Effect of tungsten-doping on the properties and photocatalytic performance of titania thin films on glass substrates

Highlights

• Increase in W-doping enhanced anatase recrystallization.

• Increase in W-doping increased grain sizes and changed grain morphology.

• W⁶⁺ ions enhanced Ti-vacancy formation, leading to lattice contraction.

• W-doping lowered the band gap owing to a positive shift in the conduction band.

• The best photocatalytic performance was observed at 0.01 mol% W.

Abstract

Highly-transparent TiO₂ thin films with varying W-dopant levels (0.010–0.100 mol%) were deposited on soda-lime-silica glass substrates by spin coating, followed by annealing in air at 450 °C for 2 h. Characterization was by GAXRD, AFM, SIMS, XPS, UV–vis, and MB degradation. The films were comprised of single-crystal anatase grains. Increasing doping levels increased the crystallite sizes, grain sizes, and film thicknesses; they also changed the grain morphology from anhedral to euhedral. W⁶⁺ ions formed a substitutional solid solution in anatase and probably exhibited charge compensation by Ti vacancy formation, which led to lattice contraction. Contamination from the glass substrates was homogeneously distributed on the grain boundaries. Doping decreased the optical indirect band gap. However, at the highest doping level, it is not certain if this was an artificial effect of precipitation from overdoping or a true effect of lattice stress. The 0.010 mol% W sample showed the best performance probably owing to its formation of a homogeneous solid solution within the range of semiconducting properties, which could have narrowed the band gap or introduced mid-gap states. The 0.100 mol% W sample may have been overdoped and caused WO₃ precipitation. If so, the presence of the TiO₂–WO₃ couple may have enhanced charge separation and hence photocatalytic performance.

Introduction

Photocatalytic materials [1], [2], [3] are being used currently for self-sterilizing surfaces [4], air purification [5], [6], water purification [7], [8], [9], self-cleaning surfaces [10], [11], [12], and solar cells [13], [14], [15]. Titanium dioxide (TiO₂) is one of the most promising photocatalysts owing to its low cost, non-toxicity, and chemical and thermal stability [16], [17], [18]. Work by Fujishima and Honda [19] on water splitting using single-crystal TiO₂ showed that high-energy radiation, such as UV, is required to excite electrons from the valence to the conduction band owing to the relatively large band gap (∼3.2 eV [20]). This limits the efficiency of TiO₂ as a photocatalyst since UV radiation comprises only ∼5% of the solar spectrum [21].

Several methods have been used to enhance both the spectral response of TiO₂ in the visible region and its photocatalytic activity [22], [23], [24]. The most common is the addition of dopants, which are intended to create mid-gap states in the electronic structure, thereby narrowing the band-gap [25], [26]. The dopants also act as trapping centers that separate the hole and electron pairs, thereby extending the lifetime of these charge carriers and enhancing photocatalytic activity [26]. Several dopants, including metals [27], [28], [29], [30] and non-metals [31], [32], [33], have been trialed for this purpose.

Doping with transition metal ions can help to retard the recombination of photogenerated electrons and holes, which thus increases the lifetime of charge carriers [34]. Moreover, tungsten (W) additions are believed to extend the spectral response of TiO₂ into the visible region through a positive shift in the conduction band [12], [13]. However, the limited studies on W-doped TiO₂ have focused on powders. León-Ramos et al. [35] reported that 0.03 mol% W-doped TiO₂, produced hydrothermally, displayed higher photodegradation compared to undoped TiO₂. Li et al. [36] demonstrated that 3 mol% WO₃-doped TiO₂ powders, prepared by sol–gel, were most effective in degrading organic solutions, while Couselo et al. [37] suggested that the optimal performance of 3 mol% W-doped TiO₂ powders resulted from two competing factors, namely increasing surface area and decreasing specific activity per adsorption site with increased doping. Recently, Hua et al. [38] reported that sol–gel-derived TiO₂ nanoparticles doped with 0.4–1.0 wt% W resulted in enhanced capability of degradation of methyl orange solution owing to the role of W ions in shifting the absorption edge of TiO₂ to the visible range. Wu et al. [39] showed that W-doped TiO₂ particles produced by the solvothermal method displayed enhanced near-infrared absorption performance and similarly enhanced photocatalytic activity to that of commercial Degussa P25 TiO₂ in the degradation of NO_x gas. Further, Sathasivam et al. [40] demonstrated that 2.25 at% W-doped TiO₂ films prepared by aerosol-assisted chemical vapor deposition were effective in degrading resazurin redox dye since the dopants contributed to an increase in the surface area and limitation in the relative grain boundary regions, which act as charge carrier recombination sites.

Prior studies have shown that doping with W⁶⁺ has the potential to modify the photocatalytic properties of TiO₂ since the similarity of its crystal radius to that of Ti⁴⁺ [41] suggests the potential for substitutional solid solubility. Further, the fact that its crystal radius is smaller than that of the two interstices adjacent to the central Ti ion in the elongated TiO₆ octahedron [42] suggests the potential for interstitial solid solubility. Both types of solid solubility would require charge compensation in the form of titanium vacancies or oxygen injection (with ionic compensation) or electrons (with electronic compensation); Frenkel defect formation also is possible. All of these would be expected to have significant effects on the photocatalytic performance.

Most prior work has focused on reducing the optical band gap of TiO₂ through W additions, without extended consideration of the effects of doping on the mineralogical, microstructural, and photocatalytic properties. Further, in other studies, the doping levels used have tended to be high (2–20 mol%) and well above the typical semiconducting addition range. This is particularly the case for W-doped thin films, as shown in Table 1 [35], [36], [37], [38], [39], [40], [43], [44], [45]. Previous work by the authors [46] has led to the observation that additions of transition metal dopants typically in excess of ∼1 mol% (metal basis) are capable of exceeding the solubility limits, resulting in precipitation on the grain boundaries. This effect lowers the band gap of thin films but it cannot alter that of the doped TiO₂ further than that of the TiO₂ at the solubility limit. Hence, the band gap of the thin film and that of the TiO₂ comprising the film are different. Further effects may include lattice distortion and partial amorphization. All of these effects can alter the photocatalytic efficiency of the doped films, generally deleteriously.

There are many variables that affect the photocatalytic performance of doped TiO₂, including crystal structure, lattice deformation, defect equilibria (structural variables), grain size, surface area, and topography (microstructural variables) [46]. Since previous studies have taken a more focused approach to this range of variables, the present work aims to describe a more comprehensive investigation into the effects of doping on the mineralogical, microstructural, optical, chemical, and photocatalytic performance of TiO₂ thin films deposited on the common substrate soda-lime-silica glass. In the present case, doping with W in a compositional range (0.010–0.100 mol%) that is consistent with the formation of a homogeneous solid solution and semiconducting dopant levels is designed to systematize the structural variables while avoiding the effects of mechanistic changes in the microstructural variables, all of which affect the photocatalytic properties. In particular, since the band gap is used widely in the assessment of photocatalytic performance, in the absence of a WO₃-TiO₂ phase diagram or other solubility data, the low levels of dopant are intended to avoid precipitation, significant modification of the microstructure from precipitation and liquid formation, and artificial alteration of the band gap. In the present work, the recognized potential to influence the photocatalytic performance by W-doping of TiO₂, summarized above, has been investigated by the fabrication of thin films by spin coating, followed by annealing at 450 °C for 2 h. The thin films were characterized using different techniques to investigate the effects of W-doping on the mineralogical, morphological, topographical, chemical, optical, and photocatalytic properties.