Table of Contents

Reaction of Water with Well-Defined Iron Oxide Surfaces P. Liu,(a) E. Nelson,(a) T. Kendelewicz,(a,b) G. E. Brown, Jr.,(a,b) Y. J. Kim,(a,c) and S. A. Chambers Supported by OBER/OBES EMSP and NSF-MRSEC. (a) EMSL User. (b) Department of Geological and Environmental Sciences, Stanford University, and Stanford Synchrotron Radiation Laboratory. (c) Taejon National University of Technology, South Korea. Understanding contaminant-mineral interactions at a molecular level requires carefully controlled experiments that employ well-defined materials. The probes of modern surface science are powerful for elucidating structure, composition, and dynamics at interfaces, but most require surfaces that are uniquely defined over the length scale of the probe, which can be as large as a few millimeters. In order to accommodate this requirement, single crystals on well-defined structures and composition are needed. We have developed and utilized powerful thin-film growth techniques designed to enable a wide range of oxides and minerals to be grown to produce such surfaces with a great deal of control. Our growth technique is oxygen-plasma-assisted molecular beam epitaxy, and we have used it to grow all of the stable phases of the iron oxides with several crystal orientations (Kim et al. 1997; Gao et al. 1997, 1998; Gao and Chambers 1997). The iron oxides and oxyhydroxides constitute important redox-active secondary mineral phases in a number of geologic formations. Having these materials, which are not available as pure, bulk crystals, allows us to examine their heterogeneous chemistry with important contaminants at an unprecedented level of molecular detail. The first step in this process is to understand the interaction of the iron oxides with pure water. In order to do so, we transport single-crystal iron oxide films grown at PNNL to Stanford Synchrotron Radiation Laboratory (SSRL) for water dosing and photoemission spectroscopy experiments. Here, we expose clean iron oxide surfaces to water vapor at various pressures for a fixed amount of time and measure surface-sensitive O1s core-level spectra to see how much water sorbs, and what chemical state the water is in on the surface. The presence of molecular water on oxide surfaces produces an O1s peak shifted to higher binding energy from the lattice oxygen peak by ~3 eV, whereas water that has dissociatively chemisorbed to produce OH shows a chemical shift of +1-2 eV relative to lattice oxygen. We generally find that water is dissociatively chemisorbed at room temperature on these surfaces. By measuring the amount of O1s OH relative to lattice oxygen as a function of water exposure pressure, we can gain insight into the mechanism of hydroxylation. One such curve is shown in Figure 2.6 for water on a -Fe2O3(0001) (Liu et al. 1998a). There is rather low uptake until a pressure of ~10-4 torr is reached, at which point the amount of sorbed OH increases many fold. The amount of OH on the surface below 10-4 torr is constant with pressure and is approximately equal to the step density on the surface (~0.1 monolayer). The observed behavior leads us to conclude that hydroxylation occurs in two distinct regimes: 1) at steps (P < ~1x10-4 torr), and 2) at terraces (P ³ ~1x10-4 torr). P = ~1x10-4 torr appears to constitute a threshold pressure, below which hydroxylation of terraces is thermodynamically forbidden. Steps, being electronically and chemically more active than terraces, are hydroxylated at lower pressures. Interestingly, exposure of this and other iron oxide surfaces to pressures below the threshold value for much longer periods of time, such that the (pressure)x(time) product exceeds that which results in terrace hydroxylation above the threshold value, does not result in terrace hydroxylation. This key result strongly suggests that the process is thermodynamically rather than kinetically controlled. Application of a simple thermodynamic model to the process allows the threshold pressure for hydroxylation of terraces to be predicted. The only input is the standard free energy change for the hydroxylation process. However, the end hydroxylation product and its standard free energy of formation must be known in order to use this model. Unfortunately, the most well understood and seemingly likely products, FeOOH and Fe(OH)3, do not produce satisfactory agreement with experiment. This result suggests that another less well understood reaction product, such as Fe2O3•2FeOOH•2.6H2O (ferrihydrite) may be forming. Application of this simple thermodynamic model to less complex oxides, such as MgO(001), CaO(001), and a -Al2O3(0001), produces very good agreement with experimental data (Liu et al. 1998a, 1998b, 1998c). We thus have confidence in the model, but its application to the hydroxylation of iron oxides must await better thermodynamic data on the less common oxyhydroxides of iron. Figure 2.6. Photoemission intensity ratio for OH O1s relative to lattice O1s vs. water exposure pressure for a -Fe2O3(0001). References Gao., Y., and S. A. Chambers, J. Crystal Growth 174, 446 (1997). Gao, Y., Y. J. Kim, and S. A. Chambers, J. Mat. Res. 13, 2003 (1998). Gao, Y., Y. J. Kim, G. Bai, and S. A. Chambers, J. Vac. Sci. Technol. A 15, 332 (1997). Kim, Y. J., Y. Gao, and S. A. Chambers, Surf. Sci. 371, 358 (1997). Liu, P., T. Kendelewicz, G. E. Brown, Jr., and G. A. Parks, Surf. Sci. 412/413, 287 (1998b). Liu, P., T. Kendelewicz, G. E. Brown, Jr., and G. A. Parks, Surf. Sci. 416, 326 (1998c). Liu, P., T. Kendelewicz, G. E. Brown, Jr., E. J. Nelson, and S. A. Chambers, Surf. Sci. 417, 53 (1998a). William R. Wiley Environmental Molecular Sciences Laboratory Feedback: webmaster@emsl.pnl.gov Revised: June 12, 2001 Security & Privacy PNNL-13147