Chemical Modeling of Arsenate Adsorption on Aluminum and Iron Oxide Minerals

The constant capacitance model was used to describe arsenate adsorption on goethite, gibbsite, and amorphous Al hydroxide. Because the model assumes a ligand exchange mechanism it is appropriate for describing the specific adsorption of arsenate anions. Arsenate surface complexation constants were fitted to the experimental data using a nonlinear least squares optimization technique. The constant capacitance model was able to represent arsenate adsorption over the pH range 4.5 to 9 on each oxide using the same set of surface complexation constants. However, the values of the surface complexation constants varied with the mineral considered. The model was able to describe arsenate-phosphate competition on gibbsite quantitatively over the pH range 4.5 to 9 using the same set of anion surface complexation constants. A set of surface complexation constants obtained for one ternary system could be used to predict competitive phosphate and arsenate adsorption on the same material for other ternary systems containing different amounts of total anions in solution. Additional Index Words: phosphate adsorption, ligand exchange, surface chemistry, constant capacitance model, FITEQL. Goldberg, S. 1986. Chemical modeling of arsenate adsorption on aluminum and iron oxide minerals. Soil Sci. Soc. Am. J. 50:11541157. A is TOXIC to both plants and animals and can accumulate in agricultural soils. Previous researchers have studied the adsorption of As by soil materials (Jacobs et al., 1970; Livesey and Huang, 1981). Jacobs et al. (1970) suggested that As was adsorbed preferentially by the oxalate-extractable amorphous Al and Fe compounds in 24 Wisconsin soils. Livesey and Huang (1981) also found significant correlation between these materials and As adsorption maxima of four soils from Saskatchewan, Canada. Various researchers have previously investigated the adsorption of arsenate on Al oxides (Kingston et al., 1971; Anderson et al., 1976; Malotky and Anderson, 1 Contribution from the U. S. Salinity Laboratory, USDA-ARS, 4500 Glenwood Drive, Riverside, CA 92501. Received 21 Oct. 1985. 2 Soil Scientist. 1976; Anderson and Malotky, 1979) and Fe oxides (Kingston 1970; Kingston et al., 1971; Harrison and Berkheiser, 1982; Lumsdon et al., 1984). Arsenate adsorption on goethite, gibbsite, and amorphous Al hydroxide exhibited a maximum in the pH range 3 to 4 followed by a gradual decline with increasing pH (Kingston, 1970; Kingston et al., 1971; Anderson et al., 1976). Kingston et al. (1971) described their adsorption data using the Langmuir equation and obtained good fits. The mechanism of specific arsenate adsorption on Al and Fe oxides is considered to be ligand exchange with surface hydroxyls and/or surface aquo groups (Kingston et al., 1971; Anderson et al., 1976; Anderson and Malotky, 1979; Harrison and Berkheiser, 1982; Lumsdon et al., 1984). Direct evidence for ligand exchange of arsenate using infrared spectroscopy has been provided by Lumsdon et al. (1984) for goethite, and by Harrison and Berkheiser (1982) for hydrous Fe oxide. Specific anion adsorption produces a shift in the zero point of charge (ZPC) of the adsorbent. Such a shift was observed by Anderson et al. (1976) and by Anderson and Malotky (1979) for arsenate adsorption on amorphous Al hydroxide indicating that specific adsorption had taken place. The constant capacitance model (Stumm et al., 1970; Schindler and Gamsjager, 1972; Stumm et al., 1976; Stumm et al., 1980) describes adsorption using a ligand exchange mechanism. It is a chemical model that explicitly defines surface complexes and chemical reactions and includes the effect of the pH variable on adsorption. The model has been used previously to describe the specific adsorption of fluoride, phosphate, silicate, selenite, and borate anions by Al and Fe oxide minerals (Sigg and Stumm, 1981; Goldberg and Sposito, 1984a; Goldberg, 1985; Goldberg and Glaubig, 1985). Since arsenic acid is a triprotic acid whose dissociation constants are very similar to those of phosphoric acid and since both anions adsorb via ligand exchange it is reasonable to expect similar amounts of adsorption for both anions. Indeed, Kingston et al. (1971) observed a close correspondance in both the GOLDBERG: CHEMICAL MODELING OF ARSENATE ADSORPTION 1155 shape of the curve relating adsorption to pH and the amounts of these anions adsorbed on goethite. Livesey and Huang (1981) found that the presence of phosphate substantially reduced arsenate adsorption in their soil system. The ability of the constant capacitance model to predict competitive anion adsorption on goethite from aqueous solutions containing two anions using the surface complexation constants obtained from single anion systems has been tested previously (Goldberg, 1985). The model could describe phosphate-selenite and phosphate-silicate competition qualitatively by reproducing the shape but not the magnitude of the adsorption curves. Site heterogeneity was suggested to explain the inability of the model to describe these anion competition data. Chu and Sposito (1981) have shown that ternary exchange systems cannot in principle be predicted based solely on binary exchange data. The purpose of this study is to test the ability of the constant capacitance model to describe arsenate adsorption on Al and Fe oxide minerals. In addition, direct optimization of the ternary data will be carried out to test whether the model can describe arsenatephosphate competition in ternary systems.