Single crystal antiferromagnetic USb2 was studied at 15K by angle-resolved photoemission with an overall energy resolution of 24 meV. The measurements unambiguously show the dispersion of extremely narrow bands situated near the Fermi level. The peak at th
situated at higher binding energies than for photon energy 34 eV and 43 eV. Its energy position varies from 575 meV (θ = 00) to 337 meV (θ = 100).
The position and intensity of peak B is different for the three photon energies investigated in Figures 2a, 3 and 4. This partially derives from the fact that we probe different parts of the Brillouin zone. The set of spectra shown in Fig. 2 are taken near the Z point in the Brillouin zone, whereas those in Fig.3 and 4 are both taken close to the Γpoint. In Fig.4 the peak is more pronounced that in Fig.3 and also appears at a higher binding energy, even though at both these photon energies we probe the vicinity of the Γpoint. We believe that this is a matrix element effect which has stronger influence on the spectra taken for hν = 43 eV. In the case of hν = 60 eV the final states are more free-electron like and hence the photoemission spectra are less influenced by matrix element effects.
Photoemission is a surface-sensitive experiment. Therefore there always exists a question as to whether features observed in the valence band derive from the bulk electronic structure or the surface electronic structure, which may differ considerably. We investigated the surface-bulk problem by means of a controlled surface termination experiment (see Fig. 5). We cleaved a USb2 sample and exposed it to up to 1x10-8 Torr of argon for 20 seconds and observed changes in the valence band by taking EDCs at a photon energy of 22eV. The effect was mostly attenuation of the valence band features. After exposure for 100 seconds (p=10-8 Torr of Ar) the photoemission from the valence band has almost disappeared, and the Fermi level peak disappears at the same rate. We carried out the surface termination and inert absorbate experiment for multiple angles to see if the changes are fundamental or are a result of k scattering. We observed that the changes are similar for different angles i.e. in the investigated part of the Brillouin zone the strong Fermi level peak is not surface related. We noticed that at 22 eV, photon exposure was causing the sample to revert to the non-absorbate surface. Therefore we warmed the sample to 40K. We observed that the Ar3p peak disappeared and the valence band returned to an attenuated version of the baseline spectra. The controlled absorbate experiment showed that the shape of the valence band photoemission spectra near the
8