Journal of Environmental Radioactivity 136 (2014) 181e187
Contents lists available at ScienceDirect
Journal of Environmental Radioactivity journal homepage: www.elsevier.com/locate/jenvrad
Surface complexation modeling of americium sorption onto volcanic tuff M. Ding a, *, S. Kelkar a, A. Meijer b a b
MS J966, Los Alamos National Laboratory, P.O. Box 1663, Los Alamos, NM 87545, USA GCX Inc., 1389 E. Stoney Canyon Cr., Tuscon, AZ 85737, USA
a r t i c l e i n f o
a b s t r a c t
Article history: Received 15 October 2013 Received in revised form 27 May 2014 Accepted 4 June 2014 Available online
Results of a surface complexation model (SCM) for americium sorption on volcanic rocks (devitrified and zeolitic tuff) are presented. The model was developed using PHREEQC and based on laboratory data for americium sorption on quartz. Available data for sorption of americium on quartz as a function of pH in dilute groundwater can be modeled with two surface reactions involving an americium sulfate and an americium carbonate complex. It was assumed in applying the model to volcanic rocks from Yucca Mountain, that the surface properties of volcanic rocks can be represented by a quartz surface. Using groundwaters compositionally representative of Yucca Mountain, americium sorption distribution coefficient (Kd, L/Kg) values were calculated as function of pH. These Kd values are close to the experimentally determined Kd values for americium sorption on volcanic rocks, decreasing with increasing pH in the pH range from 7 to 9. The surface complexation constants, derived in this study, allow prediction of sorption of americium in a natural complex system, taking into account the inherent uncertainty associated with geochemical conditions that occur along transport pathways. Published by Elsevier Ltd.
Keywords: Americium Sorption Surface complexation model Volcanic tuff
1. Introduction Sorption of dissolved radionuclides onto rock surfaces can dramatically impede downstream movement of radionuclides. It is thus a critical process in assessing the performance of geomedia as a natural barrier at geological repository sites for nuclear wastes. In this study we consider data from the Yucca Mountain, Nevada, previously studied as a potential geologic repository for the disposal of spent nuclear fuel and high-level radioactive waste in the United States (National Research Council, 1990; McCombie, 1997). Considerable experimental effort has been devoted over the last two decades to the measurements of sorption distribution coefficients (Kd, L/Kg) for various radionuclides in rock samples from the vicinity of the Yucca Mountain including americium (Triay et al., 1991). Radionuclides, including americium occur in high-level radioactive wastes generated from nuclear fuel cycles (Ko et al., 2002; Crowley, 1997). As one of the actinides and fission products of environmental concern, americium behavior in the environment has been studied extensively primarily focusing on its chemistry (Mondal et al., 1987; Stadler and Kim, 1988; Moulin
* Corresponding author. Fax: þ1 505 606 2258. E-mail addresses: [email protected], [email protected] (M. Ding). http://dx.doi.org/10.1016/j.jenvrad.2014.06.007 0265-931X/Published by Elsevier Ltd.
et al., 1988; Silva and Nitsche, 1995a,b), speciation (Nash et al., 1998; Degueldre et al., 2004; Stumpf et al., 2006; Choppin, 2007), and interaction with rocks/minerals (Shanbhag and Morse, 1982; Rai et al., 1981; Degueldre and Wernli, 1993; Degueldre et al., 1994; Lujaniene et al., 2006; Yu et al., 2012). A number of studies have reported laboratory determined Kd of americium under conditions applicable to a volcanic tuff geological repository (Allard and Beall, 1979; Allard et al., 1980; Beall et al., 1986; Triay et al., 1991). Kd as a phenomenological empirical modeling parameter has been widely used to develop reactive transport models including radionuclide migration and simulation in performance assessment for nuclear waste repository (Triay et al., 1996; Kelkar et al., 2010). In such transport models, the distribution coefficient Kd is considered a property of the geological medium. In reality, the experimentally derived Kd value is a function of the system chemistry; It is sensitive to spatial variability in the composition of solid phases or for variable chemical conditions, such as pH, ionic strength, alkalinity, or concentrations of complexing ligands that are encountered along a groundwater flow path. These variabilities introduce uncertainties which are difficult to assess quantitatively without thermodynamic constants derived from activity coefficients of surface species. A surface complexation model
182
M. Ding et al. / Journal of Environmental Radioactivity 136 (2014) 181e187
(SCM) which describes sorption behavior as a function of aqueous and mineral surface properties is more suitable for providing sorption parameters with quantifiable uncertainty for performance assessment (Turner et al., 2006). Such geochemical modeling allows extrapolation and interpretation of sorption mechanisms, hence justifying and defending the chosen ranges of Kd values for further model development (Kelkar et al., 2010). To account mechanistically for the observed Kd's, we applied an SCM (Dzombak and Morel, 1990; Davis and Kent, 1990.) to describe the sorption of americium onto volcanic rocks from Yucca Mountain. These rocks primarily consist of devitrified and zeolitic tuffs (Broxton et al., 1986). Two groundwaters, designated J-13 and p#1, representing two distinct groundwater chemistries within the Yucca Mountain are used in the model simulation. The surface complexation constants of the two principal surface reactions involving an americium sulfate and an americium carbonate complex are derived by fitting experimental data from Beall et al. (1986) for americium sorption on quartz as function of pH in synthetic groundwater. To quantify the interacting effects of other dissolved species, the SCM developed in this study included the commonly present cations such as calcium, magnesium, sodium, and aluminum. The computer code PHREEQC v2.3 (Parkhurst and Appelo, 1999) and the thermodynamic database PHREEQCDATA025.DAT were used for our geochemical modeling. The thermodynamic database llnl.dat included in the current PHREEQC-2.17.4799 software was used for comparison as we shall elaborate in the discussion of this paper. We shall demonstrate that the americium reactive transport data in volcanic tuff under dilute groundwater conditions can be inverted, and that the two species derived from the Kd of a quartz system suffice to account for sorption with complex chemistry and mineralogy.
2.1. Sorption process in geomedia and sorption distribution coefficient The following general 1-D continuity equation describes reactive transport of constituent i in a porous medium: (Freeze and Cherry, 1979)
(1)
Where vCi =vt is the concentration change of constituent i in solution per unit of time, Ci is the concentration of constituent i in solution, q is the porosity of solid, D is the effective diffusion coefficient, r is the density of solid, n is the local velocity of pore fluid, Si is the concentration of constituent i sorbed on solid, and vSi =vt is the concentration change of constituent i on solid per unit of time. Eq. (1) describes the spatial and temporal evolution of constituent i in pore fluid due to advective transport, dispersive transport, and solidesolution interactions such as adsorption. Commonly, the sorption distribution coefficient Kd is used to describe the partition of constituent i in solid and solution and is defined as: (Longmuir, 1997)
Kd ¼ Si =Ci
Cinit: Ceq: V M Ceq:
(3)
Where, V and M are the volume of solution and the mass of solid used in the test, respectively. Cinit. is the initial concentration/activity of sorbate in the solution, and Ceq. is the final (equilibrium) concentration/activity of sorbate. 2.2. Surface complexation and sorption of americium on quartz Surface complexation reactions consider mineral surfaces as being composed of specific chemical functional groups that react with dissolved solutes to form surface complexes (Dzombak and Morel, 1990; Davis and Kent, 1990). The equilibria of adsorption reactions between mineral phases and solutes can be described by mass action equations. These equations may be applied to account for variable electrostatic energy, using electrical double layer theory. Electrical charge at the mineral surface is determined by the chemical reactions of the mineral functional groups and the formation of coordinative complexes at the surface. The generalized surface complexation sorption model involves 1) surface proton exchange, and 2) cation sorption on to a neutral surface hydroxyl species ≡SOH. These reactions can be described as: 1) surface proton exchange: ≡SOH ⇔ ≡SO þ Hþ Log Kdeproton
(4)
2) cation sorption ≡SOH þ Me2þ ⇔ ≡SOMeþ þ Hþ Log K
(5)
In Eq. (5), Me2þ represents the sorption of a bivalent metal cation and K the surface complexation constant. In a nonelectrostatic model (Dzombak and Morel, 1990), K can be expressed as:
2. Surface complexation sorption model description
vCi v2 C vC r vSi ¼ D 2i v i vt vx q vt vx
Kd ¼
(2)
Traditionally, Kd tends to be determined in laboratory batch type sorption tests carried out at a fixed liquid volume (L) to solid mass (Kg) ratio. Based on its definition (Eq. (2)), experimentally derived Kd can be expressed as:
K¼
þ ≡SOMeþ f≡SOMeþ H $ $ f≡SOH ½≡SOH Me2þ
(6)
Where {} terms are aqueous activities and [ ] terms are concentrations. f is the activity coefficient. Note that in contrast to models using constant Kd, the thermodynamic constant K in SCM has the capability of describing changes in contaminant adsorption as chemical conditions and aqueous speciation vary. The rational for selection of quartz as adsorbent are two fold. One is that quartz is the major mineral and makes up more than 90 weight percent in devitrified tuff and more than 50 weight percent in zeolitic tuff at Yucca Mountain (Broxton et al., 1986). Most importantly, quartz is a weak adsorbent relative to iron oxides, clays, and zeolites present in trace amount in devitrified and zeolitic tuffs at Yucca Mountain. For performance assessment study purposes, it is customary to take a ‘conservative’ approach using parameter values from the weakest adsorbent resulting in the fastest travel times, i.e. the worst case scenario. The pH dependency of site type, site capacity, protolysis of relevant surface complexation reactions on quartz are summarized in Table 1. 2.3. PHREEQC The computer software PHREEQC (version 2.3) was used to model all the sorption data presented in this study. It is a widely used computer program for speciation, batch-reaction, one-
M. Ding et al. / Journal of Environmental Radioactivity 136 (2014) 181e187 Table 1 Site type, site capacity, protolysis and surface complexation constants of co-existing cations used in the model calculations. Site type
Site density
Reference
Qtz_OH
Pabalan et al., 1998
Qtz_OH ¼ Qtz_O þ Hþ
2.3 sites/ nm2 Log Kprotolysis 7.2
Metal cations sorption Qtz_OH þ Naþ ¼ Qtz_ONa þ Hþ Qtz_OH þ Mg2þ ¼ Qtz_OMgþ þ Hþ
Log K 2.2 6.8
Qtz_OH þ Ca2þ ¼ Qtz_OCaþ þ Hþ
7.8
Qtz_OH þ Al3þ þ 2H2O ¼ Qtz_OAl(OH)2 þ 3Hþ
9.6
Surface protolysis reactions
Dixit and Van Cappellen, 2002 Marmier et al., 1999 Dixit and Van Cappellen, 2002 Dixit and Van Cappellen, 2002 Dixit and Van Cappellen, 2002
dimensional transport, and inverse geochemical calculations (Parkhurst and Appelo, 1999). We selected the non-electrostatic model, one of the three surface-complexation models included in the current code. In this model, there is no explicit calculation of the diffuse-layer composition and no consideration of the effects of surface charge development on the formation of surface complexes. Thus, surface complexes are treated mathematically similar to aqueous complexes without activity coefficient terms. We used the thermodynamic database PHREEQCDATA025.DAT in all calculations. This database was developed by translating the qualified EQ3/6 database DATA0.R4.DAT, which is not accessible to outsiders, for the Yucca Mountain Project into the geochemical equilibrium simulation code PHREEQC. The thermodynamic database llnl.dat, included in the current PHREEQC-2.17.4799 software was used for comparison and is presented in the discussion of this paper.
As mentioned before, there are two dominant rock types in the saturated zone of Yucca Mountain along potential flow paths: devitrified and zeolitic tuff. Devitrified tuff is composed primarily of silica phases (quartz and cristobalite) and alkali feldspar (Broxton et al., 1986). It may also contain minor to trace amounts of minerals such as mica, hematite, calcite, tridymite, kaolinite, and hornblende together with minor amounts of smectite and/or zeolite. Although these generally have higher sorption capacity than quartz and feldspar, the distribution of these trace amounts of altered minerals along the flow path is not well enough established to reliably incorporate them into the field model (Kelkar et al., 2010). We assigned 2.8 m2/g as specific surface area for devitrified tuff in our PHREEQC model calculations. This value is at the lower end of the range of values for 20 tuff samples from Yucca Mountain (Triay et al., 1996), and represents a conservative estimate of the sorption coefficients. Zeolitic tuff contains more than 50% zeolite, with the balance made up of silica phases, alkali feldspar, glass, and trace amounts of clays. As representative specific surface area value we assigned 28 m2/g, representing the low end range of values for them (Triay et al., 1996). Data input to PHREEQC requires values of surface site concentration (GSOH, mol sites/L), specific surface areas (SA, m2/g), and weight of sorbent in contact with a liter of solution (CS, g/L) to determine the surface properties of a sorbent. GSOH is defined as: (Longmuir, 1997)
Ns SA CS NA
abilities of a given weight of adsorbent are proportional to its surface area (SA, m2/g) and surface e site density (NS, sites/m2). We took the surface sites of devitrified tuff to be equivalent to those of quartz, i.e., Ndevitrified (sites/m2) ¼ Nquartz (sites/m2). For zeolitic tuff, we also used the site density of quartz, to simplify the modeling and to compare its sorptive properties to that of devitrified tuff, i.e., Nzeolitic (sites/m2) ¼ Nquartz (sites/m2). Based on Eq. (7), required values of site concentrations of each rock type can be derived for a given SA (m2/g) and CS (g/L). The rock types and their surface parameters used for data input to PHREEQC model calculation are summarized in Table 2. 4. Water compositions used in modeling In the saturated zone at Yucca Mountain, there are two distinct water types in the ambient system (Meijer, 2002). One is the water from well UE-25 J-13 (designated J-13 in Table 3), located on the east side of Fortymile Wash. The other is from well UE-25 P#1 (designated p#1 in Table 3), located to the west of the Fortymile Wash. Details on the geohydrologic setting of Yucca Mountain, can be found elsewhere (Stuckless and Dudley, 2002). We selected J-13 and p#1 water in our calculation as two end-member compositions, intending to bracket the impact of water composition on sorption coefficients. Table 3 presents their chemical composition. Data were also available for the case of a synthetic groundwater, designated ORNL in Table 3, with americium sorption on quartz (Beall et al., 1986). We modeled this experimental result with SCM. It resulted in deriving two surface complexation constants that describe the two surface reactions governing americium sorption on quartz in diluted groundwater. 5. Results 5.1. Derivation of surface complexation constants for americium on quartz
3. Rock types and their surface parameters
GSOH ¼
183
(7)
Where NA is Avogadro's number (6.022 1023 sites/mol of sites), and NS is the surface-site density (sites/m2). Clearly, the sorptive
Scarcity of data necessitated deriving surface complexation constants for americium species on quartz by fitting experimental data reported in the literature (Beall et al., 1986) for the sorption of americium on quartz. As shown in Fig. 1, the sorption coefficients for americium on quartz as a function of pH can be fitted to the experimental data using the following two surface reactions involving an americium sulfate and an americium carbonate complex: þ ð1Þ Qtz OH þ Am3þ þ SO2 Log K ¼ 4:2 4 ¼ Qtz OAmSO4 þ H
(8) þ ð2Þ Qtz OH þ Am3þ þ CO2 Log K ¼ 5:2 3 ¼ Qtz OAmCO3 þ H
(9) Based on Eq:2; Kd ¼
½Qtz OAmSO4 þ ½Qtz OAmCO3 Am3þ
(10)
The modeled curve in Fig. 1 indicates that the Kd for americium Table 2 Summary of rock types, and their surface parameters used in PHREEQC. Rock types
Quartz 2
Devitrified tuff Zeolitic tuff
2.8 2.8 Specific surface area (SA, m /g) Sorbent concentration (CS, g/L) 10 50 4 Site concentration (GSOH, mol sites/L) 1.07 10 5.35 104
28 50 5.35 103
184
M. Ding et al. / Journal of Environmental Radioactivity 136 (2014) 181e187
Table 3 Chemical compositions of ORNL, J-13 and p#1 groundwater used in the calculations. Chemical constituents J-13 (mg/L) p#1 (mg/L) ORNL (mg/L) (Beall et al., 1986) Ca2þ Mg2þ Naþ Kþ SiO2 Cl F SO2 4 HCO 3 pH
11.5 1.8 45.0 5.3 64.2 6.4 2.1 18.1 128.0 6.9
37 10 92.0 5.6 49.0 13.0 3.4 38.0 344.0 6.8
18 4.3 65 3.9 12 70 3.8 9.6 123 8.2
sorption onto quartz peaks at about pH 6, in accordance with the experimentally observed data. At a pH below 6, Kd decreases due to the decrease in americium carbonate complex concentration in that pH region. At pH above 6, the Kd slowly decreases with increasing pH, due to the formation of americium (III) hydroxyl and negatively charged carbonate complexes (Silva and Nitsche, 1995a,b). The repulsion between the negative americium carbonate complex and the negatively charged quartz surface, which has a low Point of Zero Charge (PZC) in the range between 2 and 4 (Kosmulski, 2004; Ding et al., 2000; Ding and de Jong, 2007, 2010), is in all likelihood responsible for this decrease of Kd.
5.2. Americium sorption on devitrified tuff A large number of americium K'ds on devitrified tuff from Yucca Mountain has been measured using J-13 water spiked with americium (Triay et al., 1991). These Kd's range from 100 to 10,000 ml/g. The scattering in Kd's observed in these multiple experiments with the same type of rock sample may reflect the heterogeneity of field samples, differences in solution chemistry, and experimental techniques. To determine the reaction mechanism between americium and devitrified tuff and to address the potential impact of variations in water chemistry on its sorption behavior, we carried out surface complexation modeling to calculate americium Kd's on devitrified tuff in J-13 and p#1, using 2.8 m3/g as specific surface area and the complexation constants from Equations (8) and (9). Our results are shown in Fig. 2. The calculated Kd's for americium on devitrified tuff range from 100 to 5, 000 (L/kg), which fall within the range of experimentally measured Kd values from 100 to 10, 000 (L/Kg). Fig. 2 also shows
Fig. 2. Modeled americium Kd's on devitrified tuff as a function of pH.
that the Kd for americium on devitrified tuff in J-13 reaches a maximum at a pH of 6.3, similar to that for americium sorption onto quartz in ORNL. This can be ascribed to the similarity in chemical composition of ORNL to that of J-13 as listed in Table 3. In contrast, Kd's on devitrified tuff in p#1 decrease with increasing pH in the pH range from 6 to 10, the cutoff at pH 6 being predicated by the arid nature of the study region. As groundwater at Yucca Mountain has a pH in the range between about 8 and 9, our calculations suggest that experimentally derived Kd's should be at the low end of the observed range (100e1000 L/kg). The larger Kd's observed from experiments performed on field samples may be due to interaction of americium with trace and minor amount of more reactive minerals such as clays and iron oxides in the field samples. 5.3. Americium sorption on zeolitic tuff As with devitrified tuff, sorption experiments for americium on Yucca Mountain zeolitic tuff were only carried out with J-13 (Triay et al., 1991). We developed an SCM to investigate the impact of variations in water chemistry on the Kd's of americium for zeolitic tuff. This model is similar to that for devitrified tuff except for using 28 m2/g as the representative surface area value (Table 2). Our calculated Kd results for americium sorption as a function of pH are presented in Fig. 3. Note in Fig. 3, that between a pH from 6 to 10, the Kd of americium on zeolitic tuff in p#1 decreases with increasing pH. In J-13 on the other hand, the americium Kd on zeolitic tuff reaches a maximum at about pH 7. At a pH below 7, Kd decreases with decreasing pH. At a pH above 7, Kd decreases with increasing pH. Fig. 3 clearly shows that variation in water chemistry significantly impacts americium sorption on zeolitic tuff at a pH between 6 and 8. Note also that the range of americium sorption Kd's for zeolitic tuff is substantially larger than that for devitrified tuff, varying between 1000 and 100,000 L/kg. Recalling that the groundwater at Yucca Mountain has pH between 8 and 9, our calculations indicate that experimentally derived Kd's should be at the low end of the range of values to be expected in the field. 6. Discussion 6.1. Impact of the thermodynamic database on the modeling results
Fig. 1. Americium Kd on quartz from the literature (Beall et al., 1986) and our PHREEQC model fit (solid line).
As mentioned previously, we used PHREEQCDATA025.DAT in all calculations. To present this study to the scientific community, we have compared the modeling results using this database to the
M. Ding et al. / Journal of Environmental Radioactivity 136 (2014) 181e187
185
Fig. 4. Distribution of aqueous species of americium in ORNL. Fig. 3. Modeled americium Kd's on zeolitic tuff as a function of pH.
thermodynamic database llnl.dat included in the current PHREEQC-2.17.4799 software. Table 4 presents modeled americium Kd's on quartz in ORNL using the two different databases. The results from Table 4 indicate no significant difference between calculated Kd's using these two databases. Thus, the results calculated using PHREEQCDATA025.DAT are defendable and reproducible.
6.2. Americium speciation and distribution coefficients on quartz and volcanic tuff The nature of interactions between a solute and a solid depends on the surface properties of the solid and the aqueous species of the solute. Here we discuss the effect of americium speciation on its Kd using SCM results of americium sorption on quartz in ORNL. Fig. 4 presents the relative distribution of aqueous species of americium in ORNL in contact with quartz. This Figure illustrates that at pH below 5.5, Am3þ and AmSOþ 4 are the two major solution species. Hence, the americium Kd on quartz at a pH below 5.5 is primarily due to a Qtz_OAmSO4 surface complex as shown in Fig. 5. At a pH above 5.5, the AmCOþ 3 concentration rises with increasing pH. At a pH of about 7, the AmCOþ 3 concentration reaches a maximum, and is the dominant species in solution. Above pH 7, the AmCOþ 3 concentration in solution decreases, whereas the concentrations of americium hydroxyl and negatively charged carbonate increase. Thus, the repulsion between the negative americium carbonate complex and negatively charged quartz surface, is presumably responsible for the experimentally observed and calculated decrease of Kd at a pH above 6, as shown in Figs. 1 and 5. Fig. 5 depicts the quantitative relationship of the americium Kd on quartz and the contribution of the individual americium surface complexes in ORNL with varying pH. Fig. 5 suggests that at a pH below 6, the americium Kd on quartz is solely due to the Qtz_OAmSO4 surface complex. At a pH above 6, the contribution of the Qtz_AmSO4 surface complex to the Kd decreases, whereas the contribution of the Qtz_AmCO3 surface complex to the Kd increases. Thus, at a pH of about 6 both surface complexes contribute to the distribution coefficient resulting in the maximum in Kd value.
Similarly, we calculated Kd's of individual americium species on Yucca Mountain devitrified and zeolitic tuff in J-13 and p#1 with varying pH. Table 5 lists our calculational results. These results summarized in Table 5 suggest that at low pH, the americium Kd is primarily due to interaction between tuffs and an americium sulfate complex. On the other hand at high pH, the americium Kd is primarily due to interaction between tuffs and an americium carbonate complex. As pH increases, the interactions between tuffs and an americium carbonate complex become significant. Our results presented in Figs. 2 and 3 as well as in Table 5 also suggest that the americium Kd calculated in p#1 is similar in value to coefficients calculated in J-13 for devitrified tuff. However, the americium Kd on zeolitic tuff shows a significant difference between J-13 and p#1 at a pH between 6 and 8, indicating that variations in water chemistry affect mainly americium sorption on minerals with a larger surface area. 7. Conclusions Laboratory determined pH dependent sorption of americium in dilute groundwater on the devitrified and zeolitic tuff, which are the two major rock types at the Yucca Mountain, Nevada, can be modeled with two surface reactions. The surface reactions involve an americium sulfate and an americium carbonate complex, taking the specific surface area for the devitrifed tuff to be identical to quartz and that for the zeolitic tuff a factor 10 larger. The surface complexation constants of these two surface complexes are derived by fitting experimental Kd data reported by Beall et al. (1986) with our model calculations. We used two groundwaters, designated J-
Table 4 Comparison of modeled Kd's for americium on quartz in ORNL. pH
4
5
6
7
8
9
10
Kd (L/kg) using PHREEQCDATA025.DAT 57 535 2248 1255 771 452 112 Kd (L/kg) using llnl.dat 54 514 2206 1231 741 430 106
Fig. 5. Contributions of the two americium surface complexes in ORNL to total Kd as a function of pH. Here Total Kd ¼ Kd_Qtz_OAmSO4þKd_Qtz_OAmCO3.
186
M. Ding et al. / Journal of Environmental Radioactivity 136 (2014) 181e187
Table 5 Kd's (L/kg) for individual americium surface species on Yucca Mountain devitrified and zeolitic tuff in J-13 and p#1 groundwaters at different pH. pH
K'ds
of Am(III) on devitrified tuff in J-13
Kd_Qtz_OAmSO4 Kd_Qtz_OAmCO3
K'ds
of Am(III) on devitrified tuff in p#1
Kd_Qtz_OAmSO4 Kd_Qtz_OAmCO3
K'ds of Am(III) on zeolitic tuff in J-13
Kd_Qtz_OAmSO4 Kd_Qtz_OAmCO3
K'ds of Am(III) on zeolitic tuff in p#1
Kd_Qtz_OAmSO4 Kd_Qtz_OAmCO3
(%) (L/Kg) (%) (L/kg) (%) (L/kg) (%) (L/kg) (%) (L/kg) (%) (L/kg) (%) (L/kg) (%) (L/kg)
13 and p#1, as two end-member representatives of Yucca Mountain groundwater compositions. Our model results suggest that for J-13 water, Kd of americium on volcanic tuffs reaches a maximum at a pH between 6 and 7. At a pH above 7, Kd's of americium on volcanic tuffs in J-13 water decrease with increasing pH. For p#1 water, Kd's of americium on the same rocks decrease with increasing pH within the whole modeled range from 6 to 10. The interactions between americium and volcanic rocks decrease, as implied by decreasing in Kd values, likely due to the repulsion between the negatively charged americium carbonate complex and rock surfaces in the system. Such repulsion also explains the larger Kd in J13 water relative to the p#1 water at a specific pH for both rocks (Figs. 2 and 3), as p#1 water has a much higher bicarbonate content than J-13 water (Table 3). Calculated Kd of americium were in the range of 100 ml/g to 10,000 ml/g for devitrified tuff, and 1000 ml/g to 100,000 ml/g for zeolitic tuff in the pH range from 6 to 10. The SCM constants obtained in this study allow prediction of sorption of americium in a wide range of conditions under various scenarios considered in safety assessments. This study convincingly demonstrates that even in the worst case scenario, i.e. in the presence of very weakly adsorbing surfaces, americium is virtually immobile as reflected by its very high Kd value. Acknowledgment The authors thank Florie Caporuscio for verification of thermodynamic database PHREEQCDATA025.DAT assistance, and June Fabryka-Martin for quality assurance reviewing of reported experimental data. This study was funded by the Yucca Mountain Site Characterization Office as part of the Civilian Radioactive Waste Program, US Department of Energy. We are grateful for the comments of two anonymous reviewers and the editor, which substantially enhanced the cogency of this work.
References Allard, B., Beall, G.W., 1979. Sorption of americium on geological media. J. Environ. Health A14 (6), 507e518. Allard, B., Beall, G.W., Krajewski, T., 1980. The sorption of actinides in igneous rocks. Nucl. Technol. 49 (3), 474e480. Beall, G.W., Lee, W.W., Van Luik, A.E., 1986. Americium speciation and distribution coefficients in a granitic ground water. In: Scientific Basis for Nuclear Waste Management IX, Symp. Proceeding, vol. 50, pp. 501e508. Broxton, D.E., Warren, R.G., Hagan, R.C., Luedemann, G., 1986. Chemistry of Diagenetically Altered Tuffs at a Potential Nuclear Waste Repository, Yucca Mountain Nye County, Nevada LA-10802-MS. Los Alamos National Laboratory.
6
7
8
9
10
99.8 4695 0.2 8 99.8 5090 0.2 13 99.9 66,947 0.1 30 99.8 69,191 0.2 142
95.8 2863 4.2 125 93.8 1731 6.2 115 98.8 92,654 1.2 1087 96.4 34,218 3.7 1297
65.9 962 34.1 498 56.2 426 43.8 332 87.8 36,731 12.3 5086 69.2 10,304 30.8 4579
16.7 102 83.3 506 12.2 30 87.8 213 42.9 31,945 57.1 4245 19.7 917 80.3 3745
3.0 4 97.0 131 2.3 1 97.7 44 10.2 92 89.8 808 4.0 38 96.0 920
Choppin, G.R., 2007. Actinide speciation in the environment. J. Radioanal. Nucl. Chem. 273 (3), 695e703. Crowley, K.D., 1997. Nuclear waste disposal: the technical challenges. Phys. Today 50, 32e39. Davis, J.A., Kent, D.B., 1990. Surface complexation modeling in aqueous geochemistry. Mineral-Water Interface Geochem. Rev. Mineral. Ser. 23, 177e260 (Mineralogical Society of America). Degueldre, C., Wernli, B., 1993. Association behavior of 241Am(III) on SiO2(amorphous) and SiO2(quartz) colloids. J. Environ. Radioact. 20, 151e167. Degueldre, C., Ulrich, H.J., Silby, H., 1994. Sorption of 241Am onto montmorillonite, illite and hematite colloids. Radiochim. Acta 65, 173e179. Degueldre, C., Reed, D., Kropf, A.J., Mertz, C., 2004. XAFS study of americium sorbed onto groundwater colloids. J. Synchrotron Rad. 11, 198e203. Ding, M., de Jong, B.H.W.S., Roosendaal, S.J., Vredenburg, A., 2000. XPS studies on the electronic structure of bonding between solids and solutes: adsorption of arsenate, chromate, phosphate, Pb2þ, and Zn2þ on amorphous black ferric oxyhydroxides. Geochim. Cosmochim. Acta 64 (7), 1209e1219. Ding, M., de Jong, B.H.W.S., 2007. Characterization of the surface chemistry of oxides with X-ray photoelectron spectroscopy: assessment regarding surface oxygen valence charge and acid-base properties. In: Sackar, D., Datta, R., Haninggan, R. (Eds.), Developments in Environmental Science, vol. 5. Elsevier. Ding, M., de Jong, B.H.W.S., 2010. Characterizing the surface chemistry of oxides with O1s X-ray photoelectron spectroscopy: comparison of O valence charge in octahedrally and tetrahedrally O coordinated ions. Geochimica 39, 280e295. Dixit, S., Van Cappellen, P., 2002. Surface chemistry and reactivity of biogenic silica. Geochim. Cosmochim. Acta 66 (14), 2559e2568. Dzombak, D.A., Morel, F.M., 1990. Surface Complexation Modeling: Hydrous Ferric Oxide. John Wiley & Sons, New York. Freeze, R.A., Cherry, J.A., 1979. Groundwater. Prentice-Hall, Englewood Cliffs, N.J. Kelkar, S., Ding, M., Chu, S., Robinson, B.A., Arnold, B., Meijer, A., Eddebbarh, A., 2010. Modeling solute transport through saturated zone ground water at 10km scale: example from the Yucca Mountain license application. J. Contam. Hydrol. 117, 7e25. http://dx.doi.org/10.1016/j.jconhyd.2010.05.003. Ko, W.I., Kim, H.D., Yang, M.S., 2002. Radioactive waste arisings from various fuel cycle options. J. Nucl. Sci. Technol. 39 (2), 200e210. Kosmulski, M., 2004. pH dependent surface charging and points of zero charge II update. J. Colloid Interfacial Sci. 275, 214e224. Longmuir, D., 1997. Aqueous Environmental Geochemistry. Prentice-Hall, Inc. Lujaniene, G., Sapolaite, J., Amulevi cius, A., Ma zeika, K., 2006. Retention of cesium, plutonium and americium by engineered and natural barriers. Czechoslov. J. Phys. 56 (Suppl. D), D103eD110. e, A., Fromage, F., 1999. Surface complexation modeling of Yb(III) Marmier, N., Delise and Cs(I) sorption on silica. J. Colloid Interface Sci. 212 (2), 228e233. McCombie, C.,1997. Nuclear waste management worldwide. Phys. Today 50 (6), 56e62. Meijer, A., 2002. Conceptual model of the controls on natural water chemistry at Yucca Mountain, Nevada. Appl. Geochem. 17, 793e805. Mondal, J.U., Raschella, D.L., Haire, R.G., Peterson, J.R., 1987. The enthalpy of solution of 243Am metal and the standard enthalpy of formation of Am3þ (aq). Thermochim. Acta 116, 235e240. Moulin, V., Robouch, P., Vitorge, P., Allard, B., 1988. Environmental behavior of americium (III) in natural waters. Radiochim. Acta 44/45, 33e37. Nash, K.L., Cleveland, J.M., Rees, T.F., 1998. Speciation patterns of actinides in natural waters: a laboratory investigation. J. Environ. Radioact. 7, 131e157. National Research Council, 1990. Rethinking High-level Radioactive Waste Disposal. National Academy Press, Washington DC. Pabalan, R.T., Turner, D.R., Bertetti, F.P., Prikryl, J.D., 1998. Uranium (VI) Sorption onto selected mineral surface, key geochemical parameters. In: Jenne, E. (Ed.), Adsorption of Metals by Geomedia. Academic Press, San Diego, CA, pp. 99e130.
M. Ding et al. / Journal of Environmental Radioactivity 136 (2014) 181e187 Parkhurst, D.L., Appelo, C.A.J., 1999. User’s Guide to PHREEQC (Version 2) e a Computer Program for Speciation, Batch-reaction, One-dimension Transport, and Inverse Geochemical Calculations. U.S. Geological Survey Water-Resources Investigations Report 99-4259. Rai, D., Strickert, R.G., Moore, D.A., Serne, R.J., 1981. Influence of an americium solid phase on americium concentrations in solutions. Geochim. Cosmochim. Acta 45, 2257e2265. Shanbhag, P.M., Morse, J.W., 1982. Americium interaction with calcite and aragonite surfaces in seawater. Geochim. Cosmochim. Acta 46, 241e246. Silva, R.J., Nitsche, H., 1995. Actinide environmental chemistry. Radiochim. Acta 70/ 71, 377e396. Stadler, S., Kim, J.I., 1988. Hydrolysis reactions of Am(III) and Am(V). Radiochim. Acta 44/45, 39e44. Stuckless, J.S., Dudley, W.W., 2002. The geohydrologic setting of Yucca Mountain, Nevada. Appl. Geochem. 17, 659e682.
187
€nel, T., Stumpf, S., Dardenne, K., Henning, C., Foerstendorf, H., Klenze, R., Fangha 2006. Sorption of Am(III) onto 6-line-ferrihydrite and its alteration products: investigation by EXAFS. Environ. Sci. Technol. 40, 3522e3528. Triay, I.R., Meijer, A., Cisneros, M.R., Miller, G.G., Mitchell, A.J., Ott, M.A., Hobart, D.E., Palmer, P.D., Perrin, R.E., Aguilar, R.D., 1991. Sorption of americium in tuff and pure minerals using synthetic and natural groundwaters. Radiochim. Acta 52/ 53, 141e145. Triay, I.R., Cotter, C.R., Huddleston, M.H., Leonard, D.E., Weaver, S.C., Chipera, S.J., Bish, D.L., Meijer, A., Canepa, J.A., 1996. Batch Sorption Results for Neptunium Transport through Yucca Mountain Tuffs, LA-12961-MS. Los Alamos National Laboratory. Turner, D.R., Bertetti, F.P., Pabalan, R.T., 2006. In: Lützenkrichen, J. (Ed.), Chapter 21: Applying Surface Complexation Modeling to Radionuclides Sorption. Surface Complexation Modeling. Elsevier Ltd., pp. 553e604. Yu, T., Wu, W.S., Fan, Q.H., 2012. Sorption of Am(III) on Na-bentonite: effect of pH, ionic strength, temperature and humic acid. Chin. Chem. Lett. 23, 1189e1192.
Contents lists available at ScienceDirect
Journal of Environmental Radioactivity journal homepage: www.elsevier.com/locate/jenvrad
Surface complexation modeling of americium sorption onto volcanic tuff M. Ding a, *, S. Kelkar a, A. Meijer b a b
MS J966, Los Alamos National Laboratory, P.O. Box 1663, Los Alamos, NM 87545, USA GCX Inc., 1389 E. Stoney Canyon Cr., Tuscon, AZ 85737, USA
a r t i c l e i n f o
a b s t r a c t
Article history: Received 15 October 2013 Received in revised form 27 May 2014 Accepted 4 June 2014 Available online
Results of a surface complexation model (SCM) for americium sorption on volcanic rocks (devitrified and zeolitic tuff) are presented. The model was developed using PHREEQC and based on laboratory data for americium sorption on quartz. Available data for sorption of americium on quartz as a function of pH in dilute groundwater can be modeled with two surface reactions involving an americium sulfate and an americium carbonate complex. It was assumed in applying the model to volcanic rocks from Yucca Mountain, that the surface properties of volcanic rocks can be represented by a quartz surface. Using groundwaters compositionally representative of Yucca Mountain, americium sorption distribution coefficient (Kd, L/Kg) values were calculated as function of pH. These Kd values are close to the experimentally determined Kd values for americium sorption on volcanic rocks, decreasing with increasing pH in the pH range from 7 to 9. The surface complexation constants, derived in this study, allow prediction of sorption of americium in a natural complex system, taking into account the inherent uncertainty associated with geochemical conditions that occur along transport pathways. Published by Elsevier Ltd.
Keywords: Americium Sorption Surface complexation model Volcanic tuff
1. Introduction Sorption of dissolved radionuclides onto rock surfaces can dramatically impede downstream movement of radionuclides. It is thus a critical process in assessing the performance of geomedia as a natural barrier at geological repository sites for nuclear wastes. In this study we consider data from the Yucca Mountain, Nevada, previously studied as a potential geologic repository for the disposal of spent nuclear fuel and high-level radioactive waste in the United States (National Research Council, 1990; McCombie, 1997). Considerable experimental effort has been devoted over the last two decades to the measurements of sorption distribution coefficients (Kd, L/Kg) for various radionuclides in rock samples from the vicinity of the Yucca Mountain including americium (Triay et al., 1991). Radionuclides, including americium occur in high-level radioactive wastes generated from nuclear fuel cycles (Ko et al., 2002; Crowley, 1997). As one of the actinides and fission products of environmental concern, americium behavior in the environment has been studied extensively primarily focusing on its chemistry (Mondal et al., 1987; Stadler and Kim, 1988; Moulin
* Corresponding author. Fax: þ1 505 606 2258. E-mail addresses: [email protected], [email protected] (M. Ding). http://dx.doi.org/10.1016/j.jenvrad.2014.06.007 0265-931X/Published by Elsevier Ltd.
et al., 1988; Silva and Nitsche, 1995a,b), speciation (Nash et al., 1998; Degueldre et al., 2004; Stumpf et al., 2006; Choppin, 2007), and interaction with rocks/minerals (Shanbhag and Morse, 1982; Rai et al., 1981; Degueldre and Wernli, 1993; Degueldre et al., 1994; Lujaniene et al., 2006; Yu et al., 2012). A number of studies have reported laboratory determined Kd of americium under conditions applicable to a volcanic tuff geological repository (Allard and Beall, 1979; Allard et al., 1980; Beall et al., 1986; Triay et al., 1991). Kd as a phenomenological empirical modeling parameter has been widely used to develop reactive transport models including radionuclide migration and simulation in performance assessment for nuclear waste repository (Triay et al., 1996; Kelkar et al., 2010). In such transport models, the distribution coefficient Kd is considered a property of the geological medium. In reality, the experimentally derived Kd value is a function of the system chemistry; It is sensitive to spatial variability in the composition of solid phases or for variable chemical conditions, such as pH, ionic strength, alkalinity, or concentrations of complexing ligands that are encountered along a groundwater flow path. These variabilities introduce uncertainties which are difficult to assess quantitatively without thermodynamic constants derived from activity coefficients of surface species. A surface complexation model
182
M. Ding et al. / Journal of Environmental Radioactivity 136 (2014) 181e187
(SCM) which describes sorption behavior as a function of aqueous and mineral surface properties is more suitable for providing sorption parameters with quantifiable uncertainty for performance assessment (Turner et al., 2006). Such geochemical modeling allows extrapolation and interpretation of sorption mechanisms, hence justifying and defending the chosen ranges of Kd values for further model development (Kelkar et al., 2010). To account mechanistically for the observed Kd's, we applied an SCM (Dzombak and Morel, 1990; Davis and Kent, 1990.) to describe the sorption of americium onto volcanic rocks from Yucca Mountain. These rocks primarily consist of devitrified and zeolitic tuffs (Broxton et al., 1986). Two groundwaters, designated J-13 and p#1, representing two distinct groundwater chemistries within the Yucca Mountain are used in the model simulation. The surface complexation constants of the two principal surface reactions involving an americium sulfate and an americium carbonate complex are derived by fitting experimental data from Beall et al. (1986) for americium sorption on quartz as function of pH in synthetic groundwater. To quantify the interacting effects of other dissolved species, the SCM developed in this study included the commonly present cations such as calcium, magnesium, sodium, and aluminum. The computer code PHREEQC v2.3 (Parkhurst and Appelo, 1999) and the thermodynamic database PHREEQCDATA025.DAT were used for our geochemical modeling. The thermodynamic database llnl.dat included in the current PHREEQC-2.17.4799 software was used for comparison as we shall elaborate in the discussion of this paper. We shall demonstrate that the americium reactive transport data in volcanic tuff under dilute groundwater conditions can be inverted, and that the two species derived from the Kd of a quartz system suffice to account for sorption with complex chemistry and mineralogy.
2.1. Sorption process in geomedia and sorption distribution coefficient The following general 1-D continuity equation describes reactive transport of constituent i in a porous medium: (Freeze and Cherry, 1979)
(1)
Where vCi =vt is the concentration change of constituent i in solution per unit of time, Ci is the concentration of constituent i in solution, q is the porosity of solid, D is the effective diffusion coefficient, r is the density of solid, n is the local velocity of pore fluid, Si is the concentration of constituent i sorbed on solid, and vSi =vt is the concentration change of constituent i on solid per unit of time. Eq. (1) describes the spatial and temporal evolution of constituent i in pore fluid due to advective transport, dispersive transport, and solidesolution interactions such as adsorption. Commonly, the sorption distribution coefficient Kd is used to describe the partition of constituent i in solid and solution and is defined as: (Longmuir, 1997)
Kd ¼ Si =Ci
Cinit: Ceq: V M Ceq:
(3)
Where, V and M are the volume of solution and the mass of solid used in the test, respectively. Cinit. is the initial concentration/activity of sorbate in the solution, and Ceq. is the final (equilibrium) concentration/activity of sorbate. 2.2. Surface complexation and sorption of americium on quartz Surface complexation reactions consider mineral surfaces as being composed of specific chemical functional groups that react with dissolved solutes to form surface complexes (Dzombak and Morel, 1990; Davis and Kent, 1990). The equilibria of adsorption reactions between mineral phases and solutes can be described by mass action equations. These equations may be applied to account for variable electrostatic energy, using electrical double layer theory. Electrical charge at the mineral surface is determined by the chemical reactions of the mineral functional groups and the formation of coordinative complexes at the surface. The generalized surface complexation sorption model involves 1) surface proton exchange, and 2) cation sorption on to a neutral surface hydroxyl species ≡SOH. These reactions can be described as: 1) surface proton exchange: ≡SOH ⇔ ≡SO þ Hþ Log Kdeproton
(4)
2) cation sorption ≡SOH þ Me2þ ⇔ ≡SOMeþ þ Hþ Log K
(5)
In Eq. (5), Me2þ represents the sorption of a bivalent metal cation and K the surface complexation constant. In a nonelectrostatic model (Dzombak and Morel, 1990), K can be expressed as:
2. Surface complexation sorption model description
vCi v2 C vC r vSi ¼ D 2i v i vt vx q vt vx
Kd ¼
(2)
Traditionally, Kd tends to be determined in laboratory batch type sorption tests carried out at a fixed liquid volume (L) to solid mass (Kg) ratio. Based on its definition (Eq. (2)), experimentally derived Kd can be expressed as:
K¼
þ ≡SOMeþ f≡SOMeþ H $ $ f≡SOH ½≡SOH Me2þ
(6)
Where {} terms are aqueous activities and [ ] terms are concentrations. f is the activity coefficient. Note that in contrast to models using constant Kd, the thermodynamic constant K in SCM has the capability of describing changes in contaminant adsorption as chemical conditions and aqueous speciation vary. The rational for selection of quartz as adsorbent are two fold. One is that quartz is the major mineral and makes up more than 90 weight percent in devitrified tuff and more than 50 weight percent in zeolitic tuff at Yucca Mountain (Broxton et al., 1986). Most importantly, quartz is a weak adsorbent relative to iron oxides, clays, and zeolites present in trace amount in devitrified and zeolitic tuffs at Yucca Mountain. For performance assessment study purposes, it is customary to take a ‘conservative’ approach using parameter values from the weakest adsorbent resulting in the fastest travel times, i.e. the worst case scenario. The pH dependency of site type, site capacity, protolysis of relevant surface complexation reactions on quartz are summarized in Table 1. 2.3. PHREEQC The computer software PHREEQC (version 2.3) was used to model all the sorption data presented in this study. It is a widely used computer program for speciation, batch-reaction, one-
M. Ding et al. / Journal of Environmental Radioactivity 136 (2014) 181e187 Table 1 Site type, site capacity, protolysis and surface complexation constants of co-existing cations used in the model calculations. Site type
Site density
Reference
Qtz_OH
Pabalan et al., 1998
Qtz_OH ¼ Qtz_O þ Hþ
2.3 sites/ nm2 Log Kprotolysis 7.2
Metal cations sorption Qtz_OH þ Naþ ¼ Qtz_ONa þ Hþ Qtz_OH þ Mg2þ ¼ Qtz_OMgþ þ Hþ
Log K 2.2 6.8
Qtz_OH þ Ca2þ ¼ Qtz_OCaþ þ Hþ
7.8
Qtz_OH þ Al3þ þ 2H2O ¼ Qtz_OAl(OH)2 þ 3Hþ
9.6
Surface protolysis reactions
Dixit and Van Cappellen, 2002 Marmier et al., 1999 Dixit and Van Cappellen, 2002 Dixit and Van Cappellen, 2002 Dixit and Van Cappellen, 2002
dimensional transport, and inverse geochemical calculations (Parkhurst and Appelo, 1999). We selected the non-electrostatic model, one of the three surface-complexation models included in the current code. In this model, there is no explicit calculation of the diffuse-layer composition and no consideration of the effects of surface charge development on the formation of surface complexes. Thus, surface complexes are treated mathematically similar to aqueous complexes without activity coefficient terms. We used the thermodynamic database PHREEQCDATA025.DAT in all calculations. This database was developed by translating the qualified EQ3/6 database DATA0.R4.DAT, which is not accessible to outsiders, for the Yucca Mountain Project into the geochemical equilibrium simulation code PHREEQC. The thermodynamic database llnl.dat, included in the current PHREEQC-2.17.4799 software was used for comparison and is presented in the discussion of this paper.
As mentioned before, there are two dominant rock types in the saturated zone of Yucca Mountain along potential flow paths: devitrified and zeolitic tuff. Devitrified tuff is composed primarily of silica phases (quartz and cristobalite) and alkali feldspar (Broxton et al., 1986). It may also contain minor to trace amounts of minerals such as mica, hematite, calcite, tridymite, kaolinite, and hornblende together with minor amounts of smectite and/or zeolite. Although these generally have higher sorption capacity than quartz and feldspar, the distribution of these trace amounts of altered minerals along the flow path is not well enough established to reliably incorporate them into the field model (Kelkar et al., 2010). We assigned 2.8 m2/g as specific surface area for devitrified tuff in our PHREEQC model calculations. This value is at the lower end of the range of values for 20 tuff samples from Yucca Mountain (Triay et al., 1996), and represents a conservative estimate of the sorption coefficients. Zeolitic tuff contains more than 50% zeolite, with the balance made up of silica phases, alkali feldspar, glass, and trace amounts of clays. As representative specific surface area value we assigned 28 m2/g, representing the low end range of values for them (Triay et al., 1996). Data input to PHREEQC requires values of surface site concentration (GSOH, mol sites/L), specific surface areas (SA, m2/g), and weight of sorbent in contact with a liter of solution (CS, g/L) to determine the surface properties of a sorbent. GSOH is defined as: (Longmuir, 1997)
Ns SA CS NA
abilities of a given weight of adsorbent are proportional to its surface area (SA, m2/g) and surface e site density (NS, sites/m2). We took the surface sites of devitrified tuff to be equivalent to those of quartz, i.e., Ndevitrified (sites/m2) ¼ Nquartz (sites/m2). For zeolitic tuff, we also used the site density of quartz, to simplify the modeling and to compare its sorptive properties to that of devitrified tuff, i.e., Nzeolitic (sites/m2) ¼ Nquartz (sites/m2). Based on Eq. (7), required values of site concentrations of each rock type can be derived for a given SA (m2/g) and CS (g/L). The rock types and their surface parameters used for data input to PHREEQC model calculation are summarized in Table 2. 4. Water compositions used in modeling In the saturated zone at Yucca Mountain, there are two distinct water types in the ambient system (Meijer, 2002). One is the water from well UE-25 J-13 (designated J-13 in Table 3), located on the east side of Fortymile Wash. The other is from well UE-25 P#1 (designated p#1 in Table 3), located to the west of the Fortymile Wash. Details on the geohydrologic setting of Yucca Mountain, can be found elsewhere (Stuckless and Dudley, 2002). We selected J-13 and p#1 water in our calculation as two end-member compositions, intending to bracket the impact of water composition on sorption coefficients. Table 3 presents their chemical composition. Data were also available for the case of a synthetic groundwater, designated ORNL in Table 3, with americium sorption on quartz (Beall et al., 1986). We modeled this experimental result with SCM. It resulted in deriving two surface complexation constants that describe the two surface reactions governing americium sorption on quartz in diluted groundwater. 5. Results 5.1. Derivation of surface complexation constants for americium on quartz
3. Rock types and their surface parameters
GSOH ¼
183
(7)
Where NA is Avogadro's number (6.022 1023 sites/mol of sites), and NS is the surface-site density (sites/m2). Clearly, the sorptive
Scarcity of data necessitated deriving surface complexation constants for americium species on quartz by fitting experimental data reported in the literature (Beall et al., 1986) for the sorption of americium on quartz. As shown in Fig. 1, the sorption coefficients for americium on quartz as a function of pH can be fitted to the experimental data using the following two surface reactions involving an americium sulfate and an americium carbonate complex: þ ð1Þ Qtz OH þ Am3þ þ SO2 Log K ¼ 4:2 4 ¼ Qtz OAmSO4 þ H
(8) þ ð2Þ Qtz OH þ Am3þ þ CO2 Log K ¼ 5:2 3 ¼ Qtz OAmCO3 þ H
(9) Based on Eq:2; Kd ¼
½Qtz OAmSO4 þ ½Qtz OAmCO3 Am3þ
(10)
The modeled curve in Fig. 1 indicates that the Kd for americium Table 2 Summary of rock types, and their surface parameters used in PHREEQC. Rock types
Quartz 2
Devitrified tuff Zeolitic tuff
2.8 2.8 Specific surface area (SA, m /g) Sorbent concentration (CS, g/L) 10 50 4 Site concentration (GSOH, mol sites/L) 1.07 10 5.35 104
28 50 5.35 103
184
M. Ding et al. / Journal of Environmental Radioactivity 136 (2014) 181e187
Table 3 Chemical compositions of ORNL, J-13 and p#1 groundwater used in the calculations. Chemical constituents J-13 (mg/L) p#1 (mg/L) ORNL (mg/L) (Beall et al., 1986) Ca2þ Mg2þ Naþ Kþ SiO2 Cl F SO2 4 HCO 3 pH
11.5 1.8 45.0 5.3 64.2 6.4 2.1 18.1 128.0 6.9
37 10 92.0 5.6 49.0 13.0 3.4 38.0 344.0 6.8
18 4.3 65 3.9 12 70 3.8 9.6 123 8.2
sorption onto quartz peaks at about pH 6, in accordance with the experimentally observed data. At a pH below 6, Kd decreases due to the decrease in americium carbonate complex concentration in that pH region. At pH above 6, the Kd slowly decreases with increasing pH, due to the formation of americium (III) hydroxyl and negatively charged carbonate complexes (Silva and Nitsche, 1995a,b). The repulsion between the negative americium carbonate complex and the negatively charged quartz surface, which has a low Point of Zero Charge (PZC) in the range between 2 and 4 (Kosmulski, 2004; Ding et al., 2000; Ding and de Jong, 2007, 2010), is in all likelihood responsible for this decrease of Kd.
5.2. Americium sorption on devitrified tuff A large number of americium K'ds on devitrified tuff from Yucca Mountain has been measured using J-13 water spiked with americium (Triay et al., 1991). These Kd's range from 100 to 10,000 ml/g. The scattering in Kd's observed in these multiple experiments with the same type of rock sample may reflect the heterogeneity of field samples, differences in solution chemistry, and experimental techniques. To determine the reaction mechanism between americium and devitrified tuff and to address the potential impact of variations in water chemistry on its sorption behavior, we carried out surface complexation modeling to calculate americium Kd's on devitrified tuff in J-13 and p#1, using 2.8 m3/g as specific surface area and the complexation constants from Equations (8) and (9). Our results are shown in Fig. 2. The calculated Kd's for americium on devitrified tuff range from 100 to 5, 000 (L/kg), which fall within the range of experimentally measured Kd values from 100 to 10, 000 (L/Kg). Fig. 2 also shows
Fig. 2. Modeled americium Kd's on devitrified tuff as a function of pH.
that the Kd for americium on devitrified tuff in J-13 reaches a maximum at a pH of 6.3, similar to that for americium sorption onto quartz in ORNL. This can be ascribed to the similarity in chemical composition of ORNL to that of J-13 as listed in Table 3. In contrast, Kd's on devitrified tuff in p#1 decrease with increasing pH in the pH range from 6 to 10, the cutoff at pH 6 being predicated by the arid nature of the study region. As groundwater at Yucca Mountain has a pH in the range between about 8 and 9, our calculations suggest that experimentally derived Kd's should be at the low end of the observed range (100e1000 L/kg). The larger Kd's observed from experiments performed on field samples may be due to interaction of americium with trace and minor amount of more reactive minerals such as clays and iron oxides in the field samples. 5.3. Americium sorption on zeolitic tuff As with devitrified tuff, sorption experiments for americium on Yucca Mountain zeolitic tuff were only carried out with J-13 (Triay et al., 1991). We developed an SCM to investigate the impact of variations in water chemistry on the Kd's of americium for zeolitic tuff. This model is similar to that for devitrified tuff except for using 28 m2/g as the representative surface area value (Table 2). Our calculated Kd results for americium sorption as a function of pH are presented in Fig. 3. Note in Fig. 3, that between a pH from 6 to 10, the Kd of americium on zeolitic tuff in p#1 decreases with increasing pH. In J-13 on the other hand, the americium Kd on zeolitic tuff reaches a maximum at about pH 7. At a pH below 7, Kd decreases with decreasing pH. At a pH above 7, Kd decreases with increasing pH. Fig. 3 clearly shows that variation in water chemistry significantly impacts americium sorption on zeolitic tuff at a pH between 6 and 8. Note also that the range of americium sorption Kd's for zeolitic tuff is substantially larger than that for devitrified tuff, varying between 1000 and 100,000 L/kg. Recalling that the groundwater at Yucca Mountain has pH between 8 and 9, our calculations indicate that experimentally derived Kd's should be at the low end of the range of values to be expected in the field. 6. Discussion 6.1. Impact of the thermodynamic database on the modeling results
Fig. 1. Americium Kd on quartz from the literature (Beall et al., 1986) and our PHREEQC model fit (solid line).
As mentioned previously, we used PHREEQCDATA025.DAT in all calculations. To present this study to the scientific community, we have compared the modeling results using this database to the
M. Ding et al. / Journal of Environmental Radioactivity 136 (2014) 181e187
185
Fig. 4. Distribution of aqueous species of americium in ORNL. Fig. 3. Modeled americium Kd's on zeolitic tuff as a function of pH.
thermodynamic database llnl.dat included in the current PHREEQC-2.17.4799 software. Table 4 presents modeled americium Kd's on quartz in ORNL using the two different databases. The results from Table 4 indicate no significant difference between calculated Kd's using these two databases. Thus, the results calculated using PHREEQCDATA025.DAT are defendable and reproducible.
6.2. Americium speciation and distribution coefficients on quartz and volcanic tuff The nature of interactions between a solute and a solid depends on the surface properties of the solid and the aqueous species of the solute. Here we discuss the effect of americium speciation on its Kd using SCM results of americium sorption on quartz in ORNL. Fig. 4 presents the relative distribution of aqueous species of americium in ORNL in contact with quartz. This Figure illustrates that at pH below 5.5, Am3þ and AmSOþ 4 are the two major solution species. Hence, the americium Kd on quartz at a pH below 5.5 is primarily due to a Qtz_OAmSO4 surface complex as shown in Fig. 5. At a pH above 5.5, the AmCOþ 3 concentration rises with increasing pH. At a pH of about 7, the AmCOþ 3 concentration reaches a maximum, and is the dominant species in solution. Above pH 7, the AmCOþ 3 concentration in solution decreases, whereas the concentrations of americium hydroxyl and negatively charged carbonate increase. Thus, the repulsion between the negative americium carbonate complex and negatively charged quartz surface, is presumably responsible for the experimentally observed and calculated decrease of Kd at a pH above 6, as shown in Figs. 1 and 5. Fig. 5 depicts the quantitative relationship of the americium Kd on quartz and the contribution of the individual americium surface complexes in ORNL with varying pH. Fig. 5 suggests that at a pH below 6, the americium Kd on quartz is solely due to the Qtz_OAmSO4 surface complex. At a pH above 6, the contribution of the Qtz_AmSO4 surface complex to the Kd decreases, whereas the contribution of the Qtz_AmCO3 surface complex to the Kd increases. Thus, at a pH of about 6 both surface complexes contribute to the distribution coefficient resulting in the maximum in Kd value.
Similarly, we calculated Kd's of individual americium species on Yucca Mountain devitrified and zeolitic tuff in J-13 and p#1 with varying pH. Table 5 lists our calculational results. These results summarized in Table 5 suggest that at low pH, the americium Kd is primarily due to interaction between tuffs and an americium sulfate complex. On the other hand at high pH, the americium Kd is primarily due to interaction between tuffs and an americium carbonate complex. As pH increases, the interactions between tuffs and an americium carbonate complex become significant. Our results presented in Figs. 2 and 3 as well as in Table 5 also suggest that the americium Kd calculated in p#1 is similar in value to coefficients calculated in J-13 for devitrified tuff. However, the americium Kd on zeolitic tuff shows a significant difference between J-13 and p#1 at a pH between 6 and 8, indicating that variations in water chemistry affect mainly americium sorption on minerals with a larger surface area. 7. Conclusions Laboratory determined pH dependent sorption of americium in dilute groundwater on the devitrified and zeolitic tuff, which are the two major rock types at the Yucca Mountain, Nevada, can be modeled with two surface reactions. The surface reactions involve an americium sulfate and an americium carbonate complex, taking the specific surface area for the devitrifed tuff to be identical to quartz and that for the zeolitic tuff a factor 10 larger. The surface complexation constants of these two surface complexes are derived by fitting experimental Kd data reported by Beall et al. (1986) with our model calculations. We used two groundwaters, designated J-
Table 4 Comparison of modeled Kd's for americium on quartz in ORNL. pH
4
5
6
7
8
9
10
Kd (L/kg) using PHREEQCDATA025.DAT 57 535 2248 1255 771 452 112 Kd (L/kg) using llnl.dat 54 514 2206 1231 741 430 106
Fig. 5. Contributions of the two americium surface complexes in ORNL to total Kd as a function of pH. Here Total Kd ¼ Kd_Qtz_OAmSO4þKd_Qtz_OAmCO3.
186
M. Ding et al. / Journal of Environmental Radioactivity 136 (2014) 181e187
Table 5 Kd's (L/kg) for individual americium surface species on Yucca Mountain devitrified and zeolitic tuff in J-13 and p#1 groundwaters at different pH. pH
K'ds
of Am(III) on devitrified tuff in J-13
Kd_Qtz_OAmSO4 Kd_Qtz_OAmCO3
K'ds
of Am(III) on devitrified tuff in p#1
Kd_Qtz_OAmSO4 Kd_Qtz_OAmCO3
K'ds of Am(III) on zeolitic tuff in J-13
Kd_Qtz_OAmSO4 Kd_Qtz_OAmCO3
K'ds of Am(III) on zeolitic tuff in p#1
Kd_Qtz_OAmSO4 Kd_Qtz_OAmCO3
(%) (L/Kg) (%) (L/kg) (%) (L/kg) (%) (L/kg) (%) (L/kg) (%) (L/kg) (%) (L/kg) (%) (L/kg)
13 and p#1, as two end-member representatives of Yucca Mountain groundwater compositions. Our model results suggest that for J-13 water, Kd of americium on volcanic tuffs reaches a maximum at a pH between 6 and 7. At a pH above 7, Kd's of americium on volcanic tuffs in J-13 water decrease with increasing pH. For p#1 water, Kd's of americium on the same rocks decrease with increasing pH within the whole modeled range from 6 to 10. The interactions between americium and volcanic rocks decrease, as implied by decreasing in Kd values, likely due to the repulsion between the negatively charged americium carbonate complex and rock surfaces in the system. Such repulsion also explains the larger Kd in J13 water relative to the p#1 water at a specific pH for both rocks (Figs. 2 and 3), as p#1 water has a much higher bicarbonate content than J-13 water (Table 3). Calculated Kd of americium were in the range of 100 ml/g to 10,000 ml/g for devitrified tuff, and 1000 ml/g to 100,000 ml/g for zeolitic tuff in the pH range from 6 to 10. The SCM constants obtained in this study allow prediction of sorption of americium in a wide range of conditions under various scenarios considered in safety assessments. This study convincingly demonstrates that even in the worst case scenario, i.e. in the presence of very weakly adsorbing surfaces, americium is virtually immobile as reflected by its very high Kd value. Acknowledgment The authors thank Florie Caporuscio for verification of thermodynamic database PHREEQCDATA025.DAT assistance, and June Fabryka-Martin for quality assurance reviewing of reported experimental data. This study was funded by the Yucca Mountain Site Characterization Office as part of the Civilian Radioactive Waste Program, US Department of Energy. We are grateful for the comments of two anonymous reviewers and the editor, which substantially enhanced the cogency of this work.
References Allard, B., Beall, G.W., 1979. Sorption of americium on geological media. J. Environ. Health A14 (6), 507e518. Allard, B., Beall, G.W., Krajewski, T., 1980. The sorption of actinides in igneous rocks. Nucl. Technol. 49 (3), 474e480. Beall, G.W., Lee, W.W., Van Luik, A.E., 1986. Americium speciation and distribution coefficients in a granitic ground water. In: Scientific Basis for Nuclear Waste Management IX, Symp. Proceeding, vol. 50, pp. 501e508. Broxton, D.E., Warren, R.G., Hagan, R.C., Luedemann, G., 1986. Chemistry of Diagenetically Altered Tuffs at a Potential Nuclear Waste Repository, Yucca Mountain Nye County, Nevada LA-10802-MS. Los Alamos National Laboratory.
6
7
8
9
10
99.8 4695 0.2 8 99.8 5090 0.2 13 99.9 66,947 0.1 30 99.8 69,191 0.2 142
95.8 2863 4.2 125 93.8 1731 6.2 115 98.8 92,654 1.2 1087 96.4 34,218 3.7 1297
65.9 962 34.1 498 56.2 426 43.8 332 87.8 36,731 12.3 5086 69.2 10,304 30.8 4579
16.7 102 83.3 506 12.2 30 87.8 213 42.9 31,945 57.1 4245 19.7 917 80.3 3745
3.0 4 97.0 131 2.3 1 97.7 44 10.2 92 89.8 808 4.0 38 96.0 920
Choppin, G.R., 2007. Actinide speciation in the environment. J. Radioanal. Nucl. Chem. 273 (3), 695e703. Crowley, K.D., 1997. Nuclear waste disposal: the technical challenges. Phys. Today 50, 32e39. Davis, J.A., Kent, D.B., 1990. Surface complexation modeling in aqueous geochemistry. Mineral-Water Interface Geochem. Rev. Mineral. Ser. 23, 177e260 (Mineralogical Society of America). Degueldre, C., Wernli, B., 1993. Association behavior of 241Am(III) on SiO2(amorphous) and SiO2(quartz) colloids. J. Environ. Radioact. 20, 151e167. Degueldre, C., Ulrich, H.J., Silby, H., 1994. Sorption of 241Am onto montmorillonite, illite and hematite colloids. Radiochim. Acta 65, 173e179. Degueldre, C., Reed, D., Kropf, A.J., Mertz, C., 2004. XAFS study of americium sorbed onto groundwater colloids. J. Synchrotron Rad. 11, 198e203. Ding, M., de Jong, B.H.W.S., Roosendaal, S.J., Vredenburg, A., 2000. XPS studies on the electronic structure of bonding between solids and solutes: adsorption of arsenate, chromate, phosphate, Pb2þ, and Zn2þ on amorphous black ferric oxyhydroxides. Geochim. Cosmochim. Acta 64 (7), 1209e1219. Ding, M., de Jong, B.H.W.S., 2007. Characterization of the surface chemistry of oxides with X-ray photoelectron spectroscopy: assessment regarding surface oxygen valence charge and acid-base properties. In: Sackar, D., Datta, R., Haninggan, R. (Eds.), Developments in Environmental Science, vol. 5. Elsevier. Ding, M., de Jong, B.H.W.S., 2010. Characterizing the surface chemistry of oxides with O1s X-ray photoelectron spectroscopy: comparison of O valence charge in octahedrally and tetrahedrally O coordinated ions. Geochimica 39, 280e295. Dixit, S., Van Cappellen, P., 2002. Surface chemistry and reactivity of biogenic silica. Geochim. Cosmochim. Acta 66 (14), 2559e2568. Dzombak, D.A., Morel, F.M., 1990. Surface Complexation Modeling: Hydrous Ferric Oxide. John Wiley & Sons, New York. Freeze, R.A., Cherry, J.A., 1979. Groundwater. Prentice-Hall, Englewood Cliffs, N.J. Kelkar, S., Ding, M., Chu, S., Robinson, B.A., Arnold, B., Meijer, A., Eddebbarh, A., 2010. Modeling solute transport through saturated zone ground water at 10km scale: example from the Yucca Mountain license application. J. Contam. Hydrol. 117, 7e25. http://dx.doi.org/10.1016/j.jconhyd.2010.05.003. Ko, W.I., Kim, H.D., Yang, M.S., 2002. Radioactive waste arisings from various fuel cycle options. J. Nucl. Sci. Technol. 39 (2), 200e210. Kosmulski, M., 2004. pH dependent surface charging and points of zero charge II update. J. Colloid Interfacial Sci. 275, 214e224. Longmuir, D., 1997. Aqueous Environmental Geochemistry. Prentice-Hall, Inc. Lujaniene, G., Sapolaite, J., Amulevi cius, A., Ma zeika, K., 2006. Retention of cesium, plutonium and americium by engineered and natural barriers. Czechoslov. J. Phys. 56 (Suppl. D), D103eD110. e, A., Fromage, F., 1999. Surface complexation modeling of Yb(III) Marmier, N., Delise and Cs(I) sorption on silica. J. Colloid Interface Sci. 212 (2), 228e233. McCombie, C.,1997. Nuclear waste management worldwide. Phys. Today 50 (6), 56e62. Meijer, A., 2002. Conceptual model of the controls on natural water chemistry at Yucca Mountain, Nevada. Appl. Geochem. 17, 793e805. Mondal, J.U., Raschella, D.L., Haire, R.G., Peterson, J.R., 1987. The enthalpy of solution of 243Am metal and the standard enthalpy of formation of Am3þ (aq). Thermochim. Acta 116, 235e240. Moulin, V., Robouch, P., Vitorge, P., Allard, B., 1988. Environmental behavior of americium (III) in natural waters. Radiochim. Acta 44/45, 33e37. Nash, K.L., Cleveland, J.M., Rees, T.F., 1998. Speciation patterns of actinides in natural waters: a laboratory investigation. J. Environ. Radioact. 7, 131e157. National Research Council, 1990. Rethinking High-level Radioactive Waste Disposal. National Academy Press, Washington DC. Pabalan, R.T., Turner, D.R., Bertetti, F.P., Prikryl, J.D., 1998. Uranium (VI) Sorption onto selected mineral surface, key geochemical parameters. In: Jenne, E. (Ed.), Adsorption of Metals by Geomedia. Academic Press, San Diego, CA, pp. 99e130.
M. Ding et al. / Journal of Environmental Radioactivity 136 (2014) 181e187 Parkhurst, D.L., Appelo, C.A.J., 1999. User’s Guide to PHREEQC (Version 2) e a Computer Program for Speciation, Batch-reaction, One-dimension Transport, and Inverse Geochemical Calculations. U.S. Geological Survey Water-Resources Investigations Report 99-4259. Rai, D., Strickert, R.G., Moore, D.A., Serne, R.J., 1981. Influence of an americium solid phase on americium concentrations in solutions. Geochim. Cosmochim. Acta 45, 2257e2265. Shanbhag, P.M., Morse, J.W., 1982. Americium interaction with calcite and aragonite surfaces in seawater. Geochim. Cosmochim. Acta 46, 241e246. Silva, R.J., Nitsche, H., 1995. Actinide environmental chemistry. Radiochim. Acta 70/ 71, 377e396. Stadler, S., Kim, J.I., 1988. Hydrolysis reactions of Am(III) and Am(V). Radiochim. Acta 44/45, 39e44. Stuckless, J.S., Dudley, W.W., 2002. The geohydrologic setting of Yucca Mountain, Nevada. Appl. Geochem. 17, 659e682.
187
€nel, T., Stumpf, S., Dardenne, K., Henning, C., Foerstendorf, H., Klenze, R., Fangha 2006. Sorption of Am(III) onto 6-line-ferrihydrite and its alteration products: investigation by EXAFS. Environ. Sci. Technol. 40, 3522e3528. Triay, I.R., Meijer, A., Cisneros, M.R., Miller, G.G., Mitchell, A.J., Ott, M.A., Hobart, D.E., Palmer, P.D., Perrin, R.E., Aguilar, R.D., 1991. Sorption of americium in tuff and pure minerals using synthetic and natural groundwaters. Radiochim. Acta 52/ 53, 141e145. Triay, I.R., Cotter, C.R., Huddleston, M.H., Leonard, D.E., Weaver, S.C., Chipera, S.J., Bish, D.L., Meijer, A., Canepa, J.A., 1996. Batch Sorption Results for Neptunium Transport through Yucca Mountain Tuffs, LA-12961-MS. Los Alamos National Laboratory. Turner, D.R., Bertetti, F.P., Pabalan, R.T., 2006. In: Lützenkrichen, J. (Ed.), Chapter 21: Applying Surface Complexation Modeling to Radionuclides Sorption. Surface Complexation Modeling. Elsevier Ltd., pp. 553e604. Yu, T., Wu, W.S., Fan, Q.H., 2012. Sorption of Am(III) on Na-bentonite: effect of pH, ionic strength, temperature and humic acid. Chin. Chem. Lett. 23, 1189e1192.
