Environ. Sci. Technol. 2001, 35, 121-126
Cadmium Reactivity in Metal-Contaminated Soils Using a Coupled Stable Isotope Dilution-Sequential Extraction Procedure
estimate the reactivity of pools of trace elements in soils are continually being developed and assessed in an effort to gain a clearer understanding of element mobility and bioavailability and to develop successful metal remediation methods (1-3). However, comparatively few studies have focused on long-term trace element dynamics in soils and on the reactivity of the more refractory forms that may replenish the labile pool upon its depletion, for example, during chemical or biological extraction as a means for soil remediation (4).
ZOE ANN S. AHNSTROM AND DAVID R. PARKER* Soil and Water Sciences Section, Department of Environmental Sciences, University of California, Riverside, California 92521
Chemical fractionation methods have been used innumerable times to allocate trace metals into operationally defined pools of putatively distinct biogeochemical form (adsorbed, carbonate, organic, hydrous oxide, etc.) (5, 6). Most commonly, these are sequential extraction procedures (SEPs) where increasingly “harsh” reagents are used in a stepwise fashion to extract increasingly refractory forms of the metal(s) of interest. Commonly, the first one, two, or occasionally three extracted fractions are taken to be the labile pool of the metal that is bioavailable and potentially mobile. Much less frequently, this assumption is actually tested by comparing the extracted pools with, for example, plant uptake of the metal (see ref 7 and references therein).
To better understand the intrinsic reactivity of Cd, four soils having diverse sources of Cd contamination (total Cd: 22-34 mg kg-1) were investigated using a stable isotope dilution-sequential extraction procedure (SID-SEP) during a 59-week incubation. Samples were spiked with carrierfree 111Cd and periodically extracted into five operationally defined fractions; the 111Cd:110Cd ratios were measured by inductively coupled plasma-mass spectrometry. The total labile pool of Cd (E value) was calculated as well as the labile Cd within each extracted fraction. Results for three of the soils (Cd sources: natural, sewage sludge, smelter emissions) were quite similar. The overall %E after 2-week equilibration was 35-49% of total soil Cd. Within fraction 2 (sorbed/carbonate), 70-75% of the Cd was isotopically labile, while within fraction 3 (oxidizable) only 3541% of the Cd was labile within 2 weeks. The fourth (reducible) and fifth (residual) fractions were dominated by nonlabile Cd. Although all E values increased somewhat from 2 to 59 weeks, none of the extracted fractions reached isotopic equilibrium with the soluble/exchangeable Cd extracted during step 1. Because fractions 2 and 3 dominated the native Cd in all three soils, the total labile pool was contributed primarily (85-98%) by these two fractions. A fourth soil (mine spoil-contaminated) was demonstrably different: after 2 weeks, the overall %E was just 13 and, although 82% of the total Cd was present in the oxidizable fraction, just 2% of that was isotopically labile. The nonlability of Cd in this soil could be ascribed to the predominance of inorganic forms, most likely occluded Cd in sphalerite. No single Cd fraction from the SEP nor any combination of fractions showed a good correspondence with the size of the isotopically labile pool. Our results suggest that conventional SEPs may be of limited utility for predicting bioavailability, for example, during ecological risk assessment.
Introduction As soils increasingly become repositories for metalliferous wastes, the bioavailability, mobility, and fate of toxic trace elements in soils are increasingly critical issues. Methods to * Corresponding author e-mail:
[email protected]; phone: (909)787-5126; fax: (909)787-3993. 10.1021/es001350o CCC: $20.00 Published on Web 11/29/2000
2001 American Chemical Society
Isotopic dilution methods have been used for many years to probe the reactivity of various elements in soils, with an emphasis on the essential plant nutrients (e.g., refs 8-11). A smaller number of studies have examined the intrinsic lability of toxic trace elements, including Cd. Approaches include direct measurement of the isotopically exchangeable Cd pool (the E value) via a 109Cd spike followed by measurement of the specific activity of the aqueous phase (12-15). In addition, 109Cd-spiked soils have been used to rear plants in pot studies, and the plant’s specific activity has been used to compute the bioavailable soil Cd (the L value) (14, 16, 17). Comparison of E values with simple chemical extractants has yielded mixed results. Reasonable agreement was obtained between Cd extracted in dilute EDTA or DTPA and isotopically labile Cd in four soil samples (12). Nakhone and Young (13) found that 50 mM EDTA consistently overestimated the E value in 33 diverse soils, while dilute simple salts extracted highly variable but generally very little Cd as compared to the E values. But in a more recent study, Young et al. (15) found a very close agreement between E values and the Cd extractable with 1 M CaCl2. In these kinds of studies, however, element lability is typically evaluated over the short term (several days to a few weeks), leaving the reactivity of less labile fractions largely unexplored. Only rarely have isotopic tracers been combined with complete SEPs to probe the relationships between reactivity and chemical extractability of soil trace elements. Goldberg and Smith (18) examined the effects of different drying regimens on the distribution of 54Mn in five operationally defined soil fractions. Using a partial, three-step SEP, McLaren and Crawford (19) showed that, although the pyrophosphate-extractable Cu (organic) dominated the total radiolabile pool in six soils, the majority of the Cu extracted was not isotopically exchangeable. More recently, 109Cd was used in long-term soil incubation experiments to investigate Cd exchange among four fractions extracted using a modification of the widely used Tessier (20) extraction method (21). The first two pools exhibited very rapid isotopic exchange, while the second two were largely nonlabile. However, only two soils were studied, and they contained almost background levels of Cd; extrapolation to more Cdcontaminated soils and sediments cannot be safely made. VOL. 35, NO. 1, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
121
TABLE 1. Selected Chemical and Physical Properties of Four Soil Materials Studied (25) sample
Cd source
Cd (mg/kg)
pHa
organic C (g/kg)
carbonateb (g/kg)
USDA textural class
Ramona Penn Mine Millsholm Palmerton
sewage sludge Pb-Zn mine spoil natural smelter emissions
34.0 21.8 26.1 30.5
6.0 3.5 7.3 7.0
77 8 20 37
2.3 0.0 6.2 1.3
loam loam silt loam loam
a
1:1 in H2O.
b
Values are reported in calcite equivalents.
With the advent of inductively coupled plasma-mass spectrometry (ICP-MS), stable isotopes can now be more conveniently used in studies of elemental labilities and transformations in soils and may offer several advantages over radioisotopes (22-24). Our objective here was combine a Cd SEP that was previously optimized for Cd recovery and selectivity (25) with a stable isotope dilution (SID) method to investigate long-term Cd pool dynamics. Specifically, we wished to probe the relationships between the isotopic lability and the chemical extractability of the operationally defined soil Cd fractions, using reaction times ranging from days to >1 yr.
Experimental Section Materials. All mineral acids were trace-metal grade, the reagents used in the soil extraction procedures were ACS grade, and these materials were used without further purification. Enriched 110Cd (96.3% atom abundance) and 111Cd (96.7% atom abundance) were obtained as Cd metal (U.S. Services, Inc., Summit, NJ). We selected 111Cd (12.80% natural abundance, ref 26) as the tracer isotope and 110Cd (12.49%) as the reference isotope (see Supporting Information). Chelex-100 (Bio-Rad Laboratories, Richmond, CA) was used to purify NH4OAc and pyrophosphate buffers and for ion-exchange chromatography. High-purity water (18 MΩ) was used throughout. All glass- and plasticware were acidwashed in 1.6 M HCl for a minimum of 3 h and thoroughly rinsed with high-purity water. We used four soils having diverse sources of Cd contamination and physicochemical properties that were previously characterized in detail (25); the most salient characteristics are summarized in Table 1. Short-Term Incubations. Seven sets of duplicate 2.0-g (oven-dry-weight equivalent) portions of each soil sample were weighed into 50-mL polyethylene centrifuge tubes containing the111Cd tracer isotope dissolved in ∼0.25 mL of water. The spike was nominally 1% of the total Cd content for each soil. The samples were briefly (∼1 min) centrifuged at 1225g to introduce the tracer to the sample. The gravimetric water content was adjusted to 0.15-0.18 g/g, and the samples were incubated at 25 °C with the lids slightly ajar. The moisture content was monitored daily and readjusted as needed. At 2 and 24 h and 2, 4, 7, and 14 d, two duplicate samples were extracted twice in succession with 15 mL of 0.1 M Sr(NO3)2 (2 h shaking at 40 oscillations min-1) followed by a rinse of 5 mL of 0.1 M NaCl (the same as the first extraction step in the SEP; ref 25 and below). The solutions were decanted following centrifugation (1225g) for 10 min, and the rinses were pooled with the extracts. For each soil, the seventh set of duplicate samples was retained for HNO3HCl digestion (see below) and direct measurement of the new 111Cd:110Cd equilibrium isotope ratio (EQIR; see below). Long-Term Incubations. Here, 300 g (oven-dry-weight basis) of each sample was thinly spread over wax paper, and 40-mL spike solutions containing 111Cd tracer isotope equivalent to ∼1% total Cd were sprayed over each soil and thoroughly mixed using plastic spoons. The spiked samples were quantitatively transferred to 500-mL polycarbonate storage bottles, adjusted to moisture contents of 0.15-0.18 g g-1, and incubated as above. The moisture content was adjusted every second or third day as needed. 122
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 1, 2001
TABLE 2. Instrumental Operating Parameters for Isotope Ratio Determinations by ICP-MS forward RF power, W reflected power, W plasma Ar flow rate, L min-1 auxiliary Ar flow rate, L min-1 nebulizer Ar flow rate, L min-1 solution uptake rate, mL min-1
1350
