VZJ Download to Citation Manager
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 17 May 2007
Published in Vadose Zone J 6:354-362 (2007)
DOI: 10.2136/vzj2006.0072
© 2007 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
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 Similar articles in this journal
Right arrow Similar articles in Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Web of Science (1)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Serkiz, S. M.
Right arrow Articles by Clark, S. B.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Serkiz, S. M.
Right arrow Articles by Clark, S. B.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Serkiz, S. M.
Right arrow Articles by Clark, S. B.
Related Collections
Right arrow Fate
Right arrow Radionuclides
Right arrow Geochemical Processes

SPECIAL SECTION: SAVANNAH RIVER SITE

Environmental Availability of Uranium in an Acidic Plume at the Savannah River Site

S. M. Serkiza,*, W. H. Johnsonb, L. M. Johnson Wilec and S. B. Clarkd

a Savannah River National Laboratory, Westinghouse Savannah River Company, Aiken, SC 29808
b Nuclear Engineering and Health Physics Programs, School of Mechanical Engineering, Georgia Institute of Technology, Atlanta, GA 30332
c Environmental Engineering and Science Dep., Clemson Univ., Clemson, SC 29634
d Dep. of Chemistry, Washington State Univ., Pullman, WA 99164-4630
* Corresponding author (steven.serkiz{at}srnl.doe.gov).

All rights reserved. No part of this periodical may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher.




   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 Methods
 Results
 Conclusions and Implications
 REFERENCES
 
Uranium partitioning in soils collected from an acid and U impacted sandy Coastal Plain aquifer at the Savannah River Site (SRS) was investigated. The influences of hydrologic regime (vadose zone or saturated zone), proximity to the source input (impacted or unimpacted soils), and soil weathering (field or laboratory-spiked soils) on the environmental availability U were examined. Environmental availability (availability for groundwater transport) was operationally defined using a sequential extraction technique and was applied to vadose zone, saturated zone, and background soils. For saturated zone locations, matched porewater–soil sets of field samples were collected, and data generated from these samples were used to examine U partitioning under field conditions. Laboratory batch sorption studies of uranyl ion to background soils were conducted as a function of pH. Subsequently, the soil used in the sorption study was subjected to sequential extraction to investigate the environmental availability in laboratory spiked samples. Based on sequential extraction behavior of U-impacted soils and background soils and the acidic plume chemistry, U concentrations in the first three sequential extraction steps [deionized water, CaCl2, and acetic acid/Ca(NO3)2] were operationally defined as available, and the final two extraction steps (crystalline iron oxide and residual extraction steps) were operationally defined as unavailable. Based on this operational definition, soils impacted by the acidic U plume exhibited a greater fraction of available and total U. Vadose-zone soils had a smaller fraction of available U than corresponding saturated zone samples. Sequential extractions of U sorbed to background soils in a short-term laboratory experiment showed greater U availability compared with field soils collected within the contaminant plume. Field-derived Kd values ranged from 0.1 to 300 L kg–1 and were highly correlated with porewater pH.

Abbreviations: CEC, cation exchange capacity; ICP–MS, inductively coupled plasma (argon)–mass spectrometry; PZSE, point of zero salt effect; SCM, surface complexation model; SRS, Savannah River Site; TIC, total inorganic carbon; USDOE, United States Department of Energy.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
Methods
 Results
 Conclusions and Implications
 REFERENCES
 
A variety of laboratory and field techniques have been used to investigate contaminant partitioning and availability in environmental systems. Availability, for the purposes of this paper, is defined as "the ability of a soil to maintain an aqueous concentration of U in the soil solution" (Amonette et al., 1994). Partitioning data (e.g., the partition coefficient, Kd) are generally designed to investigate equilibrium processes, whereas the concept of availability encompasses not only equilibrium partitioning but also the rate at which contaminants are released to or removed from the porewater. Connecting the concepts of partitioning and availability, however, has been a difficult task, and our work attempts to explain field availability using partitioning data acquired by sequential extraction. Because extraction conditions (e.g., temperature, extract chemistry, and contact duration) do not exactly match field conditions, data derived from this approach are necessarily operational in nature. This, however, is not necessarily negative because conducting extractions over a range of geochemical conditions or target mineral phases provides predictive capability when geochemical conditions in the field are changing.

Laboratory techniques to measure the partitioning of contaminants to soils and mineral phases include batch and column studies. These techniques have been applied in the investigation of U sorption to a wide variety of model phases (Turner, 1995; Waite et al., 1994; Leiser et al., 1992) and geologic materials (Ticknor, 1994). Despite their wide application, laboratory approaches are not without problems. Freeze and Cherry (1979) listed the major disadvantages of batch sorption studies as "sample disturbance and the lack of representation of field flow conditions," which "detract from the validity of the results in the analysis of field situations." Numerous investigations have shown that batch studies tend to systemically overestimate the ability of a system to retard contaminant transport. This is, at least in part, because batch methods increase the surface area of the soil sample, thereby exposing aqueous-phase contaminants to new sorption sites that are not present or assessable under field conditions (Ryan, 1982; Ivanovich, 1993). Because of these problems, Kd values reported from batch studies may actually provide an upper limit of the Kd of a system (Ryan, 1982). Ryan (1982) stated that "undisturbed soil columns should give more reliable and conservative information than batch methods can about radionuclide migration velocities in the natural groundwater systems." Column studies, however, often fail to provide an accurate estimate of contaminant retardation because of difficulties in obtaining an undisturbed soil sample and inherent problems with channeling. Flow through the column is usually faster than in natural systems, and experimental flow rates and column dimensions have been shown to affect retardation data generated using this approach (Relyea, 1982).

For batch and column studies, partition data are necessarily valid only for the chemical conditions (e.g., pH, ionic strength, and concentration of competing dissolved species) under which the data were collected. Furthermore, because these methods are generally designed to be conducted under conditions approaching equilibrium, kinetic information is generally not obtained by these methods.

Field techniques for analyzing radionuclide partitioning include the "in situ Kd" approach (i.e., trace element analysis of matched porewater–sediment samples) (Jackson and Inch, 1989; Ivanovich, 1991; Landström et al., 1982) and field-derived contaminant transport rates (Coles and Ramspott, 1982). The shortcomings of the in situ Kd approach have been discussed in detail in McKinley and Alexander (1993) and include the inability to distinguish the bonding strength and/or mechanism (e.g., sorbed, precipitated, or matrix bound) in solid-phase samples. Additionally, other authors have noted that the process of coprecipitation and the presence of colloids in field samples may make the interpretation of these data difficult. McKinley and Alexander (1993) indicated that few field sites are amenable for developing pH-edge or isotherm sorption data (i.e., constant total contaminant concentration) from field samples because it is difficult to control sample condition (e.g., pH and contaminant concentration) for field samples.

In an attempt to characterize reactivity of the contaminant pool in soils, some researchers have turned to selective extraction (e.g., sequential extraction) techniques (Amonette et al., 1994; Beckett, 1989; Chao, 1984). As noted by many researchers (Tessier et al., 1979; Rapin et al., 1986; Kheboian and Bauer, 1987; Nirel and Morel, 1990; Xiao-Quan and Bin, 1993), these techniques are also not without their experimental limitations. Problems with sequential extractions have been well documented and include incomplete extraction of trace elements or target phases, nonselectivity, and readsorption of extracted contaminants onto existing or newly created surfaces. Thus, individual sequential extraction phases may not adequately represent the discrete soil phase to which the contaminant is bound. The actual concentration extracted may only be assumed to be associated with a given operationally defined phase (Quevauviller et al., 1994). Therefore, to stress the operationally defined nature of the approach, we refer to the sequential extraction phases by the extraction step number (1–8) rather than the intended extraction phase.

Because of the limitations of individual techniques, applying multiple techniques to validate results is often more meaningful. This study presents a comparison of laboratory data for U availability derived from batch sorption experiments using an uncontaminated background soil with that of field samples impacted by an acidic plume. Field samples consisted of matched porewater–soil sets collected from an acid-impacted waste site at the United States Department of Energy (USDOE) Savannah River Site (SRS) near Aiken, SC. The distribution of U associated with eight operationally defined soil phases was estimated using sequential extraction techniques (Clark et al., 1996). This sequential extraction approach provides an estimate of the fraction of U that is potentially available for transport in the subsurface. Sequential extraction data collected from field samples were also used to examine the influence of geochemical conditions of study site soils (water saturation, soil pH, and total U concentration) on the distribution of U between each of the extraction steps. Additionally, uranyl ion was sorbed to background soils in a series of laboratory batch sorption studies, and subsequently, sequential extractions were employed on the solids from this experiment to examine the ability of this laboratory technique to represent U partitioning and availability under field conditions.


   Methods
TOP
 ABSTRACT
 INTRODUCTION
 Methods
Results
 Conclusions and Implications
 REFERENCES
 
Field-Site Description
From 1955 until 1988, the F-Area Seepage Basins at the SRS received process wastewaters and stormwater runoff from the F-Area tritium and plutonium separation facilities. Theses three unlined basins in the F-Area were designed to allow the natural processes of evaporation and infiltration to dispose of effluent streams. Throughout the history of the basins, normal operation and unusual events resulted in the addition of large amounts of NaOH and HNO3 to the basins. These additions caused fluctuations in basin aqueous pH from a value of 0.6 to 13.2 (Fenimore and Horton, 1972). During routine operation, the technical standards for basin operation required that waste solution pH be maintained between 3 and 10; the basins were generally operated at the low end of this range (Fenimore and Horton, 1972). These operations caused groundwater in the vicinity of these basins to become acidic (see Fig. 1B) and levels of metals, radionuclides, and nitrate to become elevated. Groundwater monitoring data for this site during the time frame samples were collected for this study are summarized in Table 1. As a result of basin operations, soils downgradient of the basins have been significantly altered through accelerated acid weathering. In 1988 discharge to the basins was terminated, liquid was removed, basins were filled with a gravel bed topped with layers of calcium carbonate and blast furnace slag, and a multilayered cap was placed over the basins.


Figure 1
View larger version (33K):
[in this window]
[in a new window]

  		FIG. 1. (A) Sampling transect plan view, (B) pH, (C) 238U aqueous, and (D) 238U soil.

 

View this table:
[in this window]
[in a new window]

  		TABLE 1. Average and range of reported groundwater contaminant concentrations in the Savannah River Site F-Area between the first quarter of 1990 until the third quarter of 1992 (from Johnson, 1995). Includes all data from all aquifers monitored by downgradient wells during this period.

 
Kaplan et al. (1994, 1995) examined colloid formation and colloid facilitated transport at this site. They isolated colloidal material from three groundwater monitoring wells hydraulically downgradient from one of the F-Area Seepage Basins (A-1 Basin) and a background well not impacted by the operations of the seepage basins. They used filtration studies as well as microprobe elemental analysis of isolated colloids to investigate contaminant association with colloids. The authors found no evidence for colloidal transport of contaminants in the two wells immediately downgradient of the seepage basin. As the pH of the system moved to more neutral conditions in the most downgradient well, however, evidence for increased colloidal association was observed (e.g., 89% of the U was associated with filterable-size fractions). Microprobe elemental analysis indicated a strong association of contaminants with colloids containing Fe and Ti. Colloid counts in the plume were reported to be four orders of magnitude lower in plume than in the control well.

Johnson (1995) used the data from soil and porewater samples collected in 1993, described below, to evaluate sorption and one-dimensional transport models for Cd, Cs, and U at the subject waste site. He concluded that the U concentrations in sequential extraction Steps 7 and 8 were relatively constant and represent naturally occurring U in SRS soils and, additionally, that the U in these two phases did not participate in exchange reactions with groundwater. Surface complexation models (SCMs) (constant capacitance, double-layer model, and triple-layer model) were fit to laboratory-batch sorption data using the cation exchange capacity (CEC) of the background soil as the total binding site concentration. The fits of all three SCMs were not very good for lab and field U sorption data. Additionally, an empirical fit of Kd as a function of pH was developed from the matched soil–porewater samples described below. The mathematical form of the function was an exponential rise, and fitting parameters from this exercise were used in a one-dimensional transport model of U through background soil. These modeling results were then compared with the results of laboratory column experiments. As expected, the transport of U was controlled by the porewater pH, and although the model took into account potonation–deprotonation reactions on the soil surface (i.e., soil buffering), it overpredicted the buffering response of the soil and hence underpredicted U transport relative to the laboratory studies.

Johnson (1996) conducted a semiquantitative clay mineralogy study of soil samples collected at the waste site in 1993 and found that kaolinite was the dominant clay mineral (~90%), with a lesser amount of illite (~10%) also present (see Table 2 for a summary of individual sample mineralogical data). Particle-size analyses showed an upward fining (i.e., greater clay content nearer the surface). An analysis of the quality of the kaolinite present showed an increase in the crystallinity of kaolinite with decreasing porewater pH. Goethite (FeOOH) was determined to be the primary metal oxide present on the surface of most soils; noticeably absent in these soil samples was gibbsite [Al(OH)3]. Because of the amorphous nature of goethite, this methodology did not use quantification by X-ray diffraction analysis. Alternatively, the results of extractable amorphous Fe, extraction Step 6 from the sequential extraction methodology described below ("Sequential Extraction") were used to quantify amorphous Fe and Al mineral phases in these soils (see Table 2). Generally, the amorphous Fe and Al extracted in Step 6 were greatest for samples collected immediately downgradient of the basin and were inversely correlated to porewater pH. Furthermore, samples from near the basin appear to be enriched in amorphous Fe relative to Al.


View this table:
[in this window]
[in a new window]

  		TABLE 2. Semiquantitative clay mineralogy and extractable iron and aluminum data for A-transect soil samples.

 
Field-Derived Partitioning Data
Matched sets of soil and porewater samples were collected at the same time, depth, and location in the summer of 1993 under inert conditions from the saturated zone downgradient of the site using an electric friction-cone penetrometer system (Ebasco Services, 1993). The cone penetrometer fitted with a hydrocone was used to collect porewater samples and a geocone for the associated soil sample. Before sampling, the hydrocone was purged using N2 gas to minimize gas exchange between the sample and the atmosphere that could affect water chemistry measurements (e.g., Eh and pH). All sample transfers from this device were made under N2 gas. Vadose zone soil samples were collected using the same method except that only soil cores were collected. Six sampling transects (identified as A–F) were selected to provide samples spanning a range of geochemical conditions (e.g., contaminant concentration and pH). The location for one of the sampling transects is shown in Fig. 1A (plan view) and Fig. 1B, C, and D (cross-section). Groundwater flow is generally downward and to the right in the region depicted in Fig. 1B, C, and D. In sampling transect A, five locations (A1–A5) were sampled at discrete depths. Samples are identified by sampling transect, surface location, and depth. (e.g., identification A23 is the third vertical sample from surface location 2 of transect A).

Porewater samples were analyzed for temperature, pH, redox potential, and conductivity immediately upon sampling and before filtering through a 0.45-µm filter. Total organic and inorganic carbon concentrations were measured using an automated carbon analyzer. Elemental composition of porewater samples was analyzed quantitatively using inductively coupled plasma (argon) mass spectrometry (ICP–MS) for 43 isotopes (including 235U and 238U) representing 28 elements. The complete results of these analyses are reported in Boltz et al. (1995).

Characterization of the physical properties of soil samples consisted of particle-size analysis (Gee and Bauder, 1986), CEC (Rhoades, 1982), organic carbon content, and acid-base titration to determine the point of zero salt effect (PZSE) (Uehara and Gillman, 1981). Elemental composition of the soil samples was determined by hydrofluoric acid total digestion (Lim and Jackson, 1982) and analysis of the digestion solution by ICP–MS as described above.

Sequential Extraction
Uranium distribution between eight operationally defined soil phases was estimated using the sequential extraction technique summarized in Table 3. This method, adapted from Miller et al. (1986) and described in greater detail by Clark et al. (1996), was developed for southeastern coastal plain soils containing metal oxy-hydroxide coatings (e.g., hydrous ferric oxides) similar to those found at the SRS. The method involved sequentially leaching soils from the study site with increasingly aggressive chemicals to remove contaminants from both operationally defined binding sites (Steps 1–3) and specific mineral phases within the soil (Steps 4–8). Soils were extracted in a 1:40 soil to solution ratio (g soil:mL extraction solution) with the following eight extraction steps:

  1. Water soluble constituents removed by shaking in deionized water for 16 h.
  2. Easily exchangeable constituents removed by shaking for 16 h in 0.5 M Ca(NO3)2 solution at pH 5.5.
  3. Specifically sorbed constituents removed from the soil surfaces by a solution of 0.44 M CH3COOH and 0.1 Ca(NO3)2 (8 h; pH 2.5).
  4. Contaminants associated with easily reducible metals such as manganese oxides removed with a solution of 0.01 M NH2OH–HCl and 0.1 M HNO3 (30 min at 50°C; pH 1.1).
  5. Organically bound contaminants extracted with 0.1 M Na4P2O7 (24 h; pH 10).
  6. Contaminants associated with poorly crystalline aluminosilicates and hydrous oxides extracted using a 0.175 M (NH4)2C2O2–0.1 M C2H2O4 solution in the dark (4 h; pH 3.5).
  7. Contaminants associated with crystalline iron oxides removed by reducing Fe3+ to Fe2+ with 0.75 g Na2S2O4 and complexed with a citrate buffer made with 0.15 M Na3C6H5O7 and 0.05 M HOC(CH2CO2H)2CO2H (30 min at 80°C; pH 5.5).
  8. Residual constituents solubilized by digestion in 1:10 mixture of hydrofluoric acid and aqua regia in an acid bomb at 105°C for 3 h.


View this table:
[in this window]
[in a new window]

  		TABLE 3. Summary of soil sequential extraction experimental procedure.

 
Soils and extraction solution were separated by centrifugation at 10,000 rpm (13,000 g) for 10 min; the supernate was pipetted off and filtered through a 0.45 µM cellulose nitrate syringe filter and acidified to 1% v/v with 70% ultrapure HNO3. After each extraction step, with the exception of Step 1, the soil residue was rinsed with a 0.01 M Ca(NO3)2 wash solution. This was intended to prevent released contaminants from becoming rebound to the soil. Between extractions for Steps 1 and 2, a deionized water rinse was used as the rinse solution.

All digestions and sequential extractions steps were conducted in triplicate. ICP–MS analyses of the individual extracts were conducted three times, and extract concentrations were converted to soil concentrations based on the sample weight and extract volume. Concentrations of individual replicates were averaged using a weighted average method (Knoll, 1989).

Background Soil Batch Sorption Studies
Batch sorption studies of uranyl ions to an uncontaminated background soil were conducted as part of this study. An uncontaminated background soil sample was collected from the same lithological unit upgradient of the area of known contamination. To develop an uranyl sorption edge as a function of pH for this soil, 1.04 g of this soil was allowed to react with 42 mL of 10–5 M UO2(NO3)2 solution that was pH-adjusted to between 3 and 12 by the addition of HNO3 or NaOH. For each initial pH, two soil-solution mixtures and one blank were equilibrated overnight in a shaking water bath at 25°C. After equilibration the soil solutions were centrifuged at 10,000 rpm (13,000 g) for 10 min. Liquids from centrifugation were filtered through a 0.45-µm filter. After measuring the equilibrated solution pH, samples were acidified to 1% v/v with ultrapure HNO3 and analyzed for U concentration by ICP–MS.

Soil samples where no acid or base was added (final pH of 4.2) were subjected to sequential extraction using the method described above. Soil extracts were analyzed by ICP–MS as already described. These data were compared to sequential extraction results generated on soils collected from the U-contaminated field site to examine the affects of contaminant aging in a qualitative manner.


   Results
TOP
 ABSTRACT
 INTRODUCTION
 Methods
 Results
Conclusions and Implications
 REFERENCES
 
We focus here on (i) field-derived data (e.g., major ion chemistry of the porewater, soil mineralogy and chemistry, and 238U concentration distribution in both porewater and soil); (ii) U-availability data from sequential extraction of field samples; (iii) analysis of the influence of the geochemical conditions (e.g., vadose vs. saturated, pH, and [238U]) of the field site on the availability of U; and (iv) the estimation of contaminant transport factors (i.e., Kd) from field-derived porewater and availability data.

Field-Derived Data
Porewater chemistries spanned a wide range of geochemical conditions. Measured pH varied from 3.1 to 7.1, while Eh values were between +41 and +442 mV. The major ion chemistry in groundwater at the site is consistent with the dissolution of clay minerals (e.g., kaolinite) and existing surface mineral coatings [e.g., iron (oxy)hydroxides] as a result of the acidification of the aquifer. As a point of comparison, equilibrium aluminum (Al) concentrations were calculated for porewater at the pH of each of the field samples in contact with an infinite source of kaolinite [Al2Si2O5(OH)4] and gibbsite [Al(OH)3]. This calculation was made using the solubility constants for these two minerals contained in the USEPA's MINTEQA2 thermodynamic metal speciation code (Allison et al., 1991) without activity correction and assuming congruent dissolution. The results, along with measured porewater Al concentrations, are reported in Table 4 and indicate that Al concentrations are more consistent with kaolinite dissolution than equilibrium with gibbsite. Consistent with this interpretation is the mineralogical analysis by Johnson (1996) that detected gibbsite in only a small number of the samples. Total inorganic carbon (TIC) content was found to be below the method detection limit of 1 mg L–1 in all samples. The TIC results indicate that carbonate complexation of uranyl is probably not important in the speciation of U in this system as expected due to the low pH of the system.


View this table:
[in this window]
[in a new window]

  		TABLE 4. Aqueous phase aluminum and uranium data for A-transect groundwater samples.

 
Particle-size analysis of study soils resulted in classification of 73% of the soils as sand or loamy sand and 25% sandy loam or sandy-clay loam, with the remainder being sandy clay. Mean CEC of the soils was 9.6 ± 6.3 mmol charge kg–1 soil. The total carbon content was determined for 27 of the soil samples, and all results were less than the method detection limit of 0.05% w/w carbon. The net surface charge of all soils measured was found to be low, less than 0.5 mmol charge kg–1 soil, with a representative PZSE of 3.8.

238U concentrations in porewater samples were above the detection limit in 43 of 54 samples collected for this study and ranged from 8.2 x 10–5 to 3.2 mg L–1. 238U concentrations were above the detection limit in all soil samples and ranged from 0.49 to 19 mg kg–1. Data for 238U in the "A" sampling transect are summarized in Table 4.

Field Samples: Uranium Availability
Contaminant availability has been defined as the ability of a source (e.g., soil) to provide contaminants to porewater (Amonette et al., 1994). Environmental factors such as quantity of reactive mineral phase (e.g., concentration and porosity), their quality (e.g., types of minerals present), porewater chemistry (e.g., pH, redox potential, the presence of competing ions, ionic strength), and total contaminant concentrations all influence the equilibrium partitioning of metals in subsurface systems. Rates of mineral-phase dissolution–precipitation, sorption–desorption, and contaminant diffusion all affect the observed contaminant availability in subsurface systems that are not at equilibrium or steady state. Given the large number of processes and their generally high degree of complexity, we and other researchers have attempted to approximate availability of contaminants operationally through the use of sequential extraction techniques, where the operational natures of these extraction schemes include variations of experimental conditions such as extraction time, solution chemistry (e.g., pH, competing ions, ionic strength, and presence of chelating agents and reductants), and solid-to-liquid ratio. We have chosen an eight-step sequential extraction (see Table 3) to operationally define availability in the study system. In this extraction scheme, the first three extraction steps are designed to extract sorbed contaminants. Steps 4–7 target contaminants associated with characteristically reactive (e.g., Step 4, easily reducible) or specific mineral phases (Step 5, organically bound; Step 6, amorphous oxides; and Step 7, crystalline iron oxides). Finally, Step 8 is designed to quantify contaminants that are residual in the soil matrix.

Defining "environmentally available" and "refractory" fractions of a contaminant pool based on sequential extraction results is also operational. As described in greater detail below, we chose this distinction based on (i) similarities between the geochemistry of the contaminant plume and the extraction chemistry for the first three steps of the sequential extraction method; (ii) the U availability observed by sequential extraction for a background sample; and (iii) the U extraction behavior of background soils subjected to uranyl batch sorption experiments.

Our sequential extraction approach defined naturally occurring U in the study-site soils as that associated with the soil matrix crystalline iron oxides (Step 7) and (Step 8, residual). This fraction of the U pool in our soils is clearly refractory and, therefore, assumed to be unavailable for release to the groundwater. Steps 7 and 8 should not be used in the calculation of available concentration of the source term or in use in the calculation of transport factors (e.g., Kd), and we have operationally defined this fraction of the U pool as unavailable. A plot of total soil 238U concentration versus the sum of 238U soil concentrations in Steps 7 and 8 is shown in Fig. 2. The lower the slope of the line between the origin and any point in this plot, the greater the fraction of available U in the sample. The samples with the greatest fraction of available U fraction are those with the highest total U concentrations. With the exception of the vadose sample A31, the vadose zone samples exhibited among the lowest available U fraction. This is thought to be due to leaching of the more mobile U by rain infiltration in vadose soils.


Figure 2
View larger version (10K):
[in this window]
[in a new window]

  		FIG. 2. Total 238U vs. sum of 238U soil concentrations in extraction Steps 7 and 8.

 
We demonstrated previously that the sorption behavior of U at this site can be largely estimated by the aqueous pH (Johnson et al., 1995), which is consistent with this dataset (see Fig. 3). As the geochemistry of the aquifer and soils becomes more acidic, U sorption is significantly reduced. In the sequential extraction method used for this study, the pH range of the first three extracts is 2.5 to 5.5, which spans the range of pH observed in the field samples from the acid-impacted contaminant plume. Thus, it is reasonable to assume that sorbed U removed in the first three extraction steps may act as a source term for U transport under acidic subsurface conditions. The U associated with the intermediate extraction steps (Steps 4–6) is much less straightforward based on our results. Ultimately, the sum of the first three extraction phases was used to calculate sorbed U concentration. This assumption results in conservatively low Kd values estimated from these data. If, however, the calculation of a source term from soil concentration is of interest, then a definition of available that included some or all of the intermediate extraction steps would be more conservative.


Figure 3
View larger version (8K):
[in this window]
[in a new window]

  		FIG. 3. 238Uranium Kd as a function of pH from field-derived data.

 
Figures 4A and 4B show the results from the extraction of an uncontaminated background soil and a soil downgradient of the basins that has been impacted by the acid plume but not by the U plume. The distribution of U in the sequential extraction steps is similar in these two samples, and, as would be expected, the majority of the U is associated with the later, more refractory, extraction steps. Furthermore, the acid-impacted sample (Fig. 4B) has a slightly smaller available fraction than the background soil (Fig. 4A). This suggests that some of the natural U has been leached from the acid-impacted sample.


Figure 4
View larger version (19K):
[in this window]
[in a new window]

  		FIG. 4. 238Uranium sequential extraction percentages for (A) background soil, (B) vadose zone unimpacted soil, and (C) batch sorption to background soil.

 
Laboratory-Spiked Soil: Uranium Availability
Perhaps the most compelling evidence for defining the first three extracts as the available fraction can be gleaned from sequential extractions on background soils that were subjected to laboratory sorption studies (see Fig. 4C). The equilibration time for U sorbed to the background soils in the laboratory experiments was short (24 h) relative to U sorbed under field conditions from the contaminant plume. Thus, removal of U from these soils was not expected to require aggressive extractants. The results of this extraction (Fig. 4C) show that unavailable concentration of U associated with Steps 7 and 8 is relatively constant between the background soil (Fig. 4A) and the spiked soil (Fig. 4C), indicating, as expected, that the short time period used in the laboratory treatment is not sufficient to associate contaminants with the unavailable fraction. The majority of the U, however, remains sorbed with water and Ca(NO3)2 extracts (Steps 1 and 2), and is not removed until Step 3 involving extraction with acetic acid and Ca(NO3)2. This suggests that U observed in Step 3 is a source term for transport in an acidic plume. About 85% of the U added in the batch study is associated with this available fraction. The remaining 15% of added U is associated with the intermediately available fraction (Steps 4–6). Compared with field samples (Fig. 5A and 5B), the batch sorption soil exhibited a greater available fraction, and the extraction profile was more similar to contaminated saturated than to contaminated vadose soils. These results indicate that in the time frame of hours, high aqueous spike concentrations can access lower availability and stronger binding sites within study soils, and, most likely due to the short equilibration times, the relative distribution of contaminant availability has not reached equilibrium (i.e., it does not match field sample availability). Caution must therefore be exercised when attempting to use laboratory-manipulated background soils to represent contaminant availability in contaminated systems.


Figure 5
View larger version (27K):
[in this window]
[in a new window]

  		FIG. 5. 238Uranium sequential extraction percentages for (A) vadose zone impacted soil and (B) saturated zone impacted soil.

 
Influence of Site Geochemistry on Uranium Availability
Soils impacted by the U plume have also been affected by the acid plume and generally exhibit lower pH values and higher U concentrations. The fraction of available U is greater in soils impacted by the U plume (Fig. 5A and 5B) than either of the soils not impacted by the U plume (Fig. 4A and 4B). The concentration of the refractory U (total U – available U) is, however, relatively constant in both cases. These data further support the definition of extraction Steps 1–3 as the available fraction.

The effect of soil moisture content on U availability was investigated. A comparison of the extraction results for two vadose zone samples that were impacted by the U plume (Fig. 5A) with two saturated zone samples also impacted by the U plume (Fig. 5B) shows that the fraction of available U is much greater for saturated zone samples than for those collected in the vadose zone. This greater fraction of available U in the saturated zone samples was observed for both samples impacted and not impacted by the U plume.

Field-Derived Partition Data
Analysis of experimental data and modeling of this system suggests that adsorption is the primary mode of U partitioning to the soils (Johnson et al., 1995; Johnson, 1995). The concentration of sorbed 238U in soil samples (defined as that extracted by Steps 1, 2, and 3) ranged from 171 to 3072 µg kg–1. Expressed as a percentage of the total soil 238U concentration, this fraction ranged from 9 to 99% The fraction sorbed was calculated by

Formula
where (U)sorbed is the sorbed concentration of U, (U)aq is the aqueous concentration of U, {rho} is the soil bulk density (1.2 g cm–3), and {eta} is the porosity (0.25). Uranium distribution coefficients (Kd) for each sample were calculated by dividing the sorbed U concentration (in µg kg–1) by the U concentration of associated porewater sample (in µg L–1). Kd values ranged from 0.1 to 300 L kg–1 and are presented in Table 4, along with sorbed and aqueous U concentrations and aqueous pH. Kd values exhibit a marked increase above pH 3.5 (see Fig. 3), suggesting that U partitioning differences between porewater and soils at this waste site can be explained primarily by changes in aqueous pH and, presumably, the associated pH-dependent changes in U solution speciation and soil surface charge.


   Conclusions and Implications
TOP
 ABSTRACT
 INTRODUCTION
 Methods
 Results
 Conclusions and Implications
REFERENCES
 
Our work shows that contaminant partitioning data can be generated directly from matched soil–porewater field samples. In this and a previous study of the site (Johnson et al., 1995), U sorption could be explained largely in terms of aqueous sample pH. The field-observed sorption data has a sorption edge in the pH range from about 3 to 4.5, with Kd values ranging from 0.1 to 300 L kg–1.

Care must be taken, however, to account for the amount of refractory contaminant that is unavailable for reaction with the aqueous phase. Results of sequential extractions indicated that the U extracted in the first three phases of the Miller et al. (1986) method was available for reaction under acidic geochemical conditions.

Sequential extractions of U from soils with different geochemical regimes show three general trends. Soils impacted by the U plume had a greater fraction of available U than either unimpacted or background soils. The fraction of available U in saturated zone soils was greater than in vadose zone soils for soils, regardless of impact by the U plume. Finally, sequential extractions of batch sorption studies with background soil do not provide a good prediction of U availability for the study area soils.


   ACKNOWLEDGMENTS
 
This research was supported by the USDOE under contact DE-AC09-89SR18035, administered by Westinghouse Savannah River Company, contract DE-AC09-76SR00819, administered by the University of Georgia's Savannah River Ecology Laboratory and by an appointment by W.H. Johnson to the Nuclear Engineering and Health Physics Fellowship program administered by Oak Ridge Institute for Science and Education. Special thanks are extended to Ms. Mira Malek and Ms. Frances Wakefield for laboratory assistance.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 Methods
 Results
 Conclusions and Implications
 REFERENCES
 





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 Similar articles in this journal
Right arrow Similar articles in Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Web of Science (1)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Serkiz, S. M.
Right arrow Articles by Clark, S. B.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Serkiz, S. M.
Right arrow Articles by Clark, S. B.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Serkiz, S. M.
Right arrow Articles by Clark, S. B.
Related Collections
Right arrow Fate
Right arrow Radionuclides
Right arrow Geochemical Processes

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