A Chemical Model of Phosphate Adsorption by Soils: II. Noncalcareous Soils

The Constant Capacitance model provided a quantitative description of o-phosphate adsorption by 44 noncalcareous soils whose pH values ranged from 4.9 to 7.6. The intrinsic surface protonationdissociation constants, capacitance density, and phosphate packing area parameters required by the model were adopted from model calculations on reference hydrous oxide minerals. The intrinsic phosphate surface complexation constants were calculated through the application of a nonlinear least squares fitting program to the soil o-phosphate adsorption data. Two of these intrinsic constants were found to be independent of pH over the range investigated, as required by the model. However, the intrinsic constant for the formation of the neutral o-phosphate surface species exhibited a statistically significant dependence on pH. The Constant Capacitance model was best able to describe o-phosphate adsorption by noncalcareous soils, including pH effects, if a soil-specific set of intrinsic phosphate surface complexation constants was employed. Additional Index Words: anion adsorption, ligand exchange, surface chemistry. Goldberg, S., and G. Sposito. 1984. A chemical model of phosphate adsorption by soils. II. Noncalcareous soils. Soil Sci. Soc. Am. J. 48:779-783. I Part I of this study (Goldberg and Sposito, 1984), the Constant Capacitance model (Stumm et al., 1980) was applied successfully to describe o-phosphate adsorption by aluminum and iron oxide minerals. This model is based on a ligand exchange mechanism of adsorption, which is considered to be appropriate for o-phosphate on hydrous oxide surfaces (Mott, 1981). Both the charge on the adsorbent surface and the charge on the adsorptive anion are considered in the model, which then is able to predict phosphate adsorption under changing conditions of pH value and o-phosphate concentration. The model produced good fits to experimental adsorption data (conventional isotherms and "adsorption envelopes") for both crystalline and amorphous aluminum and iron hydrous oxides. In this paper, the Constant Capacitance model is extended for the first time to o-phosphate adsorption by noncalcareous mineral soils. This application of the model is predicated on four basic assumptions: (i) The principal phosphate adsorption mechanism is ligand exchange with surface hydroxyl groups bound to metal cations (Berkheiser et al., 1980; Mott, 1981); (ii) The protonation-dissociation constants for the surface hydroxyl groups reacting with o-phosphate do not depend on the kind of metal (i.e., Al or Fe) to which the hydroxyl groups are bound (Goldberg and Sposito, 1984); (iii) The surface area occupied by an adsorbed o-phosphate ion may be represented by the average 1 Contribution from the Dep. of Soil and Environmental Sciences, Univ. of California, Riverside, CA 92521. Received 25 Oct. 1983. Approved 14 Mar. 1984. 2 Former Graduate Research Assistant and Professor of Soil Science, respectively. The senior author is presently Soil Scientist, U.S. Salinity Lab., 4500 Glenwood Drive, Riverside, CA 92501. value of 0.6 nm; (iv) The phosphate surface complexation constants do not depend on the kind of metal to which the reactive surface hydroxyl groups are bound. These four assumptions are essentially uniformity hypotheses concerning the reactions of o-phosphate with soils. Assumption (i) should be applicable to noncalcareous soils that do not contain large amounts of allophanic minerals. Soils containing allophane as a significant component are excluded, since phosphate sorption on these soils can involve a precipitation mechanism (Veith and Sposito, 1977). Calcareous soils are excluded because the mechanism of o-phosphate adsorption may involve ligand exchange with surface carbonate groups as well as surface hydroxyl groups. Assumptions (ii) and (iv) are special cases of the hypothesis that, to a first approximation, the protonation-dissociation and o-phosphate exchange reactions of surface hydroxyl groups in soils are independent of the bulk solid phase to which the hydroxyl groups are bound. In support of this assumption, Goldberg and Sposito (1984) found that the common logarithms of the intrinsic protonation-dissociation constants, K+ (int) and K(int), and of the phosphate surface complexation constants for aluminum and iron oxide minerals were not significantly different. Assumption (iii) simplifies the model calculation and will be discussed under Data and Methods.