Assessing the solubility controls on vanadium in groundwater, northeastern San Joaquin Valley, CA

摘要

The solubility controls on vanadium (V) in groundwater were studied due to concerns over possible harmful health effects of ingesting V in drinking water. Vanadium concentrations in the northeastern San Joaquin Valley ranged from <3 lg/L to 70 lg/L with a median of 21 lg/L. Concentrations of V were highest in samples collected from oxic groundwater (49% > 25 lg/L) and lowest in samples collected from anoxic groundwater (70% < 0.8 lg/L). In oxic groundwater, speciation modeling (SM) using PHREEQC predicted that V exists primarily as the oxyanion H_2VO_4~-. Adsorption/desorption reactions with mineral surfaces and associated oxide coatings were indicated as the primary solubility control of V~(5+) oxyanions in groundwater. Environmental data showed that V concentrations in oxic groundwater generally increased with increasing groundwater pH. However, data from adsorption isotherm experiments indicated that small variations in pH (7.4–8.2) were not likely as an important a factor as the inherent adsorption capacity of oxide assemblages coating the surface of mineral grains. In suboxic groundwater, accurate SM modeling was difficult since Eh measurements of source water were not measured in this study. Vanadium concentrations in suboxic groundwater decreased with increasing pH indicating that V may exist as an oxycationic species [e.g. V(OH)_3~+]. Vanadium may complex with dissolved inorganic and organic ligands under suboxic conditions, which could alter the adsorption behavior of V in groundwater. Speciation modeling did not predict the existence of V-inorganic ligand complexes and organic ligands were not collected as part of this study. More work is needed to determine processes governing V solubility under suboxic groundwater conditions. Under anoxic groundwater conditions, SM predicts that aqueous V exists as the uncharged V(OH)_3 molecule. However, exceedingly low V concentrations show that V is sparingly soluble in anoxic conditions. Results indicated that V may be precipitating as V~(3+)- or mixed V~(3+)/Fe~(3+)-oxides in anoxic groundwater, which is consistent with results of a previous study. The fact that V appears insoluble in anoxic (Fe reducing) redox conditions indicates that the behavior of V is different than arsenic (As) in aquifer systems where the reductive dissolution of Fe-oxides with As adsorbed to the surface is a well-documented mechanism for increasing As concentrations in groundwater. This hypothesis is supported by the relation of V to As concentrations in oxic versus anoxic redox conditions. Sequential extraction procedures (SEP) applied to aquifer material showed that the greatest amount of V was recovered by the nitric acid (HNO_3) extract (37–71%), followed by the oxalate-ascorbic acid extract (19–60%) and the oxalate extract (3–14%). These results indicate that V was not associated with the solid phase as an easily exchangeable fraction. Although the total amount of V recovered was greatest for the HNO_3 extract that targets V adsorbed to sorption sites of crystalline Al, Fe and Mn oxides, the greatest V saturation of sorption sites appeared to occur on the amorphous and poorly crystalline oxide solid phases targeted by the oxalate and oxalate-ascorbic acid extracts respectively. Adsorption isotherm experiments showed no correlation between V sorption and any of the fractions identified by the SEP. This lack of correlation indicates the application of an SEP alone is not adequate to estimate the sorption characteristics of V in an aquifer system.