Published online 23 January 2008
Published in Vadose Zone J 7:67-78 (2008)
DOI: 10.2136/vzj2007.0076
© 2008 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
ORIGINAL RESEARCH
Effect of Particle-Scale Heterogeneity on Uranium(VI) Transport in Unsaturated Porous Media
D. M. Wellmana,*,
A. P. Gamerdingerb,
D. I. Kaplanc and
R. J. Sernea
a Pacific Northwest National Lab., 902 Battelle Blvd., P.O Box 999, Richland, WA 99354
b Formerly of Dep. of Soil Water and Environmental Science, 429 Shantz Building no. 38, 1200 E. South Campus Dr., P.O. Box 210038, Univ. of Arizona, Tucson, AZ 85721-0038
c Savannah River National Lab., Building 773-43A, Room 215, Aiken, SC 29808
* Corresponding author (dawn.wellman{at}pnl.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.
Received 19 April 2007.
 |
ABSTRACT
|
|---|
Uranium(VI) sorption and transport was evaluated in mixtures of silt loam and coarse sand sediments using traditional static batch sorption, saturated column, and unsaturated centrifugation experiments to evaluate the association of mobile and immobile water domains with particles of different size and surface reactivity. Exclusion of conservative tracers and a decrease in U sorption compared with what was predicted by the mass-averaged equilibrium distribution coefficient (Kd-mass-avg) was observed in sediment mixtures where the mass fraction of silt loam was 10%. This is consistent with behavior that was previously reported for coarse and fine sand separates. No exclusion of the conservative tracer, as predicted for the moderate water content range, was measured during unsaturated transport in sediment mixtures that contained 30% or more silt loam by mass. Sorption under unsaturated conditions was greater than predicted based on the batch sorption measurement of Kd-mass-avg value, however, which suggests that the fine-textured silt was in contact with the mobile water domain. This is the first evidence linking sorption to transport in a particular water domain. Results of this investigation demonstrate that the interaction between the geochemical and hydrodynamic processes has a profound effect on transport in unsaturated sediments. Definition of the fraction of mobile water was especially important for defining the front of the breakthrough curve, which is integral to predicting the arrival time of solutes at a particular depth and location in the sediment.
Abbreviations: BTC, breakthrough curve • MSC, coarse sand sediment • S/L, silt loam sediment
|
INTRODUCTION
|
|---|
Anionic radionuclides, 129I, 75Se, and 99Tc, and anionic complexes of 238U have been identified as the key long-term dose contributors in low-activity waste (Mann et al., 2001; Wood et al., 1995). Previous analyses for proposed disposal actions on the Hanford Site, southeastern Washington, have shown that groundwater transport presents the greatest potential for long-term dose uptake by humans (Mann et al., 2001). Geochemical conditions within the Hanford subsurface, typical of the arid western United States, are dominated by sands and gravel, which afford little surface area. Hanford groundwater is dominated by Ca, Na, and SO4, and has a pH of 
8.5 with dissolved CO32– concentration of 1.13 x 10–3 mol L–1 (Kaplan and Serne, 1995). Uranium, as uranyl UO22+, is predicted to form carbonate complexes in Hanford groundwater, 27% as UO2(CO3)22–, 68% as UO2(CO3)34–, 3% as UO2(OH)20 and 2% as UO2(OH)31– (Kaplan et al., 1996). These conditions limit the sorption of neutral and anionic aqueous complexes, making these radionuclides key risk drivers at the proposed disposal site due to their potential for migration through the surrounding subsurface environment via mass flow or diffusion (Serne et al., 1989, 1992, 1993, 1995).
Within the USDOE complex, U has been recognized as one of the two most frequently occurring radionuclides in groundwater and the most frequently occurring radionuclide in soils and sediments (Riley and Zachara, 1992). Accurate prediction of the long-term fate of U and efficient remediation strategies require a thorough understanding of its behavior. Uranium, present as uranyl (UO22+), is generally retained in sediments through ion exchange or surface complexation mechanisms (Duff and Amrhein, 1996; McKinley et al., 1995; Turner et al., 1996; Waite et al., 1994). The mobility of U under water-saturated conditions has been shown to be highly dependent on geochemical and mineralogical conditions (Barnett et al., 2000; Duff and Amrhein, 1996; Giblin et al., 1981; Hsi and Langmuir, 1985; Kohler et al., 1996; Rovira et al., 2000; Tripathi, 1983; Voudrias and Means, 1993). Considerable effort has been expended during the past few decades to quantify the mobility of nuclear wastes and determine factors that influence radionuclide mobility (Ames et al., 1976; Relyea and Serne, 1979; Serne et al., 1977, 1993; Wolfsberg, 1978). The average moisture content of Hanford vadose zone sediments ranges from 4 to 7% (v/v), however, which is equivalent to 10 to 20% saturation (Gee and Heller, 1985; Schalla et al., 1988). At low water contents, stagnant or immobile water domains develop and result in "mobile–immobile" water and "two-region" transport (Biggar and Nielsen, 1962; Gaudet et al., 1977; Nielsen and Biggar, 1961). Immobile water may exist in thin liquid films around soil particles and dead-end pores, or as relatively isolated regions associated with unsaturated flow (Nielsen et al., 1986). The impact of mobile–immobile water domains on sorption during transport in unsaturated sediments has not been adequately reported in the literature.
Previous research (Kaplan et al., 1996) illustrated that U sorption, as characterized by an apparent distribution coefficient (Kd-ap), varied as a function of water content; however, the trend was not consistent among sediment types. In a coarse-grained sediment, the Kd-ap values decreased with decreasing water content. In two fine-grained sediments, the Kd-ap values increased with decreasing water content. A conceptual model was proposed to explain these differences based on the mineral distribution and the extent to which fine-grained particles form hydraulically continuously connected pores. Upon lowering the water content in the coarse-grained sediment, the water may have "bypassed" the smaller pores, creating regions of immobile water, because they were not continuous throughout the length of the entire column. These smaller pores are created from smaller clay materials that have high densities of adsorption sites. In the finer grained sediment, the Kd values may have increased when water content was lowered because a greater percentage of the water traveled through the smaller pores that were continuous through the entire column. These results, as well as those of Nielsen and Biggar (1961) and Biggar and Nielsen (1962), contradict the assumption that the degree of adsorption remains constant as a function of saturation.
Subsequent research demonstrated an increase in the proportion of immobile or stagnant water with decreasing moisture saturation for Hanford sediments, particularly in coarse and fine sands (Gamerdinger and Kaplan, 2000). Additional data for U sorption confirmed the prior results, indicating a decrease in sorption (Kd-ap value) with decreasing moisture in coarse and fine sands (Gamerdinger et al., 1998). The change in sorption was correlated with the development of mobile–immobile water at lower water contents. Results suggested that (i) hydrodynamic conditions resulting in two-region transport restricted access to a fraction of the reactive sites associated with the immobile water domain, and (ii) formation of immobile water effectively increases the velocity of water in the mobile domain as a result of a reduction in the effective pore volume and corresponding decrease in sorption of U(VI) in silt-textured (Gamerdinger et al., 2001b) and coarse-textured sand (Gamerdinger et al., 2001a) sediments.
This investigation was motivated by previous work indicating that U sorption varied as a function of water content, i.e., sorption during transport in an unsaturated sediment was not predicted by sorption that was measured in a nonflowing batch or saturated column systems (the two most common techniques to measure Kd). We hypothesized that changes in sorption with water content are in response to hydrodynamic changes and that solute accessibility to reactive surfaces changes as the transport regime changes from one region to two region, to two region with isolated water. This research focused on physical mechanisms by which sorption may change with water content. A series of sediments with textures ranging from coarse sand (MSC) to silt loam (S/L) was used to assess changes in sorption during transport under unsaturated conditions. Experimental methods used included nonflowing batch systems, saturated columns, and unsaturated columns within a centrifuge. Once sorption data were collected on individual sediments with distinct particle-size distributions, sediment mixtures were created by combining the coarse sand with the silt loam. Hypotheses were developed to link the observed changes in sorption to hydrodynamic changes that occurred as sediments were progressively desaturated and to link these changes to the dominant particle texture.
|
Conceptual Models for Hydrodynamic Conditions at Various Water Contents and Supporting Laboratory Results
|
|---|
Several mechanisms may cause sorption to vary during transport in unsaturated sediments. The term ideal transport behavior is used to describe cases where sorption is rapid, reversible, and described by a linear isotherm, and where all of the water is in the mobile domain (one-region flow). Sorption can be affect by nonideal transport behaviors including isotherm nonlinearity, rate-limited sorption, and two-region transport. Isotherm nonlinearity indicates that the sorbed and solution phase concentrations do not sustain a constant ratio; during transport, isotherm nonlinearity results in breakthrough curve (BTC) asymmetry. Breakthrough curve asymmetry can also result from rate-limited sorption (i.e., slow kinetics or mass transfer). For rate-limited sorption during transport, a two-site sorption model often applies (Cameron and Klute, 1977; Selim et al., 1976). According to the two-site model, a fraction of sorption attains a rapid equilibrium and a fraction is rate limited or time dependent. Rate-limited sorption has been previously discussed in detail (Gamerdinger et al., 2001b). Nonlinear and rate-limited sorption are not expected to vary with moisture content; however, it is necessary to consider the possible contributions of these processes to changes in U sorption during transport in sediments with varying water content.
The changes in the water flow regime that were noted above for unsaturated sediments may affect the accessibility of mobile solutes to reactive surface sites. In a saturated sediment, all of the pore space is water filled; the pores are continuous or "connected," and, generally, all pores are water conducting. The water is considered "mobile" and solutes in the water are transported by the mechanisms of advection and dispersion. Stagnant regions, where water is held tightly by the matrix, can exist and are known to occur in sediments containing clays and aggregates. In this situation, the transport of solutes to the stagnant or "immobile water" domain is controlled by diffusion. As the water content is lowered in an initially saturated sediment, the large pores drain first and air becomes a barrier to water flow. Water flow in unsaturated sediments may occur as film flow along the particle surface or as "matrix flow" through smaller, water-filled pores. In very dry sediments, there is an increase in the stagnant or immobile water domain and a smaller fraction of the sorption sites contact the mobile water domain. Thus, solutes undergoing advective transport will not be exposed to the sediment surfaces of the stagnant water domain; solute transfer to these interstitial regions and surfaces is limited to diffusion processes from mobile to immobile water.
Gamerdinger et al. (2001a,b) have previously shown that the fraction of immobile water increased from essentially zero in saturated sand or silt loam sediment to 0.8 in an unsaturated sand at 14% moisture saturation. These physical changes in water distribution and the mobile water domain may affect sorption to the solid phase. Results presented in this study were collected across a wide range of water contents under well-defined hydrodynamic conditions. These conditions provide a mechanistic basis for evaluating the observed changes in sorption. Based on these results, the conceptual framework, water domains, and water content ranges for which different transport models (i.e., one-region and two-region models) will apply will be presented in a future study. Briefly, "high," "moderate," and "low" water content ranges are proposed that correspond to different hydrodynamic conditions for which different transport models are appropriate. The typical approach to transport (one-region model) corresponds to the higher water content range, the well-established phenomenon of mobile–immobile water (two-region model) corresponds to the moderate water content range, and a variation of the two-region model that includes isolated water (two-region with isolated water) corresponds to low water content conditions. Additionally, it is important to note that the proposed models presented here assume the "fine-grained" size fraction of the sediment (e.g., silt and clay) constitutes the largest sorption capacity and smallest pore sizes and could be associated with either the mobile or immobile water domains, depending on the proportion of fines in the sediment and the water content of the bulk sediment. As a smaller fraction of a mineral surface is exposed to the mobile aqueous phase, the effective sorption capacity of the sediment may change.
|
Theoretical Methods and Process Descriptions
|
|---|
We use the terms hydrodynamic changes, changes with water content, and decreasing water content to refer to changes in the transport regime that occur as water content changes from saturated to progressively unsaturated conditions. The term saturation refers to moisture saturation (volume fraction of a pore or void space that contains water) unless otherwise specified. The governing equations for sorption and transport are presented below.
Sorption and Retardation
The simplest representation of the sorption process is the linear equilibrium distribution model:
 | [1] |
where se and ce are the sorbed and solution concentrations at equilibrium, respectively, and Kd is the equilibrium distribution coefficient. Uranium sorption on a mixture of two sediments can be determined from the equilibrium distribution determined for each independent sediment:
 | [2] |
where Kd-mass-avg is the mass-averaged Kd value, M is the mass of sediment, and the subscripts 1 and 2 designate the two individual sediments. Equation [2] assumes that all surfaces for sorption remain available to U in solution. The applicability of Eq. [2] was verified by measuring sorption in well-mixed batch experiments (further discussed below).
In modeling contaminant transport in subsurface sediments, the retardation factor, R, encompasses sorption and reflects the velocity of the solute relative to that of water. An R value of 1 indicates that the solute is traveling at the average pore water velocity, R < 1 indicates that the solute is moving faster than the average pore water velocity, and R > 1 indicates retardation due to sorption. The retardation factor is calculated in two ways:
 | [3a] |
where Ref distinguishes "effective" retardation that is calculated directly from the BTC based on the average pore water velocity, v; the subscripts w and i represent water and the solute of interest, respectively; R is typically calculated from the equilibrium distribution coefficient for sorption, Kd, where 
b is the bulk density and 
is the volumetric water content of the sediment. The extent to which contaminants adsorb to sediments is reflected in the Kd value and assumes a constant ratio between the sorbed and solution concentration (linear sorption isotherm, Eq. [1]) and the rate of sorption is equal to the rate of desorption. If sorption attains equilibrium, retardation during transport should equal retardation calculated from the Kd value:
 | [3b] |
This research focused on sorption during transport in unsaturated sediments, specifically on identifying mechanisms to explain observations where the effective retardation measured during transport in unsaturated sediments, Ref, was not predicted from Kd values that were determined using batch incubation techniques:
 | [3c] |
Apparent Kd values, Kd-ap, can be calculated from Ref:
 | [4] |
and are distinguished from Kd values that are measured directly using the batch incubation technique.
One-Region Transport
One-region transport applies to solutes in a single water or flow domain. Under hydrodynamic conditions where one-region transport applies (and the velocity is sufficiently slow to allow diffusion to all sites), solutes have access to all reactive surfaces and sorption is well predicted by the batch Kd value. Solute transport models based on the convection–dispersion equation (applied to advective transport) were used to interpret effluent BTCs, which depict the dimensionless effluent concentration, c/co, as a function of cumulative pore volumes eluted. In dimensionless terms, when sorption is linear, the equation is
 | [5a] |
where
 | [5b] |
and co and c are the solute input and effluent concentrations, respectively; t is time; L is the column length (sediment bed length); x is distance from the input; and D is the hydrodynamic dispersion coefficient. The retardation factor, R, is based on the center-of-mass of the BTC and defines the velocity of water relative to that of the sorbing constituent (U in this study), and sorption is quantified, as defined in Eq. [1] and [3a]. The Peclet number, P, includes the term for hydrodynamic dispersion, D.
Two-Region Transport
A two-region transport model is used when water exists in two domains, i.e., mobile (subscript m) and immobile water (subscript im). Advective transport is restricted to the mobile water phase, while transport in and out of the immobile water domain is diffusion limited. Thus, access to reactive surfaces associated with the immobile water domain is diffusion limited. Sorption is much more dependent on velocity than during one-region transport. Sorption and transport of a solute in a two-region flow regime is described by the following equations (van Genuchten, 1981; van Genuchten and Wierenga, 1976):
 | [6a] |
where c is the solute concentration, t is time, x is distance, is the average volumetric water content, f is the fraction of sorption occurring in the mobile domain, b is the soil bulk density, K is the equilibrium distribution coefficient for sorption, v is the average pore water velocity, and 
is the mass transfer coefficient between the mobile and immobile water domains. The terms on the left depict sorption to sediments in contact with the mobile and immobile water domains. With this formulation, a single distribution coefficient, K, applies to both domains; in the application of the model presented here, K represents Kd or Kd-ap. Reactive surfaces (sorption "sites") in the immobile domain are accessible by diffusion; the presence of immobile water does not change total sorption (i.e., K is constant). The following dimensionless terms are introduced:
 | [6b] |
and the equations simplify to the form
 | [6c] |
and
 | [6d] |
where 
m is the fraction of mobile water and co is the concentration of the step input of solute.
|
Experimental Methods
|
|---|
Batch Sorption Isotherm
A nonflowing batch method (Relyea et al., 1980) was used to determine sorption isotherms for U on the MSC and S/L sediments (referred to as separates) and mixtures of these two separates. The total mass of sediment, M, and volume, V, of solution were 3 g and 7 to 8 mL, respectively. The mass/volume ratio was selected to achieve a measurable decrease in the U concentration after contact with the sediment and to retain sufficient U in solution for accurate analytical determinations. Successful attainment of these goals was verified in a single-point trial for U sorption on each of the four sediments or mixtures before performing the full study.
The sediment mixtures for which isotherm data are available are abbreviated as S/L(0.73) and S/L(0.27), which had 73.3 and 26.7% (w/w), respectively, of the silt loam combined with MSC. The separates and mixture sediments were first pre-equilibrated with uncontaminated Hanford groundwater. This was accomplished by adding groundwater to the solids, shaking the suspensions overnight, centrifuging, measuring the pH, and then pouring out the supernatant. This pre-equilibration was repeated twice, such that the pH of the groundwater solution did not change after it was combined with the equilibrated sediment. The purpose of this pre-equilibration was to isolate the radionuclide sorption reaction from other reactions that may occur while sediments and aqueous solutions equilibrate.
After pre-equilibration with Hanford groundwater and removal of the solution, fresh Hanford groundwater containing UO22+ as UO2(NO3)2·6H2O, was added. Five separate initial U concentrations, co, were used, approximately 16, 58, 110, 270, and 550 µg L–1. Each U concentration–sediment mixture was run in triplicate; separate sediment-free and radionuclide-free samples were also prepared in triplicate. After shaking for 1 wk, the suspensions were centrifuged and the equilibrium U concentration in the supernatant, ce, was determined. The sorbed concentration of U was determined by difference, assuming that the difference between co and ce was due to U sorption to the sediment. The soil-free control vials indicated that there was no loss of U to the labware. The U-free suspensions indicated that the co solution was the only source of U, and there was no measurable amount of U leached from the separates and the sediment mixtures. Aqueous U was quantified using inductively coupled plasma–mass spectrometry.
Saturated and Unsaturated Column Methods for Transport Experiments
General column methods (i.e., saturated) are described first, followed by the centrifuge method. Miscible displacement column techniques have been used extensively in our research and the methods have been previously reported (Gamerdinger et al., 1994). Briefly, columns were packed with dry sediment and saturated with Hanford groundwater; a minimum of 10 pore volumes of solution were passed through the column before initiating tracer injection experiments. A pore volume is defined as the volume of water held within the packed sediments' pore space. Sediment bulk density, b, and the average volumetric water content, , were determined by the mass of the sediment or water and the bulk volume. The moisture saturation percentage was calculated from the ratio of the volumetric water content (water-filled porosity) to the total porosity, E, estimated from the bulk density and particle density. Steady-state flow and water content were established before tracer injection, which is important for mechanistic investigations and when interpreting BTCs.
For the standard approach, a step input (pulse) of solution containing the chemical tracer (e.g., U, Br–, or pentafluorbenzoic acid) is applied for a number of pore volumes that is dependent on the sorption properties of the tracer. The column is then flushed (at the same volumetric water content) with tracer-free aqueous (background) solution. Effluent samples are collected throughout the experiment, analyzed, and dimensionless BTCs are developed. The BTC summarizes the temporal concentration distribution (normalized effluent concentration vs. the pore volumes eluted). Breakthrough curves are then analyzed with appropriate transport models.
A similar approach was used for unsaturated columns. The suitability of the centrifuge method for assessing solute transport has been established; details of the experimental system and advantages and disadvantages of the method have been discussed in the context of assessing solute transport in porous media (Gamerdinger and Kaplan, 2000, 2001; Gamerdinger et al., 2001a,b). The centrifuge method uses an unsaturated column system that consists of two volumetric infusion pumps (AVI 210A, 3M, St. Paul, MN) and an "unsaturated flow apparatus" (L8-UFA, Beckman Coulter, Fullerton, CA). The L8-UFA includes a rotating seal, rotor, and what we term the column assembly. The column assembly includes all of the fittings for containing a sediment sample, attaching the effluent collection cup, and connecting them to the rotor and rotating seal. The rotating seal is at the center of the rotor and is the conduit through which fluid is delivered from the pump to the columns. The L8-UFA rotating seal has two independent flow paths such that two sediment columns can be run at the same time.
The column dimensions accommodated by the centrifuge are 6 cm in length, L, with a radius, r, of 2.25 cm, and a bulk volume, V, of 95.43 cm3. Columns are saturated before connecting and establishing unsaturated steady-state flow within the L8-UFA. With the centrifuge method, water content is controlled by centrifugal force (controlled by the rotation speed of the centrifuge) and the fluid flux, q (cm h–1), where q is equal to the flow rate, F (cm3 h–1), divided by the column cross-sectional area, A (cm2). The flow rate is controlled with an infusion pump. Water that is forced from the column by centrifugal force is replaced by fluid delivery via the pump (water is held in the sediment column by matric potential). Miscible displacement or unsaturated transport experiments (as described above) are initiated when the sediment columns reach a steady-state average water content. The average water content is determined by weighing the column at each sampling time; the cumulative effluent volume is determined by summing the mass of each effluent sample and dividing by the specific density of the influent solution. The centrifuge method is especially suited to this research, which was directed to testing specific hypotheses using packed and disturbed or model sediments under a variety of moisture conditions. One potentially confounding factor with the centrifuge method is a bias in the kinetic measurements because of flow interruption for sampling (the centrifuge must be stopped to access the sampling cup to remove the effluent solutions). The magnitude of this problem depends on the ratio of the stop interval to the flow rate (Gamerdinger et al., 2001b) and was minimized for these tests.
Approach for Interpreting Breakthrough Curves
The approach used here for predicting U sorption during transport and for interpreting BTCs is based on well-established methods from the scientific literature; adaptations that were made during the course of this research have been described previously (Gamerdinger and Kaplan, 2000; Gamerdinger et al., 2001a,b). The approach is reviewed here.
Recovery of the solute in the effluent was determined from the area under the experimental BTC. Sorption during transport, or retardation, was determined by two procedures: R was calculated from the batch Kd or Kd-mass-avg value using Eq. [3a], or Ref was calculated from the experimental BTC using moment analysis, which is based on the center-of-mass and area under the BTC (Valocchi, 1985).
Simulation refers to an independent or semi-independent prediction of the BTC where all of the model parameters are fixed. Independent refers to cases where the parameters are known from separate experiments or estimations and do not rely on the experimental BTC. Semi-independent refers to cases where the effective retardation factor, Ref, is fixed in the simulation, but the value is determined by an independent analysis of the BTC (i.e., moment analysis). The hydrodynamic and other parameters were fixed in all simulations of U transport.
Curve fit refers to an optimization procedure that is used to estimate unknown parameters. Linear regression analysis is the simplest curve-fitting procedure to determine the slope of a straight line and was used to determine batch Kd values. Nonlinear procedures were used to determine hydrodynamic parameters for nonsorptive (i.e., conservative) tracers. For one-region transport, the Peclet number, P, was determined by curve fitting with the numerical CFITIM3 code (University of Florida). For BTCs where two-region transport applied, P was determined from the slope of the BTC (van Genuchten and Wierenga, 1986), and the hydrodynamic parameters β and 
were determined by curve fitting using the CFITIM3 code or CXTFIT (Toride et al., 1995).
When applying hydrodynamic parameters with the two-region model, it was assumed that the diffusional mass transfer coefficient, , was linearly related to velocity (Coats and Smith, 1964). It was further assumed that f = m = β, meaning that the sorption domains are distributed in the same manner as the water domains. In other words, if 40% of the water was in the mobile domain, it was assumed that the fraction of sorption sites in contact with the mobile domain was 0.4. The assumption that sorption sites change proportionally to water content relies on similar mineralogy in different pore domains. For example, in the systems investigated here, the coarse-grained sand is dominated by quartz and feldspars with some degree of surface coatings. The fine-textured silt is comprised of the same primary minerals, but also contains 2:1 clay minerals and metal oxides. Addressing this complexity is beyond the scope of this study. For computational purposes, a uniform surface site type was assumed across the different pore domains. A single value of K assumes that the nature of the sorption interaction does not vary between mobile and immobile water domains. Criteria for distinguishing one-region from two-region transport were detailed previously (Gamerdinger and Kaplan, 2000; Gamerdinger et al., 2001a,b).
For cases where experiments with U were terminated before the distal portion of the BTC reached zero, the BTC was extrapolated to zero in an attempt to determine the full BTC. This procedure was validated in replicate experiments where mass balance was attained for one of the replicates (Gamerdinger et al., 2001b). In previous studies it was noted that, in some cases, nonequilibrium or "two-site" sorption confounded interpretation of the experimental BTCs. The characteristics of two-site sorption are identical to those of two-region transport (early breakthrough [left shift], asymmetry, and tailing of the BTC). To distinguish the two phenomena, tracer data were evaluated to determine if the two-region model is applicable. When the conservative tracer data follow one-region transport and U exhibits nonequilibrium characteristics, a two-site (as opposed to two-region) model is appropriate (Gamerdinger et al., 2001a,b). When the conservative tracer data indicate two-region transport, a two-region model is applied to describe U transport. The various hydrodynamic models and their applicability across a range of water contents are summarized above.
|
Results
|
|---|
To facilitate referencing previous results (Gamerdinger et al., 2001a,b), the abbreviations for designating a particular experiment have been retained. Transport experiments are designated by the abbreviated sediment type, saturation percentage, and average pore water velocity. For example, MSC-12-39 indicates an experiment conducted with MSC sediment at 12% saturation and 39 cm h–1 average pore water velocity. For the sediment mixtures, the mass fraction of silt loam is indicated in parentheses following the S/L abbreviation: S/L(0.1)-18-2 indicates a mixture of 10% silt loam with 90% MSC at 18% saturation and 2 cm h–1 average pore water velocity.
The transport of a conservative, nonsorptive tracer (Br– or pentafluorobenzoic acid) was determined at each water content on each separate and sediment mixture to assess the hydrodynamic properties. The distribution coefficients determined by batch sorption experiments are presented in Table 1
, but are discussed with the results of U transport (Table 2
).
View this table:
[in this window]
[in a new window]
|
TABLE 1. Measured and predicted values for U sorption on each separate and mixtures of silt loam (S/L) and coarse sand (MSC) sediments.
|
|
Batch Sorption Isotherms and Equilibrium Distribution Coefficient Values
Sorption isotherms for U on each of the separate S/L and MSC sediments and two mixtures are reported in Fig. 1
. In all cases, sorption followed a linear isotherm across the concentration range used in flow-through column transport experiments with correlation coefficients R > 0.98. The solid lines in Fig. 1 were determined by linear regression analysis based on Eq. [1]. Measured values of the equilibrium distribution coefficient, Kd, are summarized in Table 1 along with values for the mixtures that were calculated using Eq. [2]. The 95% confidence interval for the Kd-mass-avg values was calculated using the lower and upper estimates for each of the MSC and S/L sediments. The 95% confidence intervals for the measured Kd values on the sediment mixtures overlap with the predictions. This suggests that mixing the sediments does not affect sorption. Small differences between the measured and predicted sorption on the sediment mixtures are probably due to experimental error in preparing the mixture or failure to obtain a representative sample when preparing the mixture.
View larger version (16K):
[in this window]
[in a new window]
|
FIG. 1. Sorption isotherms for U on separate silt loam (S/L) and coarse sand sediments [S/L(0.0)], and mixtures of the S/L and coarse sand sediments, S/L(0.7) and S/L(0.3), where the mass fraction of S/L is indicated in parentheses, se is the sorbed U equilibrium concentration, and ce is the aqueous U concentration.
|
|
Silt Loam, 10% Mass Fraction
Similar to the sediment separates (Gamerdinger et al., 2001a,b), a one-region model based on the convection–dispersion equation described solute transport under saturated conditions on the S/L(0.1) sediment mixture [S/L(0.1)-98-2, Table 2]. Batch data for a representative 10% S/L mixture is not available; however, a mass-averaged Kd value of 1.18 was predicted based on mass averaging the batch data for the "pure" separates (Table 1). Sorption during transport through a saturated sediment was quantified at the relative velocity condition of 2 cm h–1 [S/L(0.1)-98-2]. Variations in the flow rate during the experiment contributed to some uncertainty in the estimate of retardation from the experimental BTC. The calculated U distribution coefficient from the BTC ranged from 1.13 to 1.23, overlapping the mass-averaged Kd prediction based on batch data. This result is consistent with predictions for the saturated water content, where it was expected that all of the sorption sites would remain accessible. Note, however, that a two-site sorption or nonequilibrium sorption model was necessary to best describe the U BTC as utilized and described previously in Gamerdinger et al. (2001a,b).
Unsaturated transport experiments were conducted in the low water content range; saturation was 20 and 22% for the conservative tracer [S/L(0.1)-20-18 and S/L(0.1)-22-21, Table 2], and 24 and 18% for U on the same columns [S/L(0.1)-24-1.6 and S/L(0.1)-18-2, Table 3
]. The tracer transport data for S/L(0.1)-22-21 indicated solute exclusion (R = 0.85, Table 2); however, the data were not of sufficient quality to derive two-region transport parameters from the conservative tracer BTC. A potential limitation was the failure to fully saturate the column before placing it in the centrifuge to force unsaturated water content using centrifugal force. Initial saturation is done to avoid problems with water hysteresis that can arise from wetting, draining and drying, and rewetting a column. In short, a different distribution of pore sizes could be filled. Starting with a fully saturated column and then draining provides a consistent path for attaining various water contents. At first glance, failure to fully saturate a column before attaining 18% moisture saturation may seem insignificant (i.e., 82% of the pore space is void). It is important to recognize, however, that some of the pore space associated with the immobile and isolated water domains may not be readily filled when attempting to saturate a column. If not initially filled, water that may become isolated in these pores on desaturation is not initially present. This precludes analysis of the effect of isolated water and testing hypotheses regarding hydrodynamic changes and solute accessibility to reactive surfaces. Therefore, a second set of unsaturated column experiments, S/L(0.1)-20-18, was conducted and care given to saturate the column to steady state at >90% saturation (based on a theoretical estimate of saturation).
The characteristic shape of the tracer BTC for S/L(0.1)-20-18 suggests two-region transport, and moment analysis indicates exclusion of the conservative tracer at this low water content (R = 0.71, Table 2, Fig. 2
). This is consistent with results for the MSC sediment separate and may indicate an isolated water regime. Hydrodynamic parameters were determined with the two-region model, β = 0.452 and = 0.101, indicating a significant fraction of immobile water (0.55) and rate-limited mass transfer between mobile and immobile water regimes.
View larger version (13K):
[in this window]
[in a new window]
|
FIG. 2. Conservative tracer transport in a mixture of 10% silt loam (S/L) and 90% coarse sand sediments at 20% saturation and an average pore-water velocity of 20 cm h–1. The experimental breakthrough curve is represented by the open symbols; c/co is the effluent concentration. The solid line represents a curve fit with the two-region model, where the retardation factor R was fixed at 0.71.
|
|
Sorption and transport of U was determined on the S/L(0.1)-24-1.6 column (Table 3, Fig. 3
) at the slower velocity. The effective retardation and apparent sorption (Ref = 17.1, Kd-ap = 0.96) were less than predicted from batch sorption data (R = 20.7, Kd-mass-avg = 1.18) by 18% for Kd. This is consistent with the hypothesis for sediments that are dominated by coarse-textured materials (solutes do not have access to a portion of the sorption sites). The decrease in sorption is less than the fraction of pore space from which solutes were excluded (0.29, based on R = 0.71 for the tracer), however, indicating that sites for sorption are not uniformly distributed between mobile and immobile water regimes. The data for U transport on the S/L(0.1)-18-2 column at low water content were successfully modeled with the hydrodynamic parameters that were established with the S/L(0.1)-24-1.6 column series of experiments (Table 2, Fig. 4
). Effective retardation Ref = 24.0 was determined based on the effective Kd-ap = 0.96 that was determined with the S/L(0.1)-24-1.6 column data. This is important because it demonstrates reasonable prediction of U transport once hydrodynamic processes and their effects on sorption are understood.
View larger version (15K):
[in this window]
[in a new window]
|
FIG. 3. Uranium transport in a mixture of 10% silt loam (S/L) and 90% coarse sand sediments at 24% saturation and an average pore velocity of 1.6 cm h–1. The experimental breakthrough curve is represented by the open symbols; c/co is the effluent concentration. A semi-independent prediction is shown for the two-region model where effective retardation Ref = 17.1 and the apparent equilibrium distribution coefficient Kd-ap = 0.96.
|
|
View larger version (16K):
[in this window]
[in a new window]
|
FIG. 4. Uranium transport in a mixture of 10% silt loam (S/L) and 90% coarse sand sediments at 18% saturation and an average pore velocity of 2 cm h–1. The experimental breakthrough curve is represented by the open symbols; c/co is the effluent concentration. An independent prediction is shown for the two-region model where effective retardation Ref = 23.95 and the apparent equilibrium distribution coefficient Kd-ap = 0.96.
|
|
Silt Loam, 30% Mass Fraction
In saturated columns, transport was evaluated at a single velocity (2 cm h–1). Transport of the nonreactive (conservative) tracer was described with the one-region model [Table 2, S/L(0.27)-102-2]. The P value of 18 for S/L(0.27) is very low compared with saturated "pure" S/L and MSC columns (P > 100, Table 1), indicating greater dispersion. It is not clear if this is a characteristic of the mixture (i.e., due to greater heterogeneity) or due to the column packing challenges; however, R = 0.98 was as expected, indicating that there was no apparent effect on the BTC for the conservative tracer. Uranium sorption on S/L(0.27) in the nonflowing batch system was described with a Kd value of 1.94, about 5% greater than predicted from the mass average for the S/L and MSC sediments (Kd-mass-avg = 1.84, Table 1). The 95% confidence interval for the measured and predicted values overlap and the difference is probably due to variation in the sediment mixture. Sorption during transport through the saturated sediment was determined at a velocity of 2 cm h–1 (S/L(0.27)-102-2, Table 2). The apparent distribution coefficient (Kd-ap = 1.29) was less than determined in the batch experiment and was attributed to nonequilibrium sorption, as previous suggested by Gamerdinger et al. (2001b).
Unsaturated transport experiments on a S/L(0.3) mixture were conducted in the moderate water content range; saturation was 34% for the conservative tracer and 32% for U on the same column [S/L(0.3)-34-13, Table 2; S/L(0.3)-32-2, Table 3]. Conservative tracer transport on this sediment mixture was evaluated at a velocity of 13 cm h–1. Recovery was 98% and the retardation factor was 0.95, which approximates the ideal value of R = 1.0 (within 5% error) that is expected for the moderate water content range. This conservative tracer BTC shows asymmetry that is characteristic of two-region transport. A Peclet number of 4.3 was determined by the slope method, and a curve fit with the two-region model gave β = 0.29 and = 2.2 as the "best fit."
Unsaturated U transport was determined on the same column [S/L(0.3)-32-2, Table 3, Fig. 5
] as the conservative tracer [S/L(0.3)-34-13, Table 1]. Recovery of U in the effluent was >99% and sorption was greater than predicted (Ref = 39, Kd-ap = 2.52) compared with the Kd-mass-avg = 1.84 for this mixture. This represents a 27.3% increase in Kd-ap and indicates that, although immobile water was present, U retained access to the more reactive sorption sites associated with the silt loam fraction of the sediment. Indeed, it is consistent with the theory that predicts that the smaller sized pores associated with the smaller silt loam particles would retain water, thereby controlling transport. Using the hydrodynamic parameters established with the conservative tracer [S/L(0.3)-34-13, Table 2], U transport was successfully modeled (semi-independent prediction) with the two-region model (Fig. 5).
View larger version (14K):
[in this window]
[in a new window]
|
FIG. 5. Uranium transport in a mixture of 30% silt loam (S/L) and 70% coarse sand sediments at 32% saturation and an average pore velocity of 2 cm h–1. The experimental breakthrough curve is represented by the open symbols; c/co is the effluent concentration. A semi-independent prediction is shown for the two-region model where effective retardation Ref = 39.0 and the apparent equilibrium distribution coefficient Kd-ap = 2.52.
|
|
A second unsaturated column was prepared that exhibited excess dispersion. When there is excess dispersion, the conservative tracer BTC data can be described with more than one combination of parameters. When extending the hydrodynamic description to sorptive solutes, precise knowledge of the hydrodynamic parameters, especially the fraction of mobile water (β = m) is essential for determining the breakthrough. It is important to recognize that once the parameters to describe heterogeneity (exhibited as excess dispersion) are established in controlled experiments, the approach used here can handle variability in column packing and the natural heterogeneity that is inherent in the field. In other words, once the parameters from well-controlled experiments are known, they can be applied to more heterogeneous situations. This was demonstrated above for unsaturated U transport on the S/L(0.1) mixture, where well-defined hydrodynamic parameters were used to describe U transport in the column with heterogeneous flow.
Figure 6
illustrates an incomplete BTC for U transport, S/L(0.27)-31-1.9, where recovery of U in the effluent was only 52%. Retardation, based on hydrodynamic values determined from column S/L(0.3)-34-13 (Table 3), was determined by calculating a 27.3% increase in the apparent Kd (Kd-ap = 2.34) from the Kd-mass-avg value of 1.84 (Table 1). In spite of the low recovery, variation in saturation during the experiment, and the excess dispersion, U transport was reasonably described by applying the hydrodynamic parameters that were established with the S/L(0.3)-34-13 column (Fig. 6).
View larger version (10K):
[in this window]
[in a new window]
|
FIG. 6. Uranium transport in a mixture of 27% silt loam (S/L) and 73% coarse sand sediments at 31% saturation and an average pore velocity of 2 cm h–1. The experimental breakthrough curve is represented by the open symbols; c/co is the effluent concentration. An independent prediction is shown for the two-region model where effective retardation Ref = 42.60 and the apparent equilibrium distribution coefficient Kd-ap = 2.34.
|
|
Silt Loam, 70% Mass Fraction
Under near-saturated conditions using conventional columns, transport was evaluated at a single velocity (2 cm h–1) and sorption during transport followed the same trends that were reported for the "pure" silt loam sediment (Gamerdinger et al., 2001b). Transport of the nonreactive tracer was described with the one-region model [S/L(0.7)-92-2, Table 1]. The P value of 49.6 for S/L(0.7) was less than that for the silt loam (S/L-89-26, P = 134) (Gamerdinger et al., 2001b), indicating greater dispersion.
Batch experiments indicated that U sorption on S/L(0.7) was described with a Kd value of 3.50 L kg–1, about 6% less than predicted from the mass-average for the S/L and MSC sediments (Kd-mass-avg = 3.71 L kg–1). The 95% confidence interval for the measured and predicted values overlap, and the difference is probably due to variations in the sediment mixture. Uranium(VI) sorption during transport in the nearly saturated sediment mixture was determined at a velocity of 2 cm h–1 [S/L(0.7)-92-2, Table 3]. The apparent distribution coefficient (Kd-ap = 2.57 L kg–1) was less than the value determined in the batch experiment due to nonequilibrium sorption, as described above for the S/L sediment. Although unsaturated transport experiments were conducted in the moderate water content range for this S/L(0.7) mixture, quantification of hydrodynamic parameters from conservative tracer transport was not reliable due to the slow velocity and excess dispersion.
Transport of U in the S/L(0.7)-40-2 column was conducted at 40% moisture saturation (Table 2). Recovery of U in the effluent was 74%, and it is apparent from the BTC (Fig. 7
) that sorption was greater than anticipated; not all of the U eluted from the column. Analysis of the incomplete BTC underestimates retardation and gives R = 30.4 and Kd-ap = 2.69 L kg–1. This is less than predicted from the Kd-mass-avg for the mixture (3.71 L kg–1) but is consistent with the low recovery. If the experiment had continued and all of the applied U had eluted from the column, the center of mass would shift to the right and give a greater measure of retardation. The hypothesis for transport on 70% silt loam was that sorption as indicated by Kd-ap would be greater than predicted by Kd-mass-avg. Again, this is because it was expected that water would be retained in the small pores associated with the silt loam fraction and that solutes in the mobile water domain would retain access to these more reactive surfaces. Simulations were undertaken to determine if sorption and transport behavior was consistent with this hypothesis and predicted from sorption parameters that were determined for the silt loam separate.
View larger version (18K):
[in this window]
[in a new window]
|
FIG. 7. Uranium transport in a mixture of 70% silt loam (S/L) and 30% coarse sand sediments at 40% saturation and an average pore velocity of 2 cm h–1. The experimental breakthrough curve is represented by the open symbols; c/co is the effluent concentration. An independent prediction is shown for the two-region model where effective retardation Ref = 53.1 and the equilibrium distribution coefficient Kd = 4.77.
|
|
The two-region model simulation in Fig. 7 is an independent prediction, where parameters were based on hydrodynamic and sorption parameters determined for the silt loam separate (Gamerdinger et al., 2001b): Kd = 4.77 L kg–1 (Table 3), R = 53.1, P = 7.6, β = 0.425, and = 0.91 (Tables 1 and 3). The simulation provides a reasonable description of the data, although sorption may be somewhat less than observed for the silt loam separate [observed curve for S/L(0.7) is shifted left compared with the simulation]. This result is consistent with the hypothesis that the mobile water fraction is associated with the finer textured silt loam particles and controls sorption and transport such that sorption can be greater than predicted based on the Kd-mass-avg value.
|
Summary and Conclusions
|
|---|
Knowledge of hydrodynamic parameters is essential for predicting sorption during unsaturated transport. It is especially important for predicting the "front" or initial breakthrough of the solute. When the hydrodynamic parameters were clearly determined from conservative tracer data, U transport was predicted well. When retardation was significant (i.e., R > 25), knowledge of the hydrodynamic parameters was as important as the value for R in determining the initial breakthrough and peak concentration of the U.
Results presented here support the hypotheses that changes in sorption with water content are in response to hydrodynamic changes and that solute accessibility to reactive surfaces changes as the transport regime changes from one region, to two region, to two region with isolated water.
Experiments with the sediment mixtures were conducted to test hypotheses regarding the association of the mobile domain with the "fine" vs. "coarse" particle-size fraction, where S/L corresponds to the fine-textured particles with the more reactive surfaces (Kd = 4.77 L kg–1) compared with the coarse sand (Kd = 0.78 L kg–1). It was hypothesized that, for mixtures where the fine-grained S/L dominated the mass of the mixture: (i) the pores in contact with the smaller particles (S/L) would be associated with the mobile water domain (i.e., remain water conducting so that flow is through the fine-grained matrix); (ii) solutes would maintain access to the more reactive fine-grain surfaces; and (iii) sorption would be greater than or equal to sorption predicted by the batch Kd value. The experimental data for U transport on mixtures composed of 70 and 30% silt loam by mass, S/L(0.7) and S/L(0.3 and 0.27), clearly support this hypothesis.
Under hydrodynamic conditions where one-region transport applies, contaminant breakthrough data support the hypothesis that solutes maintain access to all the available reactive sediment surfaces and that sorption can be predicted by the classical batch Kd values. Establishing conditions under which the batch Kd value is a good predictor of sorption is important because this represents the simplest current approach for modeling all contaminant transport, including that under unsaturated flow conditions.
Under hydrodynamic conditions where two-region transport applies, contaminant access to reactive sediment surfaces associated with the immobile water domain is diffusion limited. When two-region transport applies (e.g., S/L-41-17 and S/L-38-2), sorption was shown to be much more dependent on velocity than during one-region transport (S/L-89-25 and S/L-90-2). For two-region transport conditions (including those with isolated water), we hypothesized that the hydrodynamic effects on sorption were dependent on sediment texture. In sediments that were dominated by fine-textured particles, sorption would be greater than (unsaturated conditions) or equal (saturated conditions) to sorption predicted by the batch Kd value. This is consistent with the hypothesis that solutes retained access to the reactive surfaces on fine-grained particles even under unsaturated flow conditions. This is presumably because the pores in contact with the smaller particles are associated with the mobile water domain (i.e., water that is being conducted flows through the amply present fine-grained, more surface reactive portion of the bulk porous medium instead of through the more sparsely present and drained coarse-grained, less surface reactive portion of the matrix).
The corresponding hypothesis for predominately coarse-textured sediments assumed that fewer fine-textured particles, with more reactive surfaces, were in the immobile or isolated water domains. Conversely, the mobile flow is through portions of the more ubiquitous coarse-grained matrix that contains proportionately fewer reactive surfaces. Thus, sorption for predominately coarse-grained sediments at low water contents is less than predicted by the batch Kd values, in which all surfaces are equally available to the solutes. That is, under unsaturated flow conditions within predominately coarse-grained sediments, much of the finer grained particles are isolated from the flowing water and there is limited solute access to the most reactive surfaces.
Experiments in mixtures of the silt loam and coarse sand sediment were undertaken as a means for testing hypotheses regarding the association of mobile and immobile water domains with particles of different size and surface reactivity. In sediment mixtures where the mass fraction of silt loam was 10%, we observed exclusion of conservative tracers and a decrease in U sorption compared with what was predicted by the Kd-mass-avg value. This is consistent with behavior that was previously reported for coarse and fine sand separates (Gamerdinger et al., 2001a). For unsaturated transport in sediment mixtures that contained 30% or more silt loam by mass, there was no exclusion of the conservative tracer, as predicted for the moderate water content range. Sorption was greater than predicted based on the Kd-mass-avg value. This suggests that the fine-textured silt was in contact with the mobile water domain. This is the first evidence linking sorption to transport in a particular water domain.
An important finding from our research is that the interaction between the geochemical and hydrodynamic processes has a profound effect on transport in unsaturated sediments. In particular, various analyses of the data from the experiments on sediment mixtures illustrate how the hydrodynamic conditions have a significant impact on the breakthrough of sorptive solutes. Definition of the fraction of mobile water was especially important for defining the front of the BTC, which is integral to predicting the arrival time of solutes at a particular depth or location in the sediment. It was assumed that adsorption sites changed proportionally to water content, which relies on similar mineralogy in different pore domains. For example, in the systems investigated here, the coarse-grained sand is dominated by quartz and feldspars with some degree of surface coatings. The fine-textured silt is comprised of the same primary minerals, but also contains 2:1 clay minerals and metal oxides. Addressing this complexity is beyond the scope of the present study.
In addition to the physical mechanisms explored in this research, it is possible that chemical changes at the sediment–water interface could affect sorption. Water could move as film flow at lower water contents in unsaturated systems, effectively bringing solutes in closer contact with the sediment surfaces. The water chemistry at the interface may differ from the water chemistry of the bulk solution and will be the subject of further investigations.
|
ACKNOWLEDGMENTS
|
|---|
This work was supported in part by the Immobilized Low-Activity Waste Performance Assessment, U.S. Department of Energy (DOE) and by the Environmental Remediation Sciences Program, Office of Biological and Environmental Research, U.S. Department of Energy under Contract DE-AC06-76RL01830. The support of F.M. Mann (Fluor Federal Services, Richland, WA) is appreciated. This work was conducted at Pacific Northwest National Laboratory, operated by Battelle Memorial Institute for the DOE under Contract DE-AC05-76RL0 1830. The assistance of K.E. Geiszler in conducting ICP-MS analyses is greatly appreciated. Washington Savannah River Company is operated for the DOE under Contract DE-AC09-96R18500.
|
REFERENCES
|
|---|
- Ames, L.L., D. Rai, and R.J. Serne. 1976. A review of actinide-sediment reactions with an annotated bibliography. Pacific Northwest National Lab., Richland, WA.
- Barnett, M.O., P.M. Jardine, S.C. Brooks, and H.M. Selim. 2000. Adsorption and transport of uranium(VI) in subsurface media. Soil Sci. Soc. Am. J.
64
:908–917.[Abstract/Free Full Text]
- Biggar, J.W., and D.R. Nielsen. 1962. Miscible displacement: II. Behavior of tracers. Soil Sci. Soc. Am. Proc.
26
:125–128.
- Cameron, D.R., and A. Klute. 1977. Convective–dispersive solute transport with a combined equilibrium and kinetic adsorption model. Water Resour. Res.
13
:183–188.[CrossRef]
- Coats, K.H., and B.D. Smith. 1964. Dead-end pore volume and dispersion in porous media. Soc. Petrol. Eng. J.
4
:73–84.
- Duff, M.C., and C. Amrhein. 1996. Uranium(VI) adsorption on goethite and soil in carbonate solutions. Soil Sci. Soc. Am. J.
60
:1393–1400.[Abstract/Free Full Text]
- Gamerdinger, A.P., and D.I. Kaplan. 2000. Application of a continuous-flow centrifugation method for solute transport in disturbed, unsaturated sediments and illustration of mobile–immobile water. Water Resour. Res.
36
:1747–1755.[CrossRef]
- Gamerdinger, A.P., and D.I. Kaplan. 2001. Physical and chemical determinants of colloid transport and deposition in water-unsaturated sand and Yucca Mountain tuff. Environ. Sci. Technol.
35
:2497–2504.[Medline]
- Gamerdinger, A.P., D.I. Kaplan, D.M. Wellman, and R.J. Serne. 2001a. Two-region flow and decreased sorption of uranium(VI) during transport in Hanford groundwater and unsaturated sands. Water Resour. Res.
37
:3155–3162.[CrossRef]
- Gamerdinger, A.P., D.I. Kaplan, D.M. Wellman, and R.J. Serne. 2001b. Two-region flow and rate-limited sorption of uranium(VI) during transport in an unsaturated silt loam. Water Resour. Res.
37
:3147–3153.[CrossRef]
- Gamerdinger, A.P., C.T. Resch, and D.I. Kaplan. 1998. Uranium(VI) sorption and transport in unsaturated, subsurface Hanford Site sediments—Effect of moisture content and sediment texture: Final report for subtask 2b. PNNL-11975, 28520. Pacific Northwest National Lab., Richland, WA.
- Gamerdinger, A.P., K.C.J. Van Rees, P.S.C. Rao, and R.E. Jessup. 1994. Evaluation of in situ columns for characterizing organic contaminant sorption during transport. Environ. Sci. Technol.
28
:376–382.
- Gaudet, J.P., H. Jegat, G. Vachaud, and P.J. Wierenga. 1977. Solute transfer, with exchange between mobile and stagnant water, through unsaturated sand. Soil Sci. Soc. Am. J.
41
:665–671.[Abstract/Free Full Text]
- Gee, G.W., and P.R. Heller. 1985. Unsaturated water flow at the Hanford Site: A review of literature and annotated bibliography. Pacific Northwest National Lab., Richland, WA.
- Giblin, A.M., B.D. Batts, and D.J. Swaine. 1981. Laboratory simulation studies of uranium mobility in natural waters. Geochim. Cosmochim. Acta
45
:699–709.[CrossRef][Web of Science]
- Hsi, C.-K.D., and D. Langmuir. 1985. Adsorption of uranyl onto ferric oxyhydroxides: Application of the surface complexation site-binding model. Geochim. Cosmochim. Acta
49
:1931–1941.[CrossRef][Web of Science]
- Kaplan, D.I., and R.J. Serne. 1995. Distribution coefficient values describing iodine, neptunium, selenium, technetium, and uranium sorption to Hanford sediments. Pacific Northwest National Lab., Richland, WA.
- Kaplan, D.I., R.J. Serne, A.T. Owen, J. Conca, T.W. Wietsma, and T.L. Gervais. 1996. Radionuclide adsorption distribution coefficients measured in Hanford sediments for the low level waste performance assessment project. PNNL-11485. Pacific Northwest National Lab., Richland, WA.
- Kohler, M., G.P. Curtis, D.B. Kent, and J.A. Davis. 1996. Experimental investigation and modeling of uranium(VI) transport under variable chemical conditions. Water Resour. Res.
32
:3539–3551.[CrossRef]
- Mann, F.M. M., K.C. Burgard, W.R. Root, R.J. Puigh, S.H. Finfrock, R. Khaleel, D.H. Bacon, E.J. Freeman, B.P. McGrail, S.K. Wurstner, and P.E. LaMont. 2001. Hanford immobilized low-activity waste performance assessment: 2001 version. DOE/ORP-2000-24. USDOE, Office of River Protection, Richland, WA.
- McKinley, J.P., J.M. Zachara, S.C. Smith, and G.D. Turner 1995. The influence of uranyl hydrolysis and multiple site-binding reactions on adsorption of U(VI) to montmorillonite. Clays Clay Miner.
43
:586–598.[Abstract]
- Nielsen, D.R., and J.W. Biggar. 1961. Miscible displacement in soils: I. Experimental information. Soil Sci. Soc. Am. Proc.
25
:1–5.
- Nielsen, D.R., M.Th. van Genuchten, and J.W. Biggar. 1986. Water flow and solute transport processes in the unsaturated zone. Water Resour. Res.
22
:89S–108S.[CrossRef]
- Relyea, J.F., and R.J. Serne. 1979. Waste isolation safety assessment program. Controlled sample program publication number 2: Interlaboratory comparison of batch Kd values. PNL-2872. Pacific Northwest National Lab., Richland, WA.
- Relyea, J.F., R.J. Serne, and D. Rai. 1980. Methods for determining radionuclide retardation factors: Status report. Pacific Northwest National Lab., Richland, WA.
- Riley, R.G., and J.M. Zachara. 1992. Chemical contaminant on DOE lands and selection of contaminant mixtures for subsurface science research. USDOE, Office of Energy Res., Washington, DC.
- Rovira, M., F.Z. El Aamrani, L. Duro, I. Casas, J. de Pablo, J. Bruno, C. Domènech, and C. Ayora. 2000. Experimental study and modeling of uranium(VI) transport through ferrous olivine rock columns. Radiochim. Acta
88
:665–671.[CrossRef]
- Schalla, R., R.W. Wallace, R.L. Aaberg, S.P. Airhart, D.J. Bates, J.V.M. Carlile, C.S. Cline, D.I. Dennison, M.D. Freshley, and P.R. Heller. 1988. Interim characterization report for the 300 Area process trenches. Pacific Northwest National Lab., Richland, WA.
- Selim, H.M., J.M. Davidson, and R.S. Mansell. 1976. Evaluation of a two-site adsorption–desorption model for describing solute transport in soils. p. 444–448. In Proc. Summer Comput. Simulation Conf., Washington, DC. 12–14 July 1976. Simulation Councils, La Jolla, CA.
- Serne, R.J., V.L. LeGore, K.J. Cantrell, C.W. Lindenmeier, J.A. Campbell, J.E. Amonette, J.L. Conca, and M.I. Wood. 1993. Solid-waste leach characterization and contaminant–sediment interactions. Vol. 1: Batch leach and adsorption tests and sediment characterization. PNL-8889 Vol. 1. Pacific Northwest National Lab., Richland, WA.
- Serne, R.J., R.O. Lokken, and L.J. Criscenti. 1992. Characterization of grouted low-level waste to support performance assessment. Waste Manage.
12
:271–287.[CrossRef]
- Serne, R.J., W.J. Martin, and V.L. LeGore. 1995. Leach test of cladding removal waste grout using Hanford groundwater. Pacific Northwest National Lab., Richland, WA.
- Serne, R.J., W.J. Martin, V.L. LeGore, C.W. Lindenmeier, S.B. McLaurine, P.F.C. Martin, and R.O. Lokken. 1989. Leach tests on grouts made with actual and trace metal-spiked synthetic phosphate/sulfate waste. Pacific Northwest National Lab., Richland, WA.
- Serne, R.J., D. Rai, M.J. Mason, and M.A. Molecke. 1977. Batch Kd measurements of nuclides to estimate migration potential at the proposed waste isolation pilot plant in New Mexico. PNL-2448. Pacific Northwest National Lab., Richland, WA.
- Toride, N., F.J. Leij, and M.Th. van Genuchten. 1995. The CXTFIT code for estimating transport parameters from laboratory or field tracer experiments. U.S. Salinity Lab., Riverside, CA.
- Tripathi, V.S. 1983. Uranium(VI) transport modeling: Geochemical data and submodels. Ph.D. diss., Stanford Univ., Palo Alto, CA (Diss. Abstr. 8412891).
- Turner, G.D., J.M. Zachara, J.P. McKinley, and S.C. Smith. 1996. Surface-charge properties and UO22+ adsorption of a subsurface smectite. Geochim. Cosmochim. Acta
60
:3399–3414.[CrossRef][Web of Science]
- Valocchi, A.J. 1985. Validity of the local equilibrium assumption for modeling sorbing solute transport through homogeneous soils. Water Resour. Res.
21
:808–820.[CrossRef]
- van Genuchten, M.Th. 1981. Non-equilibrium transport parameters from miscible displacement experiments. U.S. Salinity Lab., Riverside, CA.
- van Genuchten, M.Th., and P.J. Wierenga. 1976. Mass transfer studies in sorbing porous media: I. Analytical solutions. Soil Sci. Soc. Am. J.
40
:473–480.[Abstract/Free Full Text]
- van Genuchten, M.Th., and P.J. Wierenga. 1986. Solution dispersion coefficients and retardation factors. p. 1025–1054. In A. Klute (ed.) Methods of soil analysis. Part 1. Physical and mineralogical methods. SSSA Book Ser. 5. ASA and SSSA, Madison, WI.
- Voudrias, E.A., and J.L. Means. 1993. Sorption of uranium by brine-saturated halite, mudstone and carbonate minerals. Chemosphere
26
:1753–1765.
- Waite, T.D., J.A. Davis, T.E. Payne, G.A. Waychunas, and N. Xu. 1994. Uranium(VI) adsorption to ferrihydrite: Application of a surface complexation model. Geochim. Cosmochim. Acta
58
:5465–5478.[CrossRef][Web of Science]
- Wolfsberg, K. 1978. Sorption–desorption studies of Nevada Test Site alluvium and leaching studies of test debris. Los Alamos Natl. Lab., Los Alamos, NM.
- Wood, M.I., R.J. Serne, and K.J. Cantrell. 1995. Performance assessment for the disposal of low-level waste in the 218-W-5 burial ground. Westinghouse Hanford Co., Richland, WA.
This article has been cited by other articles:
|
|
|
|
 
M. A. Mayes, G. Tang, P. M. Jardine, L. D. McKay, X. L. Yin, M. N. Pace, J. C. Parker, F. Zhang, T. L. Mehlhorn, and R. Dansby-Sparks
Influence of Sedimentary Bedding on Reactive Transport Parameters under Unsaturated Conditions
Soil Sci. Soc. Am. J.,
October 21, 2009;
73(6):
1938 - 1946.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
D. M. Wellman, J. M. Zachara, C. Liu, N. P. Qafoku, S. C. Smith, and S. W. Forrester
Advective Desorption of Uranium(VI) from Contaminated Hanford Vadose Zone Sediments under Saturated and Unsaturated Conditions
Vadose Zone J.,
October 29, 2008;
7(4):
1144 - 1159.
[Abstract]
[Full Text]
[PDF]
|
|
|
TOP
ABSTRACT
INTRODUCTION
for U(VI).