JEQ Journal of Natural Resources and Life Sciences Education
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Related articles in JEQ
Right arrow Similar articles in this journal
Right arrow Similar articles in Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Web of Science (11)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Sowder, A. G.
Right arrow Articles by Morris, P. J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Sowder, A. G.
Right arrow Articles by Morris, P. J.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Sowder, A. G.
Right arrow Articles by Morris, P. J.
Related Collections
Right arrow Mixed Wastes
Right arrow Radionuclides
Right arrow Heavy Metals
Journal of Environmental Quality 32:885-898 (2003)
© 2003 American Society of Agronomy, Crop Science Society of America, and Soil Science Society of America

TECHNICAL REPORTS
Heavy Metals in the Environment

Partitioning and Availability of Uranium and Nickel in Contaminated Riparian Sediments

Andrew G. Sowdera, Paul M. Bertsch*,a and Pamela J. Morrisb

a Savannah River Ecology Laboratory, The Univ. of Georgia, Drawer E, Aiken, SC 29802
b Marine Biomedicine and Environmental Sciences, Medical Univ. of South Carolina, Charleston, SC 29425

* Corresponding author (bertsch{at}srel.edu)

Received for publication March 4, 2002.

   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
The effects of iron oxides and organic matter on the partitioning and chemical lability of U and Ni were examined for contaminated riparian sediments from the U.S. Department of Energy's Savannah River Site. In sequential extractions of four sediments that ranged from 12.7 to 82.2 g kg-1 in organic carbon, U was found almost exclusively in moderately labile fractions (93% in acid-soluble + organically bound). Nickel was distributed across all operationally defined fractions, including substantial amounts in the very labile fractions (4–15% in water-soluble + exchangeable), noncrystalline and crystalline iron oxides (38–49%), and in the nonlabile residual fraction (25–34%). Aqueous U concentrations in 1:1 sediment–water extracts were highly correlated to dissolved organic carbon (DOC) (R2 = 0.96; p < 0.0001) and ranged from 29 to 410 µg L-1. Aqueous concentrations of Ni exceeded U by two to three orders of magnitude (124–2227 µg L-1) but were not correlated with DOC (R2 = 0.04; p = 0.53). Partitioning and solubility trends suggest that Ni availability is controlled primarily by iron-oxide phases, whereas U availability is dominated by naturally occurring organic carbon. Discrete mineral phases were also identified as nonlabile reservoirs of anthropogenic metals. In spite of comparably high sediment concentrations, Ni appears to be significantly more available than U in riparian sediments and therefore warrants greater consideration in terms of environmental consequences (i.e., transport, biological uptake, and toxicity).

Abbreviations: AA, acid-soluble fraction • AO, noncrystalline iron-oxide fraction • CD, crystalline iron-oxide fraction • CN, exchangeable fraction • DW, water-soluble fraction • DOC, dissolved organic carbon • EDX, energy dispersive X-ray • HA, manganese-oxide fraction • HF, residual fraction • PP, organically bound fraction • SEM, scanning electron microscope • SOC, sediment organic carbon • SRS, Savannah River Site


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
THE DISCHARGE OF METALLURGICAL process wastewater has led to extensive contamination of ground water and riparian sediments in the vicinity of M-Area on the U.S. Department of Energy's Savannah River Site (SRS) (Pickett, 1990; Evans et al., 1992; Arnett et al., 1995). The SRS is an 800-km2 (310-mi2) former nuclear weapons production facility situated in the Upper Atlantic Coastal Plain of South Carolina along the Savannah River near Aiken, SC (Fig. 1) . In 1954, manufacturing of aluminum-clad nuclear reactor targets began in M-Area, and some associated aluminum forming and metal plating wastes were directly discharged via sewer outfall into Tims Branch, a small second-order stream that drains the upland region around M-Area into the Savannah River via Upper Three Runs Creek. A fraction of effluents was diverted to the M-Area Settling Basin (MASB) beginning in 1958, and subsequent changes in waste management practices eventually eliminated direct discharges to Tims Branch in 1982. The MASB was taken out of service in 1985 and capped in 1989 following consolidation and stabilization of contaminated soils and sludges (Arnett et al., 1992).



View larger version (33K):
[in this window]
[in a new window]
  		Fig. 1. Savannah River Site (SRS) and study site location map. The SRS lies in the upper Atlantic Coastal Plain along the Savannah River in South Carolina. Past discharges of process wastewater to Tims Branch resulted in extensive contamination of sediments downstream from the M-Area sewer outfall, including sediments within the Steed Pond basin. Sediment sampling locations for Steed Pond (SP) and uncontaminated reference sites, that is, upper Tims Branch (TB00) and Boggy Gut (BG), are indicated on the SRS map and M-Area inset.

 
Effluents released to Tims Branch and the MASB included substantial quantities of trichloroethylene (TCE), tetrachloroethylene (PCE), nickel, depleted and natural uranium, and lesser amounts of other heavy metal ions ("heavy metals" or "metals" hereafter) including copper, zinc, lead, chromium, and thorium (Pickett et al., 1987). Releases of U from M-Area accounted for greater than 97% of the gross {alpha} activity introduced to the environment from SRS operations (Evans et al., 1992). Total releases of U to Tims Branch have been estimated at 43 500 kg, 61% of which occurred from 1966 to 1968 (Evans et al., 1992). While cumulative releases of Ni and other inorganic wastes to Tims Branch are unknown, MASB contaminant estimates (Table 1) and sediment concentrations suggest Ni releases on par with U and proportionately smaller quantities of Cr, Cu, Zn, and Pb, among others (Looney et al., 1987; Pickett et al., 1987; Pickett, 1990). These wastes were primarily comprised of dissolved metal ions, acids (nitric, phosphoric, and sulfuric), sodium hydroxide, and sodium nitrate (Pickett et al., 1987). However, it is likely that effluents associated with metallurgical processes (cutting, etching, plating, cleaning, and recovery) also entrained some solid phases, including precipitates of U (sodium uranate, hydrogen uranyl phosphate, and uranium oxide) and Ni (nickel hydroxide), as well as metallic fines.


View this table:
[in this window]
[in a new window]
  		Table 1. Estimated inventory of inorganics and organics released to M-Area Settling Basin and vicinity (Looney et al., 1987).

 
Severe erosion of upland soils accompanying wastewater discharges resulted in the covering of sandy stream sediments with metal-contaminated strata enriched in clay and silt. It is believed that sequestration by settling sediments captured most of the dissolved metals released to Tims Branch in a few key depositional areas along a 5-km stretch of the riparian corridor. An estimated 70% of U released may reside in Steed Pond, an abandoned farm pond predating the SRS, which exists today as a wetland following the failure of its spillway structure in 1984 (Fig. 1 inset; Pickett, 1990; Evans et al., 1992).

Transport of U and Ni out of the Tims Branch system during stream base flow is not a concern, as routine water sampling has found dissolved metal concentrations at or below background levels (Hayes and Ouzts, 1986), and any measurable concentrations would be greatly diluted on mixing with Upper Three Runs Creek. However, Batson et al. (1996) reported a 15- to 28-fold increase in U transport out of the Tims Branch system during storm events due to sediment erosion. Batson et al. (1996) also demonstrated via sequential extractions that the chemical lability of U in suspended sediments decreased over the course of a storm event as higher flows mobilized larger particles, indicating that the availability of U (and possibly other metals) depends on sediment texture (i.e., greatest availability in the clay fraction). In subsequent studies on in situ chemical stabilization of U, Ni, and other metals in Steed Pond sediments, Arey et al. (1999) and Seaman et al. (2001) reported a strong influence of organic carbon on metal solubility, as well as the key roles secondary iron and aluminum minerals play in the redistribution of U and Ni following apatite addition. Recently, Punshon et al. (2003) observed that Ni uptake in the tissues of plants and small mammals sampled in Steed Pond exceeded that of U by two to four orders of magnitude in leaves and two orders of magnitude in muscle.

The extensive heavy metal contamination of Tims Branch sediments is exacerbated by the subsurface migration of solvents from the MASB, resulting in large-scale vadose zone and aquifer contamination (Arnett et al., 1995). The outcropping of TCE- and PCE-contaminated ground water into Tims Branch is a concern, and natural and enhanced attenuation of this plume by native microbial populations has been proposed as a cost-effective remediation option (Bowman et al., 1993; Brigmon et al., 1998; Travis and Rosenberg, 1997). Riparian and wetland systems are valued for their natural attenuation capacity resulting from the synergy of high organic matter content, diverse microbial populations, and wide range of geochemical conditions, which promote the biodegradation of organics and the biotransformation of metals (Bradley et al., 1998, 1999; Gambrell, 1994; Gilliam, 1994; Reddy and Gale, 1994; Sobolewski, 1999). However, as SRS and many other government and industrial sites are contaminated with mixtures of organics, inorganics, and radionuclides at potentially toxic levels (Riley et al., 1992; United States Department of Energy, 1997), it is important to understand the effect of heavy metals on environmental remediation strategies in co-contaminant scenarios. Moreover, as water is a primary vector for contaminant transport and biological effects (i.e., exposure, uptake, and toxicity), it is especially important to understand the link between solid-phase partitioning and aqueous-phase solubility of heavy metals with respect to sediment organic and mineral matrices, source term characteristics, and long-term aging effects.

In this paper we examine several geochemical aspects of U and Ni as sediment contaminants in the context of a reasonably well-defined source (SRS M-Area sewer discharges) and aging in a dynamic, organic-rich riparian environment (Tims Branch and Steed Pond) over the course of decades. Sequential and nonsequential extractions and electron microscopy are employed to elucidate the role of iron-oxide minerals and sediment organic carbon (SOC) in the partitioning behavior and chemical lability of U and Ni. This work represents an extension of previous studies and provides a foundation for ongoing research on the partitioning, chemical speciation, bioavailability, and microbial toxicity of U, Ni, and other heavy metals in complex environments.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Sediment Collection
Bulk quantities of contaminated sediments (SP1B1, SP1B2, SP2B1, and SP5B1) were collected from areas within the Steed Pond basin exhibiting the greatest {gamma} activity from Pa-234m, a decay daughter of U-238 (Cember, 1996). Sediments SP1B1 and SP1B2 were collected from the same location (SP1) adjacent to Tims Branch at approximately 0- to 5- and 5- to 10-cm depths, respectively, and were distinguished based on visible stratification resulting from differences in SOC content. Sediments SP2B1 and SP5B1 were collected from separate locations further away from the stream channel at 0- to 5-cm depths. Since the clay- and silt-enriched sediments of Steed Pond are atypical of riparian sediments on the SRS, a clayey reference soil (TB00) was collected in the upland headwaters of Tims Branch, upstream from the confluence of the M-Area sewer discharge (Fig. 1). This region was previously found to be uncontaminated by A/M Area activities based on U and Th sediment concentrations (Evans et al., 1992). All sediments were collected as grab samples, sealed in polyethylene bags, and stored on ice while in the field. Sediments were sieved (<2 mm) and refrigerated (4°C) in field-moist condition. Maintenance of in situ redox conditions was not attempted in this investigation; accordingly, the sediments examined in this work should be considered as oxidizing, a conservative endpoint for evaluating uranium availability and toxicity.

Sediment Characterization and Analytical Methods
The micropipet method of Miller and Miller (1987) was used to size-fractionate sediments for USDA texture classification. Separate water-dispersible clay and silt fractions for characterization were obtained via centrifugation to minimize alteration of metal distributions and destruction of sediment aggregates that can accompany chemical pretreatment, chemical dispersion, and sonication (Batson, 1994). Sediment mineralogy was characterized using X-ray powder diffractometry (XRD) on an X2 automated diffraction system (Cu K{alpha} radiation source operating at -45 kV and 40 mA; solid state detector; Scintag, Cupertino, CA) along with standard thermal and chemical pretreatment methods (Whittig and Allardice, 1986). High-resolution thermogravimetric analysis (TGA) on a Hi-Res TGA 2950 (TA Instruments, New Castle, DE) and differential scanning calorimetry (DSC) on a TA Instruments DSC 921S were also used for confirmation of some mineral phases, including goethite and gibbsite (Jackson, 1979; Tan et al., 1986). A number of samples were deposited as ultradilute slurries onto 25-mm, 0.2-µm polycarbonate membranes (Poretics Corp., Livermore, CA), dried at 22°C and ambient humidity ("air-dried" hereafter), and carbon-coated for imaging and elemental analysis with a LEO 982 field emission scanning electron microscope (SEM) equipped for energy dispersive X-ray (EDX) analysis (LEO Electron Microscopy, Ltd., Cambridge, UK).

Total SOC was measured on a CNS-2000 analyzer (LECO, St. Joseph, MI). Dissolved organic carbon (DOC) was measured with a TOC-5000A total organic carbon analyzer (Shimadzu, Kyoto, Japan). Dissolved metal concentrations were measured by inductively coupled plasma–mass spectrometry (ICP–MS) (ELAN 6100 DRC; PerkinElmer Corp., Wellesley, MA) or inductively coupled plasma–optical emission spectrometry (ICP–OES) (Optima 4300 DV, PerkinElmer Corp.). Total metal loadings in sediments were measured by ICP–MS analysis following microwave digestion (MDS-2000, CEM Corp., Matthews, NC) with HF–aqua regia in PFA-Teflon pressure vessels.

All solutions were prepared with 18 M{Omega} distilled, deionized water and reagent-grade chemicals. All samples for analysis were filtered with 0.22-µm MSI Cameo nylon 25-mm syringe filters (Osmonics, Minnetonka, MN) and acidified (pH < 2) with ultrapure HNO3 for metal analysis or refrigerated (4°C) for DOC measurements. Sediment pH was measured in 1:2 sediment–distilled, deionized water suspensions using a glass junction combination electrode with calomel reference (Accumet Model 13-620-286; Fisher Scientific, Pittsburgh, PA). Cation exchange capacity (CEC) was calculated as the sum of major cations released following equilibration with 0.1 M BaCl2 (Rhoades, 1982). One-way analysis of variance (ANOVA) and linear regressions were performed using ORIGIN Version 7.0 (OriginLab, 2002).

Sequential and Nonsequential Extractions
A sequential extraction method based on the protocol of Miller et al. (1986) was employed to examine the distribution of metals in sediments; the method is summarized in Table 2 (Batson, 1994; Batson et al., 1996). Approximately 0.75 g of air-dried sediment was contacted with 30 mL of each reagent (15 mL for citrate–dithionite reagent) in sequence and shaken at 90 cycles per minute for prescribed conditions, centrifuged at 10 000 rpm for 30 min, filtered, and acidified for dissolved metal analysis. Residual metals in sediments were measured following microwave digestion of 0.2- to 0.3-g subsamples in 10 mL 48% HF–1 mL aqua regia (HF). A 20-mL 0.025 M Ca(NO3)2 wash step was applied between extractions and discarded. The manganese-oxide extractant step employing hydroxylamine–HCl (HA; NH2OH–HCl) was eventually eliminated from the protocol based on previous results (Batson et al., 1996, Arey et al., 1999; Seaman et al., 2001) and this work (see below).


View this table:
[in this window]
[in a new window]
  		Table 2. Summary of sequential extraction protocol as adapted by Batson et al. (1996) from Miller et al. (1986).

 
Reagents from the sequential extraction protocol targeting iron-oxide phases, that is, ammonium oxalate in the dark (AO) and citrate–dithionite (CD) steps (Table 2), were also applied individually to evaluate the distribution of contaminant metals in relation to noncrystalline and crystalline iron-oxide coatings that dominate the surface chemistry of SRS soils and sediments. To obtain volumes sufficient for analyses, porewater chemistry was approximated in 1:1 sediment–water extractions by equilibrating 15 g of air-dried sediment with 15 mL of distilled, deionized water for 14 d at 22°C. Filtered aqueous samples were submitted for metal and DOC analyses, as described above. Unless otherwise noted, all extractions, sequential and nonsequential, were performed in triplicate in 50-mL polycarbonate centrifuge tubes.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 CONCLUSIONS
 REFERENCES
 
All sediments, SP1B1, SP1B2, SP2B1, SP5B1, and TB00, were acidic in nature (pH 5.0–5.6), varied in SOC from 12.7 to 82.2 g kg-1, had relatively low cation exchange capacity (Table 3), and exhibited typical mineralogy for SRS surface soils and sediments (Table 4). Similar sediment textures were measured for SP1B1 and SP1B2 (i.e., silty clay loam) and for SP2B1 and SP5B1 (i.e., clay). Nickel and uranium concentrations in Steed Pond sediments were elevated two to three orders of magnitude above background, ranging from 400 to 1200 mg kg-1 for Ni and 500 to 2500 mg kg-1 for U. The highest concentrations were found closest to the stream (i.e., SP1 samples). Background concentrations for Ni and U for sediments from an upland clay-rich soil, characteristic of the clay- and silt-enriched sediments within Steed Pond, were 26.2 ± 1.2 and 8.15 ± 0.85 mg kg-1, respectively (TB00, Table 3). Average concentrations (n = 2) of 3.2 and 13.2 mg kg-1 Ni and 1.2 and 4.3 mg kg-1 U were measured in two shallow sediments from an uncontaminated stream, Boggy Gut, located near the SRS boundary, far-removed from site activities (Fig. 1; unpublished data, 2002). Comparable background concentrations of 14 mg kg-1 Ni and 2.0 mg kg-1 U were reported previously for the SRS (Pickett et al., 1987). Fay and Hayes (1984) reported 2.2 to 34.0 mg kg-1 of U in southeastern U.S. stream sediments; higher U concentrations were attributed to the naturally occurring mineral monazite.


View this table:
[in this window]
[in a new window]
  		Table 3. Total metal concentrations and other characterization data for contaminated and uncontaminated Savannah River Site sediments (±1 standard deviation).

 

View this table:
[in this window]
[in a new window]
  		Table 4. Representative mineralogy of Savannah River Site sediments.

 
The redox status of SRS stream and wetland sediments varies seasonally and with depth. Typical redox potential (Eh) values measured in shallow riparian sediments ranged from -100 to +600 mV (Bertsch, unpublished data, 1995). However, in spite of the dynamic redox environment expected in seasonal wetlands and riparian systems, previous spectroscopic characterizations of Tims Branch sediments found U predominately (>=80%) as the hexavalent uranyl species (Bertsch et al., 1994; Hunter and Bertsch, 1998). Moreover, previous and ongoing studies suggest that a significant fraction of U in these sediments is not subject to reduction to the tetravalent state via microbial or abiotic processes (Bertsch et al., 1994; Coughlin et al., 2000; Duff et al., 2000a), contrary to results from other U-contaminated soils, sediments, and laboratory systems (Duff et al., 1997, 1999, 2000b; Frederickson et al., 2000).

Uranium and Nickel Concentrations in Whole Sediments and Size Fractions
Total U and Ni concentrations were compared for whole sediments, the sand fractions, and water-dispersible silt and clay fractions from Steed Pond with similar textures and the lowest (12.7 g kg-1, SP2B1) and highest (82.2 g kg-1, SP5B1) SOC concentrations (Fig. 2) . For the lowest organic sediment SP2B1, the concentrations of U and Ni generally increased with decreasing size fraction, that is, [U]clay {approx} [U]silt > [U]sand and [Ni]clay > [Ni]silt > [Ni]sand. However, this trend was not observed in the high organic sediment SP5B1; instead, U and Ni loadings were not significantly different among the whole sediment and individual size fractions. Comparison of means was performed by one-way analysis of variance with post hoc Tukey and Scheffe tests for a 95% confidence level ({alpha} = 0.05).



View larger version (45K):
[in this window]
[in a new window]
  		Fig. 2. Distribution of (a) U and (b) Ni in whole, sand, and water-dispersible silt and clay fractions of SP2B1 (12.7 g kg-1 sediment organic carbon [SOC]) and SP5B1 (82.2 g kg-1 SOC). Error bars are one standard deviation (n = 3). Significantly different means ({alpha} = 0.05) are distinguished by lowercase letters for SP2B1 distributions as determined by one-way analysis of variance and post hoc mean comparisons using Tukey and Scheffe tests. Uranium and nickel distributions in SP5B1 are not significantly different among the four size fractions. Legend applies to both graphs.

 
Scanning Electron Microscope Observations
Qualitative SEM–EDX examinations of whole and size-fractionated sediment samples before and after sequential extraction provided direct information on sediment morphology and composition that complemented partitioning information obtained from chemical extractions. Iron was ubiquitous in all sediments and size fractions, occurring as iron-oxide coatings and discrete clay-sized nodules. The EDX analysis revealed most Fe minerals to be highly substituted. Associations of Ti with Fe were the most frequently encountered; however, Fe–Ti–Mn associations were also common. Figure 3a is a typical SEM image of the sediment clay-sized fraction in which both iron-coated (region "Fe") and uncoated (region "k") kaolinite particles were visible. As expected, Al also appeared to substitute in Fe coatings, although distinguishing Al in coatings via EDX was made difficult by ubiquitous Al-rich matrix minerals, such as kaolinite and gibbsite.



View larger version (150K):
[in this window]
[in a new window]
  		Fig. 3. (a) Scanning electron microscope (SEM) image with corresponding energy dispersive X-ray (EDX) spectra of representative clay sized particles in SP1B1 clay size fraction. Region "Fe" illustrates a typical iron-oxide coating with substitution of Ti and Mn. Region "k" corresponds to underlying kaolinite matrix. (b) SEM image of a large aggregate encountered in SP2B1 water-dispersible silt fraction. (c) SEM image of a Ni phase with Cu- and Ni-rich region indicated. Both metals are anthropogenic in origin. Phases such as this one may represent source term artifacts, surviving decades of weathering, or the resulting secondary weathering products.

 
Measurements of elemental compositions in sediments following individual treatments from sequential extractions were consistent with the selective dissolution of various phases and elements (Table 2). Clay fractions subjected to the HA extractant exhibited a noticeable reduction in EDX signatures for Mn. Clay fractions treated with AO displayed marked decreases in Fe abundance, virtual elimination of Ti and Mn, and a general increase in the uniformity of elemental compositions of Al–Si–O. Following CD extraction, very little Fe was detectable in EDX spectra.

A second prominent feature was the frequent occurrence of large aggregates (>50 µm) of clay-sized particles observed in water-dispersible silt fractions (Fig. 3b). The aggregate pictured is a mixture of kaolinite, Ti-bearing Fe nodules, and some Ti- and Zr-rich particles that appear bright in backscattered electron images (anatase and zircon, respectively). Numerous SEM–EDX observations revealed the partitioning of Ni, Cu, Zn, Cr, and Pb in precipitates or possible metallic forms that were morphologically and compositionally distinct from naturally occurring sediment minerals. Nickel was the most frequently encountered metal in these discrete forms, although Cu and Zn were also common. Figure 3c is an example of a Ni-rich secondary mineral with localized regions also enriched in Cu, possibly representing either source term artifacts that have survived decades of weathering or their secondary weathering products. No discrete U-bearing phases were observed using SEM–EDX; however, heterogeneous distributions of U were commonly observed via spatially resolved synchrotron X-ray fluorescence (Sowder, unpublished data, 2002).

Water-Soluble Metal Ions
Concentrations of metal ions in water extracts were measured after 14-d equilibrations of sediments with distilled, deionized water at 22°C in a 1:1 g mL-1 sediment to solution ratio (Table 5). The long equilibration time and high sediment to solution ratio were intended to estimate metal concentrations in porewater immediately available for biological uptake and environmental transport. While not a true porewater extraction, this approach was chosen because of the relatively large aqueous volumes required for multiple analyses and the limited quantity of sediment samples. Consequently, the results probably underestimated actual porewater concentrations due to dilution effects.


View this table:
[in this window]
[in a new window]
  		Table 5. Results from 1:1 sediment–water equilibration (±1 standard deviation) and correlations obtained from linear regressions.

 
Metal solubility generally tracked the presence of organic carbon in the system, with SOC giving positive correlations with dissolved concentrations of Ni (R2 = 0.38; p = 0.03) and U (R2 = 0.48; p = 0.01). Aqueous U concentrations were highly correlated to DOC (R2 = 0.96; p < 0.0001) and ranged from 29 to 410 µg L-1. Concentrations of Ni in solution exceeded those for U by two to three orders of magnitude (124–2227 µg L-1), but increases in dissolved Ni concentrations were not correlated with increasing DOC levels (R2 = 0.04; p = 0.53). While high for natural waters, these concentrations represent a small fraction of the total U and Ni inventories present in sediments.

Iron-Oxide Extractions
Nonsequential iron-oxide extractions were used to estimate the chemical lability of Fe and associated matrix metals in the sediments. Apparent Fe lability was greater in Steed Pond sediments (SP1B1, SP1B2, SP2B1, and SP5B1) relative to the uncontaminated upland soil (TB00) as the ratio of AO- to CD-extractable iron increased from 0.10 for the upland site to between 0.47 to 0.63 for Steed Pond (Fig. 4) . Results for Ti mirrored Fe trends, with an AO to CD ratio increase from 0.17 for the upland soil to between 0.58 to 0.71 for Steed Pond (Fig. 4a). Changes in Mn lability were less apparent; AO to CD ratios of 0.65 versus 0.73 to 0.84 were observed for the upland and Steed Pond samples, respectively (Fig. 4a). Results for Al were unique among the matrix elements examined (Fig. 4b); these AO to CD ratios exceeded 1.00 for all four contaminated Steed Pond sediments.



View larger version (36K):
[in this window]
[in a new window]
  		Fig. 4. Ratio of metals concentrations in ammonium oxalate (AO) and citrate–dithionite (CD) extractions for (a) Fe, Mn, and Ti and (b) Al as the chief constituents of upland soil and Steed Pond sediment coatings. Error bars are one standard deviation (n = 3).

 
Sequential Extractions
Sequential extractions of sediments were performed to provide information on the chemical lability and partitioning of sediment-bound U and Ni using two versions of the Miller protocol on SP1B1 (Fig. 5) : one incorporating and one excluding the HA extraction step intended to selectively dissolve manganese oxides and associated trace metals (Table 2). The large inventory of U formerly liberated in the HA step (49%) was captured entirely in the pyrophosphate (PP) extraction for organics (i.e., 78% for PP without HA vs. 80% for the combined HA + PP steps) with little or no carryover to subsequent extraction stages (Fig. 5a). Omission of the HA step lead to the redistribution of that fraction's Ni (8%) into the AO, CD, and HF fractions (Fig. 5b).



View larger version (32K):
[in this window]
[in a new window]
  		Fig. 5. Comparison of sequential extraction profiles performed with and without hydroxylamine–HCl (HA) for (a) U and (b) Ni. Arrows indicate apparent redistributions of U and Ni to subsequent extraction stages following omission of HA step. Sequential extractions with HA step were performed in duplicate only (n = 2). Cumulative metal concentrations extracted are presented in Table 6. Error bars are one standard deviation (n = 3) for non-HA sequential extraction. Legend applies to both graphs.

 

View this table:
[in this window]
[in a new window]
  		Table 6. Cumulative metal concentrations extracted in sequential extraction steps (±1 standard deviation) for Fig. 5 and 6 with recoveries in parentheses (relative to total metal concentrations from Table 3).

 
Sequential extraction results (HA extraction step omitted) for U and Ni as well as matrix elements of interests (i.e., Fe, Mn, and Al) for the four Steed Pond sediments are presented in Fig. 6 and Table 6. Uranium (Fig. 6a) was predominately distributed in the moderately labile fractions: acid-soluble (AA) + PP reagents (93%); the organically bound fraction was particularly dominant (70–78%). Less than 1% of total U was extracted in the very labile fractions, that is, water-soluble (DW) + exchangeable (CN). A 5 to 7% shift from the AA- to PP-extractable U occurred for increasing SOC. Only 6% of U in Steed Pond sediments was extracted by the iron-oxide reagents (AO + CD fraction).



View larger version (25K):
[in this window]
[in a new window]
  		Fig. 6. Sequential extraction results for Steed Pond sediments for (a) U, (b) Ni, (c) Fe, (d) Mn, and (e) Al. Sequential extractions were performed without hydroxylamine–HCl (HA) step. Data columns within each sequential extraction fraction are presented by increasing sediment organic carbon content (i.e., SP2B1, SP1B2, SP1B1, and SP5B1). Cumulative metal concentrations extracted are presented in Table 6. Error bars are one relative standard deviation (n = 3). Legend applies to all graphs.

 
In contrast to U, Ni was distributed across all operationally defined lability classes (Fig. 6b). The largest concentration appeared in the noncrystalline iron-oxide (AO) fraction (29–42%), and Ni in combined iron-oxide fractions (AO + CD) represented up to half of its total inventory in sediments (38–49%). A substantial amount of Ni was also found in the very labile CN fraction (4–15%). Nickel extraction in the less labile and nonlabile fractions was more variable than for U.

Iron was extracted chiefly in the moderate to nonlabile fractions (Fig. 6c), including PP (8–32%), AO (5–15%), CD (30–67%), and HF (17–26%). Manganese was distributed over all fractions (Fig. 6d), including the most labile fractions DW (<1–3%) and CN (10–49%). Aluminum lability and partitioning trends fell between those of Fe (less labile) and Mn (more labile) with a 7 to 10% distribution in the moderately labile AA fraction (Fig. 6e). A substantial nonlabile fraction of Al was also observed (40–64% in HF fraction) and was consistent with its role as matrix component of aluminosilicate minerals. Cumulative extractions of Al were generally much lower than other elements analyzed, that is, only 33 to 48% of total Al was recovered (Table 6).

Sequential extraction results for the uncontaminated reference soil (TB00) were very different from those for Steed Pond sediments (unpublished data, 2002). Naturally occurring Ni appeared almost exclusively in the residual (HF) fraction (89%). Naturally occurring U was also distributed predominately in the residual (HF) fraction (53%); however, a major fraction of background U was also found in the moderately labile (AA + PP) fractions (36%).


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
CONCLUSIONS
 REFERENCES
 
Sequential Extraction Application and Artifacts
The low cost and simplicity of chemical extractions as measures of chemical lability and partitioning of heavy metals and radionuclides have led to their widespread application in research (Clark et al., 1996). Sequential extractions employ increasingly aggressive extractants to selectively dissolve, desorb, or otherwise remove analytes from target mineral phases or assemblages. The limitations of these methods for metal speciation are widely recognized (e.g., Kheboian and Bauer, 1987; Nirel and Morel, 1990; Rapin et al., 1986; Tipping et al., 1985). The specificity of extractants is often questionable, and readsorption and redistribution of contaminant metals following each extraction step are important concerns. Hunter and Bertsch (1998) reported the redistribution of U and changes in its speciation in spectroscopic observations over the course of sequential extractions performed on contaminated SRS sediments. The application of sequential extraction data for quantitative prediction of biological endpoints, such as plant uptake, has been problematic (Hinton et al., 1998). However, in spite of these limitations, sequential extractions are useful for indirectly probing the association of metals with mineral phases, comparing chemical lability of metals in soils and sediments, and indicating the potential for biological uptake of heavy metals and radionuclides.

A modified version of a standard method (Miller et al., 1986) was selected based on its applicability to iron-rich soils and previous use for SRS soils and sediments (Arey et al., 1999; Batson, 1994; Batson et al., 1996; Clark et al., 1996; Seaman et al., 2001). High extraction efficiency for U in SRS soils and sediments using HA has been observed, raising the concern that this behavior may mask differences in U partitioning. Consequently, following the evaluation of the Miller method with and without the HA extraction, the step was eliminated. This modification can be justified on the basis of site-specific soil–sediment mineralogy. The relative proportion of Mn in the sediments is minor compared with Fe (Table 3). Furthermore, Mn does not occur in identifiable separate phases in SRS soils and sediments (Clark et al., 1996). It is found instead as a constituent of iron-oxide coatings as indicated in SEM–EDX inspection of sediment particle surfaces (e.g., Fig. 3) and in association with soil organic matter as indicated in increased lability for Mn in the high SOC sediment sequential extraction profile (Fig. 6d). Qiang et al. (1994) questioned the effectiveness of the HA reagent hydroxylamine hydrochloride as a Mn-specific extractant in the presence of organic coatings. Accordingly, the incorporation of a separate Mn extraction step was determined to be counterproductive for the comparison of elemental partitioning in Steed Pond sediments.

In the absence of the HA step, U was simply extracted by the next reagent, PP (Fig. 5a). Consequently, the HA step appeared to serve as a second, nonspecific, and more aggressive acid extraction for U, and the HA to CD extraction interval appeared much less specific for U versus other elements, including Ni. The nonspecificity of some extractants for U is understandable given that the method was originally developed for other metals such as Fe, Mn, and Cu (Miller et al., 1986), and the aqueous geochemistry of U is dramatically different from that of the transition elements and other heavy metals (Amonette et al., 1994; Langmuir, 1997). Especially important is the high solubility of U(VI) under alkaline conditions due to the formation of stable aqueous carbonate complexes. Accordingly, the high pH of the PP extractant (pH 10) may render it more aggressive and less selective for U than intended. The extraction of 26% of naturally occurring U from the low organic, uncontaminated reference soil (TB00) by PP is consistent with this nonspecific behavior.

Elimination of the HA step resulted in a more complex distribution of Ni into the AO, CD, and residual HF fractions, bypassing the PP fraction. This pattern suggests a specific Ni association with some HA-extractable phase, as both AO and CD reagents will extract HA-fraction metals (Means et al., 1978). The omission of the HA step and the application of a transition metal–oriented sequential extraction method for U represent important compromises. However, these compromises are offset by the convenience and utility of employing the same protocol for comparing the partitioning of U and Ni among several sediments.

The issue of scale is important when interpreting results from any extraction method for the purpose of estimating and comparing the partitioning and availability of metals. While it is tempting to dismiss a relatively small fractional recovery as insignificant, concentrations of U and Ni in sediments on the order of 100 to 1000 mg kg-1 can translate extracted fractions of a few percent into significant porewater concentrations. Accordingly, the extraction of Ni in the DW fraction (<1–2%) for Steed Pond sediments has important implications for transport, bioavailability, and toxicity, as demonstrated in 1:1 sediment–water equilibrations (Table 5).

Role of Iron Oxides
Soils and sediments of SRS are characterized by ubiquitous ferrihydrite and goethite coatings on soil mineral surfaces (Bertsch and Seaman, 1999). Colloidal Fe minerals have also been identified as important mobile phases in ground and surface waters (Seaman et al., 1997; Bertsch and Seaman, 1999). The importance of iron oxides in controlling trace element mobility in the environment is extensively documented (Bruemmer et al., 1986; Evans, 1989; Jenne, 1968; Means et al., 1978; Sposito, 1989; Stumm and Morgan, 1981). In a previous study of Steed Pond sediments, Bertsch et al. (1994) found approximately 75% of U in the sand fraction associated with Fe- and Mn-rich regions, and correlation of Ni with Fe was also strong.

In this study, an estimate of sediment iron-oxide crystallinity was obtained by comparing Fe lability in nonsequential extractions using the AO and CD reagents for noncrystalline and total extractable Fe, respectively. Upland SRS soils, the source of the clay and silt present in the shallow sediments of the Tims Branch system, exhibited less than 10% "amorphous" or poorly crystalline iron oxides (Fig. 4a). The contaminated sediments, in contrast, had 50 to 60% of the total Fe in noncrystalline form. This is expected for wetland and riparian sediments, in which seasonal variations in redox conditions lead to dissolution and reprecipitation of iron oxides with substantial substitution by structurally compatible metals such as Al, Ti, and Mn (Gambrell, 1994). The SEM–EDX analyses of Fe-rich mineral surfaces revealed a large fraction of iron-oxide coatings substituted with these elements (Fig. 3a). Of these three metals, Ti lability closely mirrored that of Fe in AO vs. CD extractions while Mn lability remained relatively constant. The correlation between Fe and Ti lability is consistent with the prominence of Ti in most iron-oxide coatings. X-ray diffraction, TGA, and previous SRS soil characterization (Arey et al., 1999; Batson, 1994; Batson et al., 1996; Ruhe and Matney, 1980; Seaman et al., 2001) indicate that goethite is the dominant crystalline iron-oxide mineral. However, appreciable amounts of colloidal and noncrystalline iron oxides should also be present in the Steed Pond sediments given favorable conditions for Fe dissolution and reprecipitation and based on experimental data from sequential and nonsequential extractions. Extensive removal of Fe, Ti, and Mn coatings during AO treatments supports the importance of noncrystalline iron oxides.

The AO–CD and sequential extractions suggested that more Al was available than accounted for in iron oxides. This pattern may represent the influence of noncrystalline aluminosilicates, which often control Al3+ solubility in acidic soils (McBride, 1994). Cumulative releases of Al to Tims Branch are unknown; however, it is likely that quantities were substantial given the prominent role of Al in M-Area processes (Pickett et al., 1987). This anthropogenic Al may also explain the high lability of Al observed in contaminated sediments.

Overall, the dissolution–precipitation of iron oxides represents an important aging mechanism for decreasing the lability of contaminant metals over time as they become structurally incorporated or occluded over many dissolution–precipitation cycles. Consequently, the dynamics of iron oxides and other surface coatings may explain, in part, the effect of aging on metal lability in Tims Branch and Steed Pond sediments. Changes in the reversibility of metal sorption on iron oxides with time, elevated temperatures, pH changes, wetting and drying cycles, and other phenomena have been reported in many studies (Ainsworth et al., 1994; Bell et al., 1991; Ford et al., 1997, 1999a; Rigol et al., 1999; Schultz et al., 1987).

Elemental properties, such as ionic radii and surface binding constants, also influence aging behavior (Ainsworth et al., 1994; Ford et al., 1997, 1999a). The crystalline radius of Ni2+ (0.069 nm) is comparable with other ions such as Mn4+ (0.060 nm) and Ti4+ (0.068 nm) that commonly substitute for Fe3+ (0.064 nm) in iron-oxide structures (Weast et al., 1983). Consequently, substitution of Ni into mineral coatings of sediments is favored and explains the observed association of Ni in noncrystalline and crystalline iron-oxide target fractions (i.e., AO + CD). Uranium (VI) is a larger ion, both as a free cation (0.080 nm) (Weast et al., 1983) and as the linear uranyl dioxo-cation UO22+ (Burns et al., 1996). Accordingly, U is not expected to substitute into iron-oxide matrices to the same degree as Ni, although recent studies have indicated some incorporation of U(VI) into iron oxides in a uranate-like coordination environment (Coughlin et al., 2000).

Iron-oxide coatings can also affect the distribution of heavy metals as a function of particle size. It is commonly assumed that the clay fraction in soils and sediments exerts a disproportional influence on metal binding due to the presence of negatively charged, highly reactive mineral surfaces with large specific surface areas (Sposito, 1984). However, coating of larger, less reactive phases by iron oxides and the formation of aggregates via cementation of smaller particles (e.g., Fig. 3b) can lead to significant departures from expected trends (Bertsch and Seaman, 1999). Iron oxides have also been proposed as reactive surface sites preconcentrating dissolved uranium to form uranyl mineral precipitates under solution conditions far below thermodynamic solubility limits (Murakami et al., 1997; Sato et al., 1997).

Surface precipitation of transition metals on clay mineral surfaces has been proposed as an important process resulting in long-term stabilization of heavy metals in soils. Surface precipitation of mixed Al–transition metal phases has been demonstrated for many clay mineral systems including Ni on hydroxy interlayered vermiculite (HIV)–kaolinite (Roberts et al., 1999) and pyrophyllite (Ford et al., 1999b; Scheidegger et al., 1996a,b, 1998); Ni, Cu, Zn, and Cd on montmorillonite (Lothenbach et al., 1997); and Co on kaolinite (Thompson et al., 1999a). The proposed mechanism involves the release of structural Al followed by the formation of hydrotalcite-like minerals (Bertsch et al., 1989; Ford et al., 1999b; Roberts et al., 1999; Thompson et al., 1999b). The abundance of Al in the Steed Pond system, both as a matrix constituent of aluminosilicates and as an anthropogenic contaminant, suggests a similar process may contribute to the stabilization of Ni and other compatible heavy metals in these contaminated sediments.

Role of Organic Carbon
The influence of organic carbon in sediments, porewaters, and the water column on the partitioning and availability of heavy metals is as important as it is complex (Evans, 1989; Gambrell, 1994; Sobolewski, 1999). Wetland and riparian environments are characterized by high organic carbon concentrations in the form of animal and plant detritus and humic and fulvic acids. In sediments, organic matter can have significant effect on the environmental mobility and bioavailability of metals due to the formation of soluble aqueous complexes, stationary coatings on soil mineral surfaces, aggregates, and mobile colloidal phases.

The importance of dissolved organic matter (i.e., humic and fulvic acids) in the geochemical cycling of metals in the surface waters of the southeastern USA is well established (Alberts and Giesy, 1983; Alberts et al., 1984). Giesy et al. (1986) identified humic and fulvic acids as the principal ligands for UO2+2 in southeastern streams and rivers, with approximately 25% of U occurring as a humic–fulvic complex and more than 40% as the free ion. For Steed Pond sediments, the solubility of U and Ni in water extracts generally increased with SOC (Table 5), and aqueous U concentrations were highly correlated with DOC. Nickel concentrations, in contrast, were poorly correlated with DOC. This trend suggests other geochemical factors, in addition to organic carbon, influence the Ni partitioning in Steed Pond sediments, consistent with the apparent role of iron oxides and refractory forms indicated in sequential extractions.

Sequential extraction results for the similarly textured low- and high-organic sediments, SP2B1 and SP5B1, respectively, revealed opposite trends for U and Ni as a function of SOC (Fig. 6). Increasing SOC resulted in a 5 to 7% shift of U from the AA fraction to the less labile PP fraction. Nickel exhibited an overall enhancement of its lability, as seen in increases in the CN and PP fractions at the expense of the less labile iron oxide (AO + CD) and residual (HF) fractions.

Metal analyses for whole sediments and water-dispersible clay, silt, and sand fractions point to the organic cementation of smaller particles, as do frequent SEM observations of sediment aggregates (Fig. 3b). The distribution of U and Ni in the low organic SP2B1 (Fig. 2) reflects the expected trend in metal loadings for each fraction, that is, [M]clay > [M]silt > [M]sand, which assumes increasing surface area with decreasing particle sizes and greater reactivity of clay minerals results in greater specific sorptive capacity for metals (Sposito, 1984). However, increased SOC in the similarly textured SP5B1 appears to counteract this trend. The formation of large aggregates of finer particles by organic matter cementation has been observed in estuarine and marine sediments as the result of drying treatment artifacts (Krumgalz, 1989) and natural processes (Hamilton, 1989), leading to concentration of metal contaminants in larger size fractions.

While the importance of organics for immobilizing metals in soils and sediments and mobilizing metals in solution is widely recognized, the convoluted interaction of organic carbon with other key controls such as Fe and Mn geochemistry makes it difficult to separate the effects of one from another. Moreover, increasing mobility and solubility accompanying organic complexation of metals does not automatically translate into increasing bioavailability or toxicity. To the contrary, organic complexation of metals often decreases toxicity (Campbell, 1995). Increasing sorption of metal–organic complexes to soil mineral surfaces with decreasing pH also occurs (Davis et al., 1993) and complicates overall sorption patterns.

Source Term Effects
Research on heavy metal–amended sludges has shown that the nature of the contaminant source term can have significant influences on heavy metal mobility and bioavailability in agricultural soils (Bell et al., 1991; Korcak and Fanning, 1985; Kuo et al., 1985). Source term effects have also been observed at many industrial and defense sites. For example, characterization of contaminated soils at sites associated with the production and testing of nuclear weapons showed that uranium and plutonium occur in a wide array of physical and chemical forms intimately linked to source terms and dis-charge histories (Bertsch et al., 1994; Bondietti and Tamura, 1980; Buck et al., 1995, 1996; Morris et al., 1996; Morse and Choppin, 1991; Tamura, 1976).

In Steed Pond sediments, discrete secondary phases were observed in many sediment samples, size fractions, and sequential extraction stages. Many of these phases incorporated Ni along with other anthropogenic metals (i.e., Cu, Cr, and Zn) (Fig. 3c) and were refractory. These distinct phases may represent artifacts of the initial source term or secondary source terms (i.e., precipitates or metallic forms) that survived intact or were altered in the environment. Such forms potentially explain the distribution of Ni in the less- and non-available fractions of sequential extractions and the poor or variable recovery of anthropogenic metals in some sequential extractions. Partitioning of heavy metals in discrete secondary phases represents a refractory inventory in sediments that must be considered when interpreting sequential extraction data and drawing conclusions regarding a contaminant's mobility and bioavailability in the environment.


   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
REFERENCES
 
Substantial differences in the partitioning and chemical lability of U and Ni were observed in contaminated sediments from a riparian–wetland environment on the SRS. In spite of comparable sediment loadings, Ni was much more available than U, exceeding U concentrations in sediment–water extracts by two orders of magnitude or more. Sequential extractions revealed that U was predominately distributed in the moderately labile acid-extractable and organically bound fractions. Nickel, in contrast, was distributed across all fractions including substantial amounts in the very labile water-soluble and exchangeable fractions, as well as in the refractory residual fraction. Such distinctions are critical for understanding the potential for environmental transport and biological uptake, and this information is in turn required for evaluating ecological and human risk, tailoring remediation methods, and implementing prudent environmental management strategies.

With the aid of detailed sediment characterization data and an understanding of process, waste, and discharge histories, it is possible to attribute control of environmental availability to one or more dominant factors at work in the complex, heterogeneous system presented by the contaminated sediments of Tims Branch and Steed Pond. Not surprisingly, key geochemical controls include the abundant iron oxides and naturally occurring organic carbon. Each can serve to either immobilize heavy metals as stationary sinks, or to enhance mobility through colloidal transport or increasing metal solubility via the formation of stable aqueous complexes. Both Fe and natural organic matter are important for U and Ni cycling, but not to equal degrees.

Nickel association with extractable iron oxides in sequential extractions far exceeded that for U. On the other hand, U recovery in organically bound fractions dominated its distribution in Steed Pond sediments, and dissolved U concentrations in sediment–water extractions were highly correlated with measured DOC concentrations, unlike Ni. Accordingly, iron oxides appear to provide a compatible matrix for Ni substitution or coprecipitation, which is consistent with physical arguments and experimental evidence. Strong associations of U with naturally occurring organics appear to control U availability in one or more possible ways including immobilization as a uranyl species sorbed onto SOC, complexation in solution as a stable uranyl–humic or –fulvic acid complex, or sorption onto mineral surfaces as a uranyl–organic complex. Nickel geochemistry also appears to be directly linked to primary or secondary source term effects as indicated in frequent SEM observation of refractory Ni phases.

Metal partitioning in sediments from Steed Pond is consistent with subtle processes that, over the course of decades, can result in greater incorporation of sorbed species into mineral matrices, formation of more stable surface complexes and precipitates, and decreasing reaction kinetics for desorption and dissolution. Further delineation of metal distribution and speciation in these sediments is the subject of ongoing investigations applying more direct aqueous- and solid-phase chemical speciation methods. As little work has been conducted on aged, co-contaminated systems, these results are valuable for understanding the biogeochemical cycling of Ni and U in riparian and wetland environments. However, understanding the geochemistry of metal contaminants is only one piece of the bigger picture required to describe the effects of metals on biological communities and ecosystems. Real progress in understanding the environmental fate and consequences of metals and radionuclides follows the integration of geochemistry, biology, and ecological research in addressing the complexity of chemically and biologically driven processes in natural systems. Consequently, continuing studies are also focused on establishing clear links between chemical speciation, bioavailability, and biological endpoints such as toxicity and trophic transfer at micro- and macroscopic scales.


   ACKNOWLEDGMENTS
 
This research was made possible by Grant ER62696-1011950-0003828 from the EPA/DOE/NSF/ONR Joint Program on Bioremediation administered by the U.S. DOE Biological and Environmental Research (BER) Program and was supported in part by Financial Assistance Award no. DE-FC09-96SR18546 from the U.S. DOE to the University of Georgia Research Foundation. We thank Paul Bayer at DOE-BER for his interest in and support of this project. Dr. Mark Farmer and Dr. John Shields of the University of Georgia Center for Advanced Ultrastructural Research provided access to and assistance with electron microscope facilities. The helpful comments of Dr. C. Strojan and J. Van Nostrand on earlier manuscript drafts are also appreciated. The manuscript was greatly improved by the contributions of three anonymous reviewers and the associate editor, Dr. Michael Ebinger of Los Alamos National Laboratory.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 


Related articles in JEQ:

This Issue in Journal of Environmental Quality

JEQ 2003 32: 745-750. [Full Text]  



This article has been cited by other articles:


Home page
 J. Environ. Qual.Home page
T. Punshon, A. Lanzirotti, S. Harper, P. M. Bertsch, and J. Burger
Distribution and Speciation of Metals in Annual Rings of Black Willow
J. Environ. Qual., June 7, 2005; 34(4): 1165 - 1173.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Related articles in JEQ
Right arrow Similar articles in this journal
Right arrow Similar articles in Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Web of Science (11)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Sowder, A. G.
Right arrow Articles by Morris, P. J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Sowder, A. G.
Right arrow Articles by Morris, P. J.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Sowder, A. G.
Right arrow Articles by Morris, P. J.
Related Collections
Right arrow Mixed Wastes
Right arrow Radionuclides
Right arrow Heavy Metals

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Crop Science
Journal of Natural Resources
and Life Sciences Education		 Vadose Zone Journal
Soil Science Society of America Journal Journal of Plant Registrations The Plant Genome