Technetium Incorporation into Hematite (α-Fe2O3) - Environmental

Frances N. Skomurski*, Kevin M. Rosso, Kenneth M. Krupka and B. Pete McGrail ... M. AsmussenMicah D. MillerOdeta QafokuMichelle M. V. SnyderCharles T...
0 downloads 0 Views 1MB Size
Environ. Sci. Technol. 2010, 44, 5855–5861

Technetium Incorporation into Hematite (r-Fe2O3) FRANCES N. SKOMURSKI,* KEVIN M. ROSSO, KENNETH M. KRUPKA,† AND B. PETE MCGRAIL Pacific Northwest National Laboratory, Richland, Washington, U.S.A.

Received January 8, 2010. Revised manuscript received June 3, 2010. Accepted June 28, 2010.

Quantum-mechanical methods were used to evaluate mechanisms for possible structural incorporation of Tc species into the model iron oxide, hematite (R-Fe2O3). Using periodic supercell models, energies for charge-neutral incorporation of Tc4+ or TcO4- ions were calculated using either a Tc4+/Fe2+ substitution scheme on the metal sublattice, or by insertion of TcO4- as an interstitial species within a hypothetical vacancy cluster. Although pertechnetate incorporation is found to be invariably unfavorable, incorporation of small amounts of Tc4+ (at least 2.6 wt %) is energetically feasible. Energy minimized bond distances around this impurity are provided to aid in future spectroscopic identification of these impurity species. The calculations also show that Fe2+ and Tc4+ prefer to cluster in the hematite lattice, attributed to less net Coulombic repulsion relative to that of Fe3+-Fe3+. These modeling predictions are generally consistent with observed selective association of Tc with iron oxide under reducing conditions, and in residual waste solids from underground storage tanks at the U.S. Department of Energy Hanford Site (Washington, U.S.). Here, even though relatively high pH and oxidizing conditions are dominant, Tc incorporation into iron oxides and (oxy)hydroxides is prospectively enabled by prior reduction of TcO4- to Tc4+ via interaction with radiolytic species.

Introduction The 177 single- and double-shell carbon steel underground storage tanks at the U.S. Department of Energy’s (DOE) Hanford Site in southeastern Washington State contain waste liquids and solids from reprocessing of nuclear fuel rods for Pu production (1, 2). While the majority of waste will ultimately be removed from the storage tanks and vitrified, small amounts of residual waste (i.e., waste remaining after final retrieval operations) will remain due to technological limitations. The primary contaminants of concern for tank closure are nominally 99Tc, U, 129I, Cr, nitrite (NO2-), and nitrate (NO3-). The residual wastes are a potentially important source term if infiltrating water is able to leach and transport their associated contaminants into the vadose zone. Tc release is also a concern at sites related to spent-fuel reprocessing worldwide (3-6). 99 Tc is a high risk contaminant due to its relatively long half-life (2.13 × 105 years), high fission yield (∼6%), and high mobility under oxidizing conditions. Tc has many possible * Corresponding author phone: +1-509-371-6368; fax: +1-509371-6354; e-mail: [email protected]. † In memorium. 10.1021/es100069x

 2010 American Chemical Society

Published on Web 07/13/2010

oxidation states, but Tc(IV) and Tc(VII) are the most common in environmental systems. Under oxidizing conditions, Tc(VII) forms the pertechnetate oxyanion [TcO4-(aq)], a highly stable and, due to its anionic nature, mobile species in the environment. By comparison, Tc(IV) is relatively insoluble above ∼pH 3 and precipitates out as the solid TcO2 · nH2O phase, thus limiting the maximum solution concentration and mobility of Tc in reducing environments (7, 8). Estimating future 99Tc release from closed tanks requires, in part, understanding the nature of its association with residual waste solids. Previous work at Hanford has shown that 99Tc is associated in part with recalcitrant iron phases, such as Fe oxides and (oxy)hydroxides (e.g., hematite (RFe2O3), goethite (R-FeOOH), and poorly crystallized forms), identified by bulk X-ray diffraction (XRD) and/or scanning electron microscopy/energy dispersive spectrometry (SEM/ EDS) (9-13). Sequential leach tests show that typically less than 12 wt % (and often below detection limit) of the 99Tc in these residual waste solids is leachable using deionized water (pH 8.8), Ca(OH)2-near saturated solution (pH 11.5), or CaCO3-near saturated solution (pH 8.7). Recently, Cantrell, et al. (9) reported Tc detection by EDS in three Fe (oxy)hydroxide particles in samples of unleached, deionized waterleached, and CaCO3-near-saturated solution leached residual waste from tank C-103. The Tc concentrations ranged from ∼0.6 to ∼1 wt % (13, 14). The data therefore depict conditions of very strong Tc/Fe-phase association in the residual waste, consistent in general with intrinsic structural incorporation into the mineral solid itself. Given these observations, in the present study we endeavor to determine the molecular-scale plausibility of Tc incorporation into bulk iron oxides and (oxy)hydroxides. Hematite was selected as the model material for various reasons. It is known to be present in tanks C-103 and C-106, although other Fe (oxy)hydroxides have also been identified (13). Hematite is also the oxidized, end transformation product of a number of naturally occurring Fe(II) and Fe(III) (oxy)hydroxide parent phases, a process that provides a mechanistic incorporation pathway. For example, at low pH, sufficiently concentrated soluble Fe(III) can yield ferrihydrite (Fe(OH)3) precipitation, which with time and elevated temperature tends to age to more crystalline forms such as hematite or goethite (15-17). Therefore, conceptually, Tc initially associated with ferrihydrite has a chance of incorporation into crystalline Fe(III) forms (7, 18). Furthermore, the Eh-pH stability field of hematite overlaps significantly with that of both oxidized and reduced Tc forms, suggesting prolonged interaction between aging product hematite and most chemical forms of Tc over the long-term (Supporting Information (SI) Figure S1). Finally, with respect to structural compatibility in the lattice, hematite is capable of hosting a wide variety of impuritites, including both cations (e.g., Fe2+, Mg2+, Mn2+, Al3+ 19-21) and anions (e.g., OH-, PO43-, AsO4322-24). With respect to crystal structure, Tc4+ (4d3 high spin) and 3+ Fe (3d5 high spin) in 6-fold coordination have identical crystal radii (0.785 Å (25)). Hence, a direct substitution of Tc4+ for lattice Fe3+ can be readily envisioned if a reasonable charge compensation process is available. Creation of Fe2+ lattice sites is one possibility that has already been observed for other aliovalent substitutions of tetravalent metals into hematite (26, 27). Due to the high variability in tank waste composition, other cations in solution besides Fe2+ could also serve as charge-balancing species (e.g., Mn2+, Mg2+, and Cr2+ (13)); however, those cases are not directly considered VOL. 44, NO. 15, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

5855

here. Furthermore, the strength of the Tc-Fe oxide association is much stronger under reducing conditions, and mixed Tc4+-Fe (oxy)hydroxide phases or reduced TcO2 · nH2O-like phases are observed to form following the reduction of TcO4by by Fe2+ present in the mineral phases, or by excess Fe(II) in solution which sorbs to mineral surfaces (5, 16, 28-37). In contrast, structural compatibility between TcO4- and the hematite lattice is not at all clear. The affinity of TcO4for Fe-bearing minerals is known to be low under most conditions (16, 33, 38-42), and therefore it would not be expected to easily comprise a bulk impurity constituent. However, phosphate (PO43-) incorporation into hematite has been shown and structural details given (23). Based on this observation and given the overlapping thermodynamic stability fields mentioned above, we therefore deemed it sufficiently valuable to contrast Tc4+ substitution into hematite with the prospect of TcO4- incorporation as well. Incorporation energies were calculated at the quantum mechanical level using a hematite supercell model. We evaluated a charge neutral, coupled Tc4+/Fe2+ substitution for two Fe3+ as a function of impurity configuration and concentration, following a closely related case of Ti4+/Fe2+ substitution in hematite (27). We also evaluated interstitial incorporation of TcO4- into a hypothetical vacancy cluster consisting of three Fe3+ nearest neighbor vacancies (i.e., removing an Fe3O4+ unit), charge compensated by protonating under-bonded lattice oxygen anions enveloping the site. This strategy is consistent with experimentally observed charge-balancing mechanisms including creation of vacancies (19-21), and addition of electrons (i.e., forming Fe2+ (27)) or protons (i.e., forming OH- (22)). Our calculations suggest that while the incorporation of small amounts of Tc4+ in the hematite lattice is feasible, incorporation of TcO4is not. Structural analysis is provided and incorporation scenarios relevant to reducing conditions in the environment and oxidizing conditions regarding Hanford tank wastes are discussed.

mimic the polaronic properties of hematite (45, 53) and distinguish ionic constituents in terms of near-formal oxidation states. A 4 × 4 × 1 hematite supercell was chosen to provide sufficient lattice sites to accommodate Tc concentrations as low as 0.6 wt.% (SI Figure S2), consistent with experimentally observed concentrations (9, 13, 14). Energy minimizations of atomic fractional coordinates were performed with lattice parameters held fixed to experimental values to facilitate energy comparisons between models. We selected the antiferromagnetic phase of hematite for consistency with the computational temperature (0 K). Therefore, every other Fe bilayer along a c-axis traverse was assigned an opposite spin (e.g., after (54) Figure 1a). For simplicity, incorporation energies (Einc) are based on differences in total electronic energies and do not take into account possible thermochemical contributions, although we expect such contributions to be relatively minor in the chosen substitution reactions. Incorporation Energies. Unless otherwise noted, Ehem-pure tot hem-impure and Etot refer to energy minimized structures for the pure and Tc-substituted hematite structures, respectively. Lattice energy (Elat) is the difference in energy between the supercell models (Etot) and the sum of the energies of the free constituent ions, calculated in their gas phases. For pure hematite, for instance, Elat is calculated as follows: hem-pure pure ) Etot - nEFe3+ - mEO2Elat

(1)

where n and m refer to the number of Fe and O ions in the hematite supercell, respectively. Here, the incorporation energy (Einc) is defined as the amount of stabilization energy gained per Tc substitution relative to pure hematite. The following expressions are used to evaluate Einc: 4+/Fe2+

Tc Einc

hem-pure hem-impure :Etot + nEFe2+ + nETc4+ f Etot + 2nEFe3+

(2)

Materials and Methods Computational. Total electronic energy (0 K) calculations were performed at the spin-unrestricted Hartree-Fock (UHF) level using the code Crystal06 (43). Crystal uses a linear combination of atomic orbitals defined in terms of local Gaussian functions to represent crystalline orbitals in terms of space-symmetry adapted Bloch functions. The selfconsistent field calculation is performed in reciprocal space. The basis set must be carefully chosen to avoid overly diffuse functions and quasilinear dependence of the Bloch functions. The basis sets used in this study were based on the Durand21d41G electron core potential (ECP) basis sets for Fe3+ and Fe2+, and the Durand-41G ECP basis set for O2-, optimized for hematite (44-46). For H+, Pople’s 3-21 basis set was used (47). For Tc4+ and Tc7+, 21d41 electron core pseudopotential basis sets were optimized using the Hay-Wadt large core pseudopotential (48), and experimental structures of TcO2 (49) and the TcO4- anion (50), respectively (SI Table S1). The program LoptCG (51) was used for the basis set fitting process. Basis set performance was subsequently tested by full geometry optimizations (atom positions and lattice parameters) on bulk hematite, TcO2, Tc2O7, and molecular TcO4(SITable S2 and text). For this purpose, the highest appropriate space group symmetry was imposed for all periodic structures; symmetry was lowered in certain cases to accommodate appropriate magnetic structures in hematite and TcO2. Final charges and spin densities on atoms were determined based on Mulliken analysis (52). Calculated spin densities on all metal atoms were consistent with high-spin configurations. While sacrificing a description of electron correlation effects, the UHF approach enables the system to be treated with a sufficient degree of electron localization to 5856

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 15, 2010

-

+

-

TcO4 /2H hem-pure TcO4 hem-impure Einc :Etot + 2EH+ + Etot f Etot + 3EFe3+ + 4EO2-

(3)

Here, negative Einc implies a favorable reaction, whereas positive Einc implies an unfavorable one. In eq 3, the total energy of an optimized, gas-phase TcO4- molecule is used as a reference. The Einc terms defined above are equivalent to taking the difference in Elat for the impurity case minus impure pure - Elat ). the pure case (i.e., Einc ) Elat

Results Incorporation as Tc4+. Tc4+ incorporation was treated as a case where Tc4+/Fe2+ pairs directly replace two Fe3+ on the metal sublattice. Different arrangements and concentrations of Tc4+/Fe2+ impurity cations were tested (Figure 1b,c). First, single pairs of Fe2+ and Tc4+ were arranged in two ways: Within the same Fe-bilayer in the (001) plane (designated as case I; Figure 1b), and along [001] across a naturally vacant octahedral cavity (case II; Figure 1c). A third arrangement along [001] across a shared octahedral face between FeO6 groups was evaluated but determined to be unusable because of spin contamination issues that negated meangingful comparison to other models. In case I, the total energy dependence on interatomic spacing between the Tc4+ and Fe2+ impurity cations was also evaluated. Additional pairs of Fe2+ and Tc4+ cations were also added to both cases I and II to determine the effect of impurity concentration on incorporation energy. For example, it was of interest to evaluate whether or not a concentration limit could be estimated. In Table 1, calculated incorporation energies are

FIGURE 1. Side view of a defect-free hematite supercell looking down the x-axis (a). Light and dark circles represent Fe3+ with spin down and spin up electrons, respectively; large white circles are O2-. Tc4+/Fe2+-bearing hematite supercells depicting impurities (black, highlighted atoms) in the same Fe bilayer (case I; b), and across an interstitial site parallel to the z-axis (case II; c). The same antiferromagnetic magnetic structure was imposed for all models, but is not shown in b and c. normalized to the number of Tc4+/Fe2+ pairs (i.e., how much stabilization energy is gained per substitution).

Substitution of one Tc4+/Fe2+ pair in the hematite supercell in the case I configuration is computed to be energetically downhill (Einc is negative, Table 1). Given the expected level of accuracy, here we deem the sign of the energy to be more reliable than its magnitude, and relative comparisons of magnitude between models are taken as meaningful. A minimum distance of ∼5 Å (i.e., next-nearest neighbor) between the Tc4+ and the Fe2+ was found to be needed to avoid spin contamination. To determine the total energy dependence on their separation distance, the Fe site to which the extra electron was added was selected to be progressively further away from the Tc4+ site (within the same Fe bilayer) and the energy of the relaxed geometry recalculated. Incorporation energy values become less favorable but still remain negative upon increasing the Tc4+-Fe2+ distance from ∼5 up to ∼18 Å (Figure 2). These results are consistent with there being less Coulombic repulsion between Tc4+-Fe2+ cations compared to Fe3+-Fe3+, or equivalently greater net attraction between Tc4+- Fe2+ pairs. It therefore suggests such impurities would tend to cluster. Calculated Einc for one Tc4+/Fe2+ pair in the case II configuration is more energetically favorable than for case I (Table 1). Given the distance dependence discussed above, we can likewise attribute this difference to the smaller separation in case II (4.1 Å) versus case I (5.1 Å). To determine the dependence of Einc on Tc concentration, additional pairs of Tc4+/Fe2+ were added for case I and II configurations. The pairs were added to the supercell at a maximum possible distance from the original pairs (considering also the pair locations in neighboring cells) to mimic maximum dispersion and to avoid spin contamination issues. In both cases, no significant incorporation energy is gained or lost per pair upon the addition of two pairs (1.3 wt % Tc) (Table 1). However, Einc becomes less favorable for case I when up to four Tc4+/Fe2+ pairs are added to the structure (2.6 wt.% Tc). This result is consistent with a rather low concentration limit for impurity Tc4+ association with iron (oxy)hydroxides (9, 13, 14, 16), and with the lack of any known solid solution system between Tc and Fe, unlike that for Ti and hematite (i.e., ilmenite, FeTiO3 55, 56). Analysis of energy minimized metal-oxygen bond distances for Fe2+ and Tc4+ in the lattice for various models is given in Table 2. The average calculated Fe2+sO bond distance for all substitution models is 0.154 Å greater than average calculated Fe3+sO distance in a pure hematite model, in good agreement with the crystal radii difference for highspin Fe2+ and Fe3+ in 6-fold coordination (0.135 Å) (25). The experimental difference between average short and long Fe3+sO bonds is 0.170 Å; for Fe2+sO, values ranged from 0.179 to 0.186 Å in case I models, and from 0.162 to 0.172 Å in case II models. Hence, the calculations appear to capture appropriate bonding behavior for localized excess electrons (i.e., Fe2+) and local site asymmetry in hematite. For Tc4+ substituted into hematite, the calculated average Tc4+sO bond distance (2.007 Å) is close to that expected for Fe3+sO, consistent with the observation that average experimental TcsO bond distances in TcO2 are only 0.049 Å smaller than average Fe3+sO bond distances in hematite. Calculated differences in short and long Tc4+sO bonds range from 0.049 to 0.052 Å for case I models, and from 0.058 to 0.059 Å for case II models, suggesting a preference for higher site symmetry for Tc4+ in the hematite lattice relative to Fe, and in good agreement with the 0.033 Å experimental value for average short and long Tc4+sO bonds in TcO2. Only limited experimental information is available for comparisons and model validation beyond those above. Following the reaction of TcO4- and Fe(II) in solution under reducing conditions, Zachara et al. (34) observed a mixed ferrihydrite/magnetite-like phase formed in association with VOL. 44, NO. 15, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

5857

TABLE 1. Total, Lattice, And Incorporation Energies for Tc Substitution into Hematitea case 4+

weight % Tc

Etot(ha)

Elat(ha)

Einc(ha)

Einc/ no. pairs

0.6 1.3 2.6 0.6 1.3 0

–8647.0032 –8637.1095 –8617.3037 –8647.0062 –8637.1146 –8656.8970

–548.1295 –548.5603 –549.4036 –548.1325 –548.5654 –547.6988

–0.4307 –0.8615 –1.7048 –0.4337 –0.8666 na

–0.4307 –0.4308 –0.4262 –0.4337 –0.4333 na

0.6 0.6 0.6

–8604.8546 –8604.8523 –8604.8434

–540.3495 –540.3473 –540.3384

7.3512 7.3514 7.3604

na na na

2+

relative to Tc /Fe 1 Tc (case I) 2 Tc (case I) 4 Tc (case I) 1 Tc (case II) 2 Tc (case II) hematite (opt)

relative to TcO4–/2H+ 1 TcO4–(“same”) 1 TcO4-(“near”) 1 TcO4–(“far”) a

Note: 1 hartree (ha) ) 27.21 eV or 627.51 kcal/mol; na ) not applicable.

TABLE 2. Optimized Bond Lengths for Tc4+/Fe2+Substitution in Hematite model

bonding pair 3 short (Å) 3 long (Å) average (Å)

1 Tc (case I)



2.085

2.264

2.175

2 Tc (case I)



2.084 2.084

2.267 2.267

2.176 2.176

4 Tc (case 1)



2.081 2.081 2.080 2.080

2.263 2.265 2.268 2.267

2.172 2.173 2.174 2.174

1 Tc (case II)



2.099

2.261

2.180

FIGURE 2. Incorporation energy as a function of distance between Tc4+/Fe2+ pairs stepped farther away from one another in the same Fe bilayer within the (001) plane.

2 Tc (case II)



2.095 2.095

2.267 2.267

2.181 2.181

supercell



1.989

2.121

2.055

Tc4+. Fitting of extended X-ray absorption fine structure (EXAFS) data suggests that Tc is in 6-fold coordination with TcsO bond lengths of 2.01 Å, which is in good agreement with our calculated values for Tc4+ in hematite in Table 2. In a subsequent study, Peretyazhko et al. (35) sorbed TcO4and Fe(II) onto goethite, hematite, and phyllosilicate substrates under reducing conditions. EXAFS data for the Fe oxides indicates that the resulting Tc4+ is in association with a ferrihydrite-like phase with similar TcsO bond-lengths as described above (∼2.00 Å), whereas on the phyllosilicates, reduced Tc tends to polymerize to form a more TcO2 · nH2Olike phase (35). In a study where TcO4- was reacted with Fe2+-bearing green rust, a TcO2-like bonding environment was deduced by EXAFS data (i.e., 5 TcsO bonds at ∼2.00 Å) (31). While different substrates or experimental conditions may determine whether Tc is structurally incorporated as a mixed Fe-Tc4+ (oxy)hydroxide phase or precipitated at the mineral surface, our calculated incorporation energies and bond lengths appear generally consistent with available experimental data, and a picture of relatively facile Tc4+ substitution into the hematite lattice in small quantities, with charge compensating reduction of lattice iron to Fe2+. Incorporation as TcO4-. For the case of TcO4-, our incorporation models are similar to structures proposed for PO43- in hematite (23). Here a TcO4- anion occupies an interstitial site created by removal of an Fe3O4+ unit from the hematite lattice, charge compensated by locally protonating two under-bonded O2- anions. Three nearest neighbor Fe3+ cations and four O2- anions defining a tetrahedral cavity were removed to create a “docking” site for the TcO4- anion.

experimenta



1.946

2.116

2.031

2+

5858

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 15, 2010

model

bonding

3 short (Å)

3 long (Å)

average (Å)

1 Tc (case I)



1.977

2.029

2.002

2 Tc (case I)



1.978 1.977

2.028 2.028

2.003 2.003

4 Tc (case I)



1.978 1.977 1.978 1.978

2.027 2.027 2.027 2.027

2.003 2.002 2.003 2.003

1 Tc (case II)



1.975

2.033

2.004

2 Tc (case II)



1.974 1.974

2.033 2.033

2.004 2.004

experimentb



1.965

1.998

a

Experimental values for Fe2O3 from ref 60. values for TcO2 from ref 49.

1.982 b

Experimental

Two protons were added, initially located at two of the three positions where Fe3+ cations were removed, to yield a chargeneutral exchange (Figure 3). Experimentally determined interatomic OsO distances on an FeO6 octahedral face in hematite (∼3.1 Å) are similar to the optimal OsO distances for the TcO4- molecule in the gas phase (∼2.8 Å, ref 50 and references therein). Given this observation, and the observation that both the octahedral face of FeO6 and the tetrahedral

FIGURE 3. Starting and optimized models illustrating TcO4- substitution in hematite. Black circles are Tc7+ coordinated by four O2(hatched); gray circles are lattice Fe3+ and large white circles are lattice O2-. Small white circles are H+ arranged in three ways: within the “same” Fe bilayer (a), “near” to one another in neighboring Fe bilayers (b), and “far” from one another in neighboring Fe bilayers (c). Optimized models for the “same” (d), “near” (e), and “far” (f) cases are shown; dotted bond indicates possible Tc-O interaction. Only part of the 4 × 4 × 1 supercell is shown here. face of TcO4- have trigonal symmetry, the prospect of “docking” TcO4- in this configuration was deemed sufficiently intriguing to evaluate. Two of the three selected Fe3+ vacancies lie in the same Fe bilayer, and the third vacancy in an Fe bilayer directly adjacent to it; as such, there are three permutations to occupy these positions with two protons. First, two protons were placed in the two Fe3+ vacancies created in the same bilayer (labeled “same”; Figure 3a). The second and third permutations put one proton in each of the Fe bilayers. One case has the two protons relatively near to each other (labeled “near”; Figure 3b) and the other, relatively far (labeled “far”; Figure 3c). Hence, the three tested models differ primarily in the initial configuration and interatomic distances between the two added chargecompensating protons. As with the Tc4+/Fe2+ substitution scheme, the same 4 × 4 × 1 supercells were used here. Incorporation energies for these models are referenced to the free gas-phase TcO4- anion. Following energy minimization of the geometries of the various models, incorporation energies are similar between the three cases, especially for the “same” and “near” cases, where H-H distances are nearly identical (Table 1). For the “far” case, Einc is less favorable, and H-H distances are ∼1 Å greater than in the other cases (SI Table S3). In contrast to the Tc4+/Fe2+ substitution, however, all values are positive indicating that TcO4- incorporation into the hematite lattice is not energetically favorable. Analysis of energy minimized TcsO bond distances for TcO4- in the hematite supercell provides insight into why this incorporation scheme is not as stable as the Tc4+/Fe2+ substitution. Upon optimization, average TcsO bond distances increase in all three models (SI Table S3), and the Tc7+ center moves from its initial location to get closer to additional neighboring O atoms. The greatest structural distortion is observed for atoms in the “same” case (Figure 3d), where Tc achieves an apparent 5-fold coordination environment, although Tc movement is noted in all three cases (Figure 3d-f). We attribute this tendency to increase the number of coordinating oxygens to the inability of Tc7+ to gain enough electron density from the four closest, TcO4--

like oxygens, which are somewhat coordinatively oversaturated by their connection to the hematite lattice. Mulliken spin density analysis shows that Tc7+ is more electropositive in this configuration relative to that in gas-phase TcO4-. Although the TcO4- molecule appears to be located in a stable binding site, the resulting picture is consistent with the inherent stability of the free pertechnetate anion. Therefore, prospective residence of TcO4- in the hematite structure is unlikely, even in defect sites with sufficient space and charge compensation to accommodate the molecule.

Discussion Our calculations show the energetic feasibility of incorporating small quantities of Tc into the lattice of hematite in the form of Tc4+ (up to at least 2.6 wt %). Here we discuss possible mechanisms under which Tc4+ incorporation can take place, contrasting reducing conditions relevant to environmental settings and the oxidizing conditions specific to Hanford tank wastes. The proposed model of a charge-balanced Tc4+/Fe2+ substitution in the lattice of hematite is consistent with the Eh-pH conditions for the stability of Tc(IV) and Fe(II) species calculated from thermodynamic considerations. The aqueous speciation of Tc is strongly dependent on pH and Eh conditions. Under oxidizing conditions, TcO4- is the dominant species across the entire pH range (0-14) (SI Figure S3a). Tc(IV) exists as the cationic species TcO(OH)+ and TcO2+ below pH 2.2 and 1.3, respectively, and as the neutral aqueous species TcO(OH)20 (aq) at circumneutral to alkaline pH values up to pH ∼13. Above pH ∼13, the TcO(OH)3- species has been identified in solution, associated with increased Tc(IV) solubility at high pH (57). At pH values of less than ∼9.0, the stability fields for the dominant Tc(IV) aqueous species overlap with the Eh-pH conditions calculated for the stability of Fe(II) (SI Figure S3b). At higher pH values, the stability fields for Fe(II) aqueous species FeOH+ and Fe(OH)3- are approximately 100-150 mV lower than the boundary between the Tc(VII)/Tc(IV) aqueous species. Therefore, for Eh-pH conditions where dissolved Fe(II) is present, the following reaction is thermodynamically favorable (7): VOL. 44, NO. 15, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

5859

2+ TcO4 (aq) + 3Fe (aq) + 8H2O f 3Fe(OH)3(s) +

TcO(OH)02(aq) + 5H+ (4) Lieser and Bauscher (7) state that once TcO(OH)20 (aq) forms, Tc sorption onto Fe(III) hydroxide (i.e., ferrihydrite) phases or coprecipitation with Fe(III) may occur. According to Schwertmann and Murad (15), ferrihydrite will eventually age to form hematite or goethite, depending on solution conditions, thus presenting a possible pathway for Tc4+ incorporation in Fe(III) (oxy)hydroxides; however, the phase transformation of ferrihydrite to goethite and hematite is reportedly kinetically hindered in the presence of either Tc(IV) or TcO4- in solution (16). Due to the low solubility of dissolved Fe(III), the predicted Eh-pH conditions where hematite is oversaturated and likely to precipitate extends to anoxic Eh conditions where Tc(IV) and Fe(II) are stable at pH values greater than ∼7 (SI Figure S4). The situation is of course rather different for Hanford tank wastes, where oxidizing, high pH conditions are expected and where neither Tc(IV) nor Fe(II) are stable. Unlike the environmentally relevant conditions described for eq 4, tank waste conditions are extreme in terms of elevated temperatures, radiation fields, and the presence of dissolved organic compounds used in reprocessing (1, 2, 58). Hence, an alternative mechanism for Tc4+ incorporation into hematite is obtained by considering possible Tc reduction via radiolysis products. For example, Lukens et al. (59) showed that insoluble TcO2 · nH2O-like phases and some soluble Tc(IV) phases can form under oxidizing conditions in the presence of a γ radiation source, depending on the specific organic compounds present in solution. In high nitrate (NO3-) environments, radiolytic reductants (e.g., hydrated electrons, eaq-, and organic radicals) are generally scavenged by the reaction: NO3- + eaq- f NO32-, thus preventing direct radiolytic reduction of TcO4- by those species (58). However, in the presence of a γ radiation source and high (radiolytic) NO32- concentrations (e.g., where [NO32-] . [TcO4-]), TcO4was reduced to TcO2 · nH2O by NO32- even under oxidizing conditions, as long as organic radicals (e.g., aminopolycarboxylates) were present that could scavenge oxygen radicals (O-). Essentially, the reduction of TcO4- by NO32- is faster than the hydrolysis of NO32-, allowing the radiolytic NO32to serve as a reductant for TcO4- (58). Therefore, one possible incorporation scenario under these conditions is that if Tc(IV) is present via reaction with radiolysis products (e.g., as the TcO(OH)20 (aq) species at concentrations below the solubility of TcO2 · nH2O), and if precursors to crystalline Fe(III) oxides or (oxy)hydroxides are present (e.g., ferrihydrite), it may be possible to form mixed Tc(IV)-Fe(III) oxide phases through coassociation and aging. Ultimately, the result of these types of reactions could be the recalcitrant Tc-bearing Fe (oxy)hydroxide observed in the residual tank waste.

Acknowledgments This manuscript is dedicated to our late coauthor and friend, Kenneth M. Krupka. This work was supported by the U.S. Department of Energy, Environmental Management Science Program. We also gratefully acknowledge support from the U.S. Department of Energy (DOE) Office of Biological and Environmental Research under the Science Focus Area program at Pacific Northwest National Laboratory (PNNL). Computational molecular modeling was performed at the W. R. Wiley Environmental Molecular Sciences Laboratory, a national scientific user facility sponsored by the U.S. Department of Energy’s Office of Biological and Environmental Research and located at PNNL. PNNL is operated for the DOE by Battelle Memorial Institute under Contract DEAC05-76RL0 1830. 5860

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 15, 2010

Supporting Information Available Additional computational details are included, along with Eh-pH diagrams illustrating the overlapping stability fields of aqueous Tc and Fe species and their corresponding solid phases. This material is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited (1) Gephart, R. E. Hanford: A Conversation about Nuclear Waste and Cleanup; Battelle Press: Columbus, OH, 2003. (2) Mann, F. M. Appendix BsInventory” In RCRA Facility Investigation Report for Hanford Single-Shell Tank Waste Management Areas, DOE/ORP-2008-01, Revision 0; U.S. Department of Energy, Office of River Protection: Richland, WA, 2008; pp 176, http://www.wrpstoc.com/resources/rfi_report_tier_2. (3) Standring, W. J. F.; Oughton, D. H.; Salbu, B. Potential remobilization of Cs-137, Co-60, Tc-99 and Sr-90 from contaminated Mayak sediments in river and estuary environments. Environ. Sci. Technol. 2002, 36 (11), 2330–2337. (4) Fredrickson, J. K.; Zachara, J. M.; Kennedy, D. W.; Kukkadapu, R. K.; McKinley, J. P.; Heald, S. M.; Liu, C. X.; Plymale, A. E. Reduction of TcO4- by sediment-associated biogenic Fe(II). Geochim. Cosmochim. Acta 2004, 68 (15), 3171–3187. (5) Watson, J. H. P.; Ellwood, D. C. The removal of the pertechnetate ion and actinides from radioactive waste streams at Hanford, Washington, USA and Sellafield, Cumbria, UK: the role of ironsulfide-containing adsorbent materials. Nucl. Eng. Des. 2003, 226, 375–385. (6) Orre, S.; Gao, Y.; Drange, H.; Nilsen, J. E. Ø. A reassessment of the dispersion properties of 99Tc in the North Sea and the Norwegian Sea. J. Mar. Syst. 2007, 68, 24–38. (7) Lieser, K. H.; Bauscher, C. Technetium in the hydrosphere and in the geosphere I. Chemistry of technetium and iron in natural waters and influence of the redox potential on the sorption of technetium. Radiochim. Acta 1987, 42, 205–213. (8) Meyer, R. E.; Arnold, W. D.; Case, F. I.; O’Kelley, G. D. Solubilities of Tc(IV) oxides. Radiochim. Acta 1991, 55 (1), 11–18. (9) Cantrell, K. J.; Krupka, K. M.; Deutsch, W. J.; Lindberg, M. J.; Schaef, H. T.; Geiszler, K. N.; Arey, B. W. Hanford Tank 241C-103 Residual Waste Contaminant Release Models and Supporting Data.; PNNL-16738; Pacific Northwest National Laboratory: Richland, WA, 2008a; http://www.osti.gov/ energycitations/. (10) Cantrell, K. J.; Krupka, K. M.; Geiszler, K. N.; Lindberg, M. J.; Arey, B. W.; Schaef, H. T. Hanford Tank 241-S-112 Residual Waste Composition and Leach Test Data.; PNNL-17593; Pacific Northwest National Laboratory: Richland, WA, 2008b; http:// www.osti.gov/energycitations/. (11) Deutsch, W. J.; Krupka, K. M.; Lindberg, M. J.; Cantrell, K. J.; Brown, C. F.; Schaef, H. T. Hanford Tank 241-C-106: Residual Waste Contaminant Release Model and Supporting Data; PNNL15187, Rev. 1; Pacific Northwest National Laboratory: Richland, WA, 2007; http://www.osti.gov/energycitations/. (12) Deutsch, W. J.; Krupka, K. M.; Lindberg, M. J.; Cantrell, K. J.; Brown, C. F.; Schaef, H. T. Hanford Tank 241-C-106: Impact of Cement Reactions on Release of Contaminants from Residual Waste, PNNL-15544; Pacific Northwest National Laboratory: Richland, WA, 2006; http://www.osti.gov/energycitations/. (13) Krupka, K. M.; Cantrell, K. J.; Schaef, H. T.; Arey, B. W.; Heald, S. M.; Deutsch, W. J.; Lindberg, M. J. Characterization of solids in residual wastes from single-shell tanks at the Hanford Site, Washington, U.S.A. In Proceedings of the 35th International Waste Management Conference, March 1-5, 2009, Phoenix, AZ, Paper 9277; WM Symposia, Inc.: Tucson, AZ. (14) Cantrell, K. J.; Krupka, K. M.; Deutsch, W. J.; Lindberg, M. J. Contaminant Release from Residual Waste in Hanford Single Shell Tanks at the Hanford Site, Washington, U.S.A. In WM’09, March 1-5, 2009, Phoenix, Arizona, Paper 9276; WM Symposia, Inc.: Tucson, AZ. (15) Schwertmann, U.; Murad, E. Effect of pH on the formation of goethite and hematite from ferrihydrite. Clays Clay Miner. 1983, 31 (4), 277–284. (16) Walton, F. B.; Paquette, J.; Ross, J. P. M.; Lawrence, W. E. Tc(IV) and Tc(VII) interactions with iron oxyhydroxides. Nucl. Chem. Waste Man. 1986, 6, 121–126. (17) Stumm, W.; Sulzberger, B. The cycling of iron in natural environments: Considerations based on laboratory studies of heterogeneous redox processes. Geochim. Cosmochim. Acta 1992, 56, 3233–3257.

(18) Wakoff, B.; Nagy, K. L. Perrhenate uptake by iron and aluminum oxyhydroxides: an analogue for pertechnetate incorporation in Hanford waste tank sludges. Environ. Sci. Technol. 2004, 38, 1765–1771. ¨ .; McManus, J.; Palmer, (19) Berry, F. J.; Greaves, C.; Helgason, O H. M.; Williams, R. T. Structural and magnetic properties of Sn-, Ti-, and Mg-substituted R-Fe2O3: A study by neutron diffraction and Mo¨ssbauer spectroscopy. J. Solid State Chem. 2000, 151, 157–162. (20) Dieckmann, R. Point defects and transport in haematite (Fe2O3-ε). Philos. Mag. A 1993, 68 (4), 725–745. (21) Catlow, C. R. A.; Corish, J.; Hennessy, J.; Mackrodt, W. C. Atomistic simulation of defect structures and ion transport in R-Fe2O3 and R-Cr2O3. J. Am. Chem. Soc. 1988, 71 (1), 42–49. (22) Gualtieri, A. F.; Venturelli, P. In situ study of the goethite-hematite phase transformation by real time synchrotron powder diffraction. Am. Mineral. 1999, 84, 895–904. (23) Ga´lvez, N.; Barro´n, V.; Torrent, J. Preparation and properties of hematite with structural phosphorus. Clays Clay Miner. 1999, 47 (3), 375–385. (24) Violante, A.; Del Gaudio, S.; Pigna, M.; Ricciardella, M.; Banerjee, D. Coprecipitation of arsenate with metal oxides. 2. Nature, mineralogy, and reactivity of iron(III) precipitates. Environ. Sci. Technol. 2007, 41, 8275–8280. (25) Shannon, R. D. Revised effective ionic radii and systematic studies of interatomic distances in halides and chalcogenides. Acta Crystallogr. 1976, A 32, 751–767. (26) Balko, B. A.; Clarkson, K. M. The effect of doping with Ti(IV) and Sn(IV) on oxygen reduction at hematite electrodes. J. Electrochem. Soc. 2001, 148 (2), E85-E91. (27) Droubay, T.; Rosso, K. M.; Heald, S. M.; McCready, D. E.; Wang, C. M.; Chambers, S. A. Structure, magnetism, and conductivity in epitaxial Ti-doped R-Fe2O3 hematite: Experiment and density functional theory calculations. Phys. Rev. B 2007, 75, 104412. (28) Lee, S. Y.; Bondietti, E. A. In Technetium behavior in sulfide and ferrous iron systems, Scientific Basis for Nuclear Waste Management VI; Brookins, D. G., Ed.; North-Holland: Boston, MA, 1983; pp 315-322. (29) Cui, D.; Eriksen, T. E. Reduction of pertechnetate by ferrous iron in solution: Influence of sorbed and precipitated Fe(II). Environ. Sci. Technol. 1996a, 30, 2259–2262. (30) Cui, D.; Eriksen, T. E. Reduction of pertechnetate in solution by heterogeneous electron transfer from Fe(II)-containing geologic material. Environ. Sci. Technol. 1996b, 30, 2263–2269. (31) Pepper, S. E.; Bunker, D. J.; Bryan, N. D.; Livens, F. R.; Charnock, J. M.; Pattrick, R. A. D.; Collison, D. Treatment of radioactive wastes: An X-ray absorption spectroscopy study of the reaction of technetium with green rust. J. Colloid Interface Sci. 2003, 268, 408–412. (32) Livens, F. R.; Jones, M. J.; Hynes, A. J.; Charnock, J. M.; Mosselmans, J. F. W.; Hennig, C.; Steele, H.; Collison, D.; Vaughan, D. J.; Pattrick, R. A. D.; Reed, W. A.; Moyes, L. N. X-ray absorption spectroscopy studies of reactions of technetium, uranium and neptunium with mackinawite. J. Environ. Radioact. 2004, 74, 211–219. (33) Liu, D. J.; Fan, X. H. Adsorption behavior of 99Tc on Fe, Fe2O3 and Fe3O4. J. Radioanal. Nucl. Chem. 2005, 264 (3), 691–698. (34) Zachara, J. M.; Heald, S. M.; Jeon, B.-H.; Kukkadapu, R. K.; Liu, C.; McKinley, J. P.; Dohnalkova, A. C.; Moore, D. A. Reduction of pertechnetate [Tc(VII)] by aqueous Fe(II) and the nature of solid phase redox products. Geochim. Cosmochim. Acta 2007, 71, 2137–2157. (35) Peretyazhko, T.; Zachara, J. M.; Heald, S. M.; Jeon, B.-H.; Kukkadapu, R. K.; Liu, C.; Moore, D. A.; Resch, C. T. Heterogeneous reduction of Tc(VII) by Fe(II) at the solid-water interface. Geochim. Cosmochim. Acta 2008, 72, 1521–1539. (36) Llorens, I. A.; Deniard, P.; Gautron, E.; Olicard, A.; Fattahi, M.; Jobic, S.; Grambow, B. Structural investigation of coprecipitation of technetium-99 with iron phases. Radiochim. Acta 2008, 96, 569–574. (37) Morris, K.; Livens, F. R.; Charnock, J. M.; Burke, I. T.; McBeth, J. M.; Begg, J. D. C.; Boothman, C.; Lloyd, J. R. An X-ray absorption study of the fate of technetium in reduced and reoxidised

(38) (39) (40)

(41)

(42)

(43)

(44) (45) (46) (47) (48) (49)

(50)

(51) (52) (53) (54) (55) (56) (57) (58) (59)

(60)

sediments and mineral phases. Appl. Geochem. 2008, 23, 603– 617. Ambe, S.; Iwamoto, M.; Maeda, H.; Ambe, F. Multitracer study on adsorption of metal ions on R-Fe2O3. J. Radioanal. Nucl. Chem. 1996, 205 (2), 269–275. Hora´nyi, G.; Joo´, P. Radiometric study of the adsorption of TcO4on hematite. Russ. J. Electrochem. 2000, 36 (11), 1189–1194. Lieser, K. H.; Bauscher, C. Technetium in the hydrosphere and in the geosphere II. Influence of pH, of complexing agents and of some minerals on the sorption of technetium. Radiochim. Acta 1988, 44/45, 125–128. Um, W.; Serne, R. J. Sorption and transport behavior of radionuclides in the proposed low-level radioactive waste disposal facility at the Hanford site, Washington. Radiochim. Acta 2005, 93, 57–63. Kumar, S.; Rawat, N.; Tomar, B. S.; Manchanda, V. K.; Ramanathan, S. Effect of humic acid on the sorption of technetium on hematite colloids using 95mTc and 96Tc as tracers. J. Radioanal. Nucl. Chem. 2007, 274 (2), 229–231. Dovesi, R.; Saunders, V. R.; Roetti, C.; Orlando, R.; ZicovichWilson, C. M.; Pascale, F.; Civalleri, B.; Doll, K.; Harrison, N. M.; Bush, I. J.; D’Arco, P.; Llunell, M. Crystal06; University of Torino: Torino, Italy, 2006. Rosso, K. M.; Dupuis, M. Reorganization energy associated with small polaron mobility in iron oxide. J. Chem. Phys. 2004, 120 (15), 7050–7054. Iordanova, N.; Dupuis, M.; Rosso, K. M. Charge transport in metal oxides: a theoretical study of hematite R-Fe2O3. J. Chem. Phys. 2005, 122, 144305. Wander, M. C. F.; Rosso, K. M.; Schoonen, M. A. A. Structure and charge hopping dynamics in green rust. J. Phys. Chem. C 2007, 111, 11414–11423. Hehre, W. J.; Radom, L.; Schleyer, P. V. R.; Pople, J. A. Ab Initio Molecular Orbital Theory; Wiley: New York, 1986. Hay, P. J.; Wadt, W. R. Ab initio effective core potentials for molecular calculations. Potentials for the transition metal atoms Sc to Hg. J. Chem. Phys. 1985, 82 (1), 270–283. Rodriguez, E. E.; Poineau, F.; Llobet, A.; Sattelberger, A. P.; Bhattacharjee, J.; Waghmare, U. V.; Hartmann, T.; Cheetham, A. K. Structural studies of TcO2 by neutron powder diffraction and first-principles calculations. J. Am. Chem. Soc. 2007, 129 (33), 10244–10248. Cho, H.; de Jong, W. A.; McNamara, B. K.; Rapko, B. M.; Burgeson, I. E. Temperature and isotope substitution effects on the structure and NMR properties of the pertechnetate ion in water. J. Am. Chem. Soc. 2004, 126, 11583–11588. Zicovich-Wilson, C. M. LoptCG; Valencia, Spain, 1998. Mulliken, R. S. Electronic populations analysis on LCAO-MO molecular wave functions. I. J. Chem. Phys. 1955, 23 (10), 1833– 1840. Rosso, K. M.; Smith, D. M.; Dupuis, M. An ab initio model of electron transport in hematite (R-Fe2O3) basal planes. J. Chem. Phys. 2003, 118 (14), 6455–6466. Catti, M.; Valerio, G.; Dovesi, R. Theoretical study of electronic, magnetic, and structural properties of R-Fe2O3 (hematite). Phys. Rev. B 1995, 51 (12), 7441–7450. Harrison, R. J.; Redfern, S. A. T. Short- and long-range ordering in the ilmenite-hematite solid solution. Phys. Chem. Miner. 2001, 28, 399–412. Barth, T. F. W.; Posnjak, E. The crystal structure of ilmenite. Z. Kristallogr. 1934, 88 (4), 265–270. Warwick, P.; Aldridge, S.; Evans, N.; Vines, S. The solubility of technetium(IV) at high pH. Radiochim. Acta 2007, 95, 709–716. Lukens, W. W. J.; Bucher, J. J.; Edelstein, N. M.; Shuh, D. K. Radiolysis of TcO4- in alkaline, nitrate solutions: Reduction by NO32-. J. Phys. Chem. A 2001, 105, 9611–9615. Lukens, W. W. J.; Bucher, J. J.; Edelstein, N. M.; Shuh, D. K. Products of pertechnetate radiolysis in highly alkaline solution: Structure of TcO2 · xH2O. Environ. Sci. Technol. 2002, 36, 1124– 1129. Blake, R. L.; Hessevick, R. E. Refinement of the hematite structure. Am. Mineral. 1966, 51, 123–129.

ES100069X

VOL. 44, NO. 15, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

5861