L560 Surface Science 109 (1981) L560-L566 North-Holland Publishing Company SURFACE SCIENCE LETTERS PRESSURE DEPENDENT OXIDATION OF AI(I 11): A PHOTOEMISSION AND SURFACE EXAFS STUDY at R.Z. BACHRACH, G.V. HANSSON and R.S. BAUER Xerox Palo Alto Research Center, 3333 Coyote Hill Road, Palo Alto, California 94304, USA Received 30 April 1981 ; accepted for publication 5 June 1981 The initial oxidation of AI(111) has been studied with photoemission and surface EXAFS for single crystal surfaces. We find that the oxidation is pressure dependent and that at pressures below 2 X 10-7 Torr molecular oxygen chemisorbs while at pressures above 1 X 10-6 Torr the chemisorption is disassociative. These results are discussed in the context of other LEED and surface EXAFS studies. Significant interest exists in understanding the microscopic details of the initial oxidation of aluminum metal. The AI(111) face has been a particular focus since the discovery of an ordered room temperature oxygen chemisorption phase [I]. The recent studies performed with LEED [2] and surface EXAFS [3] have height- ened this interest since contradictory structural results were obtained although it appeared that AI(111) :O might represent a simple system for applying and com- paring these two techniques. In this work, the results of a photoemission and surface EXAFS study of the oxidation of Al(11 I) are presented which show for the first time that the initial room temperature oxidation is pressure dependent and that at low pressure (below 2 × 10 -7 Torr) molecular oxygen chemisorbs. Previous studies have concluded that the initial oxidation is always disassociative [1-7]. The new insight provided by this work suggests a resolution of the current structural controversy between the LEED [2] and surface EXAFS [3] measurements of the A1-O bond length. While both studies used passive doses of molecular oxygen, Johansson and Stohr [3] had in fact exposed their samples at 1 × 10 -6 Tort in the surface EXAFS study while Martinsson et al. [2] used 2 X 16 -7 Torr in their LEED study. The LEED study did not employ any subsidiary characterization, but rather depended upon previous photoemission work [ 1]. As shown here, the two experiments were not investigating the same phases. This possibility was conjectured recently by Jona and Marcus in retrospectively consider- ing possible problems with past LEED data that could produce such widely differ- ent results [8]. Although we find a bond length close to that of Martinsson et al. for the low pressure phase, our analysis does not find as short a distance obtained 0039-6028/81/0000--0000/$02.50 © 1981 North-Holland