Surface layer formation during lithium intercalation of anatase TiO2: A reflection mode EXAFS study 1
D. Lützenkirchen-Hecht , M. Wagemaker , P. Keil , U. Haake ,A. A. van Well and R. Frahm 1
Fachbereich C –Fachgruppe Physik, Bergische Universität Wuppertal, Gaußstr. 20, 42097 Wuppertal, Germany 2 Interfacultair Reactor Instituut, Mekelweg 15, 2629 JB Delft, The Netherlands
The electrochemical intercalation of Li in anatase TiO2 was investigated with reflection mode X-ray absorption spectroscopy at the Ti K-edge (4966.4 eV). The experiments were performed at beamline BW1 in a cell which enables the electrochemical processing of the samples which are sensitive towards oxidation by the air and humidity. The sample cell permits reflection mode, grazing incidence X-ray experiments after the controlled emersion of the electrodes from the electrolyte [1, 2]. Previous experiments have indicated that even Li containing electrolytes and Li intercalated electrodes can be processed inside the cell [1, 2]. It is well known that lithium intercalation in anatase TiO2 results in structural and electronic changes (see, e.g. ). Transmission mode X-ray absorption spectroscopy at the Ti K-edge as a function of the lithium content has been interpreted as a sum of two coexisting crystallographic phases (anatase and Lititanate), both in the case of chemically lithiated microcrystalline samples and electrochemically intercalated thin film samples . The changes in the near edge region of the Ti K-edge (XANES region) indicate that the charge-compensating electron of the Li-ion mainly occupies the crystal field split Ti 3d t2g orbitals positioned at the bottom of the conduction band . Surface sensitive reflection mode XAFS indicates that intercalation leads to a thin surface layer which is characterized by an effective Ti oxidation state close to 3+ . This can be seen in Fig. 1, where the experimental near edge spectra of a fully loaded anatase electrode are compared to model calculations using the Fresnel theory. For the calculations, a layered structure consisting of a 4 nm top layer of Ti2O3 on 25 nm Ti-titanate (Li0.5TiO2) and the 10 nm gold contact layer on the silicon substrate was assumed. As can be seen in this figure, a reasonable fit of the experimental data can be achieved in the XANES data for different angles and in the derivative spectra also, thus proving that the Ti4+ is reduced to a formal Ti3.5+ valence in the bulk of the intercalated electrode with a further reduction to Ti3+ in a near surface region of few nm thickness. It should be noted here that part of the Ti3+ layer is also evident during partial intercalation of the anatase electrode at lower potentials. Thus, the electronic structure of the electrolyte side of the intercalated layer seems to be different from the bulk of the intercalated anatase phase, i.e. different from Li-titanate Li0.5TiO2. 1.0
normalized derivative spectra
R0(E) - R(E) (normalized)
Θ=0.30° 0.6 0.4 Experiment Model calculation
0.2 0.0 4960
Energy / eV
0.6 0.4 0.2 Θ=0.30° 0.0 -0.2 Experiment Model calculation
-0.4 -0.6 4960
Energy / eV
Figure 1: Quasi in-situ investigations of the electrochemical intercalation of Li into anatase TiO2 electrodes. (a) Comparison of the reflectivity spectra obtained from the fully intercalated electrode (−−−) to a model calculation assuming a multilayered structure consisting of a top layer of 4 nm Ti2O3 on 25 nm Li-titanate Li0.5TiO2 on an underlying gold backing electrode (10 nm) on the Si substrate (− − −) for two incidence angles as indicated. (b) Derivative of the experimental data and the calculated spectrum for Θ=0.30°.
While the electronic structure of the intercalated anatase electrodes was probed by XANES experiments, the atomic short range order structure of the intercalated phase was derived from extended X-ray absorption spectroscopy. An example is presented in Fig. 2, where the magnitude of the Fourier-transform of a reflection mode EXAFS spectrum (Θ=0.30°) is compared to FT´s of reference samples measured in transmission. Obviously, the short range order structure of the intercalated electrode differs significantly from that of crystalline Ti2O3, but is very similar to that of the Li-titanate reference. Therefore, it is proposed that the additional layer presently found is a new Li-titanate like phase with a Ti oxidation state close to 3+ and a composition Li1TiO2, which is not observed for compositions with smaller Li insertion ratios x<0.5. It is interesting to compare these results with the intercalation at elevated temperatures by Macklin et al. , who reported a second constant potential plateau in the discharge curves (constant current setting) at x≈0.6, indicating the formation of a new phase with composition Li1TiO2. From the similarity of the discharge curve and based on X-ray diffraction at the composition x=0.7, Macklin et al. concluded that the material retained its original Li-titanate structure. In qualitative agreement with this finding, the present surface sensitive EXAFS indicates the presence of the Li1TiO2-phase, with an atomic short-range order configuration similar to Li-titanate. Here, the formation of this Li1TiO2 material is limited to a thin surface layer at room temperature, whereas at 120oC the whole material will be converted to composition x=1, as is illustrated by the results of Macklin et al. 2.0 intercalated anatase Θ=0.30° Ti2O3-reference 1.5
| FT (χ(k)*k ) | (normalized)
Figure 2: Comparison of the normalized magnitudes of Fourier transforms of the experimental transmission mode spectra of Ti2O3 and Li-titanate (Li0.5TiO2) to those calculated from a reflection mode spectrum of a fully intercalated anatase sample. Vertical bars indicate the peak positions corresponding to the first two coordination shells of Ti2O3 and Li-titanate. (The data are not phase-shift corrected. Hanning windows, krange for the FT: 2.5 Å-1 ≤ k ≤ 13.3 Å-1).
We like to thank HASYLAB for the support of our experiments. This work was supported by the IHP-Contract HPRI-CT-1999-00040 of the European Community.
References  D. Lützenkirchen-Hecht, M. Wagemaker, P. Keil, U. Haake and A. A. van Well, Hasylab Annual Report (2001) 947.  D. Lützenkirchen-Hecht, M. Wagemaker, P. Keil, A. A. van Well and R. Frahm, Surf. Sci. 538, 10 (2003).  V. Luca, T.L. Hanley, N.K. Roberts, R.W. Howe, Chem. Mater. 11, 2089 (1999).  M. Wagemaker, D. Lützenkirchen-Hecht, A. A. van Well, R. Frahm, J. Phys. Chem. B 108, 12456 (2004).  W.J. Maclin, R.J. Neat, Solid State Ionics 53-56, 694 (1992).