METAL PARTITIONING AT THE ANACONDA SUPERFUND SITE, MONTANA: A GEOCHEMICAL AND MINERALOGICAL STUDY

The former Anaconda Smelter site, in southwestern Montana is an EPA Superfund Complex of significant proportions. Nine decades of smelting operations have contaminated the local and surrounding soils (700 km²), groundwater, and the Clark Fork River. The five principal contaminants include arsenic, cadmium, copper, lead and zinc.

Sequential leaching methods have been used to determine the mobility and bioavailability of the contaminants. Specifically, the five-step Tessier sequential extraction (Analytical Chemistry, 1979) procedure was followed to evaluate the extent of contamination and the main sources and sinks of trace metals in soils at the smelter site. Four cores were collected proximally to the smelter tailings ponds extending to depths between 0.63m and 2.5m. The cores were then sampled at 30-50cm intervals. In addition to the Tessier procedure, soil samples were acid digested following EPA Method 3050B, which accounts for ‘environmentally available' metals. Bulk soil composition was determined on select samples by fusion techniques; both sieved (<0.25mm) and unsieved soil samples were fused. Inductively coupled plasma spectroscopy was used for all metal analyses. Some samples were further characterized by thin-section, to determine the soil mineralogy, and whether any association between mineralogy, grain-size and metal contamination exists.

The soils comprise plagioclase, quartz, carbonate phases, hematite and Fe-(oxy)hydroxide phases, and opaque minerals. The concentration of contaminant metals decrease slightly with distance from the smelter tailings ponds and with depth. The change in metal concentration with depth appears to correlate with a redox boundary. Contaminant concentrations (As, Pb, and Zn) in core three were anomalously high relative to the other cores, likely resulting from the third core's proximity to a constructed drainage channel. Furthermore, metal concentrations were observed to vary as a function of grain size. As anticipated, a strong correlation existed between the contaminant metals and the Fe- and Mn-oxide phases. This association was confirmed by thin-section and additional extraction procedures. A strong association between the contaminants and the Tessier defined ‘organic-bound' fraction was also observed.