ionic liquid templated tio 2 nanoparticles as a support in gold environmental catalysis

This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/copyright

http://www.elsevier.com/copyright

Author's personal copy

Ionic liquid templated TiO2 nanoparticles as a support in gold environmentalcatalysis

Rafael S. Avellaneda, Svetlana Ivanova *, Oihane Sanz, Francisca Romero-Sarria,Miguel Angel Centeno, Jose Antonio Odriozola

Instituto de Ciencia de Materiales de Sevilla, Centro mixto CSIC-Universidad de Sevilla, Avda. Americo Vespucio 49, 41092 Sevilla, Spain

1. Introduction

Titanium oxide is one of the most used oxides as a support incatalysis and is itself a powerful photocatalyst. The extendedapplication of the latter is due to its good mechanical properties,low price and not at the last place to its association to the so-calledstrong metal–support interaction (SMSI) effect when noble metalis supported [1–3]. The wide use of titanium oxide needs anoptimization of the material for each specific application whichprovokes an extended research in the field of the preparation ofdifferent forms of titania, such as nanoparticles, thin films ormesoporous sieves [4].

In the last few years, the advantages of ionic liquids (ILs) ininorganic nanomaterials synthetic procedures have received moreand more attention due to their unique physical and chemicalproperties [5]. The synthesis of the inorganic materials in ILs is arather new field, emerged in the last 10 years [6] and each methodof preparation is based on one or another specific properties of thelater, i.e. the electrodeposition synthesis on the good ionicconductivity [7,8], the thermal synthesis on the thermally stablenature of the ionic liquids [9,10], the microwave synthesis on thehigh polarizability and capacity to absorb the microwaves [11,12]or the microemulsion synthesis on the immiscibility of two ormore liquids [13,14].

In addition, the ILs can have a template effect for the formationof inorganic nanomaterials due to their property to form anextended hydrogen bond system in the liquid state and to be ratherhighly structured [5]. The usually employed sol–gel reactions toproduce titanium oxide result in a highly amorphous titania, whichhas to be calcined at least at 350 8C to obtain the anatase crystallinestructure. However, the utilization of the IL results directly inanatase or rutile structures without any calcination step [15,16].

Nakashima et al. [17] reported the utilization of the 1-butyl, 3-methyl imidazolium hexafluorophosphate for the synthesis ofhollow TiO2 microspheres in which the IL plays a role not only ofsolvent but also as a stabilizer. After, the preparation of the sizecontrolled anatase structures in the presence of ILs seems tobecome an easier task with or without microwave assisting[11,15,18].

The metal nanoparticles are known to have intermediateproperties between those of bulk and atomic or molecularstructures. They are very interesting from the catalytic point ofview, essentially because of their high surface to volume ratio,which provides more accessible active sites for the chosenreaction. However, nanoparticles are rather kinetically unstableand they should be prevented from aggregation into largerparticles. The high charge plus the steric bulk of ILs can createan electrostatic and/or a steric type stabilization of the nanopar-ticles [6].

The astonishing discovery of Haruta et al. [19] that very smallgold nanoparticles exhibit catalytic activity for the oxidation of COat sub-ambient temperature provoked an enormous scientific

Applied Catalysis B: Environmental 93 (2009) 140–148

A R T I C L E I N F O

Article history:

Received 16 July 2009

Received in revised form 10 September 2009

Accepted 14 September 2009

Available online 18 September 2009

Keywords:

TiO2 nanoparticles

Ionic liquid

Gold catalysts

CO oxidation

A B S T R A C T

This work presents the synthesis of a nanostructured titania support and its subsequent utilization for

the gold particles deposition and application in the reaction of the CO oxidation. A functionalized ionic

liquid has been used as a templating agent for the titanium oxide synthesis resulting in a high specific

surface nanostructured titania anatase. The as prepared support was then used for gold nanoparticles

deposition without ionic liquid removal in order to study the possible role of the latter in the

stabilization of the gold particles. The presence of ionic liquid in the catalysts results in an unusual

catalytic behaviour—strong dependence on the presence of CO and changed kinetics and rate of

oxidation.

� 2009 Elsevier B.V. All rights reserved.

* Corresponding author.

E-mail address: [email protected] (S. Ivanova).

Contents lists available at ScienceDirect

Applied Catalysis B: Environmental

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

0926-3373/$ – see front matter � 2009 Elsevier B.V. All rights reserved.

doi:10.1016/j.apcatb.2009.09.023


interest and makes nowadays gold the metal the most used and themost popular for this catalytic reaction. There are really a vastnumber of publications on this subject, various reviews and a bookwhich try to collect the biggest part of information [20–23] buteverybody agreed on the importance paid to (i) the nature of theactive phase gold particle size and structural geometry (existenceof defect sites), (ii) the nature of support and its interactions withgold, (iii) the pre-reactional treatment, and reaction conditions(especially presence of moisture) and not at the last place iv) thepreparation route for obtaining the gold nanoparticles [24–27].

The use of gold colloids for producing the active supportedcatalyst is not as popular as the deposition precipitation method.The advantage of using the colloidal route for preparing supportedgold catalysts lies in the great number of possibilities to change theconditions of preparation in order to obtain particles having anarrow size distribution about the desired mean, and differentparticle shape as well [15].

In this study the preparation of ionic liquid templated titaniawill be reported. However, contrary to all the other reportedstudies the ionic liquid will not be removed from the resultedmaterial and a chemical bonding between the titania structure andthe IL will be looking for. The choice of the IL was guided by the factthat the imidazolium ILs possess preorganized structures providedby a complicate net of hydrogen bonds that induces a highlystructured fluid through p–p stacking network in the liquid state,which can be described as ‘‘supramolecular’’. This structuralorganization of imidazolium ILs can provoke a spontaneous, well-defined, and extended ordering of nanostructures [6]. In addition,as the physicochemical properties of the ionic liquids stronglydepend on the nature of its anion or cation and the length of itslateral alkyl chains and its functionalization, we are going to searcha functionality which supplies the possibility for bonding the ionicliquid to the titania structure. This possibility was supplied by thesubstitution of the butyl radical in a normal 1-butyl 3-methylimidazolium salt by the carboxy–propyl radical. The as preparedmaterial will be used as a stabilizer of the gold colloids andsubsequently employed in the reaction of oxidation of carbonmonoxide.

2. Experimental

2.1. Ionic liquid synthesis

The 1(3-carboxipropyl) 3-methyl imidazolium chloride(CmimCl) was synthesized as proposed by Fei et al. [28] and itssynthesis steps (nucleophilic attack and hydrolysis) and molecularstructure are presented in Scheme 1.

A 0.2 mol (15.76 ml) of methyl imidazole and 0.2 mol(24.04 ml) of chloro-methyl butirate was introduced after N2

flushing in a three necked, round-bottomed flask equipped witha reflux condenser, a thermometer adapter and N2 inlet adapter.The mixture was then heated to 60 8C during 24 h under inert(N2) atmosphere. The resulted product, a dense yellow liquid(Scheme 1 product SE1) was submitted to purification throughmultiple extractions with 30 ml of diethyl ether. The subsequenthydrolysis step was carried out in a flask equipped with a refluxcondenser in which the product SE1 was mixed with HCl (37%)in 1:1 ratio at 100 8C for 1 h. The resulted product was thenvacuum dried and purified by alternating acetone and diethylether several times until the final production of the SE2compound.

2.2. Synthesis of titanium oxide-Cmim composite (TiIL)

The adequate quantities of titanium tetrabutoxide (Sigma–Aldrich), ethanol and CmimCl were well mixed at room

temperature for 24 h under vigorous stirring. The reactants molarratio was used as follows:

Ti(BuO)4:EtOH:CmimCl = 1:5:0.2

After, the mixture was refluxed at 92 8C for 1 h, and then, thehydrolyzing agent (H2O to Ti(BuO)4 ratio of 8) was added dropwiseand subsequently refluxed at the same temperature for another 3 h.At the end of the synthesis, the sample was dried at 80 8C for 24 h andthe resulting solid presents a light yellow color. To clean the samplefor any possible organic rests, the later was centrifuged (10,000 rpm,10 min) and washed twice with acetone and diethyl ether (the ionicliquid being not soluble in those solvents) and finally dried again inthe same conditions resulting in a white powder. The sample will benamed as TiIL20.

2.3. Gold deposition

The gold colloids were obtained by two different methods: thestandard Turkevich method [29] and reduction of the goldprecursor by sodium borohydrate [30]. In the Turkevich method,2.10�4 M aqueous solution of HAuCl4 (Alfa Aesar 49.92% metalbasis) was heated to boiling and then, the support was addedfollowed by 1 ml 34 mM solution of sodium citrate. The mixturewas vigorously stirred for 5 min and 1 ml of a 34 mM oxalic acidsolution was then added to assure the complete formation of goldcolloids in the following 5 min. The suspension was then filteredand dried at 80 8C overnight. The sample will be named as Au/TiIL20 (citrate).

As the application of this method is possible only in hot aqueoussolutions, and in order to prevent the risk of the ionic liquidremoval during the preparation, a room temperature method ofreduction has been also used. In this second method, a similar goldaqueous solution was mixed with the support and 2 min later the0.01 g of sodium borohydrate was added [30]. The reduction occursimmediately and the solution was kept at room temperature onenight. The solution was then filtered and dried at 80 8C overnight.The sample will be named as Au/TiIL20 (NaBH4).

As the room temperature reduction by sodium borohydratepermits the utilization of other solvents, an additional preparationhas been carried out by modifying the solvent, water beingreplaced by acetone. The objective of this preparation was firstly,

Scheme 1. Synthesis and molecular structure of 1(3-carboxipropyl) 3-methyl

imidazolium chloride (CmimCl).

R.S. Avellaneda et al. / Applied Catalysis B: Environmental 93 (2009) 140–148 141


to decrease the rate of the gold reduction by NaBH4 and from theother hand to preserve the ionic liquid, since it is not soluble inacetone. In this preparation, a 2.10�4 M gold acetone solution wasmixed with the support and 2 min later, the 0.01 g of sodiumborohydrate was added. After full reduction, the solution wasfiltered and dried at 80 8C overnight. The sample will be named asAu/TiIL20 (NaBH4–acetone).

For all the samples the gold loadings were fixed to 1 wt.%.

2.4. Characterization

X-ray diffraction (XRD) analysis was performed on a Siemensdiffractometer D500. Diffraction patterns were recorded with CuKa radiation (40 mA, 40 kV) over a 2u-range of 10–708 and aposition-sensitive detector using a step size of 0.058 and a steptime of 1 s.

The textural properties were studied by N2 adsorptionmeasurements at liquid nitrogen temperature. The experimentswere carried out on Micromeritics ASAP 2010 equipment. Beforeanalysis, the samples were degassed for 2 h at 150 8C in vacuum.

High-resolution scanning electron micrographs (SEM) weretaken using HITACHI S 5200 microscope at 5 kV of the powderedsamples.

Transmission electron microscopy (TEM) observations werecarried out in a Hitachi H 800 microscope operating at 200 kV. Thesamples were dispersed in ethanol by sonication and dropped on acopper grid coated with a carbon film.

The Raman spectra were recorded on a dispersive Horiba JobinYvon LabRam HR800 microscope, with a 20 mW He–Ne green laser(532.1 nm) without filter and with a 600 g mm�1 grating, 15acquisitions (10 s/scan). The microscope used a 50� objective witha confocal pinhole of 100 mm.

2.5. CO oxidation reaction conditions

The activity measurements, as described elsewhere [31], werecarried out in a flow U-shape glass reactor at atmospheric pressure.The catalyst was placed into the glass reactor and a reactive flow(3.4% CO and 21% O2 balanced by helium) was passed thorough it atroom temperature. The total flow was 42 ml min�1 and thequantitative analysis was carried out with a Balzers OmnistarBentchop mass spectrometer. The catalyst was tested in the flowduring 1 h at room temperature. Then, the system was heated to300 8C at 108 min�1 and stabilized at this temperature for 30 min.

3. Results and discussion

3.1. Support (ionic liquid containing titania) TiIL20

The resulted material presents a specific surface area of151 m2 g�1 and total pore volume of 0.2632 cm3 g�1. The N2

adsorption–desorption isotherm of (Fig. 1) clearly indicatesmesoporosity caused by the presence of cylindrical poresgenerated by agglomeration of the spherical particles.

The majority of the pores present a diameter situated between 2and 12 nm with an average pore diameter estimated to be 5 nm,showing a material with narrow mesoporous size distribution. Theas prepared material presents higher surface area than thosereported for the commercially available anatase titania.

However, as the ionic liquid is still present in the material thedecrease of the specific surface area could occur owing to the porefilling. In order to remove the loosely attached ionic liquid, awashing procedure was applied to the TiIL20. The treatmentconsists in washing with water during one night at roomtemperature followed by filtration and drying. The as treatedsample presents almost double of the specific surface and pore

volume compared to the initial sample. So, one can conclude thatthe application of this procedure removes successfully everythingwhich was not chemically bonded to the titania surface. Howevereven after this treatment the sample still presents a lower surfacearea compared to the titania anatase structure after ionic liquidremoval by calcination reported by Yoo et al. [32] and highercompared to that reported by Liu et al. [15] without any thermaltreatment. The diversity of the reported specific surface area forthis type of material is caused essentially by the variation of the IL/Ti(alkoxide)n ratios, the nature of the titanium alkoxide precursors,and ionic liquids [15,18,32].

The synthesized support (TiIL20) shows a pure anatasestructure as presented on XRD patterns in Fig. 2A. However, onecould state that the sample shows a certain amorphous character.The calculated particle size of anatase on the (1 0 1) indexed peakat 2u = 258 by Scherrer equation shows a nanometer size crystal-lites of 6 nm.

The SEM images (Fig. 3A) permitting a general observation ofthe morphology of the solid shows a homogeneous structure ofnanometer size grains and small pores. The TEM images alsoconfirmed these observations (Fig. 3 B).

The nature of the ionic liquid, its presence and/or the changesprovoked by the formed transition metal oxide as well as thestructural and electronic properties of the later were studied byRaman spectroscopy. Fig. 4 shows the comparison between theRaman spectra of the ionic liquid and the support TiIL20.

For clarity, the Raman spectra are divided in two parts. The first(100–800 cm�1) in which the signals of the titania structure should

Fig. 1. N2 adsorption/desorption isotherm and pore size distribution of TiIL20.

R.S. Avellaneda et al. / Applied Catalysis B: Environmental 93 (2009) 140–148142


appear and the second (800–4000 cm�1) in which the vibrations ofthe ionic liquids should be present.

For the ionic liquid, all the signals between 200 and 1200 cm�1

are attributed to the vibration of the carbon chains of the methyland carboxy–propyl radicals and into the imidazolium ring [33,34].The bands appearing at 1260–3000 cm�1 zone are assigned to thesymmetric and antisymetric vibrations of the CH bonds in theimidazolium ring. Those bands can appear as doublets due to theinteraction of the protons with the heteroatoms (N for example) orwith the p-electrons of the ring. Moreover, they can be affected aswell by the presence of water and by the formation of intra or intermolecular hydrogen bonds. All those changes of the closestneighbor entourage affects essentially the p-electrons alternatingthe electronic density and the behaviour of the protons of the ring[35].

The obtained solid (TIIL20) presents strong signals in 100–800 cm�1 zone which are attributed to the titania structure. Thebands at 161, 405, 523 and 648 cm�1 corresponds to the anatasestructure. However, the lowest wavenumber band is just a little bitred shifted compared to that reported in the literature [36].

In the second zone (800–4000 cm�1) the signals can beattributed only to the ionic liquid. The presence of both, solidtitania structure and the ionic liquid for TiIL20 provokes somechanges in the Raman spectra indicating interactions between thelatter. For example the bands at 1027, 1426 and 1569 cm�1

doubles the relative intensity, which can be attributed to thechange in the conformation of the imidazolium ring whichprovokes tensions in the CC single and especially in the CC doublebonds. However in the high wavenumber zone of the CHvibrations, the asymmetric CH2 vibrations are much moreintensive compared to the others present in the ring but anyshifts respecting the bare ionic liquid can be observed.

For better identification of the possible interaction between theionic liquid and the obtained titania structure is far moreinteresting to follow the changes, if any, through the bandattributed to the carboxylic group (�1700 cm�1). This band,observed in the bare ionic liquid spectra disappears when thetitania is produced. However, a new band at 1880 cm�1 appears.When, the hydroxyl group in the carboxylic group is substitutedwith a stronger Lewis acid, the C55O bond becomes more chargedand the separation of the partial charges in the group increasesdepending on the substituting group [37]. As the displacement ofthe C55O band is highly influenced one can imagine (from 1700 to1880 cm�1) in this case the strong interaction between Ti4+ (strongLewis acid) and the carboxylic group.

3.2. Deposition of gold

The specific surface areas (SSA) of the support and its porevolume before and after the gold addition are listed in Table 1.

Fig. 2. XRD patterns of (A) TiIL20, (B) Au/TiIL20 (citrate), (C) TiIL20 (NaBH4), (D) Au/

TiIL20 (NaBH4 acetone).

Fig. 3. SEM images (A, C) and TEMmages (B, D) of TiIL20.



When, the citrate was used for gold reduction, the sample presentsa superior specific surface area compared to the bare TiIL20support indicating the surface liberation of the ionic liquid. Theclose values for the Au/TiIL20 (citrate) to the washed TiIL20 samplelet the supposition that no effect of the ionic liquid will be observedfor this sample, as the remaining ionic liquid in the sample shouldbe minimal. One can conclude that the liberation of the IL in boilinggold solutions is quasi-complete, probably by the change in thestate of ionic liquid which becomes liquid at 80 8C. This was not thecase when gold is deposited from the cold gold solutions by usingNaBH4 as a reducing agent, since the support keeps the same

specific surface and pore volume as the initial TiIL20 demonstrat-ing the maintenance of the ionic liquid no matter the solvent used.For those systems, an influence of the ionic liquid on the metaldeposition and catalytic properties could be imagined.

The diffraction patterns for all the samples are shown in Fig. 2.The gold particle size was calculated by the Scherrer equation [38]for the diffraction line attributed to gold at 44.38 (2u) (JPCDS # 4-0784). The most intensive diffraction line for the Au (1 1 1) ataround 388 was not used because of the coincidence with theanatase structure diffractions.

The Au/TiIL20 (citrate) sample shows the presence of twophases, metallic gold and anatase structure. When the strongerreduction agent was used (NaBH4 instead of sodium citrate) theparticle size of gold decreases and, for the sample prepared inaqueous media goes under the detection limit of the XRD (Fig. 2C).For more precise calculation the XRD spectra was carried out in atight region (40–468).

For the sample prepared in acetone media (Fig. 2D), aside ofanatase and gold phases, the presence of rutile phase can bedetected suggesting a possible influence on the crystallization ofthe support. In addition, this sample shows the presence of sodiumchloride, the product of NaBH4 reduction in chloride media,apparently not well evacuated in acetone, due probably to thechange in the solubility when ionic liquid is present.

For all the samples, no changes of the particle size of anatasewere observed.

The morphology of the support for all gold containing samplesdoes not change but the gold particle size seems to be verydifferent which confirms the results obtained by XRD. The goldreduced by citrate resulted in big gold crystallites (Fig. 5 A) whilethe NaBH4-reduced sample shows the same particle size (Fig. 5 Band C). In addition, a difference between the samples prepared indifferent media can be observed. For the sample in acetone, itseems that the particles are simply deposited on the surface.However, for the gold deposited in aqueous media, the goldparticles seem to be covered by a material of the morphology equalto the support (Fig. 5 left hand images).

Even if the SEM images can give a good idea about the goldparticle size the TEM study is needed for calculation of the golddistribution which is presented in Table 1. The particle sizedistribution estimated by TEM agrees with XRD results (Scherrercalculations). The citrate mediated sample presents the bigger sizeof the gold crystallites probably due to the relatively severeconditions of the preparation (boiling gold solution and mixtureof reducing agents). However, for the samples prepared in theaqueous media, the gold particles appear like covered by a organiclayer, much more present for the citrate mediated reduction (Fig. 6).

The use of different solvents at room temperature and the samereducing agent (NaBH4) does not influence the gold particles size.In addition, the presence of the ionic liquid in these samples,evidenced by the BET surface area measurements, suggests apossible gold particle size stabilization by the presence of ionicliquid.

From the first sight, the Au/TiIL20 (NaBH4 acetone) sample(Fig. 6C) seems that the gold particles have different morphologies,the smallest particle looks spherical and the larger ones ratherhemispheric. A detailed observation shows that all the goldparticles are covered by some organic matter which gives thehemispherical form (Fig. 6C inset). This also was observed for Au/TiIL20 (NaBH4) sample. However, in this sample only the biggerparticles seem to be covered by the organic material. For the lasttwo samples, no matter, the quasi-equal particle size, differentmorphologies and interactions with the ionic liquid were observed,which suggests a possible difference in the CO oxidation behaviour.

The samples are studied by Raman spectroscopy and thespectra will be shown separated in two parts, one from 100 to

Fig. 4. Raman spectra in 100–800 cm�1 (left hand image) and 800–4000 cm�1 (right

hand image) ranges.

Table 1Samples characteristics.

Samples SSA,

m2 g�1

Pore

volume,

cm3 g�1

Gold particle

size, nm (XRD)

Gold particle

size, nm (TEM)

TiIL20 151 0.2632 –

TiIL20 washed 254 0.3265 – –

Au/TiIL20 (citrate) 227 0.2036 40 41

Au/TiIL20 (NaBH4) 162 0.1351a 14

Au/TiIL20

(NaBH4 acetone)

135 0.2465 26 22

a No diffraction lines are observed.



800 cm�1 where the vibrations characteristic for the titaniumoxide are present, and the other from 800 to 4000 cm�1 range, inwhich the vibrations of ionic liquid could be observed.

The corresponding Raman spectra of the gold containingsample compared to the bare TiIL20 support are presented inFigs. 7 and 8.

For all the samples, the bands at around 160, 410, 520 and645 cm�1 indicate the presence of anatase structure [36]. Au/TiIL20 (citrate) sample shows only anatase vibrations. For the othertwo samples, a change of the relative intensity of the bands, andespecially for the Au/TiIL20 (NaBH4 acetone) one, the appearing ofnew signals at 210, 430, 620 cm�1 attributed to rutile titania phase[39] are observed, confirming the observation by XRD of theanatase to rutile phase transition in the presence of strongreducing agent (NaBH4). This is the place, however to mention that,in the condition of the measure (especially without filter) the lasercan provoke the phase transition from anatase to rutile.

The zone of the major ionic liquids vibrations (800–4000 cm�1)of the gold containing samples compared to the TiIL20 support arepresented in Fig. 8. For the Au/TiIL20 (citrate) sample (Fig. 8B) agroup of strong vibrations at 1534, 1874, 2434, 2923 cm�1 wasobserved.

The apparition of these signals can be attributed to the reactionsand modifications which can occur to the sodium citrate in acidicmedia [40], resulting in sub-products of oxidation as presented inScheme 2.

As the Raman spectroscopy is extremely sensitive to the presenceof chromophore groups, for example the double conjugate bonds,the laser could influence the Raman response by increasing it tovarious orders of magnitude [41] affecting not only the double bondsignals but also the vibration of all neighboring groups. Moreover,the oxalic acid employed after the citrate in the preparation possesas well, double conjugated bonds which can contribute to theassembly of the intense signals in this range. The presence of ionicliquid signals in this case was hard to be evidenced. In this case, theorganic molecules covering the gold particles, evidenced by TEM forthis sample, can be attributed to these organic sub-products of theoxidation of citric acid.

For the Au/TiIL20 (NaBH4) and Au/TiIL20 (NaBH4 acetone)samples, a higher fluorescence was observed caused by theexistence of two phases of titania and by the presence of smallergold particles. The signals of the ionic liquid were, therefore, eithervery weak for Au/TiIL20 (NaBH4), either disturbed, like for Au/TiIL20 (NaBH4 acetone).

Fig. 5. SEM images of (A) Au/TiIL20 (citrate), (B) Au/TiIL20 (NaBH4), (C) Au/TiIL20 (NaBH4 acetone).



3.3. Catalytic activity

The application of the samples in the oxidation of CO, whichremains one of the most important reactions in gold catalysis, letthe possibility to study the influence of different parameters on thecatalytic activity, in our case more the influence of the ionic liquidas a integral element of the catalyst on the accessibility of theactive sites, the electronic distribution etc. All the samples havebeen tested in the reaction of CO oxidation with a number of

repeating cycles in which, the first started after drying at 80 8Cwithout any preliminary treatment. The total CO conversions forall the samples in the first cycle are presented in Fig. 9.

The Au/TiIL20 (citrate) shows a low activity close to 13%conversion of CO at the maximal temperature of the reaction: Inaddition a formation of CO (negative conversion) was detected bythe mass spectrometer caused by the complete or semi-completeoxidation of the organic molecules on the surface, as seen byRaman. The Au/TiIL20 (NaBH4) sample presents the best activity

Fig. 6. TEM images of (A) Au/TiIL20 (citrate), (B) Au/TiIL20 (NaBH4), (C) Au/TiIL20 (NaBH4 acetone).

Fig. 7. Raman spectra in 100–800 cm�1 range for (A) TiIL20, (B) Au/TiIL20 (citrate),

(C) TiIL20 (NaBH4), (D) Au/TiIL20 (NaBH4 acetone).Fig. 8. Raman spectra in the range 1000–3500 cm�1 for (A) TiIL20, (B) Au/TiIL20

(citrate), (C) Au/TiIL20 (NaBH4), (D) Au/TiIL20 (NaBH4 acetone).



from all the samples with a complete conversion at 190 8C. The Au/TiIL20 (NaBH4 acetone) was not active in the conditions of thereaction. As observed, the gold particle size was not very differentfrom the previous sample and could not be a reason for theinactivity of this sample no matter the temperature. Other study onionic liquid alumina supported gold particles shows that thereduced gold particles are defect-free and result in a perfect crystal[42]. The absence of defects, i.e. no active sites for oxygenactivation and CO oxidation, could explain the absence of activityfor this sample [43]. In addition, the observation by SEM leading tothe conclusion that the gold particles are simply deposited on thesupport could explain as well the inactivity of the sample as thereaction should occur on the gold titania interface which is missingin this case [44–46].

As the later and the citrate reduced sample did not show anyexclusive activity no other cycles are made. The following cyclesfor the Au/TiIL20 (NaBH4) sample are presented in Fig. 10.

For this sample, four oxidation cycles were made, in whichinstead of the improvement of the activity at each cycle aninfluence of the presence of CO in the flow during the cooling stepwas evidenced.

After the first oxidation cycle (activity at room temperature andunder heating) for which the complete conversion occurs at 190 8C,the catalyst was cooled in the flow of the reaction (CO, O2, He) afterwhich the second cycle of heating was started. For this cycle the

complete conversion occurs at 170 8C but the form of the curve isslightly different—fast oxidation till 95% of conversion at 130 8Cand slow regain of complete oxidation in the last 30 8C to thecomplete conversion. This cycle was followed by the cooling in theCO-free atmosphere (O2, He). The activity of the catalyst in thethird cycle then decreased but was still better then the first cycleactivity and repeats the form of the light-off curve of the first cycle.The last cycle was carried out after the cooling in the gas mixture ofthe reaction (like after cycle 1). The catalyst shows the sameactivity like in the second cycle.

The differences between the first and third cycle suggestactivation of the catalyst with the repeating of the cycles due onlyto the surface liberation through an oxidizing treatment [47] or bythe completing of the reduction of the gold to metallic gold [48–50]. However, the activation in the presence of the CO in thecooling gas mixture is clearly evidenced. In addition the samplepresents certain room temperature activity for which the oxidationtreatment was not responsible. The exact replica of the cycles 2 and4 are, as well, very astonishing. In this case the supposition that theactivation of the sample was caused by the reduction to metallicgold by the mixture of the reaction cannot be advanced as anexplication due not only to the different activity of the cycles but aswell to the different kinetics of the oxidation. These results suggestthe presence of CO close to the active sites through the wholeprocess. The gold–titania interface is changed by the presence of

Scheme 2. Different acids resulting from the oxidation of citric acid.

Fig. 9. CO conversion at first cycle for (A) Au/TiIL20 (citrate), (B) Au/TiIL20 (NaBH4),

(C) Au/TiIL20 (NaBH4 acetone). Fig. 10. CO oxidation cycles for Au/TiIL20 (NaBH4).



the ionic liquid, for which the existence of conjugated p-electronsshould be taken in consideration. In addition, the dissolution of COin the IL and the change of the environment could explain thepresence of the gas through the whole process close to the gold–support interfaces, which is the main reason for the differentkinetics and rate of oxidation [51,52].

Even if the importance of the presence of the CO close to theactive sites was clearly evidenced, another role of the ionic liquidshould be taken into consideration. Schrekker et al. [53], proposeda model of gold nanoparticles stabilization and coordination by theimidazolium cation in which the stabilization takes place throughthe imidazolium/Au surface interaction, as well as by the anionelectrostatic/Derjaugin–Landau–Verwey–Overbeek type stabiliza-tion by Cl� anions coming from the gold precursor.

The obtained results confirm that the role of the ionic liquid ismuch more complex starting from the gold particles stabilizationand going to the gold surroundings enrichment.

4. Conclusions

The utilization of the ionic liquid as a template for the preparationof the support results in nanostructured anatase titania with highspecific surface area. The high homogeneity of the sample isattributed to the capacity of the ionic liquid through its H-bonds netto direct the structure of the oxide. In addition, the presence of theionic liquid seems to stabilize the gold particle size in cold goldsolutions no matter the solvent used. However, the Turkevichmethod of preparation of the gold colloids resulted in large goldcrystallites for which no effect of the ionic liquid could be observed.The catalytic activity of the samples was guided by the gold particlesize in the case of citrate-reduced samples and by the gold particlesize environment for the NaBH4-reduced samples. The use ofacetone as a solvent do not results in active catalysts due to theabsence of gold titania interface and the production of the goldcrystallites free of defects. For the sample produced in aqueoussolutions the gold particles stabilization and the presence of theionic liquids at the particle/support interface results in an unusualcatalytic behaviour—strong dependence on the presence of CO in themixture and changed kinetics and rate of the reaction of oxidation.However, the role of the ionic liquid in this sample appears to be verycomplex, from the gold particles electronic configuration stabiliza-tion to the in situ promotion of the active sites.

Acknowledgements

F. Romero-Sarria and R.S. Avellaneda acknowledge MEC for itscontract Ramon y Cajal and FPU scholarship. S. Ivanova acknowl-edges CSIC for her JAE contract and all the authors acknowledgeJunta de Andalucia—‘‘TEP 106’’.

References

[1] K.I. Hadjiivanov, D.G. Klissurski, Chem. Soc. Rev. 25 (1996) 61–69.[2] G. Busca, H. Saussey, O. Saur, J.C. Lavalley, V. Lorenzelli, Appl. Catal. 14 (1985) 245–

260.[3] M.A. Centeno, M.C. Hidalgo, M.I. Dominguez, J.A. Navio, J.A. Odriozola, Catal. Lett.

123 (2008) 198–206.[4] G. Martra, Appl. Catal. A: Gen. 200 (2000) 275–285.[5] M. Antonietti, D. Kuang, B. Smarsly, Y. Zhou, Angew. Chem. Int. Ed. 43 (2004)

4988–4992.[6] Zh. Li, Zh. Jia, Y. Luan, T. Mu, Curr. Opin. Solid State Mater. Sci. 12 (2009) 1–8.[7] J.F. Huang, I.W. Sun, Adv. Funct. Mater. 15 (2005) 989–994.

[8] J.F. Huang, I.W. Sun, Chem. Mater. 16 (2004) 1829–1831.[9] E.R. Cooper, C.D. Andrews, P.S. Wheatley, P.B. Webb, P. Wormald, R.E. Morris,

Nature 430 (2004) 1012–1016.[10] X. Zhou, Z. Xie, Z. Jiang, Q. Kuang, S. Zhang, T. Xu, R. Huang, L. Zheng, Chem.

Commun. (2005) 5572–5573.[11] K. Ding, Z. Miao, Z. Liu, Z. Zhang, B. Han, G. An, S. Miao, Y. Xie, J. Am. Chem. Soc. 129

(2007) 6362–6363.[12] Z. Li, Z. Liu, J. Zhang, B. Han, J. Du, Y. Gao, T. Jiang, J. Phys. Chem. B 109 (2005)

14445–14448.[13] H. Gao, J. Li, B. Han, W. Chen, J. Zhang, R. Zhang, D. Yan, Phys. Chem. Chem. Phys. 6

(2004) 2914–2916.[14] A. Taubert, Acta Chim. Slov. 52 (2005) 168–170.[15] Y. Liu, J. Li, M. Wang, Z. Li, H. Liu, P. He, X. Yang, J. Li, Cryst. Growth Des. 5 (2005)

1643–1649.[16] H. Kaper, F. Endres, I. Djerdj, M. Antonietti, B.M. Smarsly, J. Maier, Y. Hu, Small 3

(2007) 1753–1763.[17] T. Nakashima, N. Kimizuka, J. Am. Chem. Soc. 125 (2003) 6386–6387.[18] Y. Zhou, M. Antonietti, J. Am. Chem. Soc. 125 (2003) 14960–14961.[19] M. Haruta, T. Kobayashi, H. Sano, N. Yamada, Chem. Lett. 2 (1987) 405–408.[20] G.C. Bond, D.T. Thompson, Gold Bull. 33 (2000) 41–51.[21] M. Haruta, Cattech 6 (2002) 102–115.[22] M.S. Chen, D.W. Goodman, Catal. Today 111 (2006) 22–33.[23] G.C. Bond, C. Louis, D.T. Thompson, in: G.H. Hutchings (Ed.), Catalysis by Gold,

Imperial College Press, London, 2006.[24] F. Romero-Sarria, L.M. Martınez, T.M.A. Centeno, J.A. Odriozola, J. Phys. Chem. C

111 (2007) 14469–14475.[25] M.I. Dominguez Leal, F. Romero-Sarria, M.A. Centeno, J.A. Odriozola, Appl. Catal. B:

Env. 87 (2009) 245–251.[26] W.Y. Hernandez, M.A. Centeno, F. Romero-Sarria, J.A. Odriozola, J. Phys. Chem. C

113 (2009) 5629–5635.[27] S. Ivanova, V. Pitchon, C. Petit, J. Mol. Catal. A: Chem. 256 (2006) 278–283.[28] Z. Fei, D. Zhao, T.J. Geldbach, R. Scopelliti, P.J. Dyson, Chem. Eur. J. 10 (2004) 4886–

4893.[29] J. Turkevich, Gold Bull. 18 (1985) 86–91.[30] F. Douglas, R. Yanez, J. Ros, S. Marın, A. de la Escosura-Muniz, S. Alegret, A.

Merkoci, J. Nanopart. Res. 10 (2008) 97–106.[31] F. Romero-Sarria, A. Penkova, L.M. Martinez, T.M.A. Centeno, K. Hadjiivanov, J.A.

Odriozola, Appl. Catal. B: Env. 84 (2008) 119–124.[32] K. Yoo, H. Choi, D.D. Dionysiou, Chem. Commun. (2004) 2000–2001.[33] R. Holomb, A. Martinelli, I. Albinsson, J.C. Lassegues, P. Johansson, P. Jacobsson, J.

Raman Spectrosc. 39 (2008) 793–805.[34] N.E. Heimer, R.E. Del Sesto, Z. Meng, J.S. Wilkes, W.R. Carper, J. Mol. Liq. 124 (2006)

84–95.[35] J. Dupont, P.A.Z. Suarez, R.F. DeSouza, R.A. Burrow, J.-P. Kintzinger, Chem. Eur. J. 6

(2000) 2377–2381.[36] V.V. Yakovlev, G. Scarel, S.R. Aita, S. Mochizuki, Appl. Phys. Lett. 76 (2000) 1107–

1109.[37] H. Baranska, A. Labudzinska, J. Terpinski (Eds.), Laser Raman Spectrometry:

Analytical Applications, Ellis Horwood, Chichester, UK, 1987.[38] H.P. Klug, L.E. Alexander, X-Ray Diffraction Procedure for Polycristalline and

Amorphous Materials, 2nd ed., Wiley, New York, 1974.[39] A. Ogden, J.A. Corno, J. Hong, A. Fedorov, J.L. Gole, J. Phys. Chem. Solids 69 (2008)

2898–2906.[40] K. Van Werde, D. Mondelaers, G. Vanhoyland, D. Nelis, M.K. Van Bael, J. Mullens,

L.C. Van Pouke, J. Mater. Sci. 37 (2002) 81–88.[41] J.G. Grasselli, B.J. Bulkin, Anal. Raman Spectrosc. Chem. Anal. S 114 (1991) 295.[42] S. Ivanova, R.S. Avellaneda, O. Sanz, F. Romero-Sarria, M.A. Centeno, J.A. Odriozola,

Proc. Europacat. IX, Salamanca, Spain, 2009.[43] G. Carriazo, L.M.T. Martinez, J.A. Odriozola, S. Moreno, R. Molina, M.A. Centeno,

Appl. Catal. B: Environ. 72 (2007) 157–165.[44] M. Haruta, M. Date, Appl. Catal. A: Gen. 222 (2001) 427–437.[45] F. Boccuzzi, A. Chiorino, S. Tsubota, M. Haruta, Sensors Actuat. B 24–25 (1995)

540–543.[46] F. Boccuzzi, A. Chiorino, S. Tsubota, M. Haruta, Catal. Lett. 29 (1994) 225–234.[47] S. Lee, A. Gavriilidis, Q.A. Pankhurst, A. Kyek, F.E. Wagner, P.C.L. Wong, K. Yeung, J.

Catal. 200 (2001) 298–308.[48] V. Schwartz, D.R. Mullins, W. Yan, B. Chen, S. Dai, S.H. Overbury, J. Phys. Chem. B

108 (2004) 15782–15790.[49] W. Yan, B. Chen, S.M. Mahurin, V. Schwartz, D.R. Mullins, A.R. Lupini, S.J.

Pennycook, S. Dai, S.H. Overbury, J. Phys. Chem. B 109 (2005) 10676–10685.[50] J.H. Yang, J.D. Henao, M.C. Raphulu, Y. Wang, T. Caputo, A.J. Groszek, M.C. Kung, M.

Scurrell, J.T. Miller, H.H. Kung, J. Phys. Chem. B 109 (2005) 10319–10326.[51] J. Kumelan, A. Perez-Salado Kamps, D. Tuma, G. Maurer, Fluid Phase Equilb. 228–

229 (2005) 207–211.[52] J. Kumelan, A. Perez-Salado Kamps, D. Tuma, G. Maurer, Fluid Phase Equilib. 260

(2007) 3–8.[53] H.S. Schrekker, M.A. Gelesky, M.P. Stracke, C.M.L. Schrekker, G. Machado, S.R.

Teixeira, J.C. Rubim, J. Dupont, J. Colloid Interf. Sci. 316 (2007) 189–195.


ionic liquid templated tio 2 nanoparticles as a support in gold environmental catalysis

Documents