Transmittal of the deliverables 'Thermodynamic Modeling of ...manuscripts examines the geochemical controls on radionuclide sorption and applies a methodology for simulating the effects

471 I'J'H A center of excellence in earth sciences and engineering A Division of Southwest Research Institute 6220 Culebra Road • San Antonio, Texas, U.S.A. 78228-5166 (210) 522-5160 ° Fax (210) 522-5155

September. 1, 2000 Contract No. NRC-02-97-009 Account No. 20.01402.871

U.S. Nuclear Regulatory Commission ATTN: Dr. John W. Bradbury Division of Waste Management Two White Flint North Mail Stop TD-13 Washington, DC 20555

Subject: Transmittal of the deliverables "Thermodynamic Modeling of Radionuclide Adsportion Part I"

(IM 01402.871.060) and "Thermodynamic Modeling of Radionuclide Adsorption Part II" (IM 01402.871.070)

Dear Dr. Bradbury:

This letter transmits the subject Intermediate Milestones (IM) under the publication titles "Thermodynamic Modeling

of the Adsorption of Radionuclide on Selected Minerals I: Cations" (IM 01402.871.060) and "Thermodynamic

Modeling ofthe Adsorption ofRadionuclide on Selected Minerals II: Anions" (IM 01402.871.070). These manuscripts

summarize the application of thermodynamic models for sorption of both cationic (IM 01402.871.060) and anionic

(IM 01402.871.070) radionuclides. These deliverables were identified in table 1 of the Periodic Management Progress

Report for Period 10. These manuscripts will be submitted for publication in the journal Geochimica et Cosmochimica Acta.

This work was conducted as an activity under the Radionuclide Transport (RT) Key Technical Issue (KTI). The

U.S. Department of Energy (DOE) Repository Safety Strategy, Revision 3 identifies RTthrough the unsaturated and

saturated zones as principal factors influencing the performance of the proposed repository at Yucca Mountain (YM),

Nevada. In the performance assessment (PA) analyses presented by DOE, repository performance was sensitive

to the retardation properties of the natural barrier system. In developing PA transport parameters for the license

application, DOE will need to demonstrate that it has provided an adequate abstraction ofthe sorption behavior ofthe

minerals in the YM environment. One means for doing this, as identified in the RT Issue Resolution Status Report,

is to use of process-level models similarto the ones developed in the subject deliverables. The work presented in these

manuscripts examines the geochemical controls on radionuclide sorption and applies a methodology for simulating the

effects of geochemistry on sorption of a number of key radionuclides, including neptunium, plutonium, americium,

technetium, iodine, and selenium. This provides an independent basis for evaluating DOE reliance on these processes.

SWa gton Office • Twinbrook Metro Plaza #210 12300 Twinbrook Parkway ° Rockville, Maryland 20852-1606

Dr. John W. Bradbury September. 1, 2000 Page 2

These manuscripts were prepared to document ongoing issue resolution activities in the RT KTI, particularly in the

subissues on transport through porous rock, transport through fractured rock, and transport through alluvium. These

analyses will also be used to develop the technical bases for acceptance criteria and review methods in the

Yucca Mountain Review Plan.

If you have any questions about this deliverable, please call me 210.522.5540 or Dr. David Turner 210.522.2139.

Sincerely,

English C. Pearcy, Manager Geohydrology & Geochemistry

d:\gh&gcefiscal year\letters\bradbury\thermodynamic modeling of the adsportion...

Enclosures

cc: J. Linehan D. DeMarco B. Meehan J. Greeves J. Holonich W. Reamer K. Stablein D. Brooks N. Coleman J. Ciocco T. Essig

W. Patrick CNWRA Directors CNWRA Managers CNWRA ISI Leads R. Pabalan P. Bertetti L. Browning M. Nugent T. Nagy (SwRI Contracts)

Prepared for submission to Geochimica et Cosmochimica Acta

Thermodynamic Modeling of the Adsorption of Radionuclides on Selected Minerals. H: Anions

PEIMING WANG* and ANDRZEJ ANDERKO

OLI Systems, Inc. 108 American Road, Morris Plains, NJ 07950

DAVID R. TURNER

Center for Nuclear Waste Regulatory Analyses, Southwest Research Institute

6220 Culebra Road, San Antonio, TX 78238-5166

* Author to whom correspondence should be addressed (E-mail: [email protected]).

Abstract- Because of their high solubility and reduced interaction with aquifer minerals, radionuclide

anions (I, 103", SeO 3"2, SeO4 "2, and TcO4 ") are typically identified as potentially significant contributors to

dose in high-level nuclear waste performance assessments. Adsorption of radionuclide anions on a

number of mineral surfaces has been studied using a thermodynamically consistent model that couples the

diffuse layer theory of surface complexation with an aqueous activity coefficient model based on the

B-dot equation for aqueous speciation calculations. The adsorbents include iron-containing minerals

(ferrihydrite, goethite, and hematite), silica, 7-alumina, 8-MnO 2, metal sulfides (cinnabar and chalcocite),

and aluminosilicate minerals (kaolinite, montmorillonite, and bentonite). Binding constants for sorption

of selected radionuclides on minerals have been determined. For aluminosilicate minerals, the diffuse

layer model uses the surface species --SiOH and -AIOH, in proportion to the stoichiometric formula of

the mineral. Comparison between the experimental data and the calculated results shows that the model

accurately represents radionuclide adsorption data over wide ranges of pH, solid-mass to solution-volume

ratio (m/V), and radionuclide concentration (CrN). Using binding constants determined from this study,

radionuclide adsorption has been predicted over extended ranges of system parameters such as pH, m/V,

and CRN. The surface complexes formed are largely related to the predicted aqueous speciation and the

surface site distribution, especially when stepwise protonation of the anion occurs under varying pH

conditions.

2

1. INTRODUCTION

In the geologic disposal of high-level nuclear waste (HLW), migration of radionuclides through

the geosphere dissolved in groundwater is one of the principal measures of repository performance.

Radionuclide sorption onto minerals is one type of process that may contribute to the delay of

radionuclide arrival or reduce groundwater concentrations at a receptor group location, and is of particular

interest in performance assessment (PA) calculations. Radionuclide sorption is strongly dependent on the

physicochemical conditions along the groundwater flow path. Understanding radionuclide sorption

behavior on various minerals is necessary for rationalizing parameter selection in PA and predicting the

behavior of radionuclides in natural environments.

A number of studies have provided data on radionuclide sorption over a wide range of chemical

conditions (e.g. Balistrieri and Chao, 1987, 1990; Balsley et al., 1996; Couture and Seitz, 1983; Ghosh et

al., 1994; Goldberg and Glaubig, 1988; Ikeda et al., 1994; Kang et al., 1996; Palmer and Meyer, 1981;

Parida et al., 1997; Saeki et al., 1990; Shade et al., 1984; Kent et al., 1986; Davis and Kent, 1990;

Dzombak and Morel, 1990; Payne and Waite, 1991; Venkataramani and Gupta, 1991; Bradbury and

Baeyens, 1993; Turner, 1995; Turner et al., 1996; Turner and Sassman, 1996, Pabalan et al., 1998; Turner

et al., 1998). These studies provide a strong basis for developing a uniform approach to developing

detailed geochemical sorption models to support PA calculations.

Recently, we have interpreted sorption data for radionuclides, selected for their significance in

high-level nuclear waste (HLW) disposal, using a thermodynamically consistent model that couples the

diffuse layer theory for surface complexation with an aqueous activity coefficient model based on the

"B-dot" method of Helgeson (1969) for aqueous speciation calculations (Wang et al., submitted). A

model has been developed based on the work of Turner and Sassman (1996) on defining the acid-base

behavior of minerals. In this paper, we report the application of this model to study the sorption of

radionuclides typically present in groundwater as anions. Negatively charged anions typically interact

weakly with silicate and oxide minerals; this limited sorption makes many anionic contaminants more

mobile and of particular concern in transport analyses.

3

Iodine, technetium, and selenium are potentially significant contributors to dose in HLW. In

oxidizing natural waters, these radionuclides exist as anionic species such as F, 103", SeO 3"2, SeO4 "2, and

TcO4 that can form highly soluble salts that are poorly retarded by geological barriers. Recent PA

calculations of the proposed HLW repository at Yucca Mountain, Nevada identify 79Se, 99Tc, and 1291 as

critical radionuclides for repository performance. For a regulatory timeframe of 104 y, 99Tc and 1291 are

the major contributors to dose (U.S. Department of Energy, 1998; Nuclear Regulatory Commission,

1999a,b). The sorption behavior of these radionuclides is dependent on geochemical conditions such as

pH. In addition, the redox conditions of the groundwater system are important; for example, Tc exists in

a heptavalent (7+) state under nonreducing conditions as the pertechnetate (TcO 4-) anion. Under reducing

conditions, however, Tc is present in a tetravalent (4+) state, with a lower solubility and stronger sorption

characteristics (Lieser and Bauscher, 1988; Pabalan et al., inpress). Similar redox dependency has been

discussed for selenium sorption (Davis et al., 1993).

Due to its significance as a toxic metal in agricultural and industrial processes, a relatively large

amount of sorption data has been reported for selenium, while only limited data are available for iodine

and technetium. The available literature data constitute a sound database for the application of detailed

models to evaluate the sorption behavior of anionic species in radioactive wastewaters.

The objectives of this work are:

(1) to evaluate published data on the sorption behavior of anionic radionuclide species using a

thermodynamic model based on the diffuse layer theory and aqueous speciation computations;

(2) to develop consistent model parameters which can either be used to support PA radionuclide transport

parameters, or in hydrogeochemical surface complexation calculations to account for radionuclide

sorption; and

(3) to predict the effect of system chemistry on the sorption of anionic species.

It is also of interest to compare the results obtained from this study with those obtained for

radionuclide cation adsorption (Wang et al., submitted), to draw general conclusions with regard to the

differences and similarities in sorption behavior for different radionuclides.

4

2. MODELING OF RADIONUCLIDE ANION ADSORPTION USING SURFACE

COMPLEXATION MODELING

2.1. General Description

For consistency of approach, the surface complexation model (SCM) used in the study of

radionuclide cation sorption (Wang et al., submitted) has been applied here. The model couples a

diffuse-layer theory for surface complexation with an aqueous activity coefficient model based on the

"B-dot" method of Helgeson (1969) for aqueous speciation calculations (Wang et al., submitted).

Optimal values for surface complex binding constants were determined from published experimental

sorption data using the general non-linear least squares optimization program, FITEQL version 2.0

(Westall, 1982a,b), revised to incorporate the B-dot equation for aqueous activity coefficients (Wang et

al., submitted). As appropriate, the acid-base behavior of minerals has been based on the work of Turner

and Sassman (1996). Acidity constants for additional minerals not considered by Turner and Sassman

(1996) (e.g. cinnabar and chalcocite) have been determined using the diffuse-layer model (DLM) with

published potentiometric titration data. The aqueous speciation equilibrium constants used in the

chemical model were taken from the DataO.com.V8.R6 file of the EQ3/6 thermodynamic database

(Wolery, 1992a,b), and are listed in Table 1. Details of the model have been presented in Part I of this

study in a separate paper (Wang et al., submitted).

Sorption modeling involves adjusting the binding constants for one or more postulated surface

reactions to minimize the differences between experimental and calculated values of the sorbed amount of

radionuclide at each pH, under defined physical and chemical conditions such as ionic strength (1), total

radionuclide concentration (CRN), solid-mass to solution-volume ratio (m/V), and mineral specific surface

area (A.,p). Although spectroscopic techniques such as Extended X-Ray Absorption Fine Structure

(EXAFS) potentially can be used to identify surface species, data are not available on I, Tc, and Se and

the selection of surface species is based primarily on the analysis of aqueous speciation and the results of

the sorption data regression.

5

2.2. Evaluation of Radionuclide Sorption Data for Modeling

Selected from a variety of published sources, sorption data for F, 103, Se03 2, Se0 4-2, and TcO 4 "

that cover a wide range of chemical conditions were used in the modeling. These data were measured

under specified conditions of pH, I, CRN, m/V, and As,,p. While it is difficult to assess the uncertainties of

the reported sorption data, the guidelines described by Dzombak and Morel (1990) for error estimates are

used in the present study in the parameter regression.

Frequently only one data set is available at a single set of experimental conditions for a given

radioelement-mineral system, but there are cases where two or more sources are available for a specific

system. For example, sorption data for selenite on goethite have been reported by Balistrieri and Chao

(1987), Zhang and Sparks (1990), and Parida et al. (1997). In these cases, each data set has been treated

separately to determine the optimal binding constants. Generally, binding constants determined from

multiple data sets are in reasonably good agreement in this study. However, among the three data sources

for the sorption of selenite on goethite, the data of Parida et al. (1997) show considerably lower sorption

under conditions comparable to those of Balistrieri and Chao (1987) and of Zhang and Sparks (1990).

The Parida et al. (1997) data were not considered in the determination of the binding constants for the

selenium-goethite system.

Several sets of data for iodine sorption on minerals such as hematite (Couture and Seitz, 1983),

cinnabar (Balsley et al., 1996; Ikada et al., 1994), and chalcocite (Balsley et al., 1996) are among those

selected for this study. Couture and Seitz (1983) also measured sorption for iodine on kaolinite, but

sorption was too low (e.g. <5% for the sorption of I) and the measurements contained too few points to

constrain modeling parameters. Moreover, observed sorption of iodine on silicate minerals has been

suggested (e.g. Couture and Seitz, 1983) to be actually due to the sorption by iron oxide or other

impurities. For this reason, these sorption data for iodine on kaolinite were not analyzed in this study.

Sorption data were also reported for F on TiO2 by Hakem et al. (1996), but the initial iodide concentration

was not reported, making it impossible to use these data for developing a geochemical sorption model.

6

Sorption data for technetium are relatively few compared with those for the other radionuclides.

Two data sets found for sorption of TcO4 (Shade et al., 1984; Kang et al., 1996) contain only a few data

points each. Palmer and Meyer (1981) measured pertechnetate sorption on a number of synthetic

inorganic materials, including A120 3, as well as on some naturally occurring minerals, under various ionic

strength conditions. The sorption data were presented in terms of a distribution coefficient (Kd) vs. pH.

However, the original paper did not indicate the solid concentration (m/V) used in their experiments,

which makes it impossible to convert Kd to the amount of sorbed pertechnetate which would then be used

in the modeling. Therefore, the data of Palmer and Meter (1981) were not analyzed in this study.

2.3. Determination of Acidity Constants for Metal Sulfide Minerals

Metal sulfides have been recognized as important adsorbents in nature. Although iron sulfides

are the most abundant metal sulfides in the earth's crust, their relatively high solubility compared to other

metal sulfides, and the formation of a series of metastable iron sulfide species (Anderko and Shuler,

1997), have made it difficult to characterize and identify the chemical species at their surfaces. The

surface charge development at metal sulfide surfaces can be ascribed to multiple functional groups, such

as the thiol group (-SH) and the metal hydroxide group (--Me-OH). The extreme insolubility of cinnabar

(HgS) and chalcocite (Cu2S) make them good model surfaces for iron sulfides. The study of adsorption

of radionuclides on these metal sulfides may provide valuable information for the characterization and

identification of surface species that may form on iron sulfides.

SCM approaches have been generally used in the literature to model adsorption on hydrous oxide

minerals (Davis and Kent, 1990; Dzombak and Morel, 1990). For modeling sorption of radionuclides on

cinnabar and chalcocite, surface acidity constants for these two sulfides are needed. Using the

potentiometric titration data of Balsley et al. (1996), DLM surface acidity constants for these two sulfides

were determined using FITEQL (Figure 1). Cinnabar and chalcocite exhibit negative surface charge in

the pH ranges of the potentiometric measurements (pH < 9) (Balsley et al., 1996). Thus, only the

deprotonation constants, K, were obtained. The values of these constants, together with their standard

7

deviations, are given in Table 2. It should be noted that in addition to thiol groups, hydroxyl groups may

also exist at the hydroxylated surface sites and contribute to the surface charge. However, hydroxylated

Cu(I) and Hg(II) only deprotonate in solution at very alkaline pH (Baes and Mesmer, 1976). The

deprotonation of thiol groups may outweigh that of the hydroxylated sites under common groundwater

pH conditions. The deprotonation constants obtained from our DLM approach perhaps reflect primarily

the deprotonation of thiol groups on the metal sulfide surfaces. Deprotonation constants for cinnabar and

chalcocite, together with aqueous reaction equilibrium constants, were used to determine binding

constants for iodide on these two metal sulfides.

3. RESULTS AND DISCUSSION

3.1. Representation of Experimental Data

The binding constants determined in this study for I, Se, and Tc are given in Tables 2-4. Listed in

these tables are also surface complexation reactions corresponding to the equilibrium binding constants,

and the intrinsic surface acidity constants used in the model. Standard deviations (oa) calculated by

FITEQL for binding constants are also given.

In the case of multiple data sets, binding constants were first determined for each individual data

sets. The best estimates were then determined by a weighted average of the optimum log K determined

from each individual data set, as discussed by Dzombak and Morel (1990). The standard deviation of the

averaged log K was assigned to be equal to the largest o-from the individual data regressions. While

binding constants determined from multiple data sets are in reasonably good agreement with each other in

most cases in this study, there are cases where inconsistency exists among data sets from different

authors. For example, binding constants determined from the data of Balistrieri and Chao [log

K(-XOH 2-HSeO 3°)= 18.83 (u=0.298) and log K(--XOH-HSeO 3-)= 12.99 (a=0.068)] give excellent

representation of experimental results at various selenite and solid concentrations. The same binding

constants predict sorption that is about 10-15% lower than the experimental data of Zhang and Sparks

8

(1990). When determined using the Zhang and Sparks (1990) data, these binding constants are

significantly higher [log K(=XOH2-HSeO30)=19.99 (a=0.392) and log K(=XOH-HSeO3)=14.35

(Y=0.152)]. The difference in log K values determined from the two data sets may arise from sources of

error in the two measurements due to the different experimental techniques. Binding constants

determined from the Balistrieri and Chao data are given in Table 3.

Figures 2-14 compare the experimental and calculated sorption results for F, 103, SeO 3z, SeO 42 ,

and TcO4 ". All of the lines in these figures were calculated based on the binding constants obtained from

the present study as listed in Tables 2-4. The agreement between the experimental and calculated

sorption results is good in general.

The strength of the present DLM for predicting the effect of changing system chemistry on

anionic radionuclide sorption behavior has been demonstrated in Figs. 2, 5-7, 9 and 10, where a set of

binding constants can accurately represent sorption results under varying geochemical conditions (I, m/V,

and CRN).

3.2. pH

The trend of sorption with changing pH for anion adsorption is clearly shown, i.e., anion sorption

decreases with increasing pH. This trend is consistent with that of the formation of protonated surface

site, --XOH 2+ (Wang et al., submitted) at low pH, and increasing deprotonation with increasing pH. It is

also consistent with anion sorption behavior analyzed in Dzombak and Morel (1990). This indicates that

in addition to the chemical bonding, electrostatic interactions between the anions and the protonated

mineral surface play an important role in the adsorption process.

3.3. Effects of Aqueous Speciation on Adsorption

Assuming a single surface complex typically results in a poor model fit to the observed data.

Inclusion of multiple surface complexes in the model can usually give the best fit of the anion sorption

9

data. The complexes formed on the mineral surfaces are expected to be different at different pH values

due to changes in aqueous speciation. This effect is most significant for the sorption of SeO32 which has

large association constants for the formation of HSeO 3 (log KI1=7.3) and H2SeO 3 (log K&2=9.9). The

result of the speciation analysis in aqueous selenite solution is shown in Fig. 15. Protonation of SeO 4-2 to

form HSeO 4 (log K,1=1.9) may also affect the sorption of selenate in acidic solutions (pH<3).

Protonation of pertechnetate, iodide, and iodate is negligible [e.g. log K0(HTcO 4) _ -9 and log K0(HI0 3) =

0.5], and the primary aqueous species are F, 103-, and TcO4 in their corresponding solutions.

Surface complexation reactions that involve different numbers of protons can be included in the

model to adjust the pH dependencies of the predicted sorption curves and improve the model results. This

is especially the case when stepwise protonation occurs under varying pH conditions, e.g., in the case of

selenite. The appropriate selection of surface complexes is crucial to the success of the model in

representing sorption results. Including multiple surface complexes (>3) can improve the quality of the

prediction when sorption results are presented over a wide pH range extending to highly acidic (pH-1)

and alkaline (pH-13) conditions. This is demonstrated for the sorption of selenate on y-A120 3 (Fig. 8)

and selenite on kaolinite and montmorillonite (Figs. 11 and 12), where 3, 4 and 6 surface complexes have

been included, respectively. Direct information (EXAFS) is not available on surface complexes for the

radionuclide-mineral systems studied here, and the exact form of the surface reaction is postulated to

obtain the best fit to the data. Combined analysis of sorption behavior, aqueous speciation, and surface

site distribution can assist in the selection of surface complexes. For example, the increase of SeO3"2

sorption with pH in acidic solutions (pH<5) on kaolinite and montmorillonite (Figs. 11 and 12)

(Goldberg and Glaubig, 1988) follows the trend for the formation of HSeO3 in this pH range, as shown in

the speciation diagram for Se(IV) (Fig. 15). Analysis of mineral surface site distribution (Wang et al.,

submitted) suggests that aluminosilicate surface sites are protonated in the same pH range to form a

positively charged site, =XOH 2÷. Thus, surface complexes of the type =XOH2-HSeO3° have been

included in the model, and the sorption data have been well represented at the low pH range.

10

Contribution to the total amount of adsorption from the formation of each surface complex was localized

in a relatively small pH range.

A positive correlation is obtained between the calculated binding constants for the sorption of

radionuclide cations and the hydrolysis constants; larger hydrolysis constants result in an increase in the

estimated binding constants (Wang et al., submitted). A similar trend is also noted in the radionuclide

anion sorption. Figure 16 shows an increase in binding constants with increasing first protonation

constants of the adsorbing divalent anions, A2 (where A=SeO4 and SeO3 in this study). This suggests that

anions prone to protonation would also associate with surface sites to form surface complexes. It is not

surprising that the logarithms of binding constants decrease (become less positive or more negative) with

increasing negative charge on the postulated surface complexes. Such a trend is also noted by Dzombak

and Morel (1990) for selenite and selenate and other divalent anions. Binding constants obtained from

this study are consistent with the trend obtained by Dzombak and Morel (1990), also using DLM, for the

divalent anion, as shown in Fig. 16 for surface complexes of the types --XOH-A-2 and -XOH-HA'. Due

to the limited number of data available for the radionuclide univalent anions (103, F, and TcO 4 ) and

negligible protonation associated with these anions, it difficult to perform the correlation on these anions.

3.4. Radionuclide Concentration

The effect of Cmv on sorption is shown in Fig. 7(a) for SeO 3"2 sorption on goethite and in Fig. 9

for the sorption of selenite on y-A120 3. When the calculated total surface site concentration, TxOH, is of

the same order of magnitude as the radionuclide concentration, CRN has a large observable effect on the

adsorption. For example, for the adsorption of selenite on goethite (Fig. 7(a)) (Balistrieri and Chao,

1987), the TxOH value is 5.64x10-6 mole sites/L, as calculated based on A.,.=49 m2/g, m/V=0.03 g/L and

the assumed value of site density, N, (2.3 lx 1018 sites/m 2) (Turner and Sassman, 1996). The CRN values

for the three sets of sorption data are 6.5x10 7 M, 2.83x106 M, and 5.34x10 6 M. Increased Cmv results in

reduced adsorption of selenite on goethite. Alternatively, if the total surface site concentration TxoH is far

11

in excess of CRN, the effect of CRN on the sorption is negligible. This is shown in Fig. 9 for the sorption of

selenite on 'y-A120 3 (Ghosh et al., 1994) where the TxoH value is 5.98x10 3 while the CRv values are from

5.7x10-6 to 1.33x10-4 M. The sorption curves for the four data sets are identical.

Effects of CpR on sorption behavior can also be predicted using parameters determined in this

study. For example, using parameters in Table 3, the sorption of Se(IV) as SeO 32 on goethite is

calculated as a function of the total selenite concentration at various pH values and is shown in Fig. 17.

This figure shows a reduction in adsorption with increasing radionuclide concentration, i.e., the same

trend that has been predicted for radionuclide cation adsorption (Wang et al., submitted). In contrast to

radionuclide cations for which the adsorption is more affected by Clv at higher pH, the effect of CRN on

the anion adsorption is more significant at lower pH.

Decreased adsorption with increasing CRv is attributed to a reduction in the number of open

sorption sites to levels insufficient for sorbing radionuclides as these sites are filled. When the total

concentration of available sorption sites is far in excess of the radionuclide concentration, this effect is

negligible. Part of this insensitivity is due to the assumption of a single site type in contrast to the strong

site/weak site approach of Dzombak and Morel (1990).

In HLW repository systems, concentrations for cationic radionuclides may be dilute (U.S.

Department of Energy, 1998; Nuclear Regulatory Commission, 1999a,b). Depending on the number of

sorption sites available along the flowpath, CR may have little effect on the radionuclide adsorption in

these systems. Anionic radionuclides, however, have a much higher solubility (U.S. Department of

Energy, 1998; Nuclear Regulatory Commission, 1999a,b), and may be comparable to TxOH, and the

effects of available sorption sites may be more significant.

3.5. Solid-Mass to Solution Volume Ratio

For a specific mineral, the amount (percent) of adsorbed radionuclide increases as m/V becomes

larger. This is demonstrated in Figs. 5 and 6 (selenate and selenite on ferrihydrite), Fig. 7(b) (selenite on

goethite), and Fig. 10 (selenite on 8-MnO 2). This trend is reasonable because the value of m/V is directly

12

proportional to the total available sorption sites. The greater the number of available sites, the more

effective the solid will be in adsorbing the trace concentration of radionuclides.

Using the parameters determined in this study, the present DLM predicts a sorption edge as a

function of m/V (Fig. 18). As pH increases, the sorption edge shifts to higher m/V values for the sorption

of anionic species. It has been predicted that this sorption edge shifts to lower m/Vvalues for the sorption

of cationic species, such as NpO 2+ (Wang et al., submitted). Pabalan et al. (1998) have noted that when

the sorption data are normalized to m/V and presented in terms of a distribution coefficient, Kd, sorption

becomes relatively insensitive to the change in solid concentration. This was also seen in our study for

modeling sorption of radionuclide cations (Wang et al., submitted). In the present study, the Kd values for

the sorption of selenite on goethite are predicted to be insensitive to the change in m/V above neutral pH,

but an increase of Kd with m/V becomes obvious at each lower pH where it appears to approach a stable

value, as shown in Fig. 19. As m/V increases, the total concentration of available sorption sites, TxoH,

increases. When m/V is small and the TxOH values are of the same order of magnitude as the total

radionuclide concentration (6.5x10-7 M for the case shown in Fig. 19), the greater m/V(or TxoH) makes

the solid more effective in adsorbing the radionuclide, resulting in increased Kd. As m/V increases and

TxOH becomes far in excess of the radionuclide concentration (e.g., TxoH>lO"5 mole sites/L), the increased

m/V has little influence on the sorption. A similar "threshold" effect has also been noted for cation

sorption by Pabalan et al. (1993). The greater sensitivity of the anion adsorption at lower pH to changing

m/V is consistent with the result obtained for the effect of CRN on the anion sorption, as described in the

previous section. The predicted insensitivity of the cation [e.g., Np(V)] sorption to m/V in our recent

study (Fig. 18 in Wang et al., submitted) is due to the extremely small concentration of radionuclide

(5.5x10 1 4 M) compared to the total concentration of available sorption sites, which ranges from 10-8 to

10-4 mole sites/L, and is far beyond the threshold value of the m/V. For PA, even though m/V values are

not known with precision in a radionuclide repository system, the concept of a threshold value may be

useful to constrain sorption in applying SCM approaches to field transport calculations. In the saturated

zone, values of m/V are expected to be large in rock-dominated systems to yield TxOH values that are far

13

greater than the trace amount of radionuclides. The rn/Vmay have little effect on Kd values in such

systems. In the unsaturated zone, however, wetting may be incomplete, reducing contact with the aquifer

minerals (e.g., reducing TxoH). Under these conditions, m/V may be more pronounced.

3.6. Valence States

For the radionuclides present in nuclear waste such as Np, Tc, Pu, and U, it has been found

(Meyer et al., 1985) that the formation of lower valence states could significantly increase the

retardation/adsorption of the nuclide by the host rock. Shown in Fig. 20 is the sorption of Se on

ferrihydrite (Balistrieri and Chao, 1990). It is seen that the lower valence states [Se(IV)] adsorb more

strongly than the higher valence states [Se(VI)] (Davis et al., 1993). Similar behavior is also observed for

the sorption of Se(IV) and Se(VI) on y-A120 3 (Figs. 8 and 9) (Ghosh et al., 1994). More importantly,

reducing conditions for Tc can result in changing of speciation from TcO 4 to TcO÷2 (anionic to cationic

form) which can have a significant enhancing effect on Tc-sorption (Lieser and Bauscher, 1988; Pabalan,

et al., in press). Moderately reducing conditions have been observed in saturated zone groundwaters near

Yucca Mountain (Ogard and Kerrisk, 1984; U.S. Department of Energy, 1999), but no data are available

on Tc(IV) sorption to allow model calculation.

3.7. Dependence of Binding Constants on Surface Acidity Constants and Aqueous Thermodynamic

Data

As described in a previous paper (Wang et al., submitted), DLM parameters obtained in this work

are dependent on the data used in constructing the geochemical equilibrium model for the optimization

runs. If acidity constants or aqueous thermodynamic parameters are modified, the radionuclide binding

constants must be reevaluated. For example, in the analysis of the results for selenite adsorption on

anatase (TiO2) (Gruebel et al., 1995), it was found that the use of DLM intrinsic acidity constants for

anatase (log K+=5.37 and log K_=-5.92), as determined by Turner and Sassman (1996), could not obtain a

satisfactory fit to the experimental results. Instead, when intrinsic acidity constants for rutile (TiO 2)

14

(logK+=4.23 and logK_=-7.49) were used in the model, the fit was excellent for data measured at two

different selenite concentrations. The pH corresponding to the zero point of charge, pHzpc, for anatase is

6.1 (Sprycha, 1984), while the relationship

pHzpc = (log K+ - log K.)12 (1)

predicts a much lower value of pHzpc (-5.6) based on Turner and Sassman's values for anatase. Sprycha

(1984) obtained log K, and log K_ values in the vicinity of 3.1 and -8.9, respectively; and Gruebel et al.

(1995) determined these values to be 3.6 and -7.6. It is clear that the intrinsic acidity constants for anatase

are largely uncertain, and it is reasonable to use an average of values reported by different authors before

the uncertainty is resolved. The results of the modeling for the sorption of SeO3"2 on anatase are shown in

Fig. 21. Radionuclide binding constants determined from this study will be recalculated when the

existing acidity constants are modified by incorporating new potentiometric titration data. Likewise, any

significant changes in the aqueous thermodynamic data will require a re-calculation of the necessary

parameters.

4. IMPLICATION OF THE MODELING RESULTS FOR PERFORMANCE ASSESSMENT

Anionic radionuclides are potentially important contributors to dose in HLW PA calculations.

Anion sorption may reduce radionuclide concentrations and delay arrival times at the point of exposure.

Because anion sorption behavior is strongly dependent on geochemical conditions along groundwater

flow paths, detailed process-level models can provide a level of understanding that is necessary to support

transport simulations in PA.

The success of the DLM in predicting the experimental sorption results suggests that it may be

possible to use the simple conceptual models developed in this study to extrapolate to a variety of

chemical conditions from a relatively limited data set. This is in contrast to typical empirical approaches,

where the lack of a strong theoretical basis frequently makes extrapolation beyond experimental

conditions uncertain. The DLM may be used to support Kd selection for HLW PA under site-specific

hydrochemical conditions (e.g., Turner and Pabalan, 1999). Where detailed information is not available,

15

it may also be possible to use the modeling approach outlined here to provide bounding constraints on

radionuclide anion sorption. While this is not an explicit incorporation of geochemistry in the transport

calculations, it does provide a step toward a more sound theoretical basis for sorption modeling in HLW

performance assessment.

Acknowledgment - This manuscript was prepared to document work performed for the Center for Nuclear

Waste Regulatory Analyses (CNWRA) and the U.S. Nuclear Regulatory Commission (NRC). The

manuscript is an independent product of the authors and does not necessarily reflect the views or

regulatory position of the NRC. CNWRA-generated original data contained here meet quality assurance

requirements described in the CNWRA Quality Assurance Manual. The assistance provided by Dr.

Roberto T. Pabalan of CNWRA is greatly appreciated.

REFERENCES

Anderko A. and Shuler P. (1997) A computational approach to predicting the formation of iron sulfide

species using stability diagrams. Comput. Geosci. 23, 647-658.

Baes C.F. and Mesmer R.E. (1976) The Hydrolysis of Cations. John Wiley & Sons

Balistrieri L.S. and Chao T.T. (1987) Selenium adsorption by goethite, Soil Sci. Soc. Am. J. 51, 1145

1151.

Balistrieri L.S. and Chao T.T. (1990) Adsorption of selenium by amorphous iron oxyhydroxide and

manganese dioxide. Geochim. Cosmochim. Acta 54, 739-751.

Balsley S.D., Brady P.V., Krumhansl J.L., and Anderson H.L. (1996) Iodide retention by metal sulfides:

Cinnabar and chalcocite. Environ. Sci. Technol. 30, 3025-3027.

Bradbury M.H. and Baeyens B. (1993) A general application of surface complexation to modeling

radionuclide sorption in natural systems. J Colloid Interface Sci. 158, 364-371.

Couture R.A. and Seitz M.G. (1983) Sorption of anions of iodine by iron oxides and kaolinite. Nuclear

and Chemical Waste Management 4, 301-306.

16

Davis J.A., Kent D.B., Rea B.A., Maest A.S., and Garabedian S.P. (1993) Influence of redox environment

and aqueous speciation on metal transport in groundwater: Preliminary results of trace injection

studies. In Metals in Groundwater (eds. H.E. Allen, E.M. Perdue, and D.S. Brown), pp. 223-273.

Lewis Publishers.

Davis J.A. and Kent D.B. (1990) Surface complexation modeling in aqueous geochemistry. In Mineral

water Interface Geochemistry (eds. M.F. Hochella, Jr. and A.F. White); Reviews in Mineralogy 23,

pp. 177-260. Mineralogical Society of America, Washington DC.

Dzombak D.A. and Morel F.M.M. (1990) Surface Complexation Modeling: Hydrous Ferric Oxide. John

Wiley & Sons.

Ghosh M.M., Cox C.D., and Yuan-Pan J.R. (1994) Adsorption of selenium on hydrous alumina. Environ.

Prog. 13, 79-88.

Goldberg S. and Glaubig R.A. (1988) Anion sorption on a calcareous, montmorillonitic soil - Selenium.

Soil Sci. Soc. Am. J. 52, 954-958.

Gruebel K.A., Davis J.A., and Leckie J.O. (1995) Kinetics of oxidation of selenite to selenate in the

presence of oxygen, titania, and light. Environ. Sci. Technol. 29, 5 86-594.

Hakem N., Fourest B., Guillaumont R., and Marmier N. (1996) Sorption of iodine and cesium on some

mineral oxide colloids. Radiochim. Acta 74, 225-230.

Helgeson H.C. (1969) Thermodynamics of hydrothermal systems at elevated temperatures and pressures.

Am. J. Sci. 267, 729-804.

Ikeda Y., Sazarashi M., Tsuji M., Seki R., and Yoshikawa H. (1994) Adsorption of I ions on cinnabar for

1291 waste management. Radiochim. Acta 65, 195-198.

Kang M.J., Rhee S.W., Moon H., Neck V., and Fanghanel T. (1996) Sorption of M0 4 (M=Tc, Re) on

Mg/Al layered double hydroxide by anion exchange. Radiochim. Acta 75, 169-173.

Kent D.B., Tripathi V.S., Ball N.B., and Leckie J.O. (1986) Surface-complexation modeling of

radionuclide adsorption in sub-surface environments. Department of Civil Engineering, Stanford

University, Technical Report No. 294.

17

Lieser K.H. and Bauscher C.H. (1988) Technetium in the hydrosphere and in the geosphere. II. Influence

of pH, of comp lexing agents and of some minerals on the sorption of technetium. Radiochimica Acta

44/45, 125-128.

Meyer R.E., Arnold W.D., and Case F.I. (1985) Valence effects on the sorption of nuclides on rocks and

minerals. II. Oak Ridge National Laboratory, Report No. NUREG/CR-4114, ORNL-6137.

Nuclear Regulatory Commission (1999a) NRC sensitivity and uncertainty analyses for a proposed HLW

repository at Yucca Mountain, Nevada, using TPA 3.1. Conceptual models and data. U.S. Nuclear

Regulatory Commission, Report No. NUREG-1668, Vol. 1.

Nuclear Regulatory Commission (1999b) NRC sensitivity and uncertainty analyses for a proposed HLW

repository at Yucca Mountain, Nevada using TPA 3.1. Results and conclusions. U.S. Nuclear

Regulatory Commission, Report No. NUREG-1668, Vol. 2.

Ogard A.E. and Kerrisk J.F. (1984) Groundwater chemistry along flow paths between a proposed

repository site and the accessible environment. Los Alamos National Laboratory, Report No.

LA-10188-MS.

Pabalan R.T. and Turner D.R. (1993) Sorption modeling for HLW performance assessment In NRC

High-Level Radioactive Waste Research at CNWRA (ed. B. Sagar), pp. 6-1 to 6-23. Center for

Nuclear Waste Regulatory Analyses, Report No. CNWRA 93-02S.

Pabalan R.T., Turner D.R., and Miklas M.P., Jr. (in press) Technetium-99 chemistry In reduced

groundwaters: Implications for the performance of a proposed high-level nuclear waste repository at

Yucca Mountain, Nevada. In Materials Research Society Symposium Proceedings: Scientific Basis

for Nuclear Waste Management - XXIII (eds. D. Shoesmith and R. Smith), Materials Research

Society (in press).

Pabalan R.T., Turner D.R., Bertetti F.P., and Prikryl J.D. (1998) UraniumvI sorption onto selected mineral

surfaces: Key geochemical parameters. In Adsorption of Metals by Geomedia (ed. E.A. Jenne), pp.

99-130. Academic Press, Inc.

18

Palmer D.A. and Meyer R.E. (1981) Adsorption of technetium on selected inorganic ion-exchange

materials and on a range of naturally occurring minerals using oxic conditions. J. Inorg. Nucl. Chem.

43, 2979-2984.

Parida K.M., Gorai B., Das N.N., and Rao S.B. (1997) Studies on ferric oxide hydroxides. III.

Adsorption of selenite on different forms of iron oxyhydroxides. J Colloid Interface Sci. 185, 355

362.

Payne T.E. and Waite T.D. (1991) Surface complexation modeling of uranium sorption data obtained by

isotope exchange techniques. Radiochim. Acta 52/53, 487-493.

Saeki K., Matsumoto S., and Tatsukawa R. (1990) Selenite adsorption by manganese oxides. Soil Sci.

160, 265-272.

Shade J.W., Ames L.L., and McGrarrah J.E. (1984) Actinide and technetium sorption on iron-silicate and

dispersed clay colloids. ACS Sym. Ser. pp. 67-77. American Chemical Society, Washington DC.

Sprycha R. (1984) Surface charge and adsorption of background electrolyte ions at anatase/electrolyte

interface. J. Colloid Interface Sci. 102, 173-185.

Turner D.R. (1995) A uniform approach to surface complexation modeling of radionuclide sorption.

Center for Nuclear Waste Regulatory Analyses, Report No 95-001.

Turner D.R. and Pabalan R.T. (1999) Abstraction of mechanistic sorption model results for performance

assessment calculations at Yucca Mountain, Nevada. Waste Management 19: 375-388.

Turner D.R., Pabalan R.T., and Bertetti F.P. (1998) Neptunium(V) sorption on montmorillonite: An

experimental and surface complexation modeling study. Clays Clay Miner. 46, 256-269.

Turner D.R. and Sassman S.A. (1996) Approaches to sorption modeling for high-level waste performance

assessment. J. Contam. Hydrol. 21, 311-332.

Turner G.D., Zachara J.M., McKinley J.P., and Smith S.C. (1996) Surface-charge properties and UO 2' 2

adsorption of a surface smectite. Geochim. Cosmochim. Acta 60, 3399-3414.

U.S. Department of Energy (1999) Downhole Eh and pH Measurements for UE-25 WT#17, Available at

http://m-oext.ymp.gov/html/prod/db-tdp/atdt/internet/TDIF307214.html.

19

U.S. Department of Energy (1998) Viability assessment of a repository at Yucca Mountain, Nevada. U.S.

Department of Energy. Office of Civilian Radioactive Waste Management, Report No.

DOE/RW-0508.

Venkataramani B. and Gupta A.R. (1991) Effect of anions on the sorption of uranyl ions on hydrous

oxides: Application of the surface hydrolysis model. Colloids Surf 53, 1-19.

Wang P., Anderko A., and Turner D.R. (submitted) Thermodynamic modeling of adsorption of

radionuclide on selected minerals: I. Cations. Geochim. Cosmochim. Acta.

Westall J.C. (1982a) FITEQL: A computer program for determination of chemical equilibrium constants

from experimental data, Version 1.2. Department of Chemistry, Oregon State University, Report No.

82-01.

Westall J.C. (1982b) FITEQL: A computer program for determination of chemical equilibrium constants

from experimental data, Version 2.0. Department of Chemistry, Oregon State University, Report No.

82-02.

Wolery T.J. (1992a) EQ3/6, A software package for geochemical modeling of aqueous systems: Package

overview and installation guide, Version 7.0. Lawrence Livermore National Laboratory, Livermore,

CA.

Wolery T.J. (1992b) EQ3NR, A computer program for geochemical modeling of

aqueous-speciation-solubility calculations: Theoretical manual, user's guide, and related

documentation, Version 7.0. Lawrence Livermore National Laboratory, Livermore, CA.

Zhang P. and Sparks D.L. (1990) Kinetics of selenate and selenite adsorption/desorption at the

goethite/water interface. Environ. Sci. Technol. 24, 1848-11860.

FIGURE CAPTIONS

Figure 1. Potentiometric titration of (a) cinnabar and (b) chalcocite. The lines are calculated using

parameters listed in Table 2 and experimental data are from Balsley et al. (1996) (mlV=50g/L, A,,,=1.99

m2/g for cinnabar and Ap=1. 11 m2/g for chalcocite, I=0.00 1 M NaCI).

20

Figure 2. Sorption of Ion cinnabar. Experimental data are from Balsley et al. (1996) (A, ASP=1.99 m2/g,

m/V=1.Og/L, [I]=lxl0-5 M, and ionic strength = 0.01 M NaCI) and Ikada et al. (1994) (0, A.,V=1.99 m2/g,

m/V=10 g/L, [I-]=Ix10- 6 M, ionic strength = 0.0003 M). SCM fits to the data of Balsley et al. (1996)

(dashed line) and Ikada et al. (1994) (solid line) are calculated using the parameters listed in Table 2.

Surface complexes included in the model are =HgSH-I and -HgSH 2-I°.

Figure 3. Sorption of ron chalcocite. Experimental data are from Balsley et al. (1996) (A.T= 1.11 m2/g;

[I-]total=Ix 10-5 M; m/V=1.0 g/L; I=0.01 M NaC1) and the solid line is calculated using parameters listed in

Table 2. Surface complexes included in the model are -CuSH-I and -CuSH 2-I0. Dashed lines represent

contributions from the formation of various surface complexes to the total amount of adsorption.

Figure 4. Sorption ofIO3 on hematite. Experimental data are from Couture and Seitz (1983) (Apv=10

m2/g; [103"]total=1.2x10-3 M; m/V=58.31 g/L; I=0.1 12 M) and the lines are calculated using parameters

listed in Table 2. Surface complexes included in the model are =-FeOH-IO 3- and -FeOH 2-IO30. Dashed

lines represent contributions from the formation of various surface complexes to the total amount of

adsorption.

Figure 5. Sorption of selenate on ferrihydrite. The lines are calculated using parameters listed in Tables

1 and 3. Experimental data are from Balistrieri and Chao (1990) (Asp=600 m2/g; [Se(VI)]totai=6.7x10-7 M;

I=0.1 M KC1): 0, m/V=0.0264 g/L; A, mlV=0.264 g/L. Surface complexes included in the model are

-FeO-HSeO 42 and (=FeOH 2)2-SeO 4°.

Figure 6. Sorption of selenite on ferrihydrite. The lines are calculated using parameters listed in Tables

I and 3. Experimental data are from Balistrieri and Chao (1990) (A.,P=600 m2/g; [Se(IV)]total=6.8x10-7 M;

21

I=0.1 M KCl): 0, m/V=O.0264 g/L; A, m/V=0.264 g/L. Surface complexes included in the model are

(-FeOH2)2-SeO 30 and (-=FeO) 2-SeO3.

Figure 7. Sorption of selenite on goethite. The lines are calculated using parameters listed in Tables 1

and 3. Surface complexes included in the model are -FeOH 2-HSeO30 and -FeOH-HSeO 3-. Experimental

data are from Balistrieri and Chao (1987) (A,,p=49 m2/g; I=0.1 M KC1). (a) m/V=0.03 g/L; 1', [SeO3 2]totai=

6.5x1 0-7 M (solid line); 0, [SeO3"2]total = 2.83xl 0-6 M (dash-dot line); A, [Se03 2]total = 5.34x10-6 M

(dashed line). (b) [Se03-2]total = 6.5xi0 7 M; 0, m/V=O.03 g/L (solid line); 0, m/V=0.006 g/L (dashed line);

A, m/V=O.003 g/L (dotted line).

Figure 8. Sorption of selenate on y-A120 3. The lines are calculated using parameters listed in Tables 1

and 3. Experimental data are from Ghosh et al. (1994) (A.TP=250 m2/g; m/V=1.0 g/L; [Se04-2]tota[ =

1.42xl 04 M; I=0.1 M NaCl). Surface complexes included in the model are (-AIO)2-SeO 4"4,

-A1OH 2-SeO 4 , and -A1OHz-H 2 SeO4+. Dashed lines represent contributions from the formation of

various surface complexes to the total amount of adsorption.

Figure 9. Sorption of selenite on y'-A120 3. The lines are calculated using parameters listed in Tables 1

and 3. Experimental data are from Ghosh et al. (1994) (A4=250 m2/g; m/V=6.24 g/L; I=0.1 M NaCI). 0,

[SeO3"2]total=5.x10 6 M; ", [SeO3 2]tota= 2.74x10- M; 0, [SeO3-2 ]total= 5.2x10-5 M; A, [Se03"2]totaI=

1.33x10-4 M. Surface complexes included in the model are (=AIO) 2-SeO3"4 and =AIO-HseO3"2. Dashed

lines represent contributions from the formation of various surface complexes to the total amount of

adsorption.

22

Figure 10. Sorption of selenite on 8-MnO 2. The lines are calculated using parameters listed in Tables 1

and 3. Surface complexes included in the model are =MnOH 2-SeO 3- and -MnO-HSeO 3"2. Experimental

data are from: (i) Balistrieri and Chao (1990) (A.,P=290 m2/g, [SeO 3"2]total = 6.5x10"7 M, 1=0.1 M KC1). 10,

m/V=0.03 g/L (solid line); A, m/V=0.3 g/L (dotted line); (ii) Saeki et al. (1995) (A.P=10.2 m2/g,

[Se03-2]total = 1.0x0-6 M, I=0.1 M NaCl). 0, m/V=3.33 g/L (dashed line).

Figure 11. Sorption of selenite on kaolinite. The lines are calculated using parameters listed in Tables 1

and 3. Experimental data are from Goldberg and Glaubig (1988) (A,,=20.5 m2/g; m/V=40 g/L;

[Se0 32]total=.9xl 0 M; 1=0.1 M NaCI). Surface complexes included in the model are -AIOH2-HSeO 3°,

-SiOH 2-HSeO3°, -SiO-HSeO3" 2, and -A1O-SeO 3"3. Dashed lines represent contributions from the

formation of various surface complexes to the total amount of adsorption.

Figure 12. Sorption of selenite on montmorillonite. The lines are calculated using parameters listed in

Tables 1 and 3. Experimental data are from Goldberg and Glaubig (1988) (A.P=18.6 m2/g; m/V=40 g/L;

[Se03-2]total= 1.9x 10- 5 M; I=0.1 M NaCI). Surface complexes included in the model are -AIOH 2-HSeO 30,

=SiOH2-HSeO3°, =AIO-HSeO32, -SiO-HSeO32, =AIO-SeO3

3, and (=-A1) 2 -SeO3 4. Dashed lines

represent contributions from the formation of various surface complexes to the total amount of

adsorption.

Figure 13. Sorption of pertechnetate on Mg/Al layered double hydroxide (Mg2AI20 9). The line is

calculated using parameters listed in Table 4. Surface complexes included in the model are -=AIOH-TcO 4

and =AlOH2-TcO4. Experimental data are from Kang et al. (1996) (Asp=70 m2/g; m/V=0.59 g/L;

[TcO4"]total=1.4x10- 5 M; I=0.01 M). Dashed lines represent contributions from various surface complexes

to the total amount of adsorption.

23

Figure 14. Sorption of pertechnetate on bentonite. The lines are calculated using parameters listed in

Table 4. Surface complexes included in the model are =AIO-TcO 4"2 and -SiO-TcO 4-2. Experimental data

are from Shade et al. (1984) (A.VP= 10 m2/g; m/V=0.0682 g/L; [TcO4"]total=4.2xl0"14 M; I=0.0 1 M). Dashed

lines represent contributions from various surface complexes to the total amount of adsorption.

Figure 15. Distribution of Se(IV). Lines are calculated based on thermodynamic data given in Table 1.

Figure 16. Plot of the binding constants, log K, vs. first protonation constants, log Ku,, for divalent

anions. Open symbols are from this study, solid symbols are taken from Dzombak and Morel (1990).

Figure 17. Sorption of Se(IV) on goethite as a function of total selenite concentration at various pH

(A~p=49 m2/g; m/V=0.03 g/L, 1=0.1 M KC1). All lines are calculated using parameters in Table 3.

Figure 18. Sorption of selenite on ferrihydrite as a function of W/V at various pH (A.,p=600 m2/g;

[SeO3-2]tota'=6.8x1 0-7 M; 1=0.1 M KCI). All lines are calculated using parameters in Table 3.

Figure 19. Plot of Kd vs. m/V for the sorption of selenite on goethite (A.S,=49 m2/g; ([Se03"2]total=6.5x 10-7

M; I=0.1 M KCI). The lines are calculated based on parameters in Table 3.

Figure 20. Sorption of Se(IV) and Se(VI) on ferrihydrite. Experimental data are from Balistrieri and

Chao (1990) (A.-P=600 m2/g; m/V=0.0264 g/L; [Se]totar=6.7x10- 7 M; I=0.1 M KC1). The lines are

calculated using parameters given in Table 3.

Figure 21. Sorption of selenite on anatase (TiO 2). The lines are calculated using parameters listed in

Table 1 and 3. Experimental data are from Gruebel et al. (1995) (A4,=8.6 m2/g; m/V=10 g/L; 1=0.01 M

24

NaC1): 0, [SeO 3-2 ]totai = 1.25xl 04 M (solid line); A, [SeO3-2]tota, =2.5x10-5 M (dashed line). Surface

complexes included in the model are =TiO-HSeO 3"2 and -TiOH-HSeO3.

25

1.2E-04

I.-

4 5 6 7 8

pH

Figure 1.

100

80

0 0)

60

40

20

0

.

1.2E-04

8.OE-05

4.0E-05

O.OE+O0(

-4.0E-05

-8.0E-05

-1.2E-04

6 7 8 9

pH

3 5 7

pH

Figure 2.

(b)

9 11

3 4 5 6 7 8 9 10 11

pH

Figure 3.

2 3 4 5 6 7 8 9 10

pH

Figure 4.

100

80

0D 60

o

S40

20

0 exp - cal

- -- - CuSH-I

-.........CuSH2-1

100

80

60 o 6 2.0

40

20

0

©

100

80

L0

0 4

20

0

4 5 6

pH

7 8

Figure 5.

100

80

*• 60 0

• 40

20

06 7 8 9 10 11 12

pH

Figure 6.

100

80

.~60

40 • 40

20

0

100

80

*% 60

00 •t 40

20

0

9 11

9 11

Figure 7.

5 7

pH

3 5 7

pH

0 exp -cal

--- AIOH2-H2SeO4+ -----AIOH2-SeO4------(AIO)2-SeO4---

2 3 4 5 6 7 8 9 10

pH

Figure 8.

o selenite=5.7E-6M

E3 selenite=2.74E-5M

Sselenite=5.2E-5M

A selenite=1 .33E-4M -Cal

------AlO-HSeO3

--- (AIO)2-SeO3----

3 4 5 6 7 8 9 10 11

pH

Figure 9.

100

80

60

4

20

40

100

80

n 60

40

20

20

100

80 ,

. e60

40

20

20!

4 5 6 7 8 9 10 11 12

pH

Figure 10.

100

80o exp

" ,-cal )60 ... AIOH2-HSeO3

.. .SiOH2-HSeO3 - 40 --------- SiO-HSeO3

S,,AIO-SeO3-

20

1 3 5 7 9 11 13

pH

Figure 11.

o exp -cal

-------. SiOH2-HSeO3

- -- -AIOH2-HSeO3

----------SiO-HSeO3-

AIO-HSeO3

AIO-SeO3--

-...... (AIO)2-SeO3---

1 3 5 7 9 11 13

pH

Figure 12.

0 exp -cal

- - -- XOH-TcO4

-..........XOH2-TcO4

11 12 13

0

100

80

60

40

20

0

0

100

80

0 n 0

60

40

20

010

pH

Figure 13.

0 exp -cal

-.... AIO-TcO4-

-..........SiO-TcO4--

4 5 6 7 8 9 10 11 12

pH

Figure 14.

100 \N / \ oo

80 \ /

60

40

20 / \ *

0o \

1 3 5 7 9 11 13

pH

------.SeO3--- - HSeO3

S.... H2SeO3

Figure 15.

100

80

"a) O0

60

40

20

U

30

ýIja (XOI-2)-A(0)

20 0 F- X, -H(l

A XOHI-A(-2) 10 * (XO)2-A(-4)

0O 0 A D&M; XOH-A(-2)

a D&M; XOH-HA(-1)

V~ Linear (D&M; XOH-HA(-1))

-20 0 2 4 6 8 10

log K.,

Figure 16

100

AE pH=3

80 -- H=

W pH=6.75 (D

m 60 -C"H

~ 40 ~pH=7.5

20 - pH=8.50 0 0 pH=1 1

0

1.E-09 1.E-08 1.E-07 1.E-06 1.E-05 1.E-04

[selenite]total, M

Figure 17.

4pH=4

UpH=5

*PH=6

9 G-pH=7.12

0pH-=9.12 m-+ pH=10

-pH11l

"- *- pH=12

0.001 0.01 0.1 1

inN, gIL

Figure 18.

0.1

1.9x10-5

~PH=5 ~pH=6

e pH=7

E3- pH=8

-- pH=9

*pH=10

1inN, gIL

qx10o4 TxOH, moleIL

Figure 19.

100

80

60

40

(D

0 U)

20~

0.0001

7.0

6.0

5.0

4.0

3.0

2.0,

1.0-

0,

E

0)

3 B E3B B E

1 1 1 11C14

0 Q; 0 I0.0 T

0.001

1.9x10 7,

0.01

1.9x106, 1.

100

80Se(IV)

S60 0 U) o 40 Se(VI)

20

0 4 5 6 7 8 9 10 1

pH

Figure 20.

100 A

80

0 60-]

,- 40

20

0 3 4 5 6 7 8 9 10 11 12

pH

Figure 2 1.

1

Table 1. Aqueous Reactions Associated with Iodate, Se(VI), and Se(IV) and Their Equilibrium Constants at 25°Ca

Aqueous Reactions log K

103- + H+ = H10 30 0.49

SeO4-2 + H+ = HSeO4" 1.91

SeO 3-2 + H+ = HSeO3- 7.29

SeO 3"2 + 2H+ = H2SeO 3

0 9.86

aEquilibrium constants were based on thermodynamic data from

Data0.com.V8.R6 file of the EQ3/6 database (Wolery, 1992a, 1992b).

Table 2. Summary of Parameters Used for the Determination of Binding Constants for the Sorption of I and 103" on Various Mineralsa

Radionuclide/Solid AP, m2/g m/V, g/L CN, M I, M log K, log K. log KIb log K2b ref.

IO3"/Hematite (a-Fe 20 3) 10 58.31 1.2x10"3 0.112c 7.35d -9.17d 2.00 (0.04) 8.43 (0.09) Courture and Seitz, 1983 IF/Cinnabar (HgS) 1.99 1.0 l.Oxl05- 0.01 (NaCI) n.c.0 -7.27 (0.06)' 6.86 (0.06) 9.05 (0.11) Balsley et al., 1996 I'/Chalcocite (Cu2S) 1.11 1.0 1.0xl0"5 0.01 (NaCI) n.c.' -6.95 (0.03)r 7.58 (0.05) 11.73 (0.14) Balsley et al., 1996

aA site density of 2.31 sites/nm2 (from Dzombak and Morel, 1990) was assigned to all minerals.

bBinding constants K, and K2 correspond to the following surface reactions:

-XPH° + A-z = =-XPH-Az (K1) =-XPH0 + A- '+ H+= •-XPH 2-AXý' (K2)

where A` = Ir or 103"; P=O for hematite and P=S for cinnabar and chalcocite. Values in the parenthesis are the standard deviations of the binding constants. 00.1 M NaC10 4 + 0.011 M NaHCO3 + 0.0012 M K10 3. d Acidity constants from Turner and Sassman (1996). 'Not considered. fLog K_ values determined in this study. The value in the parenthesis is the standard deviation of log K..

Table 3. Summary of Parameters Used for the Determination of Binding Constants for the Sorption of Se on Various Mineralsa

Radionuclide/Solid Asm 2/g m/VgIL CXR, M I, M log KVb log K. log K,' log K2' log K3' log KV log K5' log K6' log K7' ref.

SeO4-2/Ferrihydrite 600 0.0264,0.264 6.7x10-7 0.1 (KCI) 7.29 -8.93 0.65 (0.04) - 19.46 (0.10) - Balistrieri & Chao, 1990

SeO3-2/Ferrihydrite 600 0.0264,0.264 6.8x10-7 0.1 (KCI) 7.29 -8.93 26.06 (1.17) -6.96 (0.64) Balistrieri &Chao, 1990

SeO 3"2/Goethite 49 0.03 2.83x106 0.1 (KCI) 7.35 -9.17 12.99 (0.07) 18.83 (0.30) - - Balistrieri & Chao, 1987

SeO 4-2/Y-AI20 3 250 1.0 1.42x104 0.1 (NaCl) 6.85 -9.05 6.70 (0.15) - 16.72 (1.00) -9.32 (0.09) Ghosh etal., 1994

SeO 3"2/y-AI20 3 250 6.24 2.74x10"5 0.1 (NaCI) 6.85 -9.05 4.74 (0.10) - - - -3.70 (0.09) Ghosh etal., 1994

SeO3 2/8-MnO 2 290 0.03,0.3 6.5x10 7 0.1 (KCI) - -3.27 12.20 (0.05) 16.89 (0.66) - - - Balistrieri & Chao, 1990

SeO 3"2/Anatase 8.6 10 1.25x10 4,2.5x10 5 0.01 (NaCI) 4 .2 3d -7.49d 7.44 (0.04) 11.78 (0.47) - Gruebel etal., 1995

SeO32/Kaolinite' 20.5 40 1.9x10 5 0.1 (NaCI) 8.33 -9.73 - - 16.84 (0.08) -3.80 (0.50) - Goldberg & Glaubig, 1988 - -7.20 2.37 (0.17) - 12.54 (0.08) -

SeO32/Montmorillonite' 18.6 40 1. 9x10 5 0.1 (NaCI) 8.33 -9.73 2.65 (0.12) - 16.51 (0.07) -4.35 (0.09) -8.80 (0.05) Goldberg & Glaubig, 1988 - -7.20 1.50 (0.30) - 12.20 (0.09) -

'A site density of 2.31 sites/nm 2 (from Dzombak and Morel, 1990) was assigned to all minerals.

b Acidity constants from Turner and Sassman (1996).

'Binding constants K, through K7 correspond to the following surface reactions: --XOH0 + A' = --XOH-A' (Ki)

=-XOH + A- + WT= mXOH2 -A-"'l (K2) -XOH0 + A: + 2I-= -XOH 2-HANz+ 2

(K3) --XOH° + A-z + 3HI= =-XOH 2-H2AN3

(K 4)

=-XOH0 + A-'= =-XO-A~z' + H+ (Ks) 2(-XOH6 ) + A-z + 2H+ = (-XOH 2)2-A -+2 (K6) 2(-XOH°) + Az = (-XO) 2-Az 2 + 2H+ (K7)

where A` = SeO; 2 or SeO4"2. Values in the parenthesis are the standard deviations of the binding constants. d Acidity constants for rutile (Turner and Sassman, 1996) are used. See text in section 4.7. d Log K values in the first line for the aluminosilicate minerals are for aluminol-radionuclide surface complexes, those in the second line are for silanol-radionuclide surface complexes.

Table 4. Summary of Parameters Used for the Determination of Binding Constants for the Sorption of TcO 4- on Mineralsa

Solid Asp,m 2/g mnV,g/L CN, M 1, M log K.b log K.b log KI' log K2c log K3' ref.

Mg6A1209d 70 0.59 1.4x10"5 0.01 6.85 -9.05 5.97 (0.61) 15.33 (0.23) - Kang etal., 1996 Bentonite' 10 0.0682 4.2xI0' 4 0.01 8.33 -9.73 - -1.70 (0.08) Shade et al., 1984

- -7.20 -3.50 (0.35)

'A site density of 2.31 sites/nm2 (from Dzombak and Morel, 1990) was assigned to all minerals. b Acidity constants from Turner and Sassman (1996). 'Binding constants K, through K3 correspond to the following surface reactions:

=XOH° + TcO4 XOH-TcO4 (Kl) -=XOH0 + TcO; + HW = -=XOH2-TcO 4

0 (K2) --XOH° + TcO 4 = =-XO-TcO 4"2

+ W (K3) Values in the parenthesis are the standard deviations of the binding constants.

C 0.1 M NaC1O 4 + 0.011 M NaHCO3 + 0.0012 M KIO 3. d Mg/Al Layered double hydroxide. 'Log K values in the first line for bentonite are for aluminol-radionuclide surface species, those in the second line are for silanol-radionuclide surface species.