Adsorption of molecular nitrogen on nickel. II. Comparison of photoemission forN2/Ni(100) to CO/Ni(100) and to theory

Abstract

Valence-level and core-level photoemission results are presented for ${\mathrm{N}}_2$ molecularly adsorbed on a Ni(100) surface at 77 K. Comparison is made to the Ni(100)/CO system. Differences are interpreted in terms of differences in chemisorption bonding. The data are in agreement with theoretical calculations (▵SCF Hartree-Fock) for NiCO and Ni${\mathrm{N}}_2$ cluster models, provided the possibility of Ni $4p\pi -2{\pi }^{*}$ (adsorbate) backbonding is allowed for in those calculations. Certain features are explainable without inclusion of such backbonding, but the agreement with absolute ionization potentials (IP's) for the valence levels is worse and the core-level satellite structure effects become unexplainable. We conclude that the data demonstrate the existence of Ni $4p\pi$ to adsorbate ${\pi }^{*}$ backbonding (stronger for CO) and also justifies the use of the cluster-model calculations to interpret adsorbate photoemission data. Implicit in these conclusions is the assignment of the main line N $1s$ (or C $1s$ and O $1s$) IP's to a core-hole state in which an electron has been transferred from the Ni valence level to the adsorbate $2{\pi }^{*}$ level (shake-down process).