Pure versus metal-ion-doped nanocrystalline titania for photocatalysis

https://doi.org/10.1016/j.apcatb.2006.06.007Get rights and content

Abstract

Thin films of pure or doped nanocrystalline titania have been deposited on glass slides by using a sol–gel procedure, in the presence of surfactant Triton X-100, which acts as template of the nanostructure. Fe3+, Cr3+ and Co2+ were used as dopants while the doping extended in a broad domain from very low to very high levels. The photocatalytic efficiency of pure or doped titania was tested for discoloration of an aqueous solution of Basic Blue 41. The presence of dopants resulted in a progressive loss of total crystallinity, some transition from anatase into rutile and, in the case of Co2+, formation of the mixed oxide cobalt titanate. Loss of anatase had dramatic consequences on photocatalytic efficiency by UV–vis excitation, which decreased fast with increasing dopant concentration. Selected visible excitation of the doped titania could lead to photodegradation of the dye but to a far lesser degree than UV–vis excitation. Photosensitization by absorption of light by the dye itself loses its importance in the presence of the dopant. Thus the doped material is a visible-light photocatalyst but substantial photodegradation efficiency is achieved only at very high doping levels, for example, 20 at.% for Fe3+ doping. In any case, direct UV excitation of pure titania is a more efficient photocatalytic process than visible excitation of the doped semiconductor.

Introduction

The study of TiO2 and its photocatalytic applications occupies a large portion of Materials Science research and attracts the interest of an increasing number of researchers. Despite this intensified research, several questions asked 10–15 years ago [1], [2], [3], [4], [5] are still asked today and this keeps interest high. These questions will also emerge in the following paragraphs. Titania photocatalyst cannot be conceived today but in nanocrystalline form. Amorphous titania is not a photocatalyst. Among the three principal crystalline forms of titania (i.e. anatase, rutile and brookite), only the first two are detected in samples synthesized by usual “soft chemistry” procedures, like sol–gel. Both anatase and rutile are semiconductors with band gap located at about 3.2 and 3.0 eV (385 and 410 nm), respectively. These are indicative values but they vary in practice. Thus, when TiO2 nanocrystals are deposited in the form of a transparent nanocrystalline thin film and the absorption spectrum is recorded, it is found that the absorption onset varies as a function of the size of the nanoparticles. Larger nanoparticles have an absorption onset that lies at longer wavelength [6]. This behavior is very clearly marked when the same synthesis procedure is used to make nanocrystals of various sizes [6]. The variation of the absorption onset is not due to quantum size effects as one might first think, since most TiO2 nanocrystals synthesized by sol–gel procedures or commercial nanocrystals like Degussa P25 are relatively large particles of diameter of a few tens of nanometers (cf. Ref. [2]). It is a result of the existence of defect sites in and on the nanocrystals. Defect sites are very important in photocatalysis, as it will be also seen below, as important as crystallinity itself. Another reason for absorption onset variation is formation of a mixture of the two crystalline phases of titania, i.e. anatase and rutile. By increasing the percentage of rutile, the absorption onset moves to longer wavelengths. This red shift of light absorption has been a major quest of the TiO2-related research, in order to increase absorption in the visible. Visible light absorption is, of course, an important issue, since it is related with solar applications. Rutile does absorb some visible light, while anatase only absorbs in the UV. Unfortunately, rutile is not a good photocatalyst. There are a few reasons for this rutile deficiency: (1) rutile is usually obtained at higher temperature than anatase. Sintering then of nanoparticles increases their size at the expense of the total active surface area. (2) The electron–hole recombination rate is much higher in rutile than it is in anatase. The reason is that rutile carries a limited number of hydroxyl groups on its surface [4], [7]. Since hydroxyl groups are major hole scavengers their absence results in increasing recombination rate [1]. It has also been suggested that since rutile carries a limited number of oxygen defect sites it cannot accommodate oxygen, which is a major electron scavenger [8]. This again results in increasing recombination rates. (3) Furthermore, it has also been claimed that since the electron lying at the conduction-band edge will have a higher reductive power in anatase than in the lower-lying rutile, the photocatalytic efficiency of the latter will be lower [4]. Anatase is then acknowledged as the principal photocatalytic titania phase. However, it is also known that optimal photocatalytic efficiency is obtained with a mixture of anatase and a small percentage of rutile. This optimized behavior is due to a synergistic effect between the two nanocrystalline phases. The lower band gap of the rutile nanocrystals creates energy wells that trap electrons and prevent electron–hole recombination [9]. Other authors claim that synergy is not an one-way passive electron trapping by rutile but electrons can be also transferred from rutile to lower energy anatase trapping sites [10]. It looks like both conceptions may be valid, as long as the concentration of defects is optimized so as to avoid adverse effects. Since nanocrystals of different size have different energy gaps, a size polydispersity might in some cases be beneficial and in other cases detrimental for photocatalytic efficiency. When nanocrystalline titania is synthesized by the sol–gel procedure in the presence of surfactant or other organic templates high calcination temperature is not the only means to produce rutile nanocrystals since rutile can be obtained also at relatively low temperatures, such as 400–450 °C [11]. The great majority of works dealing with TiO2 involve sol–gel synthesis. The large diversity of synthesis conditions that this method permits results in a large range of compositions of nanocrystalline phases and, subsequently, photocatalytic efficiencies, which are justified in view of the above introduction.

The situation is further complicated when dopants are introduced in the titania structure. Dopants should be distinguished from noble metal deposits on the surface of the nanoparticle. These deposits serve as nanoscale metal–semiconductor contacts and usually act beneficially for photocatalysis by scavenging electrons [3], [6], [12], [13], [14]. Dopants should also be distinguished from metal ions co-dissolved in a photodegradable solution acting as redox species, i.e. donor–acceptors of the photogenerated charges [15]. On the other hand, dopants can be either transition metal ions [5], [16], [17], [18], [19], [20], [21], [22] or non-metals, like S, N and C [23], [24], [25], [26], [27], [28], [29]. Dopants also shift absorption into the visible as does nanoparticle size increase. The difference is that shifting from anatase to rutile or from smaller to larger nanocrystallites simply decreases energy band gap, while doping creates energy states within the band gap. Therefore, in the second case the red shift of absorption is more drastic. The introduction of dopants allows titania to absorb in the visible but this does not necessarily mean that the doped catalyst is a better photocatalyst. When the doping level overpasses an optimal limit, which usually lies at very low dopant concentration and low visible light absorbance, dopants become recombination sites and have adverse effects on photocatalysis. It must be also taken into account that electrons transferred in dopant trapping states have lower energy than anatase conduction-band edge electrons and for this reason will have lower reductive capacity. Furthermore, it is nowadays possible to construct photocatalytic water-treatment installations involving inexpensive UV-light sources [30] and optimized pure titania catalyst so that doping might be purposeless. Nevertheless, the idea of using visible (which means solar) radiation for environmental remediation (water treatment, air purification or surface self-cleaning) is always attractive. For this reason, the research on doped titania will continue for still years to come. The present work presents data, which are analyzed in the spirit of the above introduction and help in the rationalization of the doped versus undoped titania applicability in various photocatalytic processes.

Titania is studied as a photocatalyst also for solar conversion applications (for example, dye-sensitized solar cells). It must be emphasized that the above approach does not apply to solar cells. In that case, the mobility of charge carriers is a more important issue than absorption in the visible, which is made by the sensitizer anyway. Doping and creation of charge-carrier traps has adverse effects on charge mobility and cell functioning. Doping in fact is an important issue only in photodegradation processes.

Section snippets

Experimental

All reagents used were purchased from Aldrich, were of the best quality available and were used as received.

Nanocrystalline titania films were deposited on glass slides (both sides applicable) under the following procedure which is systematically employed in our laboratory (cf. Refs. [11], [30]): 1.4 g of the non-ionic surfactant Triton X-100 [polyoxyethylene-(10) isooctylphenyl ether] was mixed with 7.6 ml ethanol. Then we added 1.36 ml glacial acetic acid (AcOH) and 0.72 ml of titanium

Results and discussion

In order to study the photocatalytic degradation of an aqueous organic load we have used a model dye, namely Basic Blue 41 (BB). The structure of BB can be seen in the inset of Fig. 1. We have chosen this dye because we have repeatedly employed it in the past [6], [11], [30], [31], [32], [33] and because it fits the present purpose. The titania films made by the above procedure adsorb a high load of the dye [11] and for this reason photodegradation is rapid. The BB concentration in the aqueous

Conclusions

Pure and doped nanocrystalline titania films can be easily synthesized by using a straight forward sol–gel method. Three cationic dopants have been studied: Fe3+, Cr3+ and Co2+ while doping level ranged in a large domain. In all cases, the films, which have been deposited on glass slides, are nanostructured, as a result of the presence of surfactant Triton X-100 in the original sol. Pure titania consisted of anatase nanocrystals of about 15–20 nm particle diameter. The presence of the ions

Acknowledgements

We thank the European Social Fund (ESF), Operational Program for Educational and Vocational Training II (EPEAEK II), and particularly the Program HERAKLITOS, for funding the above work.

References (51)

  • E. Stathatos et al.

    Micropor. Mesopor. Mater.

    (2004)
  • S. Sakthivel et al.

    Water Res.

    (2004)
  • A. Patsoura et al.

    Appl. Catal. B: Environ.

    (2006)
  • M.I. Litter

    Appl. Catal. B

    (1999)
  • D.W. Bahnemann et al.

    Appl. Catal. B: Environ.

    (2002)
  • A. Di Paola et al.

    Catal. Today

    (2002)
  • D. Dvoranova et al.

    Appl. Catal. B: Environ.

    (2002)
  • S. Sato et al.

    Appl. Catal. A: Gen.

    (2005)
  • E. Stathatos et al.

    Colloid Surf. A: Phys. Eng. Aspects

    (1999)
  • P. Bouras et al.

    Appl. Catal. B

    (2004)
  • J. Arana et al.

    Appl. Catal. B

    (2001)
  • J. Zhu et al.

    J. Mol. Catal. A

    (2004)
  • J. Zhu et al.

    J. Photochem. Photobiol. A

    (2006)
  • X. Chu et al.

    Mater. Res. Bull.

    (1999)
  • J.T. Chang et al.

    Surf. Coat. Technol.

    (2005)
  • Y. Brik et al.

    J. Catal.

    (2001)
  • J.-M. Herrmann et al.

    Chem. Phys. Lett.

    (1984)
  • F. Kiriakidou et al.

    Catal. Today

    (1999)
  • M.A. Fox et al.

    Chem. Rev.

    (1993)
  • M.R. Hoffmann et al.

    Chem. Rev.

    (1995)
  • N. Serpone et al.

    Langmuir

    (1994)
  • G. Riegel et al.

    J. Phys. Chem.

    (1995)
  • K.E. Karakitsou et al.

    J. Phys. Chem.

    (1993)
  • E. Stathatos et al.

    Langmuir

    (2001)
  • Z. Ding et al.

    J. Phys. Chem. B

    (2000)
  • Cited by (0)

    View full text