Ordered Mesoporous Thin Films of Rutile TiO2 Nanocrystals Mixed with Amorphous Ta2O5

DOI: 10.1002/cphc.200700679

Ordered Mesoporous Thin Films of Rutile TiO2Nanocrystals Mixed with Amorphous Ta2O5

Jin-Ming Wu,[b] Markus Antonietti,[b] Silvia Gross,[c] Matthias Bauer,[d] and Bernd M. Smarsly*[a]

1. Introduction

Non-silica oxides with ordered mesoporous structures exhibitsuperior properties over their nonporous counterparts in appli-cations involving surface reactions requiring high specific sur-face area as well as highly accessible pores.[1] Among the vari-ous oxides, titanium dioxide possesses the combined advan-tages of low costs, chemically stability, and interesting physico-chemical properties. It therefore finds applications in cataly-sis,[2] gas sensing,[3] electrochromics,[4] solar cells,[1d,5] Li-ionbatteries,[6] membrane separations,[7] catalyst supports,[8] andbiomaterials.[9] Titania has three crystalline polymorphs—ana-tase, rutile, and brookite—each phase featuring different prop-erties, such as refractive index and chemical and photochemi-cal properties. Anatase is the most studied form, because it isgenerally assumed that it has improved properties in most ap-plications and also just because sol–gel synthesis usually pro-duces anatase. Ordered mesoporous titania films with anatasenanocrystalline walls have been fabricated through varioustechniques,[1,10] and their properties, especially photocatalyticactivity (PA),[11] have been widely studied. Rutile is chemicallyinert and possesses a higher reflective index, opacity, and ex-ceptional light-scattering efficiency. In certain cases, rutile hasbeen found to exhibit advantageous or comparable propertiesto anatase.[12] For example, rutile scatters light more efficientlyand is chemically more stable over anatase, which favors lightharvesting and hence high photon-to-current conversion effi-ciency when used as photoelectrode for a dye-sensitized solarcell.[12b] Sun et al. reported that rutile powder with a crystallitesize of 7 nm possessed higher PA to photodegrade phenolthan anatase titania with the same specific surface area.[12c]

Usually, in the fabrication of mesostructured titania films bysol–gel templating, anatase, the kinetically more stable phase,is formed independent of the Ti precursor and template. Phasetransformation of anatase to rutile occurs only upon heating

to high temperature, which usually destroys the ordered meso-porous structure. One distinct work by Li et al. reported thefabrication of mesoporous rutile powders by hydrolyzingTiOCl2 aqueous solution at low temperatures in the presenceof octyl poly(ethylene oxide) surfactant, which serves as tem-plate-directing agent.[13] This type of low-temperature-derivedmesoporous rutile powder possessed high PA for the gas-phase photooxidation of a mixture of benzene and metha-nol.[13] Evidently, such an approach seems to be unsuitable forthe fabrication of thin films of mesoporous rutile, which wouldbe an interesting complement to the corresponding anatasefilms. One of the difficulties in preparing mesostructured rutileusing sol–gel templating is the fact that for nanocrystals withsizes below about 9 nm,[14] contrary to the bulk, anatase isthermodynamically more stable than rutile. Also, during nucle-

Ordered mesoporous thin films of composites of rutile TiO2 nano-crystals with amorphous Ta2O5 are fabricated by evaporation-in-duced self-assembly followed by subsequent heat treatmentbeyond 780 8C. Incorporation of selected amounts of Ta2O5

(20 mol %) in the mesoporous TiO2 film, together with the uniquemesoporous structure itself, increased the onset of crystallizationtemperature which is high enough to ensure the crystallization ofamorphous titania to rutile. The ordered mesoporous structurebenefits from a block-copolymer template, which stabilizes themesostructure of the amorphous mixed oxides before crystalliza-

tion. The surface and in-depth composition analysis by X-rayphotoelectron spectroscopy suggests a homogeneous intermixingof the two oxides in the thin film. A detailed X-ray absorptionfine structure measurement on the composite film containing20 mol % Ta2O5 and heated to 800 8C confirms the amorphousnature of the Ta2O5 phase. Photocatalytic activity evaluation sug-gests that the rutile nanocrystals in the synthesized ordered mes-oporous thin film possess good ability to assist the photodegra-dation of rhodamine B in water under illumination by UV light.

[a] Prof. Dr. B. M. SmarslyInstitute of Physical Chemistry, University of GiessenHeinrich-Buff-Ring 58, 35390 Giessen (Germany)Fax: (+ 49) 641-9934509E-mail : [email protected]

[b] Dr. J.-M. Wu,+ Prof. Dr. M. AntoniettiMax Planck Institute of Colloids and Interfaces, Research Campus GolmAm MAhlenberg 1, 14424 Potsdam-Golm (Germany)

[c] Dr. S. GrossCNR-ISTM, Dipartimento di Scienze ChimicheUniversitC di Padova and INSTM UdR Padovavia Marzolo 1, 35131 Padova (Italy)

[d] M. BauerInstitute for Physical Chemistry, University of StuttgartPfaffenwladring 55, 70569 Stuttgart (Germany)

[+] Current address:Department of Materials Science and EngineeringZhejiang University, Hangzhou 310027 (China)

Supporting information for this article is available on the WWW underhttp://www.chemphyschem.org or from the author.

748 www.chemphyschem.org A 2008 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim ChemPhysChem 2008, 9, 748 – 757

ation and crystal growth the interaction of surfactants with thecrystallographic planes seems to favor the crystallization ofanatase. Owing to these difficulties encountered in preparingphase-pure rutile films with ordered mesopore structures, as atradeoff we developed a strategy to fabricate ordered mesopo-rous thin films of rutile nanocrystals in combination with amor-phous Ta2O5, through the so-called evaporation-induced self-assembly (EISA) technique.[1] Ta2O5 was chosen for several rea-sons. First, a stable, homogeneous solution could be preparedcontaining Ti and Ta species, which is the precondition to per-form the EISA process. Second, and more importantly, Ta2O5

does not form mixed crystals with TiO2 and possesses quite ahigh crystallization temperature, thus ensuring that TiO2 crys-tallizes first.

In addition, the photocatalytic performance of the rutilephase in the ordered mesoporous film was scrutinized. Al-though frequently stated, only few studies have indeed proventhat thin films with an ordered mesopore structure show im-proved performance. Aside from the preparation, the presentstudy thus addresses the fundamental question of how certainstructural parameters in thin films influence a macroscopicphysicochemical property. Here, the photocatalytic decay of adye was used as a model system to exemplify the influence ofdifferent parameters on catalytic activity. Mesoporous rutilefilms are compared to their counterparts composed of anataseto assess the fundamental question of whether the two phasesdo indeed differ in their catalytic performance. The relevanceof mesoporosity is studied by comparison with films preparedwithout templates. Based on such systematic comparison ap-plying a well-defined model reaction, herein we discuss thecatalytic properties of rutile films in the light of the special syn-thetic conditions and properties of thin films prepared by EISA.

Owing to the challenges in the synthesis, a suitable EISAprocess and crystallization pathway was developed. To mini-mize the kinetic effect leading to the formation of anatase, theonset of crystallization temperature (Tc) of titania is increasedto 780 8C through incorporation of tantalum pentoxide in themesoporous film. It is demonstrated that at such a high tem-perature amorphous titania crystallizes to rutile. The orderedmesostructure is maintained with the help of a novel KLEblock copolymer, H(CH2CH2CH2CHACHTUNGTRENNUNG(CH3)CH3)66ACHTUNGTRENNUNG(OCH2CH2)86H,which was developed previously by our group.[1b] This tem-plate combines good chemical accessibility and high robust-ness of the mesophase with a superb application profile. Themicelle/pore size is tuned to be large enough to be accommo-dated by rutile nanocrystals without destruction of the meso-structure at the high Tc. Hence, the present study is also dedi-cated to a more profound understanding of the crystallizationprocess within the confinement of a mesostructure.

2. Results and Discussion

2.1 Crystallization

A set of composite (TiO2)1�2xACHTUNGTRENNUNG(Ta2O5)x oxide thin films withvalues of x =0, 0.05, 0.1, 0.15, 0.2, and 0.25 (molar fraction) wasprepared through the EISA approach. Phase evolution is moni-

tored as a function of the calcination temperature by wide-angle X-ray scattering (WAXS). The temperature Tc, definedhere by the calcination temperature that gave significantWAXS reflections corresponding to crystalline TiO2, increaseswith increasing amounts of incorporated Ta2O5 (Figure 1a). This

finding is attributed to the perturbation of the crystallizationand is in accordance with previous studies.[15] More interesting-ly, it is found that an amorphous composite film with x =0.2crystallizes to yield only the rutile phase after calcination atTc=780 8C. For simplicity, the film with such a composition istermed a “rutile film” hereafter. The composite films with x =

0–0.15 exhibits peaks corresponding to anatase and that withx =0.25 crystallizes to show both orthorhombic Ta2O5 (JCPDScard 71-0639) and rutile (JCPDS card 21-1276), as indicated inFigure 1b. The mesopore structure of the films with x =0–0.15(anatase) is stable up to 850 8C, while beyond this temperaturemesostructural collapse is observed. Thus, ordered mesopo-rous films with rutile nanocrystalline walls are achieved onlythrough increasing the Tc, which is in agreement with recentstudies on oxides crystallizing from the amorphous phase athigh temperatures.[16]

Figure 1. a) Onset of crystallization temperature Tc of the (TiO2)1�2x ACHTUNGTRENNUNG(Ta2O5)x

composite film as a function of the molar fraction of Ta2O5 and b) the corre-sponding WAXS patterns after calcination at the Tc.

ChemPhysChem 2008, 9, 748 – 757 A 2008 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim www.chemphyschem.org 749

Mesoporous Rutile Thin Films

Hence, at the relatively higher calcination temperature,amorphous titania crystallizes to the rutile phase, rather thanthe generally achieved anatase. As Tc=780 8C (rutile film) is farbeyond the decomposition temperature of 450 8C for the blockcopolymer (KLE) template,[1b] these findings confirm our recentstudies addressing the role of this block copolymer.[17] In es-sence, one of the main advantages of such polymers is the sta-bilization of the ordered mesostructure during the formationof a dehydrated amorphous oxide. The crystallization then pro-ceeds as a solid–solid transition under preservation of the mes-ostructural order. Thus, the present study supports this proper-ty of the KLE template found for other mesoporous oxide filmswith high Tc, such as g-alumina[16] and HfO2.

[17]

Figure 2 shows WAXS patterns collected from the rutile filmafter calcination at 800 and 850 8C. Only reflections corre-

sponding to the rutile phase can be discerned. The grain sizeof rutile for both films is calculated to be approximately 9 nmby applying the Scherrer equation to the (101) peak. The Ta2O5

phase in the rutile film, which is calculated to be about 50% involume (the density of rutile is 4.2 gcm�3 and that of Ta2O5 isca. 8 gcm�3), remained amorphous at 800 and 850 8C, in spiteof the fact that pure Ta2O5 thin films fabricated through thesame procedure begin to crystallize at 760 8C. Therefore, thetwo oxides in the rutile film retarded the crystallization of eachother and the calcination temperature of 800 8C was highenough to initiate the crystallization of TiO2, but not enoughfor Ta2O5.

The retarded phase transformation of oxides by the otherone incorporated is generally reported. The transition tempera-ture from the triclinic to monoclinic phase in (Ta2O5)1�xACHTUNGTRENNUNG(TiO2)x

ceramics increases with increasing values of x of 0.05, 0.08, and0.11.[15a] Incorporation of oxides such as Nb2O5 in titania is alsoreported to retard the phase transition.[10e] In the current inves-tigation, we assume that the intermixing of TaO6/TiO6 octahe-dra[15a] disturbed the diffusion of Ta/Ti ions, increased the acti-vation energy of the diffusion, and in turn increased the crys-tallization temperature of the two oxides.

Interestingly, different crystal structures are obtained for anonporous thin film, which is fabricated in a similar procedurebut without the addition of the KLE template. Figure 3 showsthe WAXS curve of a non-templated (nonporous) (TiO2)0.6-ACHTUNGTRENNUNG(Ta2O5)0.2 films after calcination at various temperatures. The

oxidic matrix of this film crystallizes at a much lower tempera-ture of about 600 8C, compared to 780 8C for the mesoporousone, which is not surprising because a mesoporous structureretards crystallization of oxides. After calcination at 600 and700 8C, the amorphous TiO2 in the composite film crystallizesto anatase and the Ta2O5 fraction remains amorphous. TheWAXS peaks corresponding to both anatase TiO2 (JCPDScard 21-1272) and orthorhombic Ta2O5 are detected for thenonporous thin film after crystallization at temperaturesbeyond 800 8C. The fact that the anatase phase is obtained forthe nonporous thin film, rather than the rutile one as for themesoporous counterparts, is attributed to the significantlylower Tc of amorphous titania in the nonporous film, whichfails to provide enough driving force to overcome the kineticbarrier to achieve the rutile phase. Careful selection of Tc,which is high enough to achieve rutile but not so high to de-stroy the ordered mesoporous structure, is the key to fabricat-ing the present ordered mesoporous rutile thin film.

A further interesting aspect of the crystallization is the ob-servation of oriented nucleation/growth of the oxide. After cal-cination at 800 8C the WAXS reflection corresponding to therutile (101) plane was abnormally strong compared to a poly-crystalline sample (see the standard JCPDS card 21-1276),which suggests oriented growth of rutile in the film. The sub-sequent calcination at 850 8C enhanced this effect significantly.Brezesinski et al. have recently reported the oriented growth ofoxides assisted by surfactants.[18] These insights can explain thefact that oriented growth of rutile is observed for the mesopo-rous film, while no remarkable oriented growth of titania canbe discerned in the nonporous film (Figure 3).[18] Notably, theonset of crystallization (observable by WAXS) is higher thanthe decomposition temperature. We speculate that at lowertemperatures oriented crystallization occurs at the interface to

Figure 2.WAXS patterns of the rutile film (x =0.2) after calcination at a) 800and b) 850 8C.

Figure 3. Phase evolution of the nonporous (TiO2)0.6 ACHTUNGTRENNUNG(Ta2O5)0.2 film with thecalcination temperature.

750 www.chemphyschem.org A 2008 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim ChemPhysChem 2008, 9, 748 – 757

B. M. Smarsly et al.

the substrate, which is not detectable by WAXS due to thevery low quantities of material at the interface. Further detailedexperiments are planned to elucidate this phenomenon. Theabnormally strong rutile (101) peak was reported for submi-crometer-scale rutile rods fabricated using TiCl3 solution underhydrothermal conditions,[19] and nanometer titania rods de-rived through Ti–H2O2 reactions.[2, 20] In addition, for both filmsthe WAXS peaks corresponding to titania, either anatase in thenonporous film or rutile in the mesoporous one, shifted slight-ly to lower angles. This finding suggests the incorporation of alow fraction of Ta atoms in the respective titania crystal lattice.

2.2 Atomic Force Microscopy

Atomic force microscopy (AFM) examinations reveal an ex-tremely flat, smooth, and crack-free surface with ordered mes-opores for the rutile film obtained after calcination at 800 8C(Figure 4). Height profiles across the mesopores reveal a high

degree of lateral mesoscopic perfection. The distance betweentwo neighboring mesopores, as indicated by the two arrows inFigure 4d, is uniform, thus suggesting the homogeneity of themesoporous structure. The AFM examinations suggest that theordered mesoporous structure of the rutile film is retainedupon heating to a temperature as high as 820 8C (see Figure S1in the Supporting Information), thus exceeding the stability ofother oxides. So far, such high thermal stability of the meso-pore structure was not observed for TiO2 films.

2.3 2D Small-Angle X-ray Scattering

The ordered mesoporous structure of the rutile film, heat-treat-ed at various temperatures, is further confirmed through char-

acterization of the film by small-angle X-ray scattering (SAXS)with a 2D detector (Figure 5) using ultrathin wafers as sub-strates (thickness of 30 mm). The maxima attributable to abody-centered cubic (bcc) mesostructure in [110] orientation

can be discerned in the well-defined 2D patterns measuredwith b=108 (angle of incidence between the X-ray beam andthe sample surface), even after crystallization at 800 8C.[21] Withincreasing calcination temperature, the 2D-SAXS patternsbecame increasingly ellipsoidal, which suggests an anisotropicshrinkage of the mesostructure in the direction normal to thesubstrate. All the transmission patterns (b=908) exhibit iso-tropic rings contributing to a random orientation of mesostruc-tured domains within the plane parallel to the substrate, as re-vealed by the AFM observations (Figure 4b,c). An in-plane lat-tice parameter of a=21 nm is obtained from the transmissionpatterns. The ordered mesoporous anatase films fabricated byapplying the same KLE template and the same EISA procedurehave been reported to present a bcc Im3m structure with [110]direction normal to the substrate.[10c] Assuming that the pres-ent mesopores also formed a bcc lattice with (110) plane paral-lel to the substrate, the pore-to-pore distance along the [110]direction in the (110) plane should be

ffiffiffi2p

a, that is, approxi-mately 29 nm. This value is in good agreement with the AFMobservation shown in Figure 4d.

2.4 X-ray Photoelectron Spectroscopy and Extended X-RayAbsorption Fine Structure Measurements

The actual surface and in-depth compositions of the mixedoxide films (TiO2)0.6ACHTUNGTRENNUNG(Ta2O5)0.2 treated at 200 and 800 8C are stud-ied by X-ray photoelectron spectroscopy (XPS). Figure 6 showsthe survey spectra of the rutile film (TiO2)0.6ACHTUNGTRENNUNG(Ta2O5)0.2 treated at200 and 800 8C after 30 min of sputtering, which demonstratesthe presence of Ti and Ta for both films. As far as the composi-tion is concerned, the Ti/Ta atomic ratio, whose theoreticalvalue based on the stoichiometry of the starting solutionshould be 1.5, is about 1.2–1.3 all along the layer up to 40 minof sputtering, thus revealing a very homogeneous intermixingof the two oxides in the thin films. The deviation of the Ti/Taratio most likely results from the deviated precursor composi-tion. In the deepest layers, this ratio is reduced to about 1,

Figure 4. AFM images of the rutile film a) before and b, c) after calcinationat 800 8C. The height profile of the section indicated in (c) is shown in (d).

Figure 5. 2D SAXS pattern of the mesoporous rutile film after calcination atvarious temperatures. The scattering vector s is defined as s =2/l sinq.

ChemPhysChem 2008, 9, 748 – 757 A 2008 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim www.chemphyschem.org 751

Mesoporous Rutile Thin Films

thus indicating a local enrichment of tantalum, the reason forwhich is not clear at present.

When comparing the carbon content, it is noted that in thesample treated at a lower temperature, a higher content ofcarbon (65 at.%) is detected on the surface, which is still highafter 30 min of sputtering (32 at.%). This finding is reasonablewhen considering the temperature of annealing, which is toolow to promote the complete thermal degradation of the or-ganic part. On the contrary, in the sample treated at highertemperature, the carbon content on the surface is about 45%,which is mainly ascribed to adventitious contaminatingcarbon. Accordingly, its amount in the deeper layers is reducedto about 7% upon sputtering.

Besides the survey spectra, selected spectra of the regionsof interest are also acquired. The BEs in the annealed samplesare corrected for charging effects by assuming for the C 1stransition a value typical of adventitious carbon, that is,284.6 eV.[22] As can be seen in Figure 7, the Ti 2p peak is sym-metric and does not show further components. The BE values,corrected for charging effects, are in the range 457.9–458.2 eV,that is, values slightly lower than those of titanium(IV) in titania(458.3–459.0 eV).[22] As far as the chemical state of tantalum isconcerned, in the sample treated at the higher temperaturethe BE of the Ta 4f peak is in the range 25.0–26.0 eV, which isalso in this case lower than that reported in the literature for

tantala (26.0–27.0 eV).[22] It should, however, be pointed outthat, in the case of the Ti 2p region, the values on the surfaceare higher (i.e. 458.2 eV), whereas after sputtering they de-crease, which can be ascribed to a partial reduction of titaniumdue to the preferential sputtering of oxygen induced by argonions. This would be also be in agreement with the experimen-tal O/ ACHTUNGTRENNUNG(Ti+Ta) molar ratios, which are higher on the surface andprogressively decrease upon sputtering (due to the preferentialremoval of the lighter oxygen atoms) and with the observedbroadening of the Ti 2p peak which, after sputtering, is broadand most likely composed of different components. Theoxygen O1s peak is quite symmetric and has a BE of 529.7 eVon the surface of the sample, which is close to the literaturevalues given for TiO2 (529.7–530 eV) and for Ta2O5 (529.8–530.4 eV).[22] As Ta and Ti have the same electronegativity valueof 1.5[23] and the cation polarizabilities of the two metal cationsare practically identical (0.185 L3 for Ta in Ta2O5 and 0.184 L3

for Ti in TiO2),[24] these lower BE values cannot be ascribed to

possible polarization effects occurring in the oxide networks.Instead, the strongly modified chemical environment aroundthe metal atoms with respect to that in the pure oxide, the lat-tice distortions, or the incorporation of Ti and Ta, respectively,into Ta2O5 and TiO2 could have resulted in the lower BE valuesfor the Ti�O and Ta�O bonds. The slightly shifted peaks corre-sponding to rutile in the WAXS patterns shown in Figures 2and 3 support the latter hypothesis, assuming such incorpora-tion. Furthermore, the presence of mixed Ti�O�Ta bonds couldnot be evidenced by XPS analysis, which is also in agreementwith the XRD data, thus revealing the formation of two sepa-rate anatase and orthorhombic tantala phases.

To analyze the chemical environment of the metal atoms inthe film more closely and to investigate the short-range orderaround the Ta atoms, X-ray absorption fine structure (XAFS)measurements on the (TiO2)0.6ACHTUNGTRENNUNG(Ta2O5)0.2 sample annealed at800 8C are carried out at the Ta LIII edge. As no reference sam-ples were measured, a sound oxidation state determination byanalysis of the edge position was not possible. Nevertheless,the X-ray absorption near-edge structure (XANES) spectrumshown in Figure S2 in the Supporting Information, especiallythe shape of the first resonance after the edge, is in goodagreement with that given in the literature for Ta2O5,

[25] whichis also consistent with the WAXS results. To confirm this pre-liminary conclusion, the extended XAFS (EXAFS) spectrum ofthe (TiO2)0.6ACHTUNGTRENNUNG(Ta2O5)0.2 sample annealed at 800 8C is fitted in afirst attempt with different tantalum oxides as model struc-tures. Their structural parameters are used as input in the ad-justments, in the course of which the coordination numbersare fixed to the crystallographic values. In this case, the bestquality of fit (R =25.3) is achieved with b-Ta2O5 of space groupPccm,[26] to which the signals in the XRD pattern can also beassigned.

The experimental and calculated spectra are compared inFigure 8. When the parameters of TaO2 are used, an R factor ofonly 29.1 is observed, which corresponds to a quality of fit re-duced by 15%. The difference between the two models is sig-nificant, and together with the XRD results the presence of b-Ta2O5 as the main component is confirmed by EXAFS spectros-

Figure 6. Superimposed XPS survey spectra of (TiO2)0.6 ACHTUNGTRENNUNG(Ta2O5)0.2 samplestreated at a) 200 and b) 800 8C after 30 min of sputtering. BE=bindingenergy.

Figure 7. Ti 2p region on the surface of the (TiO2)0.6 ACHTUNGTRENNUNG(Ta2O5)0.2 sample treatedat 800 8C.

752 www.chemphyschem.org A 2008 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim ChemPhysChem 2008, 9, 748 – 757

B. M. Smarsly et al.

copy. However, the fit with a rigid Ta2O5 model could not re-produce the signal at �3.1 L in the Fourier-transformed spec-tra. Several possibilities were tested to determine the underly-ing shells. It was not possible to fit this signal with only oneshell, no matter whether oxygen, titanium, or tantalum wasused as the backscatterer. A two-shell model consisting of thethird Ta�O of b-Ta2O5 and an additional tantalum shell, whosecoordination numbers were iterated, is superior with respectto all other possible combinations of the three available typesof backscatterers. During the fit, all other coordination num-bers of the basic b-Ta2O5 model were still fixed at their crystal-lographic values to maintain a high degree of overdeterminacy.The obtained result is given in Table 1 in comparison to theXRD values of b-Ta2O5. The improved approximation of the ex-perimental spectrum by this extended model in Figure 8 isclear. It is evident that the basic b-Ta2O5 structure is still pres-

ent and dominating, but an ad-ditional tantalum shell appears,at a Ta�Ta distance that can befound in other orthorhombicTa2O5 modifications, for exam-ple, with the space groupsC2mm[27] and Pmm2.[28] There-fore, a mixture of Ta2O5 modifi-cation can be deduced, with b-Ta2O5 as the main component,which contributes to the amor-phous character of Ta2O5 deter-mined by XRD. This highlyamorphous structure is also anexplanation of the rather largedeviation of distances from thecrystallographic values beyondthe nearest-neighbor shell. Thehigh Debye–Waller-like factors,especially in the first two tanta-lum shells, support this conclu-sion. No Ta�O�Ti groups couldbe fitted to the experimentaldata, which confirms the resultsof the XPS investigations.

2.5 Photocatalysis

Figure 9 presents the UV/Vis absorbance spectrum of the rutilefilm heated at 800 8C. The spectrum of the mesoporous ana-tase film obtained by calcination at 550 8C is shown as a refer-ence. As the rutile films also contains amorphous Ta2O5, theUV/Vis spectrum of an amorphous Ta2O5 mesoporous film,heated-treated at 550 8C to remove the KLE template, is alsoincluded in Figure 9. The absorbance of the amorphous Ta2O5

film detected by the UV/Vis spectrometer is relatively lowwithin the wavelength range measured; therefore, its contribu-tion to the photon-absorption behavior of the rutile film is notsignificant. The enhanced baseline for the rutile film is ascribedto its increased surface roughness compared to that of the

other two mesoporous thinfilms. The adsorption edge forthe mesoporous rutile and ana-tase films, determined by linearextrapolation of the steep partof the UV absorption towardthe baseline, was 350 and340 nm, respectively. Kaliwohet al.[29] have reported that thebandgap for the composite Ti–Ta oxide films deposited byphotoinduced chemical vapordeposition increased with in-creasing fraction of Ta2O5, be-cause the bandgap of Ta2O5

(4.0 eV[30]) is higher than that ofTiO2 (3.2 eV for anatase and

Figure 8. Experimental (*) and theoretical (c) k3·c(k) functions (left) and their corresponding Fourier-trans-formed spectra (right) of the (TiO2)0.6 ACHTUNGTRENNUNG(Ta2O5)0.2 sample calcined at 800 8C. Top: the fit with a rigid Ta2O5 model,which leaves the signal at �3.1 L unapproximated. Bottom: the fit with the parameters given in Table 1.

Table 1. Results of the EXAFS analysis.

Sample Abs–Bs[a] N[b] (Bs) R[c] [L] s[d] [L] R factor Ef [eV]

b-Ta2O5-Pccm[e] Ta–O 4 1.93ACHTUNGTRENNUNG(XRD) Ta–O 2 2.16

Ta–O 1 3.11Ta–Ta 6 3.63Ta–Ta 2 3.90

ACHTUNGTRENNUNG(TiO2)0.6 ACHTUNGTRENNUNG(Ta2O5)0.2 Ta–O 4 1.92�0.02 0.063�0.012 21.1 1.75800 8C Ta–O 2 2.03�0.02 0.055�0.011ACHTUNGTRENNUNG(EXAFS) Ta–O 0.4�0.1 3.03�0.03 0.032�0.006

Ta–Ta 2.4�0.4 3.17�0.03 0.105�0.021Ta–Ta 6 3.73�0.03 0.112�0.022Ta–Ta 2 3.88�0.02 0.067�0.014

[a] Abs=X-ray-absorbing atom, Bs=backscatterer. [b] Coordination number; values in italic were kept fixed atthe crystallographic value during the fit. [c] Interatomic distance Abs–Bs. [d] Debye–Waller-like factor. [e] Twocrystallographic sites were averaged.

ChemPhysChem 2008, 9, 748 – 757 A 2008 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim www.chemphyschem.org 753

Mesoporous Rutile Thin Films

3.0 eV for rutile).[31] Although the present rutile film containedsignificant amounts of Ta2O5, its bandgap, which is in inverseproportion to the adsorption edge, was smaller than that ofthe pure anatase film. This feature can be ascribed to the factthat the bandgap of rutile is slightly lower than that of ana-tase, and the amorphous Ta2O5 matrix in the mesoporous filmdid not affect significantly the photon-absorption behavior ofthe rutile film. The small broad peaks observed beyond360 nm for the three thin films could be attributed to interfer-ence effects of visible light between the ITO coating and the ti-tania layer.[32]

The PA of the rutile film was evaluated through photodegra-dation of rhodamine B (RhB), axanthene dye, in water. Theblank tests revealed that RhBdecomposed only slightly underUV light in the absence of anycatalysts. The decrease in RhBconcentration due to adsorptionof RhB molecules on the presentfilm was also found to be negli-gible. Figure 10a shows thephotodegradation curves of RhBin the presence of various thin films. Mesoporous anatase thinfilms heated at 550 8C, which were found to give the best PAunder the present conditions, served as a reference. The thick-ness of the three thin films was almost the same (ca. 200 nm),estimated by scanning electron microscopy observations. As-suming that the photocatalytic reaction follows first-order reac-tion kinetics (lnc0/c=kt),[33] a linear relationship between lnc0/cand the illumination time t can be achieved (see Figure 10b).The rate constants k derived for the various thin films arelisted in Table 2.

The rutile film obtained by heating at 800 8C possessed pro-nounced PA to assist photodegradation of RhB in water, theactivity being only slightly lower than that of the optimizedanatase film with ordered 3D mesoporous structure, takinginto account that the films possess a similar thickness. Further-more, our parallel experiments revealed that Ta2O5 thin filmswith ordered mesoporous structure possess a certain PA,which, however, is substantially lower than that of ordered

mesoporous TiO2 films. Amorphous films of Ta2O5 possess evenpoorer PA due to the increased recombination of photogener-ated electron–hole pairs caused by the large number of sur-face defects.[34] Meanwhile, rutile TiO2 is generally accepted topossess lower PA than anatase TiO2 because of its lower ca-pacity to adsorb O2 molecules, which serve as effective trapsfor photogenerated electrons.[35] However, the present rutilefilms, which consist of 50% amorphous Ta2O5 and 50% rutileTiO2 by volume, possessed only slightly lower PA than anatasemesoporous films. This means that the 50% rutile TiO2 nano-crystals in the walls should have provided more, or at leastcomparable, active sites for the photocatalytic reaction thanthe anatase phase in the pure mesoporous film. Therefore, itcan be safely concluded that the rutile films with rutile nano-crystals about 9 nm in size, distributed homogeneously inamorphous Ta2O5 in the walls of the ordered mesopores thatare accessible readily from the top, possess excellent photoca-talytic properties.

Figure 10. a) Photodegradation curves and b) the corresponding fitting re-sults assuming the pseudo-first-order reaction of RhB assisted by the variousoxide films: the mesoporous rutile film heated at 800 (Ru-800) and 850 8C(Ru-850), pure mesoporous anatase film heated at 550 8C (An-550), and non-porous (TiO2)0.6 ACHTUNGTRENNUNG(Ta2O5)0.2 film heated at 800 8C (Non-800).

Table 2. PA of the various thin films.

Composition Calcination temperature [8C] Morphology PA [min�1] Correlation coefficient R

ACHTUNGTRENNUNG(TiO2)0.6 ACHTUNGTRENNUNG(Ta2O5)0.2 800 Ordered mesoporous 0.0161 0.998ACHTUNGTRENNUNG(TiO2)0.6 ACHTUNGTRENNUNG(Ta2O5)0.2 850 Mesoporous 0.0251 0.996ACHTUNGTRENNUNG(TiO2)0.6 ACHTUNGTRENNUNG(Ta2O5)0.2 800 Nonporous 0.00615 0.999TiO2 550 Ordered mesoporous 0.0203 0.994

Figure 9. UV/Vis absorbance spectra of a) the indium tin oxide (ITO) sub-strate, b) amorphous Ta2O5, and c) anatase and d) rutile thin films with or-dered mesoporous structures.

754 www.chemphyschem.org A 2008 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim ChemPhysChem 2008, 9, 748 – 757

B. M. Smarsly et al.

Sun et al. argued that well-crystallized rutile titania powderwith high specific surface area and small crystal size possessedphotocatalytic properties as high as those of anatase titaniawith the same surface area.[12c] In the current investigation,grain growth of rutile is confined within the thin walls of theordered mesopores, even when heated at a high temperatureof 800 8C, which contributed positively to the PA of the rutilefilm. The potential of the conduction band edge of Ta2O5

[�0.4 eV vs normal hydrogen electrode (NHE)[34]] is a littlemore negative than that of TiO2 (�0.2 eV vs NHE).[36] It is there-fore possible for the photogenerated electrons of Ta2O5 to beinjected into the conduction band of the surrounding rutileTiO2, thus improving the PA of the Ta2O5 fraction in the rutilefilm. The rutile nanocrystals in combination with a differentoxide lead to an increased lifetime of photogenerated elec-trons due to the separation of the electron–hole pair at the in-terface between TiO2 and Ta2O5. Such separation on the nano-meter scale and surface electron-trapping sites at the interfa-ces play a crucial role in the case of the high-performance cat-alyst “P25”.[37]

Trace Ta doping in rutile might contribute to the photocata-lytic reaction; however, it is not enough to account for the de-rived PA of those 50% rutile nanocrystals in the compositefilm. The reaction rate constant of the composite film (TiO2)0.9-ACHTUNGTRENNUNG(Ta2O5)0.05 with ordered mesoporous structure, but consistingmainly of anatase (see Figure 1 for the WAXS pattern), is ap-proximately 0.0101 min�1, which is much lower than that ofthe pure anatase counterpart. The remarkably reduced PAcannot be fully understood at the present time; however, thisfact hints that the enhanced PA of rutile nanocrystals in thecomposite film (TiO2)0.6ACHTUNGTRENNUNG(Ta2O5)0.2 cannot be explained solely bythe positive effect of the trace Ta doping, if any.

Furthermore, we speculate that the oriented growth of rutilenanocrystals in the walls of the ordered mesoporous film couldalso contribute to the enhanced PA. It is reported that therutile (001) plane possesses the best PA among the crystallineplanes.[38] The difference in the photochemical activity toreduce silver ions observed between the least active (100) andthe most active (001) orientation is approximately an order ofmagnitude for the film 200 nm in thickness.[38] The abnormallystrong (101) peak in the WAXS pattern of the present rutilefilm suggests increased amounts of rutile nanocrystals with(101) planes parallel to the substrate. Therefore, the mostactive rutile (001) plane, which inclines to the rutile (101) planewith an angle of 338, would have more chance to expose itselfto the RhB solution in the wall of spherical mesopores.

The rutile film obtained after calcination at 850 8C possesseseven higher PA than the mesoporous anatase film, in spite ofthe fact that the ordered mesoporous structure deteriorated tosome extent due to further diffusion sintering at such a hightemperature (see the AFM surface morphology in Figure S1,Supporting Information). This fact highlights the importance ofcrystallinity to the PA of the rutile nanocrystals, because ahigher calcination temperature of 850 8C results in an im-proved crystallinity while maintaining the small grain size forrutile (Figure 2). In addition, although the nonporous (TiO2)0.6-ACHTUNGTRENNUNG(Ta2O5)0.2 film heated at 800 8C consists of well-crystallized

Ta2O5 and anatase TiO2 (Figure 3), its PA was only half that ofthe mesoporous rutile film. Therefore, the mesoporous struc-ture also contributed significantly to the photodegradation ofRhB in water, owing to the enhanced specific surface area aswell as the improved accessibility of the titania nanocrystals.[11]

3. Conclusions

Rutile thin films with mesoporous structures have been fabri-cated successfully through the EISA of a mixed solution of tita-nia and tantalum pentoxide, followed by calcination in a tem-perature range of 780–820 8C. Incorporation of 20 mol% Ta2O5

in the composite film increased the onset of crystallizationtemperature, which was high enough to ensure the crystalliza-tion of the amorphous titania to rutile, but at the same timelow enough to retain the ordered mesoporous structure withthe help of a novel KLE template. Although XPS indicates onlya small deviation of the actual Ta/Ti ratio from the compositionof the starting sol–gel solution in the majority of the film, thebottom part showed an enrichment in Ta. Grain growth ofrutile was confined by the mesoporous structure to be within9 nm, which, together with the ordered mesoporous structureitself, provides a comparatively good PA. The ordered mesopo-rous rutile films achieved under such a high temperature canbe expected to be of high thermal stability, and thus mayserve as an ideal catalyst or catalyst support to be used at ele-vated temperatures.

Experimental Section

Film Fabrication: In a typical procedure to deposit the (TiO2)0.6-ACHTUNGTRENNUNG(Ta2O5)0.2 film, TiCl4 (360 mg) and TaACHTUNGTRENNUNG(OC2H5)5 (480 mg) were eachdissolved in ethanol (2 g). Double-distilled water (0.5 g) was thenadded to the TiCl4 solution under magnetic stirring, after which theTaACHTUNGTRENNUNG(OC2H5)5 solution was added dropwise. A solution consisting ofKLE (100 mg), ethanol (2 g), and THF (1 g) was added to the mixedsolution and the final precursor was further stirred for 6–10 hbefore dip-coating. To prepare the composite film with various Ta/Ti ratios, various amounts of the Ta and Ti inorganic precursorswere used, with the total amount kept constant at 3 mm. The com-posite thin films were deposited on Si wafers by dip-coating in acontrolled relative humidity of 11% with a constant withdrawalspeed of 6.5 mms�1. The films were then dried at 80 8C for 2 h inair followed by a treatment at 200 8C overnight to stabilize themesoporous structure. Crystallization of the thin films was ach-ieved through heating at a speed of 5 Kmin�1 to the desired tem-peratures, followed by immediate removal of the sample to allowfor rapid cooling. The fabrication of pure Ta2O5 films follows a pre-viously described recipe.[18]

Film Characterization: AFM investigations were conducted using aNanoscope III instrument (Digital Instruments) in the tappingmode. WAXS signals were collected in a Bruker D8 diffractometerwith an accelerating voltage of 40 kV and a current of 40 mA,using CuKa radiation. The 2D SAXS measurements were carried outusing a Nonius rotating-anode setup (CuKa radiation with l=0.154 nm) featuring a three-pinhole collimation system and a MARCCD area detector, with a sample-to-detector distance of 750 mm.Ultrathin Si wafers with thickness of approximately 30 mm wereused for the measurement. The angle b between the incident

ChemPhysChem 2008, 9, 748 – 757 A 2008 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim www.chemphyschem.org 755

Mesoporous Rutile Thin Films

beam and the substrate was set to 108 and 908. UV/Vis absorbancespectra were collected from thin films dip-coated on ITO-coatedquartz glass substrates, using a Perkin–Elmer Lambda 2 UV/Visspectrophotometer, with the substrate serving as a reference.The composition of the thin film was investigated by XPS. Spectrawere obtained on a Perkin–Elmer F5600ci spectrometer usingstandard AlKa radiation (1486.6 eV) working at 350 W. The workingpressure was <5P10�8 Pa. The spectrometer was calibrated by as-suming the binding energy (BE) of the Au 4f7/2 line at 83.9 eV withrespect to the Fermi level. The standard deviation for the BE valueswas 0.15 eV. The reported BEs were corrected for charging effects,by assigning to the C 1s line of carbon the BE value of 284.6 eV.[22b]

Survey scans (187.85 pass energy, 1 eV per step, 25 ms per step)were obtained in the 0–1300 eV range. Detailed scans (58.7 eVpass energy, 0.1 eV per step, 100 ms per step) were recorded forthe O 1s, C 1s, Ti 2p, and Ta 4f regions. The atomic composition,after a Shirley-type background subtraction,[39] was evaluated usingsensitivity factors supplied by Perkin–Elmer.[22a] Depth profiles werecarried out by Ar+ sputtering at 3 kV with an argon partial pres-sure of 5P10�6 Pa. A specimen area of 2P2 mm2 was sputtered.Samples were introduced directly, by a fast entry lock system, intothe XPS analytical chamber. The assignment of the peaks was car-ried out by using the values reported in refs. [22a,40].

EXAFS Measurements: Tantalum LIII edge (9.881 keV) XAFS meas-urements were performed at the GILDA beam line (BM08) of theEuropean Synchrotron Radiation Facility in Grenoble, France. Ta ab-sorption spectra were collected in the fluorescence mode using adynamically sagittal-focusing SiACHTUNGTRENNUNG(111) monochromator and a 13-ele-ment hyperpure Ge detector. Digital electronics was used to readthe fluorescence detector. Five spectra were averaged to increasethe signal-to-noise ratio. To avoid mistakes in the XANES regiondue to small changes in the energy calibration between two meas-urements, all spectra were corrected to the edge position of a tan-talum foil, which was measured parallel to the samples betweenthe second and third ionization chambers. Several spectra wereaveraged to achieve a better signal-to-noise ratio.Data evaluation started with background absorption removal fromthe experimental absorption spectrum by subtracting a Victoreen-type polynomial.[41] To determine the smooth part of the spectrum,corrected for pre-edge absorption, a piecewise polynomial wasused. It was adjusted in such a way that the low-R components ofthe resulting Fourier transform were minimal. After division of thebackground-subtracted spectrum by its smooth part, the photonenergy was converted to photoelectron wavenumbers k. The re-sulting c(k) function was weighted with k3 and Fourier-transformedusing a Hanning window function of DR =3 L (1.0–4.0 L). Dataanalysis was performed in k space (Dk=11 L�1) on Fourier-filtereddata to remove noise and multiple scattering contributions. The fil-tered range was chosen according to the range of significant dataand is given in Table 1 together with the results of the fitting pro-cedure. According to the data range used, the number of inde-pendent points is Nind= (2DkDR/p)=21. The maximum number ofadjusted parameters was 15; therefore a high overdeterminacycould be achieved.[42] Adjustment of the common theoreticalEXAFS expression [Eq. (1)]:

cðkÞ ¼X

j

Nj

kr2jS20ðkÞFjðkÞe�2k2s2

j e�2rj=l sin 2krj þ djðkÞ� �

ð1Þ

was carried out according to the curved wave formalism of theEXCURV98 program with XALPHA phase and amplitude func-tions.[43] The mean free path of the scattered electrons was calcu-lated from the imaginary part of the potential (VPI set to

�4.00 eV). Since XALPHA phases and amplitudes were used, anamplitude reduction factor S2

0 of 0.8 was used to account for in-elastic processes.[44] However, the error that can be introduced by awrong S2

0 value is reflected in the rather large uncertainty of thecoordination numbers and Debye–Waller-like factors s. An innerpotential correction Ef was introduced when fitting experimentaldata with theoretical models, which accounts for an overall phaseshift between the experimental and calculated spectra. The qualityof fit is given in terms of the R factor according to Equation (2):[45]

R ¼X

i

k3 cexpðkiÞ � ctheoðkiÞ�� ��

k3 cexpðkiÞj j :100% ð2Þ

PA Evaluation: The PA of thin films was evaluated using RhB as amodel organic compound. For each test, RhB aqueous solution(33 mL) with an initial concentration of 0.01 mm was illuminatedwith an 8 W UV lamp (lmax=254 nm, CAMAG, Germany) in thepresence of the thin films (4P4 cm2 in total), which were immersedjust below the solution level. The distance between the lamp andthe film was about 2 cm. The solution was stirred continuously andexposed to air during the photocatalytic reaction. The change inthe RhB concentration was monitored with a UVIKON 931 UV/Visspectrophotometer at a wavelength of 554 nm, using a quartz cuv-ette with a 1 cm optical path length. About 2 mL RhB solution wasused for each measurement, which was then put back after deter-mining the absorbance to keep a constant solution volume.

Acknowledgements

J.-M.W. thanks the Alexander von Humboldt Foundation for hisresearch fellowship. BM08 Gilda at European Synchrotron Radia-tion Facility (ESRF) (Grenoble, France) is gratefully acknowledgedfor the provision of synchrotron radiation for EXAFS measure-ments. GILDA is a project jointly financed by CNR and INFN, Italy.Mauro Rovezzi and Dr. Francesco D’Acapito are gratefully ac-knowledged for the helpful technical and scientific support.

Keywords: mesoporous materials · photocatalysis · self-assembly · thin films · titanates

[1] a) D. Grosso, F. Babonneau, C. Sanchez, G. J. de A. A. Soler-Illia, E. L. Cre-paldi, P. A. Albouy, H. Amenitsch, A. R. Balkenende, A. Brunet-Bruneau, J.Sol–Gel Sci. Technol. 2003, 26, 561–565 ; b) B. Smarsly, M. Antonietti, Eur.J. Inorg. Chem. 2006, 6, 1111–1119; c) G. Cortial, M. Siutkowski, F. Goett-mann, A. Moores, C. Boissiere, D. Grosso, P. Le Floch, C. Sanchez, Small,2006, 2, 1042–1045; d) E. Lancelle-Beltran, P. Prene, C. Boscher, P. Belle-ville, P. Buvat, S. Lambert, F. Guillet, C. Boissiere, D. Grosso, C. Sanchez,Chem. Mater. 2006, 18, 6152–6156.

[2] a) J. M. Wu, T. W. Zhang, Y. W. Zeng, S. Hayakawa, K. Tsuru, A. Osaka,Langmuir 2005, 21, 6995–7002; b) J. M. Wu, Environ. Sci. Technol. 2007,41, 1723–1728.

[3] a) O. K. Varghese, D. W. Gong, M. Paulose, K. G. Ong, E. C. Dickey, C. A.Grimes, Adv. Mater. 2003, 15, 624–627; b) A. S. Zuruzi, A. Kolmakov, N. C.MacDonald, M. Moskovits, Appl. Phys. Lett. 2006, 88, 102904.

[4] K. L. Frindell, J. Tang, J. H. Harreld, G. D. Stucky, Chem. Mater. 2004, 16,3524–3532.

[5] T. Watanabe, H. Hayashi, H. Imai, Sol. Energy Mater. Sol. Cells 2006, 90,640–648.

[6] V. Subramanian, A. Karki, K. I. Gnanasekar, F. P. Eddy, B. Rambabu, J.Power Sources 2006, 159, 186–192.

[7] T. Tsuru, D. Hironaka, T. Yoshioka, M. Asaeda, Sep. Purif. Technol. 2001,25, 307–314.

756 www.chemphyschem.org A 2008 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim ChemPhysChem 2008, 9, 748 – 757

B. M. Smarsly et al.

[8] V. Subramanian, Z. Ni, E. G. Seebauer, R. I. Masel, Ind. Eng. Chem. Res.2006, 45, 3815–3820.

[9] a) X. X. Wang, S. Hayakawa, K. Tsuru, A. Osaka, J. Biomed. Mater. Res.2000, 52, 171–176; b) J. M. Wu, S. C. Zhang, Y. W. Li, M. Wang, A. Osaka,Surf. Coat. Technol. 2006, 201, 3181–3188; c) J. M. Wu, J. F. Liu, S. Haya-kawa, K. Tsuru, A. Osaka, J. Mater. Sci. Mater. Med. 2007, 18, 1529–1536.

[10] a) P. C. A. Alberius, K. L. Frindell, R. C. Hayward, E. J. Kramer, G. D. Stucky,B. F. Chmelka, Chem. Mater. 2002, 14, 3284–3294; b) D. Grosso,G. J. de A. A. Soler-Illia, E. L. Crepaldi, F. Cagnol, C. Sinturel, A. Bourgeois,A. Brunet-Bruneau, H. Amenitsch, P. A. Albouy, C. Sanchez, Chem. Mater.2003, 15, 4562–4570; c) B. Smarsly, D. Grosso, T. Brezesinski, N. Pinna, C.Boissiere, M. Antonietti, C. Sanchez, Chem. Mater. 2004, 16, 2948–2952;d) S. Tursiloadi, Y. Yamanaka, H. Hirashima, J. Sol–Gel Sci. Technol. 2006,38, 5–12; e) A. B. Dros, D. Grosso, C. Boissiere, G. J. de A. A. Soler-Illia,P. A. Albouy, H. Amenitsch, C. Sanchez, Microporous Mesoporous Mater.2006, 94, 208–213.

[11] a) J. C. Yu, L. Z. Zhang, J. G. Yu, Chem. Mater. 2002, 14, 4647–4653; b) E.Beyers, P. Cool, E. F. Vansant, J. Phys. Chem. B 2005, 109, 10081–10086;c) Y. Sakatani, D. Grosso, L. Nicole, C. Boissiere, G. J. de A. A. Soler-Illia, C.Sanchez, J. Mater. Chem. 2006, 16, 77–82; d) J. Tang, Y. Y. Wu, E. W.McFarland, G. D. Stucky, Chem. Commun. 2006, 1670–1671.

[12] a) S. Tsai, S. Cheng, Catal. Today 1997, 33, 227–237; b) N. G. Park, J. vande Lagemaat, A. J. Frank, J. Phys. Chem. B 2000, 104, 8989–8994; c) J.Sun, L. Gao, Q. H. Zhang, J. Am. Ceram. Soc. 2003, 86, 1677–1682;d) Y. Y. Li, J. P. Liu, Z. J. Jia, Mater. Lett. 2006, 60, 1753–1757; e) G. Li, K. A.Gray, Chem. Mater. 2007, 19, 1143–1146.

[13] Y. Z. Li, N. H. Lee, E. G. Lee, J. S. Song, S. J. Kim, Chem. Phys. Lett. 2004,389, 124–128.

[14] A. S. Barnard, P. Zapol, Phys. Rev. B 2004, 70, 235403.[15] a) P. S. Dobal, R. S. Katiyar, Y. Jiang, R. Guo, A. S. Bhalla, J. Phys. Chem.

Solids 2000, 61, 1805–1808; b) L. Armelao, H. Bertagnolli, S. Gross, V.Krishnan, U. Lavrencic-Stangar, K. MTller, B. Orel, G. Srinivasan, E. Tondel-lod, A. Zattin, J. Mater. Chem. 2005, 15, 1954–1965; c) L. Armelao, H.Bertagnolli, D. Bleiner, M. Groenewolt, S. Gross, V. Krishnan, C. Sada, U.Schubert, E. Tondello, A. Zattin, Adv. Funct. Mater. 2006, 17, 1671–1681.

[16] M. Kuemmel, D. Grosso, U. Boissiere, B. Smarsly, T. Brezesinski, P. A.Albouy, H. Amenitsch, C. Sanchez, Angew. Chem. 2005, 117, 4665–4668;Angew. Chem. Int. Ed. 2005, 44, 4589–4592.

[17] T. Brezesinski, B. Smarsly, K. Iimura, D. Grosso, C. Boissiere, H. Amenitsch,M. Antonietti, C. Sanchez, Small 2005, 1, 889–898.

[18] T. Brezesinski, M. Groenewolt, N. Pinna, H. Amenitsch, M. Antonietti, B.Smarsly, Adv. Mater. 2006, 18, 1827–1832.

[19] E. Hosono, S. Fujihara, K. Kakiuchi, H. Imai, J. Am. Chem. Soc. 2004, 126,7790–7791.

[20] a) J. M. Wu, J. Cryst. Growth 2004, 269, 347–355; b) J. M. Wu, B. Qi, J.Am. Ceram. Soc. 2007, 90, 657–660.

[21] W. Ruland, B. Smarsly, J. Appl. Crystallogr. 2007, 40, 409–417.

[22] a) J. F. Moulder, W. F. Stickle, P. E. Sobol, K. D. Bomben in Handbook of X-Ray Photoelectron Spectroscopy, Perkin–Elmer, Eden Prairie, MN, 1992 ;b) M. P. Seah in Practical Surface Analysis, Vol. 1 (Eds. : D. Briggs, M. P.Seah), Wiley, 1990, p. 543.

[23] CRC Handbook of Chemistry and Physics, 84th Ed. , CRC, Boca Raton,2004.

[24] J. Portier, H. S. Hilal, I. Saadeddin, S. J. Hwang, M. A. Subramanian, G.Campet, Prog. Solid State Chem. 2004, 32, 207–217.

[25] V. M. JimUnez, A. Caballero, A. FernandUz, J. C. SVnchez-LWpez, A. R.GonzVlez-Elipe, J. F. Trigo, J. M. Sanz, Surf. Interface Anal. 1997, 25, 707.

[26] L. A. Aleshina, S. V. Loginova, Crystallogr. Rep. 2002, 47, 415–419.[27] N. C. Stephenson, R. S. Roth, Acta Crystallogr. B 1971, 27, 1037.[28] H. U. Hummel, R. Fackler, P. Remmert, Chem. Ber. 1992, 125, 551–556.[29] N. Kaliwoh, J. Y. Zhang, I. W. Boyd, Appl. Surf. Sci. 2002, 186, 246–250.[30] H. Kato, A. Kudo, Chem. Phys. Lett. 1998, 295, 487–492.[31] Q. H. Zhang, L. Gao, J. K. Guo, J. Eur. Ceram. Soc. 2000, 20, 2153–2158.[32] N. Sakai, Y. Ebina, K. Takada, T. Sasaki, J. Am. Chem. Soc. 2004, 126,

5851–5858.[33] a) Y. Ma, J. N. Yao, Chemosphere 1999, 38, 2407–2414; b) J. M. Wu, T. W.

Zhang, J. Photochem. Photobiol. A 2004, 162, 171–177.[34] H. Kominami, M. Miyakawa, S. Murakami, T. Yasuda, M. Kohno, S.

Onoue, Y. Kera, B. Ohtani, Phys. Chem. Chem. Phys. 2001, 3, 2697–2703.[35] a) M. A. Fox, M. T. Dulay, Chem. Rev. 1993, 93, 341–357; b) S. Ichikawa, R.

Doi, Catal. Today 1996, 27, 271–277; c) T. Ohno, K. Sarukawa, M. Matsu-mura, J. Phys. Chem. B 2001, 105, 2417–2420.

[36] A. Hagfeldt, M. Gratzel, Acc. Chem. Res. 2000, 33, 269–277.[37] D. C. Hurum, K. A. Gray, T. Rajh, M. C. Thurnauer, J. Phys. Chem. B 2005,

109, 977–980.[38] P. A. M. Hotsenpiller, J. D. Bolt, W. E. Farneth, J. B. Lowekamp, G. S.

Rohrer, J. Phys. Chem. B 1998, 102, 3216–3226.[39] D. A. Shirley, Phys. Rev. 1972, 55, 4709.[40] NIST XPS Database, X-ray Photoelectron Spectroscopy Database 20, Ver-

sion 3.0, National Institute of Standards and Technology, Gaithersburg,MD, http://srdata.nist.gov/XPS.

[41] a) T. S. Ertel, H. Bertagnolli, S. HTckmann, U. Kolb, D. Peter, Appl. Spec-trosc. 1992, 46, 690–698; b) M. Newville, P. Livins, Y. Yacoby, J. J. Rehr,E. A. Stern, Phys. Rev. B 1993, 47, 14126.

[42] E. A. Stern, Phys. Rev. B 1993, 48, 9825–9827.[43] S. J. Gurman, N. Binsted, I. Ross, J. Phys. C 1984, 17, 143–151.[44] M. Roy, S. J. Gurman, J. Synchr. Rad. 1999, 6, 228.[45] N. Binsted, EXCURV98: CCLRC Daresbury Laboratory computer program,

manual (Ed. : F. Mosselman), 1998.

Received: October 10, 2007

Revised: November 17, 2007

ChemPhysChem 2008, 9, 748 – 757 A 2008 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim www.chemphyschem.org 757

Mesoporous Rutile Thin Films