Environ. Sci. Technol. 2004, 38, 3153-3160
Enhanced Contaminant Desorption Induced by Phosphate Mineral Additions to Sediment DANIEL I. KAPLAN* AND ANNA S. KNOX Westinghouse Savannah River Company, Building 773-43A, Room 215, Aiken, South Carolina 29808
Apatite, Ca10(PO4)6(OH,F)2, has been successfully used as a soil amendment at numerous sites to immobilize metals and radionuclides. Such sites commonly contain multiple contaminants; the impact of apatite on these contaminants is expected to vary greatly. The objective of this study was to determine the influence of apatite on nontargeted sediment contaminants. Laboratory batch experiments were conducted under oxidized (several weekly wet/dry cycles) and reduced (water-saturated) conditions with a sediment collected from a wetland contaminated with numerous metals and radionuclides. Apatite additions resulted in the significant (p e 0.05) reduction of porewater Cd, Co, Hg, Pb, and U concentrations. However, apatite additions also resulted in the enhanced desorption of As, Se, and Th. Increases in porewater As and Se concentrations were the result of phosphate competitive exchange and not to the release of these contaminants directly from the apatite, which contained 29 mg kg-1 As and 0.2 mg kg-1 Se. Apatite additions increased porewater Th and organic C concentrations under oxidized (Eh ) 497 mV) but not reduced (Eh ) 65 mV) conditions. In the oxidized system, the leachate from the apatite treatment had a brown coloration and contained 226 mg L-1 organic C, as compared to 141 mg L-1 in the unamended samples. The desorbed organic C likely contained significant quantities of Th. This conclusion was supported by (i) the observation that porewater Th partitioned to hydrophobic resins, (ii) thermodynamic calculations which predicted that essentially all porewater Th existed as organic matter complexes, and (iii) there were significant correlations (r ) 0.91, n ) 8, p e 0.01) between porewater organic C and Th concentrations. Sediment additions of zero-valent iron particles along with the apatite eliminated the enhanced desorption of As, Se, and Th observed when only apatite was added. This study underscores the importance of monitoring the influence of sediment amendments on nontarget contaminants and provides examples of how the sediment additions of apatite can effectively immobilize some contaminants while enhancing the mobility of others.
Introduction The use of mineral amendments to immobilize metals and radionuclides in contaminated sediments is obtaining increasing public acceptance as the cost of cleaning and the * Corresponding author
[email protected]. 10.1021/es035112f CCC: $27.50 Published on Web 04/29/2004
phone:
(803)725-2363;
2004 American Chemical Society
e-mail:
number of such sites has steadily increased over the years (1). Like other in situ methods, it has the advantage that it reduces the risk of worker exposure during remediation and is typically less expensive and much less disruptive to ecosystems than conventional ex situ methods involving excavation and treatment, followed by disposal. However, it has the distinct disadvantage that the contaminants remain in place. Thus, at sites where the contaminants pose an unacceptable human or ecological risk, it is necessary to demonstrate that the sediment amendments reduce the rate and extent that contaminants enter the biosphere (i.e., convert the contaminants into a less bioavailable form). At sites where the contaminants pose an unacceptable groundwater risk, it is necessary to demonstrate that the amendments slow contaminant release for a sufficient duration. Many contaminated sites contain several metals or radionuclides, which commonly possess different chemical properties controlling their mobility and toxicity. As such, no single amendment is expected to immobilize all inorganic contaminants. For example, the use of apatite, Ca10(PO4)6(OH,F)2, as a sediment amendment has been shown to be extremely effective at immobilizing Pb and U; less effective for Cd, Co, Cu, Mn, Ni, and Zn; and ineffective with most anions (2-8). Applications of various phosphate forms for agricultural and remediation purposes have caused porewater As concentrations to increase. This has been attributed to either phosphate exchange for sorbed arsenate or the release of As directly from the phosphate amendment (7-9), which may contain percent levels of As (10). Similarly, increases in surface water U concentrations in agricultural areas of the United States have been attributed to the existence of U in phosphate fertilizers (11, 12). Some additional secondary reactions involving sediment phosphate amendments that may influence contaminant mobility include (i) increased pH as a result of apatite dissolution (a H+-consuming reaction) or ligand exchange of phosphate for hydroxyls on amphoteric mineral surfaces; (ii) increased concentrations of dissolved organic matter, a strong metal ligand, as a result of increased pH and phosphate exchange with sorbed organic anions; (iii) decreased anion exchange; and (iv) increased particle dispersion (13-15). Sufficient changes in these sediment properties could result in stronger binding of some contaminants and weaker binding of others. Iron is an excellent metallic material for environmental remediation because it is a strong reducer, nontoxic, and inexpensive and the reaction rates of interests are limited by mass transport (16, 17). Secondary Fe phases formed during remediation of contaminated water also play an important role in surface-induced reduction, in coprecipitation, and in creating additional sorption sites. Zero-valent iron, Fe(0), is almost exclusively used to treat contaminated groundwater under saturated conditions. Its proposed use as a surface soil amendment is primarily to provide a source of Fe to form iron oxyhydroxide, which may enhance contaminant removal by coprecipitation and adsorption (18, 19). A pilot-scale, nuclear research facility located on the T-Area of the Savannah River Site (Aiken, SC) introduced several radionuclides (including 228Ac, 137Cs, 228Ra, 90Sr, 228Th, and 233/234/235/238U), metals (including Cd, Co, Cr, Hg, and Pb), and metalloids (including As and Se) into an unlined seepage basin during the 1960s and 1970s (Figure 1). These materials have since migrated by overland and subsurface flow into an adjacent wetland. The forested site is within the 100-yr flood plain and is often flooded during winter and spring. The most widespread contaminant on the site is 228Th. Total Th concentrations in the sediment at the site range VOL. 38, NO. 11, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3153
FIGURE 1. Thorium sediment concentrations in the surface 20 cm at the T-Area study site (data redrawn from WSRC; 20). from 7 (background) to 924 mg kg-1 (20). A surface sediment sample collected 250-m southwest from the point source contained 194 mg kg-1 Th (Figure 1). In the natural environment, thorium exists exclusively in the +4 oxidation state, and its geochemistry is largely controlled by hydrolysis (forming Th(OH)3+, Th(OH)22+, Th(OH)3+, Th(OH)40, and several polynuclear species), low solubility, and organic matter complexation (21-23). The objective of this study was to evaluate the use of sediment amendments to immobilize the metals and radionuclides at the T-Area, thereby offering a less ecologically intrusive alternative to ex situ approaches. Because of the large number of different inorganic contaminants at the site, two sediment amendments were evaluated, Fe(0) and apatite. Given the wide range of moisture and redox conditions expected at the site, particular attention was directed at evaluating the efficacy of the amendments under oxidizing and reducing conditions. It is anticipated that the sediment amendments will be broadcast on the ground surface and backfilled into drilled 2 cm diameter × 30 cm deep holes spaced across the contaminated area.
Materials and Methods The approach used in this study was to collect a contaminated sediment from the T-Area, characterize it, equilibrate it with apatite and Fe(0), and then analyze porewater contaminant concentrations. The apatite- and Fe(0)-treated sediments were equilibrated in a 15 g:25 mL sediment:water suspension to simulate flooded conditions (low redox) and in a variable saturated system (high redox) to simulate a system undergoing repeated wet and dry cycles. Leachate samples were collected from the saturated samples after 70 d of equilibration and from the first and last wetting cycle from the wet/ dry-cycled treatments. The first cycle porewater samples were analyzed to provide insight into the system when it was furthest from steady state, a condition when contaminant mobility can be its greatest. The last sampling of the wet/ dry-cycled treatments and the only sampling of the saturated paste treatments were analyzed to provide insight into the 3154
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 11, 2004
system as it presumably approached steady state under oxidizing and reducing conditions, respectively. Sampling, Characterization, and Statistics. A water sample was collected from a surface stream located in a wetland approximately 1 km upstream from the study site. The water was passed through a 0.45-µm filter and stored at 4 °C. It was chemically characterized using standard laboratory methods. The surface water was acidic (pH of 5.1), had a low ionic strength (0.6 mM as calculated using vMINTEQA2, Version 2.15; 24) dominated by Ca and sulfate, and contained high organic C concentrations (6.1 mg kg-1 total organic C). A sediment sample was collected from a contaminated portion of the T-Area. The leaf litter on the sediment surface (the O horizon) was removed prior to collecting the surface 15 cm in plastic tubes. The sample was stored moist and in the dark at 4 °C. All characterization and subsequent equilibration tests were conducted with the sediment in the moist state. Moist samples were used in an effort to minimize experimental artifacts introduced by overdrying sediments, such as changing the reactivity of iron, aluminum, and manganese oxides or oxidizing organic matter. The sediments were characterized for pH in a 1:1 water:sediment slurry; particle-size distribution by the sieve and pipet method; cation and anion exchange capacities by the unbuffered ammonium and chloride exchange method; and free Fe by the dithionite-citrate buffer method (25). Organic matter was determined by the weight loss on ignition method conducted at 360 °C for 2 h (26). Mineralogy was determined by X-ray diffraction analyses of the
