The valence-band structure of nickel aluminum was measured by use of angle-resolved photoemission with synchrotron radiation and calculated using the local-density approximation. The overall agreement between theory and experiment is remarkably goodmuch better than for pure nickel. This means that the "self-energy" corrections are significantly less in NiA1 than in pure nickel. The core-level binding energies in NiAl are compared experimentally and theoretically with the equivalent levels in Ni and Al. Surprisingly, the Ni core shifts to higher binding energy and the Al to lower as if charge were transferred from Ni to Alopposite to the direction predicted from electronegativity. These observations are discussed in terms of bonding in NiA1. 42 1582 1990 The American Physical Society 42 ELECTRONIC STRUCTURE OF NiAl 1583 from Ni to Al), leading to a net transfer from nickel to aluminum.This is consistent with the observed corelevel shifts, which are in the direction of larger binding energy for the nickel and smaller binding energy for the aluminum. Shifts to the lower binding energy are usually attributed to an increase in electron number at a site which increases the Coulomb repulsion with the cores and reduces the binding energy.

No abstract

No abstract

No abstract

There are two possible terminations for the ideal NiAl(111) surface, i.e. , all Al or all Ni on the surface atomic layer. We investigate which termination occurs on the NiAl(111) surface by comparing the pseudopotential electronic band structure of a Ni-terminated and an Al-terminated NiAl(111) surface with the angle-resolved photoemission data on a NiAl(111) sample. The measured surface band structure shares common features partly with that of the Ni termination and also partly with that of the Al termination, which supports the theory that the real NiAl(111) surface is composed of both Ni-and Al-terminated (111)domains, as suggested by a recent low-energy electron diffraction (LEED) study. We also determine the relaxation of the two outermost atomic layers for both terminations by the pseudopotential total-energy calculations and compare them with the LEED results. We 6nd that the present results are in good agreement with the LEED analysis for Ni termination, but in qualitative disagreement for Al termination.