Arsenic speciation in multiple metal environments: I. Bulk-XAFS spectroscopy of model and mixed compounds.

X-ray absorption fine structure (XAFS) spectroscopy was employed to determine the bonding environment of As(V) in the presence of Cu(II) and Zn(II) on goethite and gibbsite. In addition, several mineral species and precipitates derived from homogeneous and heterogeneous (presence of alpha-Cr(2)O(3)) super-saturations were studied. Structural parameters were determined after resolving the broad second shells in r-space by differential k-weighting (1, 2 or 3) and k-ranging (2.5- vs 3.5-12.75 A) of the raw EXAFS functions. In precipitates, AsO(4) was incorporated in the metal-hydroxides forming clinoclase-like and koettigite-like structures in the presence of Cu(II) and Zn(II), respectively. In the presence of both Cu(II) and Zn(II), the clinoclase structure formed preferentially over the koettigite structure under homogeneous oversaturated solution conditions and in the presence of eskolaite (alpha-Cr(2)O(3)). Silica promoted the formation of koettigite-like zinc-arsenate precipitates from initial As(V) and Zn(II) solution concentrations of 500 muM. On goethite and gibbsite, 750 muM As(V) formed mainly bidentate binuclear surface species in accordance with many previous findings even in the presence of equimolar Cu(II) and or Zn(II) concentrations. Copper was readily identified in the second shell environment of As(V) sorbed on gibbsite, but not on goethite. We hypothesize that this complex formed on the basis of Cu(II)'s ability to form polymeric species in solution and at the mineral-water interface in agreement with previous studies. The effects of Zn(II) on the coordination environment of As(V) on gibbsite and goethite could not be ascertained with As K-edge EXAFS spectroscopy. In addition to bidentate binuclear surface complexes, As(V) formed edge-sharing complexes with Fe, Al, and Cu atoms, which we could differentiate on the basis of the inter-atomic distances, phase shifts between wavefunctions of Fourier-filtered peaks, and differences in amplitude of the absorption envelopes. The analyses showed that of all data reduction steps, data presented in r-space and as wavefunctions of Fourier-filtered shells offer the greatest possibility for fingerprinting and inferring the influence of co-sorbing metal cations on the As(V) sorption complex. With regards to interpretations of micro-EXAFS data by abstract factor analyses and linear least-square combination fitting, analyses of As K-edge data should not be performed on the raw chi(k) data, but rather on consistently isolated second and higher-order shell features.