Surface Complexation Modeling of Antimonate Adsorption By Variable-Charge Minerals.

The adsorption of antimonate [Sb(V)] by gibbsite, kaolinite, goethite, and birnessite was examined as a function of pH and ionic strength (0.01 and 0.1 M KNO₃), and in the presence of phosphate and sulfate. The experimental findings suggest that the retention of Sb(V) by kaolinite and gibbsite occurs via a combination of mechanisms. Electrostatic adsorption occurs throughout the pH 3 to 10 range; whereas, covalent bonding by the aluminol functional group becomes important in pH < 6 suspensions. The retention of antimonate by goethite and birnessite occurs predominately by covalent mechanisms. Antimonate adsorption by the mineral surfaces was successfully predicted by any one of several triple-layer surface complexation models, including outer-sphere, monodentate or bidentate inner-sphere, or combined outer-sphere and inner-sphere (monodentate or bidentate) models. Thus, there was no unique set of reactions that would specifically describe Sb(V) adsorption by a given surface. In general, the surface complexation models that generated the lowest goodness-of-fit parameters included both outer-sphere [≡SOH₂⁺‒Sb(OH)₆^‒] and monodenate inner-sphere [≡SOSb(OH)₅^‒] complexation reactions. The magnitude of the intrinsic constants generally reflects the observed capacities of the minerals to adsorb Sb(V). The intrinsic constants for ≡SOH₂⁺‒Sb(OH)₆^‒ and ≡SOSb(OH)₅^‒ formation (as K^int) on gibbsite and kaolinite, which have a low capacity for Sb(V) retention, are orders of magnitude lower than the constants for complex formation on goethite and birnessite, which effectively immobilize Sb(V). The ability of the surface complexation models to predict Sb(V) adsorption in competitive Sb(OH)₆ and SO₄ or Sb(OH)₆ and PO₄ systems, using the surface complexation constants obtained for the single-ligand systems, was also evaluated. In general, the reoptimization of the surface complexation models was required to satisfactorily predict ligand adsorption in the competitive systems. For the gibbsite systems, the single-ligand models provided satisfactory descriptions of Sb(V) adsorption in both the competitive SO₄ and PO₄ systems. The reoptimization of the intrinsic constants for the mixed ligand systems resulted in an improved prediction of Sb(V) adsorption in the competitive PO₄ systems, but not in the SO₄ systems. For both the kaolinite and goethite systems, the single-ligand models did not adequately predict Sb(V) adsorption in the competitive SO₄ or PO₄ systems. Reoptimization resulted in the satisfactory prediction of Sb(V) adsorption in the SO₄ systems, but not in the PO₄ systems. The satisfactory description of Sb(V) adsorption by birnessite in the competitive PO₄ systems also required reoptimization of the intrinsic constants.