Ordered Oxygen Overlayer Associated with Chemisorption State on Al(111)

Abstract

We present experimental evidence of a unique, ordered chemisorption phase in the initial interaction of oxygen with the Al(lll) surface. At high oxygen exposure or high temperature, this phase is shown to transform irreversibly to a bulklike aluminum oxide. The measured temperature dependence, as well as the low-energy electron diffraction, suggests a threefold, centered bonding site. A comparison between calculated and experimental valence-band density of states for the oxygen-covered Al(lll) surface is made for the estimated oxygen-atom-substrate-surface distance.Experimental evidence is presented showing that the initial interaction of oxygen with the (111) crystal face of aluminum is a two-step process. Oxygen atoms are shown to chemisorb first at equivalent sites on the surface in an ordered overlayer. Upon increasing the oxygen exposure or the temperature above 170°C, the chemisorbed oxygen is irreversibly transformed into a bulklike oxide film. The chemisorption phase in the aluminum-oxygen interaction was first observed for "polycrystalline" films. 1 The present work shows that the two-step oxidation process is unique to the closepacked (111) face of aluminum and reveals for the first time the presence of an ordered overlayer with oxygen on a simple fee metalc The other two faces investigated, (100) and (110), form bulklike oxide films for the lowest observable coverages. 2 The existence of a well-defined chemisorption phase on a free-electron-like metal such as aluminum is of great interest since a number of theoretical calculations using different techniques treat chemisorption of oxygen atoms on aluminum as a model system. 3 " 5 Because of the relatively simple bulk electronic structure of this system, calculations have been made self-consistent and have been performed for different adsorbate-substrate distances. Some calculations explicitly assume a specific adsorption site, but comparisons of the calculated valence-band density of states with experimental valence-band photoemission spectra have been hindered by uncertainty about the actual Al-O configurations. In the present work we deduce a specific position and estimate an adsorbate-substrate distance for oxygen atoms on the (111) surface of aluminum. Valence-band spectra obtained for the oxygen-exposed surfaces are compared with a theoretical calculation. 3 ' 6 Discussion of the valence-band spectra for the clean faces is presented elsewhere. 7 The experiments reported were performed using the monochromatized radiation from the 4° beam line at Stanford Synchrotron Radiation Laboratory (SSRL) as the excitation source. Photons at two energies were used to excite electrons: 50-eV photons to excite the valence-band region and 130-eV photons to excite the 2p core of aluminum. Incoming light illuminated the sample at 5°-10° away from grazing incidence. The photoelectrons were energy analyzed in a doublepass, cylindrical mirror analyzer having its optical axis 5°-10° off the sample surface normal; the joint energy resolution (photon plus electron) was 0....

Interaction of oxygen with Al(111) at elevated temperatures

“…This contrasts very much the nonadiabatic kinetics at 300 K. 2 The thermal growth of the (1ϫ1)O ad islands at 350-530 K demonstrates that it is possible to anneal the surface covered with small amounts of O ad without direct, full transformation of the-metastable-chemisorbed oxygen into the thermodynamically stable oxide. This had not been found in previous studies in which, upon annealing, the electronic [12][13][14][15][16] and vibrational spectra 17,18 displayed directly the characteristics of aluminum oxide. This may be explained by the larger exposures ͑30-100 L͒ applied there ͑except in one HREELS study, 17 in which the exposure scale deviates, however, from the other studies also at 300 K 2 ͒.…”

“…Using x-ray photoemission spectroscopy ͑XPS͒, the O(1s) peak 12 was found to show similar changes as when the surface was exposed to larger amounts of O 2 Al(2p) state [13][14][15][16] showed a shift by 2.7 eV towards higher binding energy characteristic for the oxide ͑which can be clearly discriminated from the 1.4 eV shifted, chemisorbed state͒, and high-resolution electron energy loss spectroscopy ͑HREELS͒ 17,18 revealed the typical oxidic three peak spectrum. All of these data seem to indicate that upon annealing the chemisorbed oxygen layer just transforms into Al oxide.…”

Effect of temperature on the initial stages of oxidation of γ-Al4Cu9(110)

“…A gap of 2.5-3.9 eV between the metal aluminium and oxidised aluminium peaks is generally reported in the literature, whereas lower values (0.5-1.5 eV) are attributed to chemisorbed phases, intermediate oxides or very thin oxide films [11,[23][24][25][26][27][28][29][30][31][32]. Although most of the work in the literature is based on the study of the Al2p peak, Rotole and Sherwood have shown, by comparing a clean aluminium plate and ␣ alumina powder, that the differences between the Al2s and Al ox 2s peaks are very similar to those observed between the Al2p and Al ox 2p peaks [33,34].…”

An x-ray spectromicroscopic study of electromigration in patterned Al(Cu) lines

“…This shift is consistent with the Al ϩ3 state in the ␣ϪAl 2 O 3 . 9 On the other hand the Al 2p binding energy found at 76.4 eV at point O is shifted by 3.2 eV towards higher binding energy with respect to metallic Al. It was proposed that the presence of Cu promotes cracking of the oxide layer or induces an intrinsic loss of electrical resistance in the oxide layer itself.…”