Electron-energy-loss-spectroscopy study of oxygen chemisorption and initial oxidation of Fe(100)

Abstract

The initial stages of the interaction of oxygen with an Fe(100) surface have been investigated at 300 K mainly by electron-energy-loss spectroscopy (EELS) with in situ combined low-energy-electron-diffraction, Auger-electron and secondary-electron-emission spectroscopy, and workfunction---change measurements ($\ensuremath{\Delta}\ensuremath{\varphi}$). From all the results, three different stages of the oxygen interaction are distinguished: (i) dissociative chemisorption stage up to \ensuremath{\sim}3L, (ii) incorporation of O adatoms into the selvedge between 3 and 20 L, and (iii) oxidation above 20 L, leading to the formation of $\ensuremath{\gamma}$-${\mathrm{Fe}}_{2}$${\mathrm{O}}_{3}$. [1 langmuir (L)\ensuremath{\equiv}${10}^{\ensuremath{-}6}$ Torr sec.] Of special interest is the change in surface electronic properties from that characteristic of the metal to that of the oxide observed by EELS. An energy-loss peak characteristic of the chemisorbed oxygen was observed at 6 eV below 3 L, being ascribed to the transition between the bonding and antibonding orbitals, and the EELS spectrum in the oxide phase was characterized by the peaks due to the ${\mathrm{O}}^{2\ensuremath{-}} 2p\ensuremath{\rightarrow}{\mathrm{Fe}}^{3+} 3d$ charge-transfer transitions. The $\ensuremath{\gamma}$-${\mathrm{Fe}}_{2}$${\mathrm{O}}_{3}$-FeO phase transition at \ensuremath{\sim}570\ifmmode^\circ\else\textdegree\fi{}C was also confirmed by monitoring an energy-loss peak due to the $d\ensuremath{\rightarrow}d$ transition, which is spin allowed for an ${\mathrm{Fe}}^{2+}$ ion in FeO. It is shown that the Fe $3{d}_{yz,zx}$ electrons play a major role in the chemisorption bond (O adatoms located in the centered hollow or bridge site), and for the incorporation process the Fe $3{d}_{{z}^{2}}$ electrons are also involved in bonding by the symmetry breaking. A direct-recombination process following the $3p\ensuremath{\rightarrow}3d$ transition is proposed for the high-energy-side peak of the Fe ${\mathrm{M}}_{2,3}\mathrm{VV}$ Auger spectrum, which---though not yet unambiguously explained---has been tentatively assigned to an Auger transition from a doubly ionized Fe core level.