The Uranium from Seawater Program at the Pacific Northwest National Laboratory: Overview of Marine Testing, Adsorbent Characterization, Adsorbent Durability, Adsorbent Toxicity, and Deployment Studies

The Pacific Northwest National Laboratory (PNNL) is evaluating the performance of adsorption materials to extract uranium from natural seawater. Testing consists of measurements of the adsorption of uranium and other elements from seawater as a function of time using flow-through columns and a recirculating flume to determine adsorbent capacity and adsorption kinetics. The amidoxime-based polymer adsorbent AF1, produced by Oak Ridge National Laboratory (ORNL), had a 56-day adsorption capacity of 3.9 +/- 0.2 g U/kg adsorbent material, a saturation capacity of 5.4 +/- 0.2 g U/kg adsorbent material, and a half-saturation time of 23 2 days. The ORNL AF1 adsorbent has a very high affinity for uranium, as evidenced by a 56-day distribution coefficient between adsorbent and solution of log K-D,K-56day = 6.08. Calcium and magnesium account for a majority of the cations adsorbed by the ORNL amidoxime-based adsorbents (61% by mass and 74% by molar percent), uranium is the fourth most abundant element adsorbed by mass and seventh most abundant by molar percentage. Marine testing at Woods Hole Oceanographic Institution with the ORNL AF1 adsorbent produced adsorption capacities 15% and 55% higher than those observed at PNNL for column and flume testing, respectively. Variations in competing ions may be the explanation for the regional differences. Hydrodynamic modeling predicts that a farm of adsorbent materials will likely have minimal effect on ocean currents and removal of uranium and other elements from seawater when farm densities are <1800 braids/km(2). A decrease in uranium adsorption capacity of up to 30% was observed after 42 days of exposure because of biofouling when the ORNL braided adsorbent AI8 was exposed to raw seawater in a flume in the presence of light. No toxicity was observed with flow-through column effluents of any absorbent materials tested to date. Toxicity could be induced with some non-amidoxime based absorbents only when the ratio of solid absorbent to test media was increased to part per thousand levels. Thermodynamic modeling of the seawater-amidoxime adsorbent was performed using the geochemical modeling program PHREEQC. Modeling of the binding of Ca, Mg, Fe, Ni, Cu, U, and V reveal that when binding sites are limited (1 X 10(-8) binding sites/kg seawater), vanadium heavily outcompetes other ions for the amidoxime sites. In contrast, when binding sites are abundant, Mg and Ca dominate the total percentage of metals bound to the sorbent.