Article pubs.acs.org/JPCC
Making Photo-selective TiO2 Materials by Cation−Anion Codoping: From Structure and Electronic Properties to Photoactivity Antonio M. Márquez,* José J. Plata, Yanaris Ortega, and Javier Fdez. Sanz Departamento de Química Física, Universidad de Sevilla, Facultad de Química, 41012 Sevilla, Spain
Gerardo Colón Instituto de Ciencia de Materiales de Sevilla, Centro Mixto CSICUniversidad de Sevilla, C/Américo Vespucio 49, 41092 Sevilla, Spain
Anna Kubacka and Marcos Fernández-García Instituto de Catálisis y Petroleoquímica, CSIC, C/Marie Curie 2, 28049 Madrid, Spain S Supporting Information *
ABSTRACT: Photoselective oxidation yielding high-added value chemicals appears as a green novel process with potential to be explored. In this study we combine spectroscopic XPS (N 1s and O 1s) and multiwavelength Raman data with density functional theory calculations to explore the structural and electronic properties of W,N-codoped TiO2 anatase surfaces and interpret the outstanding photocatalytic properties of such a system in partial oxidation reactions. Theoretical calculations allow us to examine several substitutional and N-interstitial configurations at different concentrations of the W,N dopants (similar to those experimentally found), as well as their interaction with structural point defects: Ti cation vacant sites and surface wolframyl species that are required to compensate the extra charge of the W6+ and N-containing anions. The joint use of theoretical and experimental XPS and Raman tools renders key structural information of W,N-codoped microcrystalline TiO2 solids. The incorporation of N at substitutional positions of anatase with the concomitant presence of WO species introduces localized states in the band gap, a result that is critical in interpreting the chemical behavior of the solids. The combination of the electronic and geometric structural information leads to a simple mechanism that rationalizes the experimentally observed photoactivity and selectivity in partial oxidation reactions. of the crystal surface.14 An obvious problem hindering the photochemical applications of TiO2 is its poor visible light response due to its wide band gap (∼3.2 eV on the anatase polymorph). Different approaches have been pursued to narrowing this band gap and increasing visible light absorption. One of the most common methods is the introduction of
1. INTRODUCTION Titanium dioxide (TiO2) is a material that has received considerable attention in the literature over the years, due to its versatility and the multitude of its applications in catalysis, photocatalysis, solar cells, waste remediation, and biocompatible materials.1−10 In particular, nanocrystalline TiO2 has been extensively studied due to its notable catalytic and photocatalytic activity.11 Of the different titania polymorphs, anatase appears to be the most important for these applications, as the majority of nanostructured materials display this specific structure,12,13 with the (101) face dominating more than 94% © 2012 American Chemical Society
Received: May 9, 2012 Revised: August 14, 2012 Published: August 14, 2012 18759
dx.doi.org/10.1021/jp3045143 | J. Phys. Chem. C 2012, 116, 18759−18767
The Journal of Physical Chemistry C
Article
using the generalized gradient approximation (GGA) of DFT proposed by Perdew, Burke, and Ernzerhof (PBE),36 and the electronic states were expanded using a plane wave basis set with a cutoff of 500 eV. Forces on the ions were calculated through the Hellmann−Feyman theorem, including the Harris−Foulkes correction to forces.37 This calculation of the force allows a geometry optimization using the conjugated gradient scheme. Iterative relaxation of the atomic positions was stopped when the forces on the atoms were less than 0.01 eV/Å. Full optimization of the cell parameters in some representative doped model structures was found to result in relative changes in cell volume similar in both sign and magnitude to the experimental data (∼0.8−0.9%). The corresponding changes in the computed relative energies were quite small, 0.03 eV in the biggest case. So, we have kept constant the cell parameters in these calculations given the small influence of the remaining stress in the relative energies. The anatase bulk lattice parameters are taken from previous work38 in which the same theoretical approximations of this work were used. The resulting parameters are a = 3.789 Å and c = 9.486 Å, which are within 0.2% of the experimental values (a = 3.782 Å and c = 9.502 Å).39 The anatase (101) surface models were taken from previous work and were designed so as to allow the simulation of various dopant and defect concentrations. The first model (model I, see Figure 1) is a
impurities in the TiO2 structure, either nonmetals or metals.15,16 Since the seminal work of Asahi et al.,17 N-doped TiO2 has become by far the most studied anion-doped TiO2 system. This is due to the observed lowering of the titania band gap that shifts the adsorption edge of the solid to the visible region, making it a promising photocatalytic material.18 It has been shown that band gap narrowing can take place due to the mixing of N 2p and O 2p states and/or the creation of additional localized N 2p states within the titania band gap.19,20 More recently, simultaneous N/metal TiO2 codoping has attracted considerable interest, since synergistic effects may result in peculiar properties compared to single element doping. In this regard, some studies have been reported and, in some cases, a narrowing of the band gap and improved photocatalytic activity in the visible region has been observed.21−26 To date there is not a clear picture of the enhanced photoactivity mechanism, mainly due to the lack of detailed microscopic information about the effects of codoping on the structure and electronic properties of TiO2. In particular, Kubacka et al. have synthesized microcrystalline W,N-codoped TiO2 anatase materials that showed both high activity and selectivity in the gas phase oxidation of aromatic hydrocarbons, using sunlight as excitation energy and air (e.g., molecular oxygen) as oxidant.23 The enhanced photoactivity of the resulting material was rationalized on the basis of a change in the surface properties related to local variations of the active center, band gap modification, and the maintenance of the structural ordering at the anatase structure.27,28 As mentioned above in a general way for codoped anatase materials, a complete structure−chemical function relationship is to be obtained. In this study we first address structural and electronic issues by examining the interaction between the W,N-codopants and the charge compensating point defect structures: Ti vacancies and wolframyl species, present in W-doped TiO2 materials.29,30 In particular, we focus on the anatase (101) surface because, as previously indicated, anatase is the main structural polymorph found in nanostructured TiO2 materials, and the (101) surface is the thermodynamically most stable crystal face. Thus, its structure is expected to play a fundamental role in any chemical or photochemical process that may take place on it. After briefly presenting the necessary computational details about the models and the methodology used, and the details of the experimental procedures, we will start by exploring the energetic and geometric details of the interaction between the W,N-codopants and the above-mentioned structural point defects. The structural characterization will be first done via density functional theory (DFT) calculation of different local arrangements of W,N impurities and defects. The insight obtained from these theoretical calculations will help to interpret the experimental N 1s and O 1s XPS and Raman spectra of the prepared solids. Later on the electronic structure of these systems will be explored with the help of density of states (DOS) and Bader charge analysis. These results will allow us to propose a mechanism for the enhanced photocatalytic activity and selectivity observed in partial oxidation reactions.23,27,28
Figure 1. A 1 × 2 supercell model of a TiO2 (101) anatase surface with three O−Ti−O layers. Atom labeling for the substitutional W positions: 5s, pentacoordinated surface site; 6s, hexacoordinated surface site; 6u, subsurface hexacoordinated (up) site; 6d, hexacoordinated subsurface (down) site. Atom colors: Ti, gray; O, red.
supercell, containing 72 atoms, of dimensions 1 × 2 in the [101̅] and [010] directions, respectively. In the [101] direction, the slab was three TiO2-trilayers thick (nine atomic layers), with a vacuum of 9 Å separating the slab images. The second model (model II) is obtained by doubling the previous one along the [101̅] direction. Thus, it is a 2 × 2 supercell three TiO2-trilayers thick. According to their size, structural calculations were performed in model I derived structures using a Monkhorst−Pack set of 2 × 2 × 1 k-points, while a 1 × 2 × 1Monkhorst−Pack set was used for model II derived configurations. In all cases, the atoms in the bottom layer were kept fixed to their bulk positions during geometry optimizations in order to simulate the presence of the bulk underneath. Density of states (DOS) calculations were performed using a denser 4 × 4 × 2 Monkhorst−Pack set of k-points.
2. EXPERIMENTAL SECTION 2.1. Model and Computational Details. Periodic threedimensional calculations were carried out using the VASP 4.6 code31−33 with the projector augmented wave method (PAW).34,35 In these calculations, the energy was computed 18760
dx.doi.org/10.1021/jp3045143 | J. Phys. Chem. C 2012, 116, 18759−18767
The Journal of Physical Chemistry C
Article
Figure 2. Supercell models of a W-doped TiO2 (101) anatase surface showing the defects introduced to achieve charge compensation: (a) model with two W atoms incorporating a VTi site (light blue sphere); (b) model with one W atom and an extra O included. Bond distances are in angstroms. Top: side view. Bottom: top view. Atom colors: Ti, gray; O, red; W, ochre.
centrifuged, decanted, rinsed with methanol, and dried at 20 °C for 12 h. Following the microemulsion preparation method, the solid precursors were subjected to a heating ramp (2 °C/min) in 8% (v/v) NH3/N2 up to 450 and 600 °C and treated at this temperature in 20% (v/v) O2/N2 for 2 h. Reference systems with the absence of N containing species were also synthesized from the same precursors and subjected to the same thermal treatment but always in the presence of oxygen (20% v/v O2/ N2). W/(Ti+W) and N/(N+O) composition was analyzed by using inductively coupled plasma and atomic absorption (ICP− AAS), while BET surface areas were measured by nitrogen physisorption (Micromeritics ASAP 2010). UV−visible diffuse reflectance spectroscopy experiments were performed with a Shimadzu UV2100 apparatus with a nominal resolution of ca. 1 nm using BaSO4 as reference. Band-gap values have been calculated from the UV−vis diffuse reflectance spectra shown in Figure S1 of the Supporting Information (SI). The chemical analysis and main physicochemical properties of the synthesized materials are summarized in Table S1 (see the SI). Micro-Raman measurements were performed using a LabRAM Jobin Yvon spectrometer equipped with a microscope. Laser radiation (λ = 532 nm) was used as excitation source at 5 mW. All measurements were recorded under the same conditions (1 s of integration time and 30 accumulations) using a 100× magnification objective and a 125 nm pinhole. XPS data were recorded on 4 × 4 mm2 pellets, 0.5 mm thick, prepared by slightly pressing the powered materials which were outgassed in the prechamber of the instrument at 150 °C up to a pressure