612 Conference Proceedings J. Synchrotron Rad. (1999). 6, 612-614 Identification of Cr species at the aqueous solution-hematite interface after Cr(VI)-Cr(III) reduction using GI-XAFS and Cr L-edge NEXAFS Daniel Grolimund,'" Thomas P. Trainor, a Jeffrey P. Fitts," Tom Kendelewicz," Ping Liu," Scott A. Chambers ° and Gordon E. Brown, Jr. "b aStanford University, Geological & Environmental Sciences, Stanford, CA 94305-2115, USA, bStanford Synchrotron Radiation Laboratory, SLAC, Stanford, CA 94309, USA, CPacific Northwest National Laboratory, EMSL, Richland, WA 99352, USA. Email:daniel@pangea.stanford. edu Surface-mediated redox processes and sorption reactions occurring at the solid-liquid interface are of importance to a broad variety of environmental and industrial pollution or corrosion/passivation problems. A more fundamental under- standing of the mechanisms of these reactions is required in order to increase our ability to predict interfacial phenomena. In this study, gazing-incidence XAFS (GI-XAFS) and near-edge XAFS (both L- and K-edge NEXAFS) techniques were used to inves- tigate the chemical and structural characteristics of interfacial Cr(III) and Cr(VI) species on a molecular level. In particular, products of postulated electron transfer reactions between Cr(VI) and Fe(II) at the oxide-solution interface were investigated. Both techniques confirmed the reduction of Cr(VI) to Cr(III) mediated by a partially reduced hematite (O001) surface. The observed local structure of the interracial chromium species emphasizes the need for an improved, molecular-level conceptualization of reactions occurring at the solid-solution interface. Keywords: Grazing-incidence XAFS, molecular adsorption, electron transfer reaction (redox), chromium, single crystal hematite, hematite-aqueous solution interface 1. Introduction In natural systems, the mobility as well as the potential hazard 6f contaminants is controlled to a notable extent by their affinity for solid phases and, in the case of redox sensitive pollutants, by their oxidation state (Stumm & Morgan, 1996). Consequently, physicochemical reactions occurring at the solid-liquid interface, such as surface complexation or electron transfer reactions, are of central importance to a broad variety of environmentally relevant phenomena including the dispersal of pollutants in the subsurface zone, residence times of contaminants in surface waters, waste water treatment, and bioavailability of plant nutrients. In addition, sorption and surface redox processes are important in a number of industrial areas, including heterogeneous catalysis and corrosion/passivation of metal surfaces. The identification of reaction products is required to develop a molecular-level understanding of the mechanisms of such interfacial reactions. Resulting concepts would lead to an improved ability to predict the reactivity of different interfacial systems. In this study we have used Cr K-edge GI-XAFS and Cr L-edge NEXAFS spectroscopy to investigate the types of surfaces complexes formed when aqueous Cr(VI) and Cr(IIl) react, respectively, with reduced (containing Fe(II)) and unreduced single crystal (0001) hematite surfaces. In the 6 ÷ form, aqueous Cr is both mobile and toxic to organisms, whereas in the 3 ÷ form, Cr is relatively insoluble and is less mobile and less toxic than Cr(VI). Thus, a knowledge of the speciation of Cr and the processes responsible for its transformation from Cr(VI) to Cr(lll) is essential for predicting its potential impact as an environmental pollutant. A number of previous studies have shown that Cr(VI) is readily reduced in natural environments by interactions with minerals containing Fe(II) (e.g., Peterson et al., 1997a). However, it has also been shown in model systems that the surfaces of Fe(II)-containing oxides and hydroxides can become passivated as a result of Fe(III) oxyhydroxide formation (e.g., Peterson et al., 1997b; Kendelewicz et al., 1998). In this paper, we extend this earlier work to reduced hematite surfaces and examine the reaction of aqueous Cr(VI) with synthetic, near atomically smooth (0001) single crystal hematite surfaces. Our main objective is to provide additional insights about the mechanism of the Cr(VI) -)' Cr(III) redox reaction, including the nature of reaction products formed at the hematite surface. 2. Materials and methods Thin (-350 ]k) synthetic single crystal hematite (~-Fe203) samples were grown on oriented sapphire substrates by multiple beam epitaxy, resulting in atomically smooth (2-5 Arms roughness) hematite (0001) surfaces (Kim et al., 1997). Partially reduced surfaces were produced by annealing at -500°C in vacuum. The partially reduced hematite was exposed to a Cr(VI)- containing solution (5raM Na2CrO4, pH 6.0, 0.1M NaNO3) in a N2 atmosphere and immediately transferred to a UHV chamber for Cr L-edge NEXAFS data collection. Subsequently, the same sample was investigated using GI-XAFS under ambient conditions. In addition, unreduced thin-film hematite was exposed to a Cr(III) solution (10 .5 M Cr(NO3)3, pH 4.8) in a N2 atmosphere and analyzed by GI-XAFS under ambient conditions. Experimental conditions (e.g., pH, total Cr concentration) were chosen to avoid the formation of multinuclear Cr complexes or supersaturation of Cr species in solution with respect to known hydroxides, carbonates, or basic salts (e.g., Smith & Martell, 1976). X-ray photoelectron spectroscopy (XPS) was used to check for surface cleanliness prior to the experiment and to estimate surface coverage after reaction. In addition to Fe and O, only adventitious carbon (<30 atoms/nm 2) was detected. The observed Cr surface coverage was below 0.1 monolayer for the unreduced hematite while the partially reduced hematite sample yield a coverage of -0.3 monolayer. GI-XAFS/NEXAFS experiments were performed at SSRL (3GeV and 60-100mA) on bean'dines 6-2 and 10-1. GI-XAFS data were collected using the SSRL grazing-incidence apparatus in the specular geometry with the incident angle set slightly below the critical angle of the hematite substrate at -6 keV (- 0.2 o). GI-XAFS data analysis was pertbrmed using EXAFSPAK (George & Pickering, 1995), and k3-weighted EXAFS were fit over a k-range of approximately 3-11 ,~~. Phase and amplitude functions were calculated with FEFF7 (Rehr et at., 1992) and verified by comparison with model mineral compounds. The accuracy of the optimized parameters can be estimated based on fits of crystalline model compounds (first shell: N_+15% and R__0.03A; more distant shells: N_+30% and R_+0.07]k). The Debye-Waller term (~2) tor each shell was estimated based on fits of both model compounds and single- (~ 1999 International Union of Crystallography Printed in Great Britain - all rights reserved Journal of Synchrotron Radiation ISSN 0909-0495 (~ 1999