Band structure of reconstructed Cu2O(-110)
N. Barrett, L. Zagonel, J. Leroy (CEA/SPCSI)
F. Bruneval, N. Vast, L. Reining (LSI, Ecole Polytechnique)
M. Izquierdo, C. Chauvet, F. Sirotti (Tempo Beamline, SOLEIL Synchrotron)
The objective of the project was to measure the band dispersion of the Cu2O(-110) and Cu2O(111) in order to disentangle the hybridisation band s-d that theory predicts to be situated between the Cu 3d and the O 2p bands1.
We present first results on the band dispersion measured for the first time on a reconstructed Cu2O (-110) surface, allowing, in principle, direct access to the hybridized band.
Recent angle resolved photoemission measurements have claimed to have identified such a hybridized band. In order to identify a possible 3d-4sp band, not visible in the raw data, the authors had to apply normalisation procedures which may also produce artefacts2.
Our LEED patterns suggest a (nearly) stoichometric complex reconstruction of the Cu2O(-110) surface (Fig. 1a)
The MΓM dispersion is mapped (see Fig. 1b), showing some evidence for a band dispersing between the O 2p (-6 eV) and Cu 3d (-4 eV) orbitals. Following Önsten et al this becomes unambiguous, however, not all cross-section effects are necessarily accounted for by the normalisation procedure. Using linearly polarized light (hν=60 eV) at an incidence angle of 45°, we have measured the band dispersion along the [110] and [111] high symmetry directions in the (-110) surface. In the [110] direction, some intensity (at KE ~ 56 eV) is observed even in the raw data (Fig. 1c). However, in the [111] direction, where the polarization should exalt hybridisation along the Cu-O-Cu chains, no evidence of 3d-4sp hybridization is found (Fig. 1d).
Figure 1:
a) Atomic positions of Cu (green triangles) and O (green circles) in top view of the Cu2O(-110) surface;
b) Raw (top) and normalized (bottom) dispersion along MΓM of Cu2O(-110);
c) raw and normalized dispersion along [110] showing evidence for the 3d-4sp hybridization (see KE ~ 56 eV and θ ~ 4°);
d) wide angle band dispersion around normal emission in the [111] direction showing the absence of photoelectron intensity between the Cu 3d and O 2p bands. The vertical blue line indicates normal emission.
In both c) and d) the O 2p band is split around normal emission.
In both the [110] and [111] directions, there is clear evidence for a splitting in the O 2p band around normal emission, which is not predicted by theory.
Further work is in progress using laboratory UPS and XPS to quantify the chemistry of the reconstruction, to allow a robust analysis of the 3d-4sp hybridisation and to further probe the origin of the apparent splitting in the O 2p band. It is already clear that Cu2O, considered for a long time as a model oxide, has an extremely complex and rich band structure.
1F. Bruneval, N. Vast, L. Reining, M. Izquierdo, F. Sirotti, N. Barrett, Phys. Rev. Lett. 97 267601 (2006)
2A. Önsten, M. Mansson, T. Claesson, T. Muro, T. Matsushita, T. Nakamura, T. Kinoshita, U.O. Karlsson, O. Tjernberg Phys. Rev. B 76 115127 (2007)
