For temperature, degrees Celsius (°C) can be converted to degrees Fahrenheit (°F) by using the formula °F = (1.8)(°C) + 32.Abbreviated units used in report: Kd (distribution coefficient), mg/L (milligrams per liter), mL (milliliter), mL/g (milliliters per gram), mg/kg (milligrams per kilogram), ^is/cm (microsiemens per centimeter) at 25°C, m2/g (meters squared per gram).
IV The Use of Chemical and Physical Properties for Characterization of Strontium Distribution Coefficients at the Idaho National Engineering and Environmental Laboratory, IdahoBy Jeffrey J. Rosentreter, Reinaldo Nieves, John Kalivas, Joseph P. Rousseau, and Roy C. Bartholomay
AbstractThe U.S. Geological Survey and Idaho State University, in cooperation with the U.S. Department of Energy, conducted a study to determine strontium distribution coefficients (Kds) of surficial sediments at the Idaho National Engineering and Environmental Laboratory (INEEL). Batch experimental techniques were used to determine experimental Kds of 20 surficial-sediment samples from the INEEL. The Kds describe the distribution of a solute between the solution and solid phase. Kds of the 20 surficial-sediment samples ranged from 36 to 275 milliliters per gram. Many chemical and physical properties of both the synthesized aqueous solution and sediments used in the experiments also were determined. The following solution properties were determined: initial and equilibrium concentrations of calcium, magnesium, and strontium; pH and specific conductance; and initial concentrations of potassium and sodium. Sediment properties determined were grain-size distribution, mineralogy, whole-rock major oxide, strontium and barium concentrations, and Brunauer-Emmett-Teller surface area. Multivariate-regression techniques were used to identify which of these variables or set of variables could best predict the strontium Kd values. Partial leastsquares regression was used to fit these data to an empirical model that could be used to predict strontium Kds of surficial sediments at the INEEL. The best-fit model was obtained using a four-variable data set consisting of surface area, manganese oxide concentration, specific conductance, and pH. Application of the model to an independent split of the data resulted in an average relative error of prediction of 20 percent and a correlation coefficient of 0.921 between predicted and observed strontium Kds. Chemical and physical characteristics of the solution and sediment that could successfully predict the Kd values were identified. Prediction variable selection was limited to variables which are either easily determined or have available tabulated characteristics. The selection criterion could circumvent the need for time-and labor-intensive laboratory experiments and provide an alternate faster method for estimating strontium Kds.