Journal article

Arsenic sequestration by sorption processes in high-iron sediments

High-iron sediments in North Haiwee Reservoir (Olancha, CA), resulting from water treatment for removal of elevated dissolved arsenic in the Los Angeles Aqueduct system, were studied to examine arsenic partitioning between solid phases and porewaters undergoing shallow burial. To reduce arsenic in drinking water supplies, ferric chloride and a cationic polymer coagulant are added to the aqueduct upstream of Haiwee Reservoir, forming an iron-rich floc that scavenges arsenic from the water. Analysis by synchrotron X-ray absorption spectroscopy (XAS) showed that the aqueduct precipitate is an amorphous hydrous ferric oxide (HFO) similar to ferrihydrite, and that arsenic is associated with the floc as adsorbed and/or coprecipitated As(V). Arsenic-rich floc and sediments are deposited along the inlet channel as aqueduct waters enter the reservoir. Sediment core samples were collected in two consecutive years from the edge of the reservoir along the inlet channel using 30- or 90-cm push cores. Cores were analyzed for total and extractable arsenic and iron concentrations. Arsenic and iron speciation and mineralogy in sediments were examined at selected depths by synchrotron XAS and X-ray diffraction (XRD). Sediment-porewater measurements were made adjacent to the core sample sites using polyacrylamide gel probe samplers. Results showed that sediment As(V) is reduced to As(III) in all cores at or near the sediment-water interface (0-4 cm), and only As(III) was observed in deeper sediments. Analyses of EXAFS spectra indicated that arsenic is present in the sediments mostly as a bidentate-binuclear, inner-sphere sorption complex with local atomic geometries similar to those found in laboratory studies. Below about 10 cm depth, XAS indicated that the HFO floc had been reduced to a mixed Fe(II, III) solid with a local structure similar to that of synthetic green rust (GR) but with a slightly contracted average interatomic Fe-Fe distance in the hydroxide layer. There was no evidence from XRD for the formation of a crystalline GR phase. The release of dissolved iron (presumably Fe2+) and arsenic to solution, as monitored by in situ gel probes, was variable but, in general, occurred at greater depths than arsenic reduction in the sediments by spectroscopic observations and appears to be near or below the depth at which sediment GR was identified. These data point to reductive dissolution of the sorbent iron phase as the primary mechanism of release of sorbed arsenic to solution. © 2007 Elsevier Ltd. All rights reserved.


Related material