| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Dep. of Environ. Sciences, Univ. of California, Riverside, CA 92521 USA
dparker{at}mail.ucr.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMaterials and methodsResults and discussionREFERENCES |
|---|
3%), but reached 12% for F3 of the sludge-amended soil. Quantitative, reproducible recovery of Cd (96.5 ± 2.1%) was obtained across all samples and averaged 11, 32, 40, 8, and 6% CdT in the respective five fractions. Fractionation trends reflect the Cd sources and physicochemical properties of the samples with Cd being dominant in F3 for soils high in organic matter or contaminated by metal sulfides.
Abbreviations: A254, absorbance at wavelength 254 nm • CdT, total extractable Cd • Conv., conventional sequential extraction procedure • DOC, dissolved organic carbon • FA, fulvic acid • ICP-OES, inductively coupled plasma optical emission spectrometry • ICP-MS, inductively coupled plasma mass spectrometry • Mod-F2, modified sequential extraction procedure with a lead acetate extraction inserted between F2 and F3 • Mod-F3, modified sequential extraction procedure with a lead acetate extraction step inserted between F3 and F4 • NaOAc, sodium acetate • OC, organic carbon • Pb(OAc)2, lead acetate • SEP, sequential extraction procedure • SRM, soil reference material • TMG, trace metal grade
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResults and discussionREFERENCES |
|---|
In a sequential extraction procedure (SEP), a sample is treated with a series of progressively harsher reagents to dissolve increasingly refractory forms. Ideally, the reagents are chosen to selectively attack a specific soil compartment with minimal dissolution of nontargeted phases. In practice, however, the discrete extraction of any given phase may be unachievable (Tessier et al., 1979). Because the extent of extraction depends on the method and conditions employed, the results are at best operationally defined.
Initially developed to examine trace metal fractionation in soils and sediments, sequential extraction techniques are being applied to an ever-widening spectrum of environmental matrices. Despite the expanding applications, the integrity of the methods and meaningfulness of the results are increasingly being questioned. Foremost among the criticisms are poor reagent selectivity, possible redistribution of elements during extraction, and the indiscriminate application of methods (Nirel and Morel, 1990; Kim and Fergusson, 1991).
A number of studies have examined reagent selectivity and the extent of element redistribution during sequential extraction. Whereas some studies indicated poor selectivity (Kheboian and Bauer, 1987) or extensive redistribution (Rendell et al., 1980), others have shown good selectivity and limited redistribution (Kim and Fergusson, 1991). Recent results suggest that the apparent redistribution and poor selectivity reported in some earlier studies might actually be poor extraction efficiency (Han and Banin, 1995; Yarlagadda et al., 1995). Rauret et al. (1989) showed that extraction efficiency and fractionation accuracy could be improved by monitoring selected parameters (e.g., pH, Eh, or dissolution kinetics) to establish extraction end points for each step. Repeated extractions at each step may also minimize element redistribution (Shuman, 1991).
As part of a larger study using isotopic exchange to examine the intrinsic reactivity of various Cd pools in soils, we needed an SEP having high accuracy and reproducibility for soils with diverse sources of Cd contamination and physicochemical properties. Assessment and optimization included an examination of extraction efficiency and reagent specificity at each fractionation step and an investigation of Cd redistribution for selected fractions. Our goal was to develop and adopt an SEP that could reliably apportion soil Cd into five operationally defined pools that would most accurately reflect the physicochemical forms of Cd present.
|
Materials and methods |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResults and discussionREFERENCES |
|---|
15 cm) soils (experimental) and two soil reference materials (SRMs; NIST, 1993) were used to assess and optimize SEP for Cd recovery. The experimental soils were selected for diversity of Cd source and physicochemical properties. Ramona (fine-loamy, mixed, thermic Typic Haploxeralf) was collected from the Moreno Field Station of the Univ. of California in Moreno Valley, CA, and was contaminated with Cd from applications of anaerobically digested sewage sludge. From 1976 to 1991 sludge was applied at a rate equivalent to 180 Mg ha-1 yr-1 dry wt. such that the cumulative loading totaled 176 kg Cd ha-1 (Chang et al., 1997). The second sample, Penn, was collected from spoil material at the Penn Mine, a derelict Cu–Zn mine located in Calaveras County, CA. The third sample, Millsholm (loamy, mixed, thermic Lithic Xerochrept) was collected near Point Dume, Los Angeles County, CA, and was derived from Cd-rich shale (Lund et al., 1981). Palmerton (loamy-skeletal, mixed, mesic, shallow Typic Dystrocrept) was collected from a limed pasture contaminated by airborne oxide and sulfur compounds emitted from a Zn smelter in Palmerton, PA (Buchauer, 1973). The two SRMs, 2710 and 2711, contain elevated and moderately elevated levels of trace metals, respectively. The experimental soils were crushed, sieved (2-mm screen), mixed, and stored (4°C) at field-moist conditions in plastic bags until ready for use. The SRMs were received air-dried, pulverized (<74 m), and homogenized.
Unless otherwise noted, all chemicals were American Chemical Society (ACS) reagent-grade and used without further purification. All plastic and glassware were acid-washed and thoroughly rinsed with 18-M
water. Total C was determined following ignition (Carlo Erba NA 1500, Milan, Italy) and carbonate C was determined by pressure calcimetry (Nelson, 1986). Organic C (OC) was calculated by difference. Soil pH was measured in a 1:1 soil:water slurry using a combination electrode. Exchangeable bases were determined by inductively coupled plasma optical emission spectrometry (ICP-OES; Perkin-Elmer 3000 DV, San Jose, CA) after extraction in ammonium acetate at pH 7.0 (Thomas, 1986). Total-extractable Cd (CdT) was determined following microwave digestion using trace metal grade (TMG) HNO3 and HCl (Millward and Kluckner, 1989) by inductively coupled plasma mass spectrometry (ICP-MS; VG Plasma Quad P2T, VG Elemental Inc., Franklin, MA). Particle size analysis by the pipet method (Gee and Bauder, 1986) was conducted for the experimental samples only.
Optimization and Evaluation of the Sequential Extraction Procedure
The selection of extractants used in our SEP was based largely on the reviews of Chao (1984), Beckett (1989), and Shuman (1991). The soluble/exchangeable fraction, commonly regarded as the most mobile, bioavailable form of soil elements, is comprised of free ions and soluble complexes. This fraction is typically extracted with dilute solutions of Ca or Mg salts (Shuman, 1991). Because we used Ca and Mg as indicators of reagent selectivity, we instead used 0.1 M Sr(NO3)2 for this fraction.
The use of 1.0 M sodium acetate (NaOAc) at pH 5.0 is an established method for the dissolution of soil carbonates. However, the effect of the reagent is not limited to carbonate dissolution. Considerable amounts of specifically sorbed trace metals may be solubilized at pH 5.0 (Hickey and Kittrick, 1984; Aualiitia and Pickering, 1987; Ainsworth et al., 1994). Metal complexation by acetate may also play a role (Tessier et al., 1979).
In most SEPs, trace elements bound in reducible oxides are extracted prior to organic matter/sulfide dissolution. Recent studies, however, indicate that hydroxylamine hydrochloride, the reagent most frequently used to dissolve reducible oxides, releases substantial amounts of trace elements bound to organic matter (Kim and Fergusson, 1991) and some sulfide minerals (Hall et al., 1996). Consequently, the recovery of trace elements in the oxide fraction may be overestimated at the expense of the oxidizable fraction. Furthermore, although Cd incorporates into hydrous Mn and Fe oxides under favorable experimental conditions (Ainsworth et al., 1994), the extent to which substitution occurs in soils has not been established (Gerth, 1990). Thus the role of secondary oxides in controlling Cd solubility may be overstated. For this reason, we have chosen to extract the oxidizable fraction before the reducible fraction.
A number of methods have been developed for extracting trace metals from organic matter. Foremost among these are extraction with alkaline pyrophosphate salts and oxidation using either H2O2 or NaOCl. Pyrophosphate solubilizes organic matter through complexation and dispersion and, while the reagent appears to be selective for organic matter (Papp et al., 1991), extraction efficiency is poor. Evidence shows that acidified H2O2, the most frequently cited method for the oxidation of organic matter and sulfides, attacks Mn oxides (Papp et al., 1991), Fe oxides, and silicate minerals (Douglas and Fiessinger, 1971). Conversely, minimal degradation of Mn and Fe oxides (Shuman, 1983) and silicates have been indicated for OCl-, suggesting its suitability for use earlier in the extraction sequence.
To extract Cd bound in the lattice of secondary oxides, we selected the ascorbate-oxalate method of Shuman (1982). Although acidified hydroxylamine dissolves Mn and amorphous Fe oxides, dissolution of crystalline Fe oxides is minimal (Chao and Zhou, 1983). Additionally, the citrate-bicarbonate-dithionite method (Mehra and Jackson, 1960), the standard for secondary oxide dissolution, is plagued by problems with Zn contamination and metal sulfide precipitation. Shuman (1982) examined the efficacy of ascorbic acid combined with oxalate to dissolve and complex trace metals bound in amorphous and crystalline secondary oxides and found it superior to the dithionite method.
Typically, the residual fraction is obtained by total matrix dissolution using hazardous HF and/or perchloric acids. Unless an unusually high proportion of the analyte is bound in aluminosilicates, total dissolution may not be necessary. In microwave digestion, samples are mixed with HNO3 and HCl at elevated temperature and pressure to form NOx and Cl2, strong oxidants. The method yields rapid, quantitative, and reproducible recovery of many elements with lower analytical blanks and greater precision than conventional hot plate or reflux methods (Millward and Kluckner, 1989; Nieuwenhuize et al., 1991).
Prior to extraction, portions of each sample were air-dried overnight, and ground in an agate mortar to pass through a 150-µm screen. Duplicate 2.0-g portions (oven-dried-wt. basis) were weighed into 50-mL polycarbonate centrifuge tubes and sequentially treated to obtain the following five operationally defined fractions:
Fraction 1 (F1): "soluble–exchangeable." Each sample was reacted with 15 mL of 0.1 M Sr(NO3)2 in a reciprocating shaker (2400 oscillations per h) for 2 h at room temperature. Extraction efficiency for F1 was determined by conducting three successive treatments for each soil material and determining the extent of release of Cd with each treatment.
Fraction 2 (F2): "specifically sorbed–carbonate bound." After optimizing F1, the residue was treated with 30 mL of 1.0 M NaOAc adjusted to pH 5.0 with TMG glacial acetic acid (Gibson and Farmer, 1986). The slurry was shaken as described above for 5 h. The cap was opened occasionally to expel CO2. Following extraction, the pH of the supernatant was determined and compared to that of the unreacted buffer. The need for additional treatment is indicated by a 
pH > 0.1 (Rauret et al., 1989; Han and Banin, 1995).
Fraction 3 (F3): "oxidizable." Following the optimization of F2, the residue was mixed with 5 mL of 5% NaOCl (adjusted to pH 8.5 with TMG HCl) and reacted in a water bath (90–95°C) for 30 min with the cap slightly ajar (Shuman, 1983). Every 10–15 min the cap was tightened and the slurry briefly vortex-mixed. This procedure was conducted seven times successively for each sample, the Cd concentration extracted, and the C content of the residues determined. A plot of cumulative Cd concentration and residual C vs. treatment number was used to evaluate extraction efficiency. We also monitored the pH of the supernatants throughout.
Fraction 4 (F4): "reducible." Following the optimization of F3, the residue was mixed with 20 mL of 0.2 M oxalic acid + 0.2 M NH4 oxalate + 0.1 M ascorbic acid (adjusted to pH 3.0 with TMG NH4OH). The slurry was placed in a water bath (9–95°C) for 30 min (Shuman, 1982) with periodic mixing as for F3. Each sample underwent five successive treatments and the cumulative Cd, Fe, Mn, Al, and Si extracted for each treatment was determined. As before, the extent of element release was plotted as a function of treatment number.
Fraction 5 (F5): "residual." Following the optimization of F4, the residue was oven-dried, pulverized, and mixed. Duplicate 0.20-g subsamples were treated by microwave digestion as for CdT (Millward and Kluckner, 1989).
Between each step the residue was suspended in 5 mL of 0.1 M NaCl to displace entrained solution, minimize sample dispersion, and limit Cd resorption. These rinse solutions were pooled with the preceding extracts rather than discarded. All slurries were centrifuged at 1225 g (10 min) and the supernatants filtered (Whatman #42) into 100 mL (F1, F2), 250 mL (F3, F4), or 50 mL (F5, CdT) volumetric flasks containing 10 µg In L-1 internal standard. Fractions 1 to 3 were acidified to 0.16 M HNO3. A drop of toluene was added to the F4 extracts to discourage bacterial growth.
To investigate Cd redistribution, a series of modified sequential extractions were conducted. In one series, a lead acetate [Pb(OAc)2] extraction step was inserted after the rinse between F2 and F3 and between F3 and F4. Lead was selected because of its chemical similarity to Cd (both are soft Lewis acids) and because Pb readily displaces specifically resorbed metals (Aualiitia and Pickering, 1987). Twenty mL of 0.5 M Pb(OAc)2 adjusted to pH 5.0 with TMG glacial acetic acid was reacted with the residue for 2 h in a reciprocating shaker as above. The supernatants were filtered and acidified following centrifugation. After extraction with Pb(OAc)2, the residues were treated with the remaining sequential reagents as described above.
Total dissolved solids were determined gravimetrically for F1 through F4 and samples were appropriately diluted to 
2000 mg L-1, a tolerance limit for ICP-MS. The microwave extracts (F5, CdT) were diluted to reduce acidity. All Cd analyses were conducted by ICP-MS. Matrix-matched standards were used to analyze F1, F2, and the Pb(AOc)2 extracts; all other solutions were analyzed using Cd standards prepared in 0.16 M HNO3.
To assess reagent specificity, a complete set of extracts for each sample was analyzed for Ca, Mg, Ti, Al, Fe, Mn, Cu, S, Zn, Ni, and Si by ICP-OES using matrix-matched, mixed-element standards to correct for interelement interferences. The amount of soil organic carbon dissolved in the F1 and F2 extracts was also measured. For F1, dissolved organic carbon (DOC) was measured directly using a Dohrman DC 190 C analyzer (Rosemont Analytical, Santa Clara, CA); matrix-matched C standards were prepared from potassium hydrogen phthalate. However, because the abundant acetate in the F2 extracts precluded direct DOC measurement, we determined aromaticity by measuring the absorbance at wavelength 254 nm (A254, Traina et al., 1990) using a Perkin-Elmer Lambda 2 spectrometer (Perkin-Elmer Corp., San Jose, CA).
The acidified F2 extracts showed no evidence of humic acid precipitation suggesting that the extracted soil OC, if present, was fulvic acid (FA). Fulvic acid was isolated from each of the experimental samples as described by Stevenson (1994). Briefly, 25 g of sample was reacted with 0.5 M NaOH and the centrifuged, decanted supernatant was acidified to pH 1.0 with HCl to flocculate humic acid. Following centrifugation to separate the fulvic and humic fractions, the concentration of DOC in diluted FA solutions was determined as for F1. Matrix-matched FA standards were prepared and the A254 was measured for the standards and the F2 extracts. Standard curves of A254 vs. DOC were evaluated by linear regression and the parameters used to estimate DOC in the corresponding F2 extracts. The large amount of soil material required and associated high cost prohibited organic matter extraction from the reference samples.
|
Results and discussion |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResults and discussionREFERENCES |
|---|
The soil materials studied showed a wide range of physicochemical properties (Table 1) . Total extractable Cd ranged from 22 to 42 mg kg-1; Cd levels > 0.5 mg kg-1 generally indicate polluted soils (McBride, 1994). The CdT value reported for Millsholm, a soil derived from Cd-rich shale, is in excellent agreement with Lund et al. (1981). Organic C contents ranged from 8 to 77 g kg-1 with the highest level found in Ramona, the sludge-amended sample. The low pH measured for Penn is typical for soils contaminated by metal sulfides. Reference sample 2711 had the highest carbonate content; the high exchangeable Ca may reflect some dissolution of calcite by the ammonium acetate reagent (pH 7.0).
|
Incomplete carbonate dissolution can result in erroneous metal partitioning in soils (Yarlagadda et al., 1995). The efficacy of 1 M acetate (pH 5.0) for the complete dissolution of carbonates has been examined elsewhere (Tessier et al., 1979; Han and Banin, 1995). A single treatment appears to incompletely dissolve these minerals, particularly slower-reacting dolomite, only when the buffering capacity has been exceeded. Since the carbonate contents of our samples were low (Table 1) and no pH value exceeded 5.05, a single treatment seemed sufficient for the complete dissolution of carbonates in our samples.
For all the samples, Cd extraction appeared complete for F3 after the third hypochlorite treatment despite high C in the residues of some samples (e.g., 12 g C kg-1 for Palmerton). Representative results are given for the Penn and Palmerton samples in Fig. 1
. Sodium hypochlorite (pH 8.5) is a poorly buffered solution and the pH values of the supernatants showed appreciable acidification during oxidation. For all samples, the initial pH values were low (4–5.0) but increased progressively to pH 6.0 with additional treatments (data not shown). Lavkulich and Wiens (1971) compared pH values in H2O2 and OCl- extracts and found much greater acidification with H2O2; values as low as 1.5 have been reported for soil–peroxide slurries (Douglas and Fiessinger, 1971). Others (Lavkulich and Wiens, 1971) have attributed solution acidification to metal hydrolysis, which may result in precipitation and poor extraction efficiency for some elements, such as Fe. However, precipitation in OCl- slurries has been reported only for samples having very low organic matter (Hoffman and Fletcher, 1981; Papp et al., 1991). Hypochlorite extracts are rich in aromatic and (poly)carboxylic acids (Chakrabartty et al., 1974) suggesting that complexation inhibits metal precipitation in these solutions. Conversely, peroxidation by-products consist largely of CO2 and H2O (Savage and Stevenson, 1961). Hence, metal precipitation may be more problematic for H2O2 extracts than OCl- solutions.
|
|
Evaluation of Reagent Selectivity
Extracts at F1 and F2 were slightly colored, suggesting the extraction of some soil organic matter. For F1, only small percentages (<3%) of the total soil OC were solubilized (Table 2)
. Tessier et al. (1979) reported comparable percentages for MgCl2 extracts. For F2, the proportion was similar or somewhat higher (0.7–11.2%) than for F1 (Table 2). The largest amount of DOC extracted in F2 was from the Penn sample, a low-OC sample (Table 1). The small amount of Cd found for F2 for this sample (see below) suggests that the contribution of Cd from organic matter extraction was likely insignificant. Overall the small quantities of DOC found in F1 and F2 suggest that the reagents did not result in excessive organic matter dissolution and that the integrity of the oxidizable fraction was not significantly compromised by reagents used earlier in the sequence.
|
|
The release of concomitant elements at a given step may be used as an indicator of reagent specificity for a SEP (Han and Banin, 1995). Representative results for Ca, Mg, Ti, Al, Fe, Mn, Cu, S, Zn, Ni, and Si analyses conducted for each fraction of the Penn and Palmerton samples are depicted in Fig. 3. The element concentration found in each fraction is reported as a percent of the sum for the five fractions; the sum (g kg-1) is given above the bar. Element percentages in F1 extracts were low to moderate across all samples. On the average, 26 and 5% of the Ca and Mg, respectively, were extracted. Higher Ca percentages were found in limed samples, Palmerton and SRM 2711, suggesting some calcite dissolution by the reagent. By extracting F1 with Ca(NO3)2, one could use the common-ion effect to minimize dissolution of calcite and gypsum in soils. Some Mn, Cu, Zn, and Ni (<18%) were recovered in F1, but only for the acidic samples Penn, Ramona, and 2710 (Fig. 3). The relative absence of Fe, Al, Ti, and Si indicate minimal dissolution of Fe oxides and aluminosilicates. In agreement with others (Tessier et al., 1979; Shannon and White, 1991), the generally low extracted S (<4%) indicates limited attack on OC or inorganic sulfides.
Only SRM 2711 had a significant quantity of Ca (50%) in F2. This amount of Ca (14.8 g kg-1) is equivalent to a calcite content of 37.0 g kg-1, which is in reasonably good agreement with the measured value of 44.0 g kg-1 (Table 1). Additionally, the low Mg (<4%) in this sample suggests that the carbonate mineralogy was calcic rather than dolomitic. A high proportion of Mn (30%) was also found in F2 for 2711. Li et al. (1995) reported similar Ca and Mn values for this SRM. High acetate-extractable Ca and Mn levels have been reported elsewhere for calcareous soils (Han and Banin, 1995). Calcite Mn may be chemisorbed or structural (McBride, 1994). Elevated levels of Mn, Cu, Zn, Ni, and S were found in F2 for the acidic samples Penn, Ramona, and 2710. These elevated S levels cannot be explained by the DOC levels in F2 (Table 2), suggesting that extracted S was largely surface-sorbed, possibly as ternary metal-sulfate complexes (McBride, 1994). Iron, Al, Ti, Mg, and Si quantities for all the samples were minor, indicating limited attack on Fe oxides and aluminosilicates (Tessier et al., 1979; Han and Banin, 1995; Li et al., 1995). The results suggest that 1 M NaOAc showed good selectivity for calcite and associated Mn and also extracts specifically sorbed metals.
Overall S recoveries were highest (37%) in F3 reflecting the destruction of organic matter and metal sulfides. High Cu (46%) and Ni (32%) levels in Ramona indicate the enrichment of Cu and Ni in sludge (Pérez-Cid et al., 1996) and the tendency of these two elements to form strong organic complexes. As noted earlier for Penn, the very high proportion of Zn in F3 (77%) suggests sphalerite oxidation (Fig. 3). Manganese contents in the F3 extracts ranged between 0 and 28% with a mean of 12%. Shuman (1983) reported similar values for soils treated twice in succession with hypochlorite. The selectivity and specificity of hypochlorite for organically bound Mn has been demonstrated by Shuman (1985) and Hoffman and Fletcher (1981). In both studies, the release of extremely low amounts of Mn from subsurface soils having very low organic matter contents indicated that the reagent did not dissolve Mn oxides to any great extent. Iron, Al, Ti, and Si were <1% of the total amounts extracted; Ca and Mg levels were <8%, indicating limited attack on iron oxides and silicate minerals.
As expected, the strongly hydrolyzing elements (Ti, Al, and Fe) were measured almost exclusively in F4 and F5. Whereas Ti and Al were predominately residual, the majority of Fe (>53%) and about 41% of Mn were reducible. With the exceptions noted above and in agreement with others (Shuman, 1991; Li et al., 1995), Cu, Zn, and Ni were found largely in the reducible and residual fractions. The concentrations of Si were highest (3.53–5.54 g kg-l) in F4. Silica levels of 7% have been reported for natural goethite and hematite minerals (Fordham and Norrish, 1983), suggesting that the Si recovered here may have been bound in iron oxides; however, oxalate may also enhance the solubility of aluminosilicates, particularly at elevated temperatures (Fein, 1991). For the SRMs, the Si in F4 was <2% of the certified values suggesting overall only minor attack on aluminosilicates. In summary, our results suggest that the reagents used provided good specificity and selectivity for the targeted fractions.
Cadmium in the Operationally Defined Fractions
Results for the sequential extraction of Cd using the optimized procedure are summarized in Fig. 4
. The mean recovery for three extractions completed in triplicate on different days is reported as a percentage of CdT or the certified value for the experimental and reference soils, respectively. Cadmium recovery ranged from 92 to 103%.
|
Cadmium concentrations in F3 ranged from 5.4 to 16.8 mg kg-1, representing from 14 to 77% of the Cd in these samples (Fig. 4). As stated above, Cd is commonly found in sphalerite and the comparable partitioning of Cd and Zn for Penn supports this interpretation (Fig. 3). The high percentage of Cd for the sludge-amended Ramona soil is likely due to organic matter oxidation (Ducaroir and Lamy, 1995). The relatively high proportions of Cd and S (42 and 72%, respectively) in F3 of Palmerton as compared to the low percentage of Cd in F4 (6%) suggests that site contamination is largely due to S-rich dust deposition (Buchauer, 1973). For Millsholm, 30% of the Cd was found in F3, suggesting that the Cd-rich shale parent material was of marine origin, where anaerobic conditions favored metal sulfide formation (Mason and Berry, 1968). Smaller percentages of Cd were extracted in F3 for 2710 (25%) and 2711 (14%).
Cadmium quantities in F2 tended to increase with sample pH (Fig. 4, Table 1), reflecting both pH-dependent sorption and isomorphic substitution of Cd2+ for Ca2+ in calcite (McBride, 1994). In contrast with Li et al. (1995), a large amount of Cd (32.0 mg kg-1 or 77%) was extracted in F2 for the calcareous reference sample 2711. Similar levels of acetate-extractable Cd for calcareous soils have been reported elsewhere (Han and Banin, 1995). Calcite Cd is particularly vulnerable to MgCl2 extraction (Kim and Fergusson, 1991). This may account for the comparatively high and low Cd concentrations reported for F1 and F2, respectively, by Li et al. (1995) for SRM 2711. Contrary to results obtained for smelter emissions by Hickey and Kittrick (1984), a relatively high amount (34%) of sorbed–carbonate Cd was extracted from Palmerton. This contrast likely reflects differences in sample pH with Palmerton being more alkaline. In agreement with Jeng and Singh (1993), small quantities of Cd (<16%) were found in F2 for the two most acidic samples, Penn and 2710. Conversely, considerable Cd (30%) was measured for F2 of the acidic, sludge-amended sample (Ramona), probably reflecting pH-dependent desorption from organic matter or the release of Cd from carbonates. The significance of the sorbed–carbonate fraction for Cd has been demonstrated previously for sludge-amended soils (Hickey and Kittrick, 1984).
Quantities of trace metals in the soluble–exchangeable fraction of uncontaminated soils are typically small, ranging from none to a few percent (Shuman, 1991). However, soils containing evaporites (Hall et al., 1996), phosphate fertilizers (Jeng and Singh, 1993), or contaminated with metal-rich sediments (Hickey and Kittrick, 1984) may have considerable amounts of exchangeable trace metals. In agreement with Li et al. (1995), exchangeable Cd was high for SRM 2710 (9.7 mg kg-1 or 44%), a sample contaminated with pond sediments from a nearby mine (NIST, 1993). A high percentage of exchangeable S (15%) in F1 for this sample suggests that Cd may have been bound in metal-sulfide sediments. Rapin et al. (1986) showed that samples containing sedimentary (amorphorous) metal sulfides released high levels of metals early in a SEP. Soluble–exchangeable Cd was considerably lower (<9%) in the other samples (Fig. 4).
Although the mean Cd recovery in F4 and F5 was low (8 and 6%, respectively), moderate percentages were recovered in F4 for Millsholm and in F5 for SRM 2710 (Fig. 4). These results indicate that, for most of the samples, comparatively little Cd was bound in Fe/Mn oxides or aluminosilicate minerals.
Evaluation of Possible Cd Redistribution during the Sequential Extraction
Results for selected steps of the modified and conventional SEPs in which Pb displaced (re)sorbed Cd are summarized in Table 3 . Because Cd concentrations at F1 and F5 for both procedures were not different (±1 SD) they were omitted for simplicity. As indicated by the comparable Cd partitioning for both procedures and the low Cd levels measured for the Pb extracts, little Cd redistribution (<1%) is evident for the Penn sample. Similarly, when the Pb extraction step was inserted between F3 and F4 (Mod-F3) only minor redistribution (<3%) from F3 into subsequent fractions was indicated for the Ramona, Millsholm, and Palmerton samples. However, when the Pb extraction step followed F2 (Mod-F2) for these three samples, some Cd redistribution from F2 was indicated. For Ramona and Millsholm, as much as 12 and 9% of CdT, respectively, may have resorbed to oxidizable components. For Palmerton, lower-than-typical Cd recoveries were obtained in F3 and F4 for Mod-F2, indicating possible redistribution into these two fractions. However, our method likely does not distinguish between indigenous, desorbable Cd and solublized Cd that is resorbed. Consequently our results represent "worst-case," upper limits for Cd redistribution; actual redistribution may be less. Cadmium complexation with Cl and soluble organic ligands, competitive sorption by co-extracted metals, and repeated treatments at selected steps may limit Cd redistribution with our SEP. Overall, our results agree with others (Kim and Fergusson, 1991) who have demonstrated minimal Cd redistribution during sequential extraction.
|
|
ACKNOWLEDGMENTS |
|---|
Received for publication September 4, 1998.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResults and discussionREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  A. Voegelin, G. Tokpa, O. Jacquat, K. Barmettler, and R. Kretzschmar Zinc Fractionation in Contaminated Soils by Sequential and Single Extractions: Influence of Soil Properties and Zinc Content J. Environ. Qual., May 1, 2008; 37(3): 1190 - 1200. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Wu, J. Chen, J. Ji, P. Gong, Q. Liao, Q. Tian, and H. Ma A Mechanism Study of Reflectance Spectroscopy for Investigating Heavy Metals in Soils Soil Sci. Soc. Am. J., May 16, 2007; 71(3): 918 - 926. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 W. Chen, A. C. Chang, L. Wu, and A. L. Page Modeling Dynamic Sorption of Cadmium in Cropland Soils Vadose Zone J., November 20, 2006; 5(4): 1216 - 1221. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. F. Sukkariyah, G. Evanylo, L. Zelazny, and R. L. Chaney Recovery and Distribution of Biosolids-Derived Trace Metals in a Clay Loam Soil J. Environ. Qual., September 8, 2005; 34(5): 1843 - 1850. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. T. Basta, J. A. Ryan, and R. L. Chaney Trace Element Chemistry in Residual-Treated Soil: Key Concepts and Metal Bioavailability J. Environ. Qual., January 1, 2005; 34(1): 49 - 63. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. R. Cave, A. E. Milodowski, and E. N. Friel Evaluation of a method for identification of host physico-chemical phases for trace metals and measurement of their solid-phase partitioning in soil samples by nitric acid extraction and chemometric mixture resolution Geochemistry: Exploration, Environment, Analysis, February 1, 2004; 4(1): 71 - 86. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Voegelin, K. Barmettler, and R. Kretzschmar Heavy Metal Release from Contaminated Soils: Comparison of Column Leaching and Batch Extraction Results J. Environ. Qual., May 1, 2003; 32(3): 865 - 875. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Kabala and B. R. Singh Fractionation and Mobility of Copper, Lead, and Zinc in Soil Profiles in the Vicinity of a Copper Smelter J. Environ. Qual., March 1, 2001; 30(2): 485 - 492. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D.E. Abbott, M.E. Essington, M.D. Mullen, and J.T. Ammons Fly Ash and Lime-Stabilized Biosolid Mixtures in Mine Spoil Reclamation: Simulated Weathering J. Environ. Qual., March 1, 2001; 30(2): 608 - 616. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS |
