrofloxacin and removalof its photodegradation products from water by reverse osmosis andnanofiltration membranes, Sep. Purif. Technol. 115 (2013) 1–8.

[18] V. Gianotti, M. Benzi, G. Croce, P. Frascarolo, F. Gosetti, E. Mazzucco, M. Bottaro,M.C. Gennaro, The use of clays to sequestrate organic pollutants. Leachingexperiments, Chemosphere 73 (2008) 1731–1736.

[19] P.H. Chang, Z. Li, T.L. Yu, S. Munkhbayer, T.H. Kuo, Y.C. Hung, J.S. Jean, K.H. Lin,Sorptive removal of tetracycline from water by palygorskite, J. Hazard. Mater.165 (2009) 148–155.

[20] Z. Li, P.H. Chang, W.T. Jiang, J.S. Jean, H. Hong, L. Liao, Removal ofdiphenhydramine from water by swelling clay minerals, J. Colloid Interf. Sci.360 (2011) 227–232.

[21] S. Barreca, S. Orecchio, A. Pace, The effect of montmorillonite clay in alginategel beads for polychlorinated biphenyl adsorption: isothermal and kineticstudies, Appl. Clay Sci. 99 (2014) 220–228.

[22] X.L. Hu, Y.P. Huang, A.Q. Zhang, X.R. Zhao, Adsorption of organic pollutants byferric oxide modified montmorillonite, Adv. Mater. Res. 955 (2014) 80–83.

[23] C.J. Wang, Z.H. Li, W.T. Jiang, J.S. Jean, C.C. Liu, Cation exchange interactionbetween antibiotic ciprofloxacin and montmorillonite, J. Hazard. Mater. 183(2010) 309–314.

[24] Z.H. Li, H. Hong, L. Liao, C.J. Ackley, L.A. Schulz, R.A. MacDonald, A.L. Mihelich, S.M. Emard, A mechanistic study of ciprofloxacin removal by kaolinite, ColloidSurf. B 88 (2011) 339–344.

[25] W. Yan, S. Hu, C.Y. Jing, Enrofloxacin sorption on smectite clays: effects of pH,cations and humic acid, J. Colloid Interf. Sci. 372 (2012) 141–147.

[26] 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.

[27] S. Battaglia, L. Leoni, F. Sartori, Determinazione della capacità di scambiocationico delle argille attraverso l'analisi in fluorescenza X di pasticche dipolvere, Atti Soc. Tosc. Sci. Nat. Mem. Serie A 109 (2004) 103–113.

[28] H. Van Olphen, J.J. Fripiat, Data Handbook for Clay Minerals and Other Non-metallic Minerals, Pergamon Press, Oxford, 1979.

[29] I.R. Wilson, J. Jiranek, Kaolin deposits of the Czech Republic and somecomparisons with south-west England, Proc. Ussher Soc. 8 (1995) 357–362.

[30] E. Rivagli, A. Pastorello, M. Sturini, F. Maraschi, A. Speltini, L. Zampori, M. Setti,L. Malavasi, A. Profumo, Clay minerals for adsorption of veterinary FQs:behavior and modeling, J. Environ. Chem. Eng. 2 (2014) 738–744.

[31] M. Sturini, A. Speltini, F. Maraschi, E. Rivagli, A. Profumo, Solvent-freemicrowave-assisted extraction of fluoroquinolones from soil and liquidchromatography-fluorescence determination, J. Chromatogr. A 1217 (2010)7316–7322.

[32] H.M. Ötker, I. Akmehmet-Balcıo�glu, Adsorption and degradation ofenrofloxacin, a veterinary antibiotic on natural zeolite, J. Hazard. Mater. 122(2005) 251–258.

[33] Y. Picò, V. Andreu, Fluoroquinolones in soil-risks and challenges, Anal. Bioanal.Chem. 387 (2007) 1287–1299.

[34] C. Xu, H. Wu, F.L. Gu, Efficient adsorption and photocatalytic degradation ofRhodamine B under visible light irradiation over BiOBr/montmorillonitecomposites, J. Hazard. Mater. 275 (2014) 185–192.

[35] S. Barreca, J.J.V. Colmenares, A. Pace, S. Orecchio, C. Pulgarin, Neutral solarphoto-Fenton degradation of 4-nitrophenol on iron-enriched hybridmontmorillonite-alginate beads (Fe-MABs), J. Photochem. Photobiol. A: Chem.282 (2014) 33–40.

[36] H.T. Lai, J.J. Lin, Degradation of oxolinic acid and flumequine in aquaculturepond waters and sediments, Chemosphere 75 (2009) 462–468.

[37] A. Albini, S. Monti, Photophysics and photochemistry of fluoroquinolones,Chem. Soc. Rev. 32 (2003) 238–250.

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

[39] L. Pretali, E. Fasani, D. Dondi, M. Mella, A. Albini, The unexpectedphotochemistry of marbofloxacin, Tetrahedron Lett. 51 (2010) 4696–4698.