In-situ photo assisted reduction of platinum particles supported on TiO2 followed by XAS characterization

International @ ECI2D+JCO:ILL:PAF:CGE:EPUInternational audience1IntroductionThe production of the energy is a huge challenge for our society. In this way, it appears necessary to find al-ternative to our dependence on fossil fuel. Solar energy seems to be a solution to provide sustainable energy. This one appears to be the most abundant energy source on earth. Photocatalysis is one possible way to use this kind of energy to satisfy our needs. In this regard, water splitting by powder-form photocatalysts has a lot of benefits, espe-cially H2 produced by photo-assisted process which should be a chemical energy vector.The main objective of this project is to develop photocatalyst materials by adding co-catalysts which permit the water splitting into H2 and O2 under visible light.In order to have a better knowledge of the water splitting reaction, the characterization of the nanoparticles of platinum supported on TiO2, as a reference, has been performed, by in-situ XAS during the photocatalytic reaction. Furthermore, the photoreduction of platinum on TiO2 has been studied to improve the understanding of the mechanism of the photodeposition.2ExperimentalThe experiments were carried out on the French beamline (FAME) of the European Synchroton Resarch Facility of Grenoble (France).The absorption spectra were obtained in a fluorescence mode (low metal loading) at the platinum L3-edge (11564 eV), previously calibrated with a sample of platinum foil. Two kinds of data have been recorded: XANES corresponding to X-ray Absorption Near Edge Structure and EXAFS for Extended X-ray Absorption Fine Structure. The spectra are recorded using a quick-XANES (about 3.5 minutes) and a mapping to avoid/limit the beam damage.The experimental setup is composed of three parts: (i) a system of control and regulation of the gaseous flow for purging the cell and the preparation of a flow saturated in methanol (3.8 kPa) at 273 K, (ii) the new in situ XAS cell (Fig.1) designed with 3 Kapton windows to work in transmission or fluorescence mode. The sample in pellet form is placed respecting an angle of 45° referenced to the x-rays. In front of the pellet are inserted optical fibers connected to a Xenon light source (possibility to change the wavelength and the photon flux) for the UV illumination. (iii) Finally, the composition of the gaseous flow was followed by a mass spectrometer (MS).On one hand, the oscillation part of EXAFS as a function of the X-ray photon energy was extracted as de-scribed elsewhere[2]. Analysis of the data has been done using the software IFFEFIT [3]. Phase shift and backscattering amplitudes were given by the reference compound: platinum foil for Pt-Pt (2.77Å) [4]. On the other hand, the XANES spectra was simulated using 2 functions: arctangent (constant) and Lorentzian (variable). The area of the lorentzian curve was adjusted to fit the experimental curve. This parameter is assumed to represent the variation of the number of electrons on “d” orbitals of platinum. To determinate this relationship, three references compounds have been analyzed: Pt0, Pt+II and Pt+IV.3Results and discussionThe X-ray absorption spectra were performed on different amount of H2PtCl6 (1% wt.) impregnated on TiO2 P25 (rutile – anatase).The first spectra on the different photocatalysts obtained under helium flow in the dark do not show any modification of the state of the platinum due to beam damage. The exploitation of EXAFS data highlights that the initial state of platinum is clearly dependent on the nature of the support and the crystalline structure of TiO2.Then, in the dark, a helium flow saturated in methanol (3.8 kPa) is introduced in the cell. The first oscillations are modified without any shift of the edge. The difference observed may be due to the variation of the nature of the ligand (-Cl to –OH).Finally, the UV is light on under methanol flow. The height of the white line of platinum decreases slightly as a function of irradiation time, and its position shifts to lower energy. After about one hour, a stationary state is ob-tained. The decrease of the photon flux leads to a decrease of the area of the Lorentzian curve. This modification is assumed to be a variation of the electronic density of platinum.To support the hypothesis of the modification of the electronic density, the oscillations of the EXAFS spectra were analyzed to evaluate the neighbourhood of the platinum. The parameters of a theorical curve: coordination number (CN), distance (R), Debye-Waller factor (s) and E0 correction are modified until the fit of the experimental curve in the R space between 1 and 3.5 Å. Whatever the photon flux,, the fitting curve was obtained using only Pt-Pt bond with a coordination number equal to 7.8 with a distance Pt-Pt equal to 2.75 Å. Whatever the conditions, the particles of platinum have the same size.Using the relationship between the area of the Lorentzian curve and the number of electrons in the layer “d”, the filling of the layer “d” of the platinum can be followed as a function of irradiation time (Fig 1.A). The final stationary state is also dependent on the photon flux of UV light used during the photoreduction (Fig 1.B). This innovative set up appears as powerful tool to better understand electron transfer during photocatalytic reaction.Fig 1 : Evolution of the number of electrons in layer “d” of Pt as a function of irradiation time (A) and photon flux (B)4ConclusionsIn-situ XAS characterization has been performed on TiO2 P25, impregnated with H2PtCl6 (1% wt. of Pt), dur-ing the photocatalytic reaction. Thank to this innovative technique the mechanism of the photoreduction of plati-num is better understood. A dependence of the initial state of platinum with TiO2 crystalline structure as been observed. Then, in presence of methanol, and under UV light, the reduction of platinum is observed leading to a final stationary state depending on the photon flux. Higher is the photon flux, lower is the area of the Lorentzian curve, meaning that the electronic density of platinum changes. A linear relationship has been established between the area of the Lorentzian curve and the number of “d” electrons using reference compounds to evaluate this point. This relationship allows us to evaluate the modification of the number of “d” electrons and to develop a new approach to calculate the number of electrons transferred from the semiconductor to the metal.AcknowledgementsWe would like to thank the King Abdullah University of Science and Technologies (KAUST) for their support (award n° UKC0017) on this project. We also thank Dr. Denis Testemale and Dr. Olivier Proux from FAME-ESRF for their help and for their technical advice during and after the EXAFS experiments.References1D.Yamasita et al., Solid State Ionics, 172 (2004) 591[2]O. Alexeev et al., J. Catal. 164, (1996) 1.[3][4]W.A.Spieker et al., Applied Catalysis A, 232 (2002) 219[5]Kip, B. J., et al., J. Catal. 105, (1987) 26