Photocatalytic efficiency and self-cleaning properties under visible light of cotton fabrics coated with sensitized TiO2

Page 1

Pf

RU

a

ARRAA

KSPCSS

1

ptctstpb[[t

0d

Applied Catalysis B: Environmental 104 (2011) 361–372

Contents lists available at ScienceDirect

Applied Catalysis B: Environmental

journa l homepage: www.e lsev ier .com/ locate /apcatb

hotocatalytic efficiency and self-cleaning properties under visible light of cottonabrics coated with sensitized TiO2

. Rahal, T. Pigot, D. Foix, S. Lacombe ∗

MR CNRS 5254, IPREM, Université de Pau et des Pays de l’Adour, Hélioparc, 2 Avenue Président Angot 64053, Pau cedex 09, France

r t i c l e i n f o

rticle history:eceived 11 January 2011eceived in revised form 1 March 2011ccepted 3 March 2011vailable online 10 March 2011

eywords:ensitized TiO2

hotocatalysisoated cotton textilesinglet oxygenelf-cleaning

a b s t r a c t

A simple and reproducible one-pot process for the elaboration of cotton fabrics coated with sen-sitized TiO2 was developed. A molecular precursor [Ti(OR)3(O2C-AQ)] was prepared starting withanthraquinone-2-carboxylic acid (AQ-COOH) and characterized by FTIR, CPMAS NMR and XPS. Hydroly-sis of mixtures of [Ti(OR)3(O2C-AQ)] and Ti(OR)4 at low temperature in an aqueous medium leads to paleyellow cotton fabrics together with the corresponding free P-TiO2/AQ powders. The diffuse reflectanceUV spectra confirmed the shift of absorption towards the visible range. From FTIR, CPMAS NMR and XPSanalysis of the samples (cotton pieces and powders), it was shown that AQ-COOH was not only adsorbedon Titania but tightly bond through a carboxylate complex as in the molecular precursor. Anatase poly-morph was always characterized by XRD even in the absence of a calcination step. Examination by SEM oftreated cotton tissues before and after washing showed stable and homogeneous coating of TiO2 particleson the cotton fibers.

The photocatalytic properties of the samples were investigated, with special care to visible light activa-tion. Under UV light, acetone mineralization was observed, while under filtered visible light, no acetonemineralization occurred. However efficient singlet oxygen addition to di-n-butyl sulfide was evidencedunder visible light. Sulfoxide and sulfone were obtained in better yields using sensitized TiO2 than usingun-modified TiO2 or Anthraquinone alone treated fabrics. Optimum results were obtained with low levelof sensitizing AQ-COOH relative to TiO2 (8%) and no reactivity improvement was noticed with higher

AQ-COOH levels. The cotton pieces coated with sensitized TiO2 also displayed self-cleaning propertiestowards wine stain, either under solar illumination or even in indoor light.

The better efficiency of sensitized TiO2-coated cotton is accounted for by a synergy effect between TiO2

and AQ-COOH, enhancing the formation of Reactive Oxygen Species (singlet oxygen and/or superoxideradical-anion). However, under these conditions, the production of hydroxyl radical seems to be ruled

out.

. Introduction

The development of (multi)-functionalized textiles with specificroperties is a field of growing interest. Several textile modifica-ions are possible, such as incorporation of functional additives,hemical grafting of functional additives, or post-equipping tex-iles with functional coatings. The latter method often relies onol–gel chemistry, starting with inorganic sols based on nanopar-icular modified silica and other metal oxides [1]. Recently, someapers were devoted to TiO2-functionalized textiles which cane photoactivated, inducing self-cleaning [2–14], UV-blocking

8,15], photo-oxidative [16–19] and/or bactericidal properties3,9,20]. Besides the most frequently studied photocatalytic tex-iles treated with a TiO2-based coating, textiles producing singlet

∗ Corresponding author. Tel.: +33 559 407 579.E-mail address: [email protected] (S. Lacombe).

926-3373/$ – see front matter © 2011 Elsevier B.V. All rights reserved.oi:10.1016/j.apcatb.2011.03.005

© 2011 Elsevier B.V. All rights reserved.

oxygen under irradiation due to incorporation of a photosensitizer(PS) also display interesting bactericidal [21–29] or detoxifyingproperties [30].

For the first series of textiles treated with a TiO2 coating, one ofthe main drawbacks is related to the poor overlapping of the TiO2absorption spectrum with the solar emission spectrum, leading toa maximum activation of the modified textiles in the UV-A range.On the contrary, PS-containing textiles are most often activated inthe visible range. It should be emphasized that in the case of TiO2-treated textiles, the oxidative and bactericidal properties are oftenascribed to the formation of hydroxyl HO• radicals [19–20], while inthe case of PS-containing textiles, singlet oxygen is mainly involved[21–23,28,29]. In both cases membrane damages are expected.

In order to enhance the reactivity of TiO2 in the visible range,

several chemical modifications are possible: for example cou-pling with a photosensitizer or with another narrow band-gapsemi-conductor absorbing in the visible range, doping with metalimpurities, preparing oxygen deficient TiO2 and doping TiO2 with

Page 2

3 B: Env

nr

cvoTdz(tdrkfsfs[o

CE

[dRcob

adwbacpwi

2

d

2a

o(taacaitd4ilria

62 R. Rahal et al. / Applied Catalysis

on-metal atoms (anion doping). All these methods were recentlyeviewed by Fujishima et al. [31].

In this paper, we report the elaboration of cotton samplesontaining either TiO2 or sensitized TiO2 in order to increase itsisible range efficiency. Sensitization involved the direct couplingf a sensitizer (anthraquinone 2-carboxylic acid or AQ-COOH) to aiO2 precursor, Ti(OiPr)4, by an aqueous one-step process alreadyescribed by Rahal et al. [32] for the synthesis of para-amino ben-oic acid-TiO2 hybrid nanostructures. The choice of this sensitizerAQ-COOH) was based upon its common use in textile dying indus-ry as a fixing tank agent for various usual dies. Under industrialying conditions, anthraquinone (AQ) is usually grafted to the fab-ics by a reductive/oxidative process [33]. Moreover AQ is a wellnown PS, capable of producing singlet oxygen by energy trans-er under irradiation [34]. Anthraquinone 2-sulfonate was alsohown to produce superoxide radical anion O2

•− by electron trans-er [34,35] in addition with singlet oxygen [36]. The question ofuperoxide radical anion formation sensitized by AQ was raised37–39] but should be possible considering the reduction potentialf AQ and O2 (E0

AQ/AQ•− = −0.62 V and E0O2/O2

•− = −0.33 V vs SHE in

H3CN, respectively [37,40]). It should be noted that the value of0O2/O2

•− differed significantly in previous papers (between −0.69

41] and −0.53 V [42], with values between −0.47 and −0.60 Vepending on pH [43], vs SHE in CH3CN (taking ESHE = ESCE + 0.25 V).ecently, we showed that anthraquinone, easily incorporated onotton fabrics by a reductive–oxidative process, could efficientlyxidize di-n-butylsulfide and that the treated cotton pieces had aactericidal effect when irradiated at 420 nm [44].

The characterization of these modified cotton samples waschieved by various methods in order to demonstrate both TiO2eposition and the change induced by the photosensitizer. Theyere also tested under different irradiation conditions in the visi-

le range. They were first evaluated under solvent-free conditionsgainst gaseous acetone mineralization to investigate TiO2 effi-iency and second, against alkylsulfides oxidation to check theossible formation of singlet oxygen. Their self-cleaning propertiesere also investigated. The results obtained are discussed accord-

ng to the assumed reaction mechanisms.

. Experimental

All the synthetic procedures for preparing the fabrics and pow-ers are detailed in Supplementary Information (SI).

.1. Study of the photocatalytic activity by mineralization ofcetone

This experiment was carried out by injecting a defined quantityf acetone (3 �l) through a septum into a sealed reactor (350 ml)corresponding to 2800 ppmv as the initial concentration of ace-one) containing the tissue (15 cm2). This cylindrical reactor hadcircular pyrex window on the top through which it was irradi-

ted. It was linked to a TCD (thermal conductivity detector) gashromatograph with micro-catharometers (GC-4900 VARIAN), sos to observe simultaneously the increase of CO2 concentrationnduced by acetone mineralization and the drop of acetone concen-ration during irradiation. The irradiations were carried out with aevice containing four 420 nm lamps (Rayonet RPR 4190-A) with a20 nm cut off filter (Schott GG420). A picture of the whole device

s given in Figure S1, SI. The distance between the samples and the

amps was 10 cm. The irradiance was measured with a spectro-adiometer ILT900-R from International Light Technology (totalrradiance 4.3 mW/cm2). The irradiance spectra of the lamps withnd without filter are given in Figures S2, SI.

ironmental 104 (2011) 361–372

2.2. Study of the photocatalytic activity by photooxydation ofdi-n-butylsulfide (DBS)

This experiment was carried out by adding a tissue (15 cm2) ina modified Petri dish (with a septum at the top for injection of DBS,tightly sealed in order to prevent any evaporation of reactants andproducts. 10 �l of DBS solution was injected on the tissue throughthe septum. Half of the Petri dishes were kept in the dark at 20 ◦Cand the other half were irradiated for 24 h with a device madeof four 420 nm lamps (Philips TLD-K 30 W/03) and a 420 nm cutoff filter made of a polyester/acrylic sheet (Film anti-UV SanergiesNeutral 240 C, thickness 60 �m) placed in an incubator regulated at20 ◦C. The distance between the samples and the lamps was 10 cm.The irradiance was measured with a spectroradiometer ILT900-R(total irradiance 5.1 mW/cm2). The irradiance spectra of the lampswith and without the anti-UV filter are given in Figure S3, SI. After24 h, without opening the Petri dish, 5 ml of acetonitile contain-ing cyclododecane (5 × 10−3 M) as internal standard were addedthrough the septum and the washing solutions were analysed by GC(Varian 3900) equipped with Chrompack column CPSil-5CB (30 m,0.25 mm, 1 �m). The same experiment was also performed underindoor light of the laboratory.

2.3. Discoloration of red wine stain

500 �l of red wine was pipetted on the tissues (treated or nontreated). These tissues were irradiated for 24 h with a solar simu-lator (Oriel Instruments with a 150 W Xe arc lamp) with a 420 nmcut off filter (Schott GG 420).

The same experiment was carried out under indoor light with-out any other irradiation. Half of these samples were covered withblack paper for comparison.

2.4. Materials and equipments

The diffuse reflectance spectra in the UV–visible range (DRUV)were measured at room temperature with a double beam Cary 5000spectrophotometer equipped with an 11 cm diameter integratingsphere and a home made powder holder. The diffuse reflectancespectra were corrected vs a white standard (Teflon 55 microns,Aldrich). The Kubelka–Munk model describes the light penetrationin porous media with only two parameters: an absorption coeffi-cient, k, and an isotropic scattering coefficient, s (which both haveunits of cm−1) [45].

The Fourier transform infrared (FTIR) spectra of the samplesin Nujol for (1) or in KBr pellets for the modified TiO2 powderswere recorded using a MAGNA-560 spectrometer at a resolution of4 cm−1 in an absorption mode using 400 scans.

All 1H HRMAS NMR spectra were recorded on a Bruker Avanceinstrument operating at 400.13 MHz using a 4 mm HRMAS 1H/13Cprobe head. Powdered samples were packed in a 4-mm zirconiarotors, sealed with Kel-FTM caps and spun at 7 kHz and at a con-tact time of 500 ms. Chemical shifts were determined relative totetramethyl silane (TMS) used as control.

XPS measurements were carried out with a Kratos Axis Ultraspectrometer, using a focused monochromatized Al K� radiation(h� = 1486.6 eV). The XPS spectrometer was directly connectedthrough a transfer chamber to an argon dry box, in order toavoid moisture/air exposure of the samples. For the Ag3d5/2 line,the full width at half maximum (FWHM) was 0.58 eV underthe recording conditions. The analysed area of the samples was300 �m × 700 �m. Peaks were recorded with constant pass energy

of 20 eV. The pressure in the analysis chamber was around5.10−9 mbar. The binding energy scale was calibrated from thehydrocarbon contamination using the C1s peak at 285.0 eV. Corepeaks were analysed using a nonlinear Shirley-type background

Page 3

R. Rahal et al. / Applied Catalysis B: Environmental 104 (2011) 361–372 363

(b)

+ Ti (OiPr)4Cotton pieces

Hydrolysis in 100 ml H2Ofor 3 hours

Washing

1 hour for the tissues 18 hours for the powders

Thermal treatment at 80° C

(a)

Ti(OiPr)4

18 hours RT

AQ-CO2HAQ-CO2H

18 hours RT

[Ti(OiPr)3(O2C-AQ)][Ti(OiPr)3(O2C-AQ)]

Cotton pieces

Modified cotton pieces + TiO2 powders

Path (b)TiO2/AQ(-)@T

&P-TiO2/AQ(-)

Path (a)TiO2/AQ(+)@T

&P-TiO2/AQ(+)

Fig. 1. Scheme of the synthesis of: path (a) TiO2/AQ(+)@T and P-TiO2/AQ(+); path(b) TiO2/AQ(−)@T and P-TiO2/AQ(−).

[llr

Irt(a

dcrttdtpr

eEv

3

rpl(wmb

As expected, the DRUV spectrum of crude cotton presented noabsorption band in the UV–visible range (Fig. 3). The TiO2@T tis-sue displayed the characteristic cut-off around 380 nm, usually

46]. The peak positions and areas were optimized by a weightedeast-squared fitting method using 70% Gaussian, 30% Lorentzianineshapes. Quantification was performed on the basis of Scofield’selative sensitivity factors [47].

Thermogravimetric analysis (TGA) was performed using a TAnstruments (Guyancourt, France) apparatus 2950 TGA. The purgeate was 100 ml min−1, with a flow distribution of 20 ml min−1 tohe balance chamber (nitrogen) and 80 ml min−1 to the furnaceoxygen). The heating rate was 5 ◦C min−1 and all samples werenalysed in the powder form.

Powder XRD patterns were recorded on an INEL XRG 3000iffractometer using a curved position-sensitive detector (CPS 120)alibrated with Na2Ca3Al2F14 as standard. The monochromaticadiation applied was CuK� (1.5406 A) from a long fine focus Cuube operating at 40 kV and 35 mA. Scans were preformed overhe 2� range from 5◦ to 115◦. Accurate unit cell parameters wereetermined by a least squares refinement from data collected byhe diffractometer. Basal spacing distances for the different sam-les analysed were determined from the position of the d (0 0 3)eflection.

Observation of the surface morphology was performed using annvironmental scanning electronic microscopy (ESEM Electroscan3), and the samples were analysed without metallisation, with aoltage acceleration of 25 kV at room temperature.

. Results

The methods of preparation of the various tissues are summa-ized in Fig. 1 and the names and characteristics of the preparedowders (P) and tissues (T) are summarized in Table 1. In the fol-

owing samples names, AQ(+) stands for a high AQ-COOH/Ti ratiothe only source of Ti is the precursor [Ti(OiPr)3(O2C-AQ)], Path a),hile AQ(−) stands for a low AQ-COOH/Ti ratio (the source of Ti is aixture of Ti(OiPr)4 and of the precursor [Ti(OiPr)3(O2C-AQ)], Path

).

Fig. 2. Infrared spectra (1800–650 cm−1) of AQ-CO2H, [Ti(OiPr)3(O2CAQ)] (1) andNujol.

3.1. Synthesis and characterization of modified cotton fabrics

The synthetic method for the precursor [Ti(OiPr)3(O2C-AQ)] (1)preparation was previously reported starting with another aro-matic molecule (p-aminobenzoic acid) to synthesize hybrid TiO2nanoparticles with controlled amounts of organics for cosmeticsand ionic separation [32,48]. The same method was used here, usingan anthraquinone derivative, AQ-COOH, as AQ is well-known for itsphotosensitizing properties [34–37,49] and as an industrial dyingcompound for fixation of dyes on cotton and polyester.

The key-step is the preparation of the precursor (1) as a yellow-orange powder, obtained from the reaction between equimolaramounts of Ti(OiPr)4 and anthraquinone-2-carboxylic acid (AQ-CO2H) in isopropanol at room temperature.

The comparison of FT-IR spectra of AQ-CO2H and (1) (Fig. 2)showed in both cases the presence of the C O stretching bandof the quinone moiety at 1677 cm−1. On going from AQ-CO2Hto (1), the drop of the C O stretching band at 1699 cm−1 indi-cates the disappearance of the free carboxylic moiety. New sharpbands appears, corresponding to asymmetric stretching �as(CO2

−)at 1608, 1579 cm−1 and symmetric stretching �s(CO2

−) at 1556and 1479 cm−1. The value �� = [�as(CO2) − �s(CO2)] smaller than100 cm−1 suggests a bidentate chelating/bridging coordination[50]. This precursor (1) was used for the synthesis of our modifiedtissues.

Hydrolysis of (1) was performed without any organic sol-vent: the neat powder (1) for P-TiO2/AQ(+) and TiO2/AQ(+)@T orthe solution ([Ti(OiPr)3(O2C-AQ)] + Ti(OiPr)4) for P-TiO2/AQ(−) andTiO2/AQ(−)@T, were added to boiling aqueous solution containing0.01 equivalent of NBu4Br per atom of titanium and 15 pieces ofcotton (15 cm2 each) under vigorous magnetic stirring. An imme-diate pale yellow precipitate appeared. After heating under reflux,the yellow tissues were removed and the remaining suspensionwas centrifuged.

For UV–visible diffuse reflection spectroscopy, theKubelka–Munk relation measuring K/S (K and S are respec-tively the absorption and scattering coefficients of TiO2) for thicksamples with low optical transmittance allowed the conversion ofthe reflectance (R) into the equivalent of absorption spectra viaF(R) remission function:

F(R) = K

S= (1 − R)2

2R

observed with TiO2. No significant difference between this spec-

Page 4

364 R. Rahal et al. / Applied Catalysis B: Environmental 104 (2011) 361–372

Table 1Name and characteristics of the prepared samples.

Powders (P) Cotton fabrics (T)

Pure anthraquinone-2-carboxylic acid AQ-COOH –Prepared from commercial TiO2 Degussa P25 P-TiO2P25 TiO2P25@TPrepared from a mixture of TiO2 P25 and AQ-COOH.

P-TiO2P25/AQ(−) TiO2P25/AQ(−)@TMolar ratio Ti/AQ-COOH = 60TiO2 prepared from Ti(OiPr)4 P-TiO2 TiO2@TPrepared from a mixture of the home-made precursor [Ti(OR)3 (O2C-AQ)]a with Ti(OiPr)4.

P-TiO2/AQ(−) TiO2/AQ(−)@TMolar ratio Ti/AQ-COOH = 60Prepared from the home-made precursor [Ti(OR)3(O2C-AQ)]a.

P-TiO2/AQ(+) TiO2/AQ(+)@TMolar ratio Ti/AQ-COOH = 1Prepared by reductive/oxidative coupling of anthraquinone (AQ) on cotton tissues

a R = iPr.

FT

twsrpcPtsw3stlnd

etntpetti

TD

2 2to �as(CO2

−) and �s(CO2−) vibrations. For P-TiO2/AQ(−) we sug-

gest only one structural form of the titanium–carboxylate complexwithout free carboxylic acid. For the two other samples, mixtures

4

5

6

(Ene

rgy)

]

TiO2@T

TiO2P25/AQ(-)@T

ig. 3. DRUV spectra of crude cotton samples, TiO2@T, TiO2AQ(−)@T,iO2P25/AQ(−)@T, AQ@T.

rum and those of TiO2P25@T (not shown) and TiO2P25/AQ(−)@Tas noticed. A shift towards the visible range up to 500 nm, con-

istent with the spectrum of AQ@T prepared by conventionaleductive/oxidative process, was only observed for the sample pre-ared with the home made precursor (1), TiO2/AQ(−)@T. Whenomparing the tissue TiO2/AQ(−)@T and the corresponding powder-TiO2/AQ(−) (Figure S4, SI), the spectrum of the powder appearedo be more intense than that of the tissue. Both these spectra aretrikingly different from the spectrum of AQ-CO2H: neither theeak n–�* band at 425 nm, nor the more intense �–�* band at

50 nm of AQ is observed, due to a high amount of TiO2 in theseamples (molar ratio Ti/AQ = 60). For all these samples, the spec-ra were dominated by the characteristic features of TiO2, more oress shifted towards the visible range. On the other hand, it can beoticed that the spectrum of P-TiO2/AQ(+) (not shown) was mainlyominated par AQ-COOH feature (Fig.S4, SI).

The bandgap energies Eg of the different materials may bestimated using the equation ˛ = A(h� − Eg)n/Eg, where ˛ is absorp-ion coefficient, A is a constant, h� is the energy of light and

is a constant on the nature of the electron transition [51]. Inhe case of TiO2, n = 2 assuming an indirect bandgap [52]. A isroportional to Kubelka–Munk function (F(R)) and the bandgap

nergy can be obtained from the plots of (F(R)h�)1/2 vs h� ashe intercept at (F(R)h�)1/2 = 0 of the extrapolated linear part ofhe plot (Fig. 4) [53]. The determined bandgaps are summarizedn Table 2.

able 2etermined bandgaps for TiO2 and various AQ-sensitized TiO2.

Sample Eg (eV)

TiO2/AQ(−)@T 3.14TiO2P25/AQ(−)@T 3.25TiO2@T 3.3

– AQ@T

The value obtained for TiO2@T was consistent with literaturedata (3.3 eV, [53]), with no significant change for TiO2P25/AQ(−)@T(3.25 eV). A slight shift towards lower energy (3.14 eV) was onlynoticed for the sample TiO2/AQ(−)@T, as already observed with N-or C-doped TiO2 [54], and is mainly due to AQ-COOH absorption(Figure S4-SI).

The crystallinity of TiO2 on the modified textile surface wasinvestigated by powder X-ray diffraction (PXRD) with the diffrac-tion angle (2�) in the range 20–60 ◦C. Fig. 5(left) shows thediffractograms of two samples TiO2@T and TiO2/AQ(−)@T. Broadand small peaks at � = 25◦, 38◦, 48◦ and 55◦ were observed, assignedto the anatase phase of TiO2 (anatase, JCPDS card 21-1272). Notethat these TiO2 coated textiles were made at 100 ◦C in an aque-ous medium without any chemical treatment (neither calcinationnor acid addition). The intense peak at 23◦ and the broad one at34◦ constitute the typical XRD pattern of cellulose fibers [10]. XRDdiffractograms of P-TiO2/AQ(−) and P-TiO2P25/AQ(−) are similar(Fig. 5(right)) and, as previously, the presence of AQ-CO2H doesnot affect the crystalline phase of TiO2, even if the broad peaks ofP-TiO2/AQ(−) compared to those of P-TiO2P25/AQ(−) indicate thatthe nanoparticles obtained with our method are smaller (7 nm) [32]than those of TiO2P25 (25 nm).

The IRTF spectra of the powders in KBr pellets are shown in Fig. 6.As for the precursor (1), the drop in intensity of the 1695 cm−1

band, assigned to the C O free carboxylic group, is noticed forP-TiO2/AQ(+) and P-TiO2P25/AQ(−). This band even completelydisappears for P-TiO2/AQ(−). On the contrary the quinone bandat 1672 cm−1 is observed in all the samples. Broad new bands at1639 and 1548 cm−1 observed respectively for P-TiO2/AQ(−) and P-TiO /AQ(+) are not observed for P-TiO P25/AQ(−) and are assigned

0

1

2

3

4,03,93,83,73,63,53,43,33,23,13,0

SQ

RT

[(K/S

)*

E(ev)

TiO2/AQ(-)@T

Fig. 4. Bangap determination using (F(R)*E)1/2 vs E plots for TiO2@T,TiO2P25/AQ(−)@T, TiO2/AQ(−)@T.

Page 5

R. Rahal et al. / Applied Catalysis B: Environmental 104 (2011) 361–372 365

605550454035302520

Cou

nts

(a.u

)

TiO2@T

TiO2/AQ(-)@TAnataseCotton

605550454035302520

Cou

nts

(a.u

.)

P-TiO2/AQ(-)P-TiO2P25/AQ(-)Anatase

Rutile

TiO2/

op

aaaD1rib

swa(bpAiawwAmtoaiAia(

FT

2θ

Fig. 5. PXRD diffractograms of (left): the cotton tissues TiO2@T and

f the adsorbed acid and of the titanium–carboxylate complex arerobably obtained.

CPMAS RMN 13C spectra of the two solid samples P-TiO2/AQ(−)nd P-TiO2/AQ(+) were similar and in good agreement with annthraquinonic structure. They showed two quinone carbons (C O)t 180 ppm and 179 ppm, close to the corresponding peaks inMSO-d6 solution of AQ-COOH (182 ppm). However, the peak at71 ppm attributed to the carbonyl group was significantly shiftedelative to the free acid in DMSO-d6 solution (165 ppm). This results consistent with the involvement of the carboxylate moiety in theonding with titanium atoms.

The thermogravimetric analysis and differential thermal analy-is (DTA) curves recorded in air up to 600 ◦C exhibited weight lossithin the temperature range 20–600 ◦C (Fig. 7). The two materi-

ls P-TiO2/AQ(−) and P-TiO2/AQ(+) showed one main weight loss4.9–48.8%, Table 3) above 400 ◦C corresponding to a chemicallyound complex with TiO2, as it occurs at much higher tem-erature than for free anthraquinone-2-carboxylic acid (340 ◦C).ccordingly, the first but weaker weight loss (3.1–11.2%, Table 3)

n the range 200–390 ◦C may be assigned to minor amounts ofdsorbed anthraquinone-2-carboxylic acid. The curve obtainedith P-TiO2P25/AQ(−) displays more obviously two ranges ofeight loss, indicating the presence of both free and complexedQ-COOH, consistent with the previous IR data. All these measure-ents are in agreement with the weight of introduced AQ during

he synthesis (Table 3), although some loss of introduced AQ wasbserved in a more significant extent with P-TiO2P25/AQ(−) rel-tive to P-TiO2/AQ(−). This result demonstrates a more efficientncorporation of AQ-COOH on TiO2 with our synthesis method.n accurate examination of the DTA curves (inset of Fig. 8(right))

ndicates that the temperature loss occurs for P-TiO2/AQ(−) atn intermediate temperature (420 ◦C) between that of AQ-COOH350 ◦C) and that of P-TiO2P25/AQ(−) (460 ◦C). This result is con-

ig. 6. Infrared spectra (2000–700 cm−1) of AQ-CO2H; P-TiO2P25/AQ(−);. P-iO2/AQ(−) and P-TiO2/AQ(+) in KBr pellets.

2θ

AQ(−)@T; (right): the powders P-TiO2/AQ(−) and P-TiO2P25/AQ(−)

sistent with a stronger bonding of AQ-COOH with TiO2 P25 thanwith home-made TiO2.

The morphology of the obtained samples is shown in the SEMpictures of the treated cotton pieces in comparison with untreatedsamples in Fig. 8. The presence of TiO2 is obviously observed aroundthe cotton fibers for both TiO2P25/AQ(−)@T and TiO2/AQ(−)@T.Uncontinuous TiO2 nanocristallites appear to be homogeneouslydeposited on the fibers and not between them. The picture of thefabric before and after washing (Figure S5, SI) in a machine with adetergent containing no bluing dye indicates that the TiO2 coatingis resistant to mechanical washing.

For the XPS analysis, we first describe the spectra of referenceproducts (Fig. 9), then the spectra of the modified powders (Fig. 10)and finally those of the modified tissues (Figure S6, SI). Quantitativeresults are summarized in Tables S1–S3, SI.

First, the four reference compounds P-TiO2P25, P-TiO2, AQ-COOH and the raw tissue were studied by XPS (Fig. 9). For bothP-TiO2P25 and P-TiO2, the Ti2p3/2 peak appears at 458.8 eV, ener-getic position characteristic of the Ti4+ form. The O1s peak consistsof a maximum at around 530.0 eV assigned to the O of the lattice ofTiO2, and of a shoulder at 531.8 eV assigned to the hydroxyl (–OH)of the surface.

For AQ-COOH, the C1s peak is made of a maximum at 284.6 eVcorresponding to the carbon atoms of the aromatic rings and ofsmaller peaks at 287.2 eV and 289.1 eV assigned to the carbon atomsof the C O and COOH groups, respectively. The satellite at 290.7 eVis due to the aromatic rings. The O1s highest peak at 531.3 eV isattributed to the oxygen atom of the carbonyl bonds ( O), and thesecond peak at 533.0 eV is attributed to the acid (–OH) oxygen atom.

For the tissue, 3 peaks can be distinguished in the C1s signal. Theone at 285.0 eV corresponds to hydrocarbon contamination. Thetwo other are characteristic of the tissue: the highest at 286.8 eVis attributed to the C–O–H and C–C–O bonds, while the smallestat 288.3 eV is correlated to the O–C–O environments. For the O1speak, all the oxygen atoms are equivalent (C–OH or C–O–C) andappear at 533.1 eV.

Four powders were then analysed (Fig. 10): the precursor[Ti(OR)3(O2C-AQ)], P-TiO2P25/AQ(−), P-TiO2/AQ(−), P-TiO2/AQ(+).For the precursor, the involvement of the COOH group in the bond-ing to TiO2 (forming a COO–Ti complex) may be deduced from thefollowing observations:

• besides the O1s peak of TiO2 at 529.8 eV and the peak at 531.5 eVcorresponding to the C O of AQ-COOH and to the oxygen atomsof OiPr groups, the peak at 533.0 eV (–OH) in AQ-COOH almostdisappeared. If the O-H group is now a O–Ti group, as oxygenis the first neighbour of Ti, the influence of its electropositivity

is strong. It may thus be assumed that the peak at 533.0 eV ofAQ-COOH is strongly shifted and included in the peak at 531.5 eV.

• for C1s, besides the peak of the carbon atoms of the aromaticring at 284.6 eV, and the peak at 287.3 eV assigned to the C O

Page 6

366 R. Rahal et al. / Applied Catalysis B: Environmental 104 (2011) 361–372

0

20

40

60

80

100

120

55045035025015050

Wei

ght (

%)

Temperature (°C)

AQ-CO2HP-TiO2P25/AQ(-)P-TiO2/AQ(-)P-TiO2/AQ(+)

-2,5

-2

-1,5

-1

-0,5

0

0,5

55045035025015050

Der

ivat

ive

wei

ght

(% m

in-1

)

Temperature (°C)

-0,2-0,18-0,16-0,14-0,12-0,1-0,08-0,06-0,04-0,0200,02

-2

-1,5

-1

-0,5

0

0,5

550500450400350

Fig. 7. TGA (left) and DTA curves (right) of AQ-COOH, P-TiO2/AQ(+), P-TiO2P25/AQ(−), P-TiO2/AQ(−).

Table 3Determination of the percentage weight of introduced AQ-COOH in the modified TiO2 powders deduced from the TGA analysis and comparison with the introduced amountduring the synthesis.

P-TiO2P25/AQ(−) P-TiO2/AQ(−) P-TiO2/AQ(+)

Introduced AQ (% weight) 20 10 76

•

Determined AQ (% weight) 8.9Temperature range 190–290 390–510% Weight 3.5 5.4

carbon, the last peak present for AQ-COOH at 289.1 eV (attributedto COOH) is shifted to 288.4 eV in the precursor. The influence ofTi electropositivity on the C atom of COOTi is thus slighter thanfor oxygen atoms.the atomic ratio deduced from the spectra (O: 27, C:70,

Ti:3; Table S2, SI) is consistent with the molecular formula[Ti(OR)3(O2C-AQ)].

Fig. 8. SEM of TiO2-coated cotton fa

8.3 63.3200–340 340–500 280–390 390–5403.1 5.2 11.2 52.1

For P-TiO2P25/AQ(−), the signal is dominated by the TiO2 fea-tures and AQ-COOH is hardly detectable, in agreement with thelarge Ti amount (17%, Table S2, SI):

• the O1s spectrum presents the same peak positions as P-TiO2P25(529.8 eV and 531.5 eV).

• For the C1s peak, the peak at 285.0 eV cannot be assigned to AQ-COOH (284.6 eV), but instead to hydrocarbon contamination. The

brics at various enlargement.

Page 7

R. Rahal et al. / Applied Catalysis B: Environmental 104 (2011) 361–372 367

ounds

bo

Tf

Fig. 9. C1s and O1s XPS spectra of the reference comp

two other peaks at 287.7 eV and 290.5 eV are tentatively assignedto –COO and CO3 contamination.

To summarize, for P-TiO2P25/AQ(−), the O1s signal is dominatedy the feature of TiO2, whereas the C1s signal is not characteristic

f AQ, but of hydrocarbon contamination.

For P-TiO2/AQ(−), the O1s signal is very similar of the one ofiO2. This is consistent with the large amount of TiO2 (atomic ratioor Ti atom 20%, Table S2, SI).

P-TiO2 P25, P-TiO2, AQ-COOH and raw cotton tissue.

For P-TiO2/AQ(+), the signals of O1s and C1s peaks are domi-nated by AQ-COOH features. The O1s peak is made of signals at529.8 eV (TiO2 lattice), 531.3 eV (C O of AQ-COOH) and 532.8 eV.As this latter peak is broader and at lower energy than the one of theOH of AQ-COOH at 533.1 eV, it probably involves a double environ-

ment for this carboxylate oxygen (adsorbed AQ-COOH not reactedwith TiO2, and complexed COOTi group appearing at lower bind-ing energy than 533.1 eV of COOH). Besides the C1s peak at 284.6 eV(carbon of the aromatic rings) and at 287.2 eV (carbonyl C O bond),

Page 8

368 R. Rahal et al. / Applied Catalysis B: Environmental 104 (2011) 361–372

Fig. 10. C1s and O1s XPS spectra of the precursor [Ti(OR)3 (O2C-AQ)], P-TiO2P25/AQ(−), P-TiO2/AQ(−) and P-TiO2/AQ(+).

t(a

te

he peak at 288.8 eV is larger and at lower energy than in AQ-COOH289.1 eV). This could be due to a complex environment of the C

toms of the COOH and COOTi groups as assumed previously.

The XPS spectra (O1s and C1s) of the tissues (Figure S6, SI) showhat Ti is present in the proportions given in Table S3, SI withnergetic position and width identical to the Ti2p peak of TiO2,

characteristic of Ti4+, as confirmed by the O1s peaks. The O1s andC1s signals of AQ-COOH are hidden by the characteristic features of

the tissues, and can only be deduced from the larger peak width andfrom their relative intensity. It can be noticed that the atomic ratioof Ti grafted on the surface of the tissues is higher for TiO2/AQ(−)@T(17%) than for TiO2P25/AQ(−)@T (11%).

Page 9

R. Rahal et al. / Applied Catalysis B: Environmental 104 (2011) 361–372 369

Table 4Yield of solvent-free oxidation products upon irradiation at 420 nm at 20 ◦C of modified cotton fabrics in sealed Petri dishes containing DBS (DBSO: di-n-butyl sulfoxide,DBSO2: di-n-butyl sulfone).

Crude tissue TiO2@T AQ@T TiO2P25/AQ(−)@T TiO2/AQ(−)@T

DBS 100 35DBSO 53DBSO2 12

Table 5Yield of oxidation products upon irradiation under indoor light at room temperatureof modified cotton fabrics in sealed Petri dishes containing DBS (DBSO: di-n-butylsulfoxide, DBSO2: di-n-butyl sulfone).

TiO2@T AQ@T TiO2P25/AQ(−)@T TiO2/AQ(−)@T

DBS 100 100 100 3

dt

3

so

3i

rlcflow2

e2tCc(eoetwTawnm

l

DBSO – – – 80DBSO2 – – – 17

It should be mentioned that control XPS analysis after irradiationid not show any significant modification of the signals, evidencinghe stability of the coating.

.2. Photocatalytic tests

As the photocatalytic test carried out with TiO2/AQ(+)@Thowed that these fabrics were not active, we focus in the followingn TiO2/AQ(−)@T and P-TiO2/AQ(−).

.2.1. Photocatalytic degradation of acetone under visible lightrradiation

Acetone mineralization photocatalysed by TiO2 is a well knowneaction, ultimately giving rise to CO2 [55]. Fig. 11 shows the evo-ution of the CO2 concentration (relative to the initial ambient CO2oncentration) upon irradiation with the RPR 4190A lamps of theat reactor containing 3 �l of acetone (2500 ppmv) where a piecef cotton tissue (3 cm × 5 cm) was placed. The upper pyrex windowas first covered by a Schott GG420 filter, which was removed after

5 h irradiation.

With the 420 nm cut-off filter (<0.01 mW cm−2 UVA) and what-ver the studied sample, CO2 evolution was very weak after4 h irradiation. The evolution of C (CO2 concentration) withime relative to C0 (initial CO2 concentration i.e. atmosphericO2 ∼ 400 ppmv) was not more than 1.4 with the most effi-ient TiO2/AQ(−)@T. Relative to the initial acetone concentration2800 ppmv), this means that less than 3% of acetone are min-ralized. With the sample TiO2/AQ(+)@T, no mineralization wasbserved and the fabrics was strongly colored in orange at thend of irradiation, probably due to AQ oxidation. When the fil-er was removed (0.13 mW cm−2 UVA), mineralization to CO2as more efficient with all the samples: it reached 13% with

iO2P25/@T and TiO2P25/AQ(−)@T, 8% or less with TiO2/AQ(−)@Tnd TiO2@T. These results show that TiO2 P25, sensitized or notith anthraquinone, is the most efficient under UV light, whereas

one of the TiO2-coated cotton fabrics prepared with laboratory-ade precursor mineralize acetone in significant amounts.Control GC-MS experiments under the same conditions with

abeled acetone H3C13C(O)CH3 showed that 13CO2 (m/z 45) was

72 69 223 29 34

5 2 64

produced (Figure S7, SI), indicating that carbon dioxide originatedfrom acetone, and not from the tissue or from decomposition ofAQ on the tissue. It may be noted that the relative abundance ofm/z 45/44 peaks decreased from 17% after 18 h under UVA irradi-ation (fluorescent lamps with maximum emission at 350 nm) to13% after 48 h under mainly visible irradiation (unfiltered 420 nmlamps) and to less than 1% under neat visible irradiation (420 nmlamps filtered with the Schott GG 420 cut-off filter). These resultsconfirmed that 13CO2 production was almost inefficient under vis-ible light. From these experiments and from a “blank” experimentwithout acetone where no carbon dioxide evolution was detected,it may also be concluded that no degradation of the tissues or ofanthraquinone was observed with visible light irradiation.

3.2.2. Photocatalytic oxidation of di-n-butylsulfide (DBS) undervisible light irradiation (420 nm maximum emission with 420 nmPMMA cut-off filter)

Further experiments were carried out in order to assess thephoto-activity of the coated cotton fabrics against reactants usuallysensitive to singlet oxygen, such as DBS. For these experiments thetissues were placed in modified sealed Petri dishes where the reac-tants were introduced without any solvent through a septum. After24 h irradiation with Philips TLD-K 30 W/03 lamps filtered with asheet of polyester/acrylic sheet in an incubator at 20 ◦C, 5 ml of ace-tonitrile were introduced into the Petri dishes and the extractionsolutions were analysed by GC and GC/MS. The oxidation reactiongave rise to several products (di-n-butyl sulfoxide DBSO, di-n-butylsulfone DBSO2, di-n-butyl disulfide DBDS, n-butyl n-butane thio-sulfonate DBSSO2), according to Eq. (1). The results are summarizedin Table 4.

(1)

Under these conditions with the untreated tissue or with thesame samples kept in the dark, no sulfide oxidation occurred. All thetreated tissues gave rise to oxidation, as di-n-butylsulfoxide (DBSO)and di-n-butylsulfone (DBSO2) were detected. The efficiency ofTiO2@T can be attributed to unfiltered UV wavelengths with this fil-ter (0.4 mW cm−2 UVA). With AQ@T (6% AQ) and TiO2P25/AQ(−)@T,the same results were obtained: only 30% of sulfide oxidation wasobserved in both cases, with major formation of sulfoxide. The bestefficiency was obtained with TiO2/AQ(−)@T as most of the sulfidewas mainly converted to sulfone (64%) and sulfoxide (32%) within24 h. Neither disulfide nor product arising from carbon–sulfur bondcleavage was detected in any of these experiments. The sampleTiO2/AQ(+)@T was not efficient and no oxidation product wasdetected. These results clearly indicated that singlet oxygen is moreefficiently produced from TiO2/AQ(−)@T than from the other tis-sues. They also evidenced synergetic effects between AQ and TiO2

when the tissue is prepared with the laboratory made sensitizedTiO2 and not with the modified commercial P25 TiO2. They alsoshow that small amounts (8.3% in weight relative to TiO2) of AQ-COOH are sufficient to promote this synergetic effect.

Page 10

370 R. Rahal et al. / Applied Catalysis B: Env

Fs

upt

nawo

3S

isoat

ig. 11. Curve of CO2 evolution upon irradiation at 420 nm of gaseous acetone in aealed reactor with (left) and without (right) the Schott GG420 filter.

These results were confirmed when letting the Petri dishesnder indoor light for one week (Table 5): the TiO2/AQ(−)@T sam-le was the only one able to achieve complete sulfide conversiono sulfoxide/sulfone.

Several successive irradiations under visible light did not sig-ificantly modify the photo-oxidation efficiency of the fabrics,lthough we did not perform long-term stability experiments. Thisas not the case under UV irradiation, where AQ bleaching was

bserved.

.3. Discoloration of wine stain under a solar simulator with achott 420 nm cut-off filter

The self-cleaning properties of the prepared tissues werenvestigated by monitoring wine stains discoloration under solar

imulator or in indoor light. Six drops of red wine were pipettedn the crude or coated cotton fabrics, irradiated either 24 h undersolar simulator or by natural indoor light of the laboratory for

hree weeks. Under solar illumination (Fig. 12), total discoloration

Fig. 12. Wine stain discoloration on the co

ironmental 104 (2011) 361–372

of the wine stain was achieved within 24 h with the sensitizedTiO2-coated cotton fabrics TiO2/AQ(−)@T. An interesting result wasalso obtained under indoor light illumination (without any UVA,Figure S8, SI), as the most significant discoloration of the winestain was achieved within three weeks with TiO2/AQ(−)@T. Thediscoloration was less obvious with TiO2@T and TiO2P25/AQ(−)@T.

4. Discussion

The goal of this study was to prepare cotton fabrics coated withsensitized TiO2. Enhanced properties relative to the photocatalyticTiO2 and to sensitizing anthraquinone were expected [34–38], asanthraquinone-treated cotton fabrics previously showed a smallbut noticeable oxidation efficiency under irradiation at 420 nm[44]. Even if AQ does not present the most suitable properties asa sensitizer, it is commonly used in textile industry as a tank dyingagent [33]. Starting with the published method for the synthesisof hybrid TiO2 nanostructured materials [32], we prepared TiO2sensitized with anthraquinone-2-carboxylic acid (AQ-COOH) inthe form of powders (P-TiO2/AQ(−) and P-TiO2/AQ(+)) and tissues(TiO2/AQ(−)@T and TiO2/AQ(+)@T) with different AQ/TiO2 ratio,and compared their properties with those of materials obtainedby adsorbing AQ-COOH on commercial TiO2 P25 (P-TiO2P25/AQ(−)and TiO2P25/AQ(−)@T). It is worth noting that this one-pot synthe-sis of sensitized TiO2 is carried out in water as the sole solvent anddoes not need any extensive synthetic steps. The preparation ofthese samples relies on the synthesis of the laboratory made pre-cursor [Ti(OR)3(O2C-AQ)] containing AQ-COOH. This precursor wascharacterized by IRTF, CPMAS NMR and XPS as a complex betweenthe carboxylic moiety of AQ-COOH and Ti(OR)4 used as titaniumsource.

All the coated tissues contained TiO2 well dispersed aroundthe fibers, and these coatings appeared to be resistant to wash-

ing in water (no modification of the DRUV spectrum and of thephotocatalytic properties of the fabrics after one run of machinewashing at 45 ◦C with a washing powder containing no fluores-cent whiting dye). The introduced [Ti(OR)3(O2C-AQ)] modifies the

tton fabrics under a solar simulator.

Page 11

R. Rahal et al. / Applied Catalysis B: Env

Fig. 13. Proposed mechanical scheme for the synergy effect observed in singletpc

DrtTtootP

neiUaawwidatrit

cmoTdpMtstF

THtctr(g

roduction by sensitized TiO2/AQ(−)@T (Visible light is absorbed by the AQ@TiO2

omplex).

RUV, FTIR, and XPS characteristics of the tissues TiO2/AQ(−)@Telative to TiO2P25/AQ(−)@T, prepared by simply adding AQ-COOHo TiO2 P25, while keeping the crystallographic anatase phase ofiO2. The results of FTIR and TGA demonstrate the formation of aightly bound complex between TiO2 and the carboxylate moietyf AQ. This bond induces a shift of the electronic absorption spectraf the samples (powders P-TiO2/AQ(−) and tissues TiO2/AQ(−)@T)owards the visible range (3.14 eV), less significant for the samples-TiO2P25/AQ(−) and TiO2P25/AQ(−)@T (3.25 eV).

When irradiated under filtered visible light (totally free of UVA),either of these samples exhibited a significant mineralizationfficiency against acetone (less than 3%). This efficiency slightlyncreased with unfiltered visible light containing a weak part ofVA (0.13 mW cm−2), but remained under 13% after 24 h irradi-tion. It may be noticed that when using fluorescent lamps withn emission maximum at 350 nm, total acetone mineralizationas achieved within 1 h with TiO2/AQ(−)@T. GC–MS experimentsith labeled acetone demonstrated that the evolved carbon diox-

de originated from acetone and not from the tissue or from theecomposition of anthraquinone. These results indicate that UVAre necessary to activate the TiO2 part of the complex, leadingo acetone mineralization. This reaction, most often proposed asesulting from hydroxyl radicals (HO•) formation [55], is almostnefficient under visible light with AQ sensitized TiO2, whateverhe AQ/TiO2 ratio and the mode of incorporation of AQ.

On the other hand, the TiO2/AQ(−)@T tissues were very effi-ient for di-n-butyl sulfide photooxidation under visible light,ainly leading to sulfoxide and sulfone mixtures, whereas all the

ther samples were much less efficient under similar conditions.he same results were obtained under indoor light. These resultsemonstrate efficient singlet oxygen production by TiO2/AQ(−)@Trepared from the laboratory-made precursor [Ti(OR)3(O2C-AQ)].oreover a synergy effect is noticed with sensitized TiO2 relative

o AQ alone. This result is to be compared with recent papers oninglet oxygen formation from doped TiO2 [56–60]. The scheme ofhe possible reaction path for explaining these results is given inig. 13.

Tentatively, absorption of a photon in the visible range by theiO2/AQ complex should lead to an (e−, h+) pair located on theOMO and LUMO of the AQ moiety (Eq. (2)). The energetic level of

he promoted electron being close to the conduction band of TiO2,harge transfer to TiO2 should be possible. Either the TiO2 conduc-

ion band electron or the electron promoted in AQ LUMO is able toeduce ground state 3O2 to the superoxide radical anion O2

•− (Eq.3)). This superoxide radical-anion could in turn be oxidized to sin-let oxygen by the strongly oxidizing AQ* (Eq. (4)), while strongly

ironmental 104 (2011) 361–372 371

electron donor DBS should be easily oxidized to DBS•+ (Eq. (5)).Oxidation of water to hydroxyl HO• radicals appears unfavourableunder these conditions, as no evidence for significant mineraliza-tion was obtained. The reaction between DBS/1O2 or DBS

•+/O2•−

would give rise to high yield of sulfoxide and sulfone (Eq. (6)).

TiO2/AQ−→hv

TiO2/AQ∗[h+, e−] (2)

TiO2/AQ∗[e−] + 3O2 → TiO2/AQ + O2•− (3)

TiO2/AQ∗ + O2•− → TiO2/AQ•− + 1O2 (4)

TiO2/AQ∗[h+] + DBS → TiO2/AQ + DBS•+ (5)

DBS+1O2DBS•+ + O2

•− → DBS(O) + DBS(O)2 (6)

The synergy effect observed for TiO2/AQ(−)@T could result fromthe enhancement of production of superoxide radical-anion, andto its possible easy oxidation to singlet oxygen. It may be recalledthat direct singlet oxygen production by Type II mechanism is alsoknown with AQ [34,38].

As it is also well known that superoxide radical can undergoesdisproportion reactions to produce small quantities of H2O2 andHO• radicals (Eqs. (7)–(9)), the effect of this two ROS can neverthe-less not be neglected to account for our results.

2O2•− + H+ → H2O2 + O2 (7)

H2O2 + e− → HO• + HO− (8)

H2O2 + O2•− → HO• + HO− + O2 (9)

Bleaching of red wine stains could also be due by the samephotosensitized mechanism and/or wine may behave as a pho-tosensitizer injecting electrons to TiO2 conduction band andproducing ROS able to destroy the wine components. It is not pos-sible to affirm that 100% of wine stain destruction is only producedby a photocatalytic mechanism, as some moderate bleaching is alsoobserved with un-sensitized TiO2 coated fabrics.

5. Conclusion

This study describes a simple and reproducible one-pot processfor the elaboration of cotton fabrics coated with TiO2. The modifica-tion of titanium isopropoxide with an anthraquinone derivative asa sensitizer leads to a precursor [Ti(OR)3(O2C-AQ)], characterizedby FTIR, CPMAS and XPS. Bonding between the organic moleculeand titanium atom occurs via a carboxylate complex. Hydrolysis ofmixtures of [Ti(OR)3(O2C-AQ)] and Ti(OR)4 at low temperature inan aqueous medium leads to pale yellow cotton fabrics togetherwith the corresponding free TiO2 powders. The diffuse reflectanceUV spectra confirmed the shift of absorption towards the visiblerange. From FTIR, CPMAS NMR and TGA analysis of the samples (cot-ton pieces and powders), it was shown that AQ-COOH was not onlyadsorbed on Titania but tightly bond through a carboxylate com-plex as in the molecular precursor. XPS data indicated that TiO2 wasgrafted on the surface, but AQ-COH was hardly detectable except forthe photo-chemically inactive P-TiO2/AQ(+) with large amount ofAQ-COOH (63%). Anatase polymorph of TiO2 was always character-ized by XRD even in the absence of a calcination step. Examinationby SEM of cotton tissues before and after washing showed stableand homogeneous coating of TiO2 particles on the cotton fibers.

Under UV light, acetone mineralization was observed, whileunder carefully filtered visible light, no acetone mineralization

occurred. However efficient singlet oxygen addition to di-n-butyl sulfide was evidenced under visible light. Sulfoxide andsulfone were obtained in better yields using sensitized TiO2(TiO2/AQ(−)@T) than using un-modified TiO2@T or AQ@T alone.

Page 12

3 B: Env

OCnsw

aeout

clRcf

A

PgK

A

t

R

[

[

[

[

[[

[

[[

[[

[

[

[

[

[

[[[

[

[[[

[

[

[

[

[[

[[

[[[

[

[[[[[

[[

[

[[[

[

[

72 R. Rahal et al. / Applied Catalysis

ptimum results were obtained with low level of sensitizing AQ-OOH relative to TiO2 (8%) and no reactivity improvement wasoticed with higher AQ-COOH levels. The cotton pieces coated withensitized TiO2 also displayed self-cleaning properties towardsine stain, either under solar illumination or even in indoor light.

The better efficiency of sensitized TiO2-coated cotton isccounted for by a synergy effect between TiO2 and AQ-COOH,nhancing the formation of Reactive Oxygen Species or ROS (singletxygen production and/or superoxide radical-anion). However,nder these conditions, the production of hydroxyl radical seemso be ruled out.

Further studies devoted to the bactericidal effects of this coatedotton pieces are currently being undertaken, together with theirong-term stability under irradiation. The direct characterization ofOS at the gas-solid interface would also help understanding theseomplex oxidation mechanisms and will be developed in a nearuture.

cknowledgments

The authors acknowledge FEDER and the Conseil Général desyrénées Atlantiques for funding this work (research grant), Vir-inie Pellerin for her kind recording of the SEM pictures and Abdelhoukh for the CPMAS NMR spectra.

ppendix A. Supplementary data

Supplementary data associated with this article can be found, inhe online version, at doi:10.1016/j.apcatb.2011.03.005.

eferences

[1] B. Mahltig, H. Haufe, H. Bottcher, J. Mater. Chem. 15 (2005) 4385–4398.[2] W.A. Daoud, S.K. Leung, W.S. Tung, J.H. Xin, K. Cheuk, K. Qi, Chem. Mater. 20

(2008) 1242–1244.[3] K. Qi, W.A. Daoud, J.H. Xin, C.L. Mak, W. Tang, W.P. Cheung, J. Mater. Chem. 16

(2006) 4567–4574.[4] K. Qi, X. Chen, Y. Liu, J.H. Xin, C.L. Mak, W.A. Daoud, J. Mater. Chem. 17 (2007)

3504–3508.[5] W.S. Tung, W.A. Daoud, S.K. Leung, J. Colloid Interface Sci. 339 (2009) 424–433.[6] W.S. Tung, W.A. Daoud, J. Colloid Interface Sci. 326 (2008) 283–288.[7] M.J. Uddin, F. Cesano, D. Scarano, F. Bonino, G. Agostini, G. Spoto, S. Bordiga, A.

Zecchina, J. Photochem. Photobiol. A: Chem. 199 (2008) 64–72.[8] M.J. Uddin, F. Cesano, F. Bonino, S. Bordiga, G. Spoto, D. Scarano, A. Zecchina, J.

Photochem. Photobiol. A: Chem. 189 (2007) 286–294.[9] D. Wu, M. Long, J. Zhou, W. Cai, X. Zhu, C. Chen, Y. Wu, Surf. Coat. Technol. 203

(2009) 3728–3733.10] M.I. Mejia, J.M. Marin, G. Restrepo, C. Pulgarin, E. Mielczarski, J. Mielczarski, Y.

Arroyo, J.-. Lavanchy, J. Kiwi, Appl. Catal. B: Environ. 91 (2009) 481–488.11] T. Yuranova, R. Mosteo, J. Bandara, D. Laub, J. Kiwi, J. Mol. Catal. A: Chem. 244

(2006) 160–167.12] A. Bozzi, T. Yuranova, I. Guasaquillo, D. Laub, J. Kiwi, J. Photochem. Photobiol.

A: Chem. 174 (2005) 156–164.13] A. Bozzi, T. Yuranova, J. Kiwi, J. Photochem. Photobiol. A: Chem. 172 (2005)

27–34.

14] K.T. Meilert, D. Laub, J. Kiwi, J. Mol. Catal. A: Chem. 237 (2005) 101–108.15] N. Onar, M.F. Ebeoglugil, I. Kayatekin, E. Celik, J. Appl. Polym. Sci. 106 (2007)

514–525.16] J. Grzechulska-Damszel, A.W. Morawski, Catal. Lett. 127 (2009)

222–225.

[[

[

ironmental 104 (2011) 361–372

17] P. Park, H. Na, Environ. Technol. 29 (2008) 1001–1007.18] M.I. Mejia, J.M. Marin, G. Restrepo, L.A. Rios, C. Pulgarin, J. Kiwi, Appl. Catal. B:

Environ. 94 (2010) 166–172.19] B. Tryba, J. Hazard. Mater. 151 (2008) 623–627.20] X. Zhang, H. Su, Y. Zhao, T. Tan, J. Photochem. Photobiol. A: Chem. 199 (2008)

123–129.21] J. Mosinger, O. Jirsak, P. Kubat, K. Lang, J. Mosinger, J. Mater. Chem. 17 (2007)

164–166.22] J. Mosinger, K. Lang, P. Kubat, J. Sykora, M. Hof, L. Plistil, J. Mosinger, J. Fluoresc.

19 (2009) 705–713.23] J. Mosinger, K. Lang, L. Plistil, S. Jesenska, J. Hostomsky, Z. Zelinger, P. Kubat,

Langmuir 26 (2010) 10050–10056.24] J. Bozja, J. Sherrill, S. Michielsen, I. Stojiljkovic, J. Polym. Sci. Part A: Polym. Chem.

41 (2003) 2297–2303.25] J. Sherrill, S. Michielsen, I. Stojiljkovic, J. Polym. Sci. Part A: Polym. Chem. 41

(2003) 41–47.26] M. Krouit, R. Granet, P. Krausz, Bioorg. Med. Chem. 16 (2008) 10091–10097.27] M. Krouit, R. Granet, P. Krausz, Eur. Polym. J. 45 (2009) 1250–1259.28] M. Wainwright, M.N. Byrne, M.A. Gattrell, J. Photochem. Photobiol. B: Biol. 84

(2006) 227–230.29] P.S. Golding, T.A. King, L. Maddocks, D.B. Drucker, A.S. Blinkhorn, J. Photochem.

Photobiol. B: Biol. 47 (1998) 202–210.30] S. Brewer, C. Artiles, J. Taylor, M. Dennis, Appl. Surf. Sci. 256 (2010) 1908–1912.31] A. Fujishima, X. Zhang, D.A. Tryk, Surf. Sci. Rep. 63 (2008) 515–582.32] R. Rahal, S. Daniele, L.G. Hubert-Pfalzgraf, V. Guyot-Ferreol, J. Tranchant, Eur. J.

Inorg. Chem. (2008) 980–987.33] G. Simonet, J. Meybeck, Guide Des Techniques De L’ennoblissement Textile,

SPIET, Editions de l’Industrie Textile, Paris, 1982.34] I. Gutierrez, S.G. Bertolotti, M.A. Biasutti, A.T. Soltermann, N.A. Garcia, Can. J.

Chem. 75 (1997) 423–428.35] V. Maurino, D. Borghesi, D. Vione, C. Minero, Photochem. Photobiol. Sci. 7 (2008)

321–327.36] A.E. Alegria, A. Ferrer, G. Santiago, E. Sepulveda, W. Flores, J. Photochem. Pho-

tobiol. A: Chem. 127 (1999) 57–65.37] K. Gollnick, S. Held, J. Photochem. Photobiol. A: Chem. 70 (1993) 135–145.38] V. Latour, T. Pigot, M. Simon, H. Cardy, S. Lacombe, Photochem. Photobiol. Sci.

4 (2005) 221–229.39] L. Shen, H. Ji, H. Zhang, J. Mol. Struct.: Theochem 851 (2008) 220–224.40] T. Oppenländer, Photochemical Purification of Water and Air, Advanced

Oxidation Processes: Principles, Reaction Mechanisms, Reactor Concepts,Wiley-VCH, Weinheim, 2003.

41] J. Eriksen, C. Foote, J. Am. Chem. Soc. 102 (1980) 6083–6088.42] G. Kavarnos, N. Turro, Chem. Rev. 86 (1986) 401–449.43] D.T. Sawyer, in: A.E. Martell, D.T. Sawyer (Eds.), Oxygen Complexes and Oxygen

Activation by Transition Metals, Plenum Press, New-York, 1987, pp. 131–148.44] M. Weyland, S. Lacombe, T. Pigot, Fibres Textiles Photoactives Dépolluantes Et

Désinfectantes, U.S. Patent 2924445 (A1) 2009.45] V. Weidner, J. Hsia, J. Opt. Soc. Am. 71 (1981) 856–861.46] D. Shirley, Phys. Rev. B 5 (1972) 4709.47] J. Scofield, J. Electron Spectrosc. Relat. Phenom. 8 (1976) 129–137.48] R. Rahal, S. Daniele, H. Jobic, Chem. Phys. Lett. 472 (2009) 65–68.49] K. Lang, D. Wagnerova, P. Stopka, W. Damerau, J. Photochem. Photobiol. A:

Chem. 67 (1992) 187–195.50] K.S. Finnie, J.R. Bartlett, J.L. Woolfrey, Langmuir 14 (1998) 2744–2749.51] J.I. Pankove, Optical Processes in Semiconductors, Prentice-Hall, Englewood

Cliffs, N.J., 1971.52] H. Tang, K. Prasad, R. Sanjines, P. Schmid, F. Levy, J. Appl. Phys. 75 (1994)

2042–2047.53] R. Beranek, H. Kisch, Photochem. Photobiol. Sci. 7 (2008) 40–48.54] X. Chen, C. Burda, J. Am. Chem. Soc. 130 (2008) 5018–5019.55] G. Vincent, P.M. Marquaire, O. Zahraa, J. Photochem. Photobiol. A: Chem. 197

(2008) 177–189.56] J.A. Rengifo-Herrera, J. Kiwi, C. Pulgarin, J. Photochem. Photobiol. A: Chem. 205

(2009) 109–115.57] Y. Nosaka, T. Daimon, A.Y. Nosaka, Y. Murakami, Phys. Chem. Chem. Phys. 6

(2004) 2917–2918.58] K. Hirakawa, T. Hirano, Chem. Lett. 35 (2006) 832–833.59] T. Daimon, T. Hirakawa, M. Kitazawa, J. Suetake, Y. Nosaka, Appl. Catal. A: Gen.

340 (2008) 169–175.60] T. Daimon, Y. Nosaka, J. Phys. Chem. C 111 (2007) 4420–4424.