RELEASE KINETICS OF SURFACE-ASSOCIATED MN AND NI IN SERPENTINITIC SOILS - PH EFFECTS

Serpentinitic soils have unique chemical properties. While relatively little is known about element release kinetics and exchangeable Mn concentrations of such soils, exchangeable Ni concentrations commonly indicate potential plant toxicity. This study investigated Mn and Ni release kinetics in three serpentinitic soils with properties representative of many serpentinitic soils. The study was conducted at near ambient pH values in 0.1M ammonium acetate-based solutions using batch (medium-term experiments: 2 to 112 h) and stirred-now (short-term experiments: 0 to 1.25 h) techniques. Both batch and stirred-flow experiments showed that Mn and Ni release increased with decreasing solution pH, indicating an increase in metal release rates with solution hydrogen ion concentration. Linear relationships of both log (free) Mn2+ and log Ni2+ versus pH with a range of slopes from -0.36 to -0.54 (0.96 less than or equal to r(2) less than or equal to 1.00) were found for batch equilibrations of 15 h. Medium-term Mn and Ni release at a solution pH of 5.5 was well described using the parabolic: kinetic model (r(2) greater than or equal to 0.99). More than 90% of total (after 75 min) dissolved Mg was released during the first 5 min of the stirred-flow experiments. In contrast with Mg, most of the Mn and Ni were released during the 5 to 75-min interval. Five kinetic models (zero, first and second order, parabolic, as well as three simultaneous first order models, respectively) mere tested for their ability to fit cumulative metal release versus time curves of stirred-flow experiments. As seen from the coefficients of determination, the parabolic model with its two fitting parameters simulated the data well (0.94 less than or equal to r(2) less than or equal to 1.00), generally better than zero, first, or second order models. The parabolic model led to estimates of both initial. and short-term Mn and Ni release rates, as web as the dependence of these rates on pH. Very good fits were obtained with the three simultaneous first order model (r(2) greater than or equal to 0.99). The use of this six-parameter model, however, is questioned as long as the physical importance of the individual parameters is unknown.