Behavior of Technetium in Alkaline Solution: Identification of Non- Pertechnetate Species in High-Level Nuclear Waste Tanks at the Hanford Reservation

Technetium is a long-lived (Tc: 213,000 year half-life) fission product found in nuclear waste and is one of the important isotopes of environmental concern. The known chemistry of technetium suggests that it should be found as pertechnetate, TcO4, in the extremely basic environment of the nuclear waste tanks at the Hanford site. However, other chemical forms of technetium are present in significant amounts in certain tanks, and these nonpertechnetate species complicate the treatment of the waste. The only spectroscopic characterization of these non-pertechnetate species is a series of X-ray absorption near edge structure (XANES) spectra of actual tank waste. To better understand the behavior of technetium under these conditions, we have investigated the reduction of pertechnetate in highly alkaline solution in the presence of compounds found in high-level waste. These results and the X-ray absorption fine structure (XAFS) spectra of these species are compared to the chemical behavior and XANES spectra of the actual non-pertechnetate species. The identity of the nonpertechnetate species is surprising. * Author to whom correspondence should be addressed. Email: [email protected]. Introduction The Hanford Reservation in eastern Washington State is the site of one of the largest and most costly remediation efforts in the U.S. Years of plutonium production has generated 53 million gallons of high-level nuclear waste, which is now stored in 177 underground tanks. The waste consists of three distinct fractions, supernate, saltcake, and sludge. The supernate is an aqueous solution of sodium nitrate, nitrite, and hydroxide, and various organic compounds including citrate, gluconate, formate, oxalate, EDTA, and NTA; in addition, appreciable quantities of Cs, Sr, I, Np and Tc are present in the supernate. Saltcake consists of water-soluble salts that have precipitated during reduction of supernate volume by evaporation and consists mainly of sodium nitrate and nitrite. Sludge consists of the waste components that are insoluble under strongly alkaline conditions and includes most of the fission products and actinides plus large quantities of aluminum and iron oxides and aluminosilicates. The current plan for immobilizing this waste requires separating it into high and low activity streams, which will be vitrified separately to form high and low activity waste glasses. The low activity waste stream mainly consists of supernate and dissolved saltcake, and the high activity waste stream is mainly sludge. Due to the previous performance requirements for the low activity glass, almost all of the Cs and Sr and approximately 80% of the technetium (Tc) needed to be removed from the low activity waste stream and sent to the high activity waste stream as illustrated in Scheme!1. This technetium separation was to be accomplished by ion exchange of pertechnetate, TcO4, the most thermodynamically stable form of technetium at high pH. Although ion exchange was effective for many tanks, work by Schroeder showed that it failed for Complexant Concentrate (CC) waste tanks, including tanks SY-101 and SY-103, which contain a high concentration of organic complexants including nitrilotriacetate (NTA), ethylenediaminetetraacetate (EDTA), citrate, and gluconate. In these tanks, the vast majority of technetium is present as a soluble, lower-valent, non-pertechnetate species (NPS) that is not removed during pertechnetate ion exchange. 4,5 Scheme 1. Simplified illustration of immobilization of high-level nuclear waste at the Hanford Site illustrating the role of Tc separation. The identity of this species is unknown, and its behavior has hampered removal efforts. It is not readily removed by ion exchange, and although the NPS is air-sensitive (it slowly decomposes to pertechnetate), it is difficult to oxidize in practice. The only spectroscopic characterization of the NPS is a series of Tc K-edge X-ray absorption near edge structure (XANES) spectra of CC samples reported by Blanchard (Fig. 1). Although its identity can not be determined directly from these spectra, the NPS was believed to be Tc(IV) based upon the energy of its absorption edge, 7.1 eV lower than that of TcO4. This edge shift is similar to that of TcO2, 6.9 eV lower than that of TcO4. The presence of soluble, lower-valent technetium species is unexpected in light of the known chemistry of technetium; under these conditions, insoluble TcO2•2H2O would be expected rather than soluble complexes. This work identifies the potential candidates for the non-pertechnetate species and identifies technetium complexes that have XANES spectra identical to that of the NPS shown in Fig. 1. Figure 1: Tc K-edge XANES spectra of the non-pertechnetate species (NPS) in tanks a) SY-103, b) SY-101 reported by Blanchard in Ref. . Experimental Section Procedures. Caution: Tc is a b-emitter (Emax = 294 keV, t1/2 = 2 ¥ 10 years). All operations were carried out in a radiochemical laboratory equipped for handling this isotope. Technetium, as NH4TcO4, was obtained from Oak Ridge National Laboratory. The solid NH4TcO4 was contaminated with a large amount of dark, insoluble material. Prolonged treatment of this sample with H2O2 and NH4OH did not appreciably reduce the amount of dark material. Ammonium pertechnetate was separated by carefully decanting the colorless solution from the dark solid. A small amount of NaOH was added to the colorless solution, and the volatile components were removed under vacuum. The remaining solid was dissolved in water, and the colorless solution was removed from the remaining precipitate with a cannula. The concentration of sodium pertechnetate was determined spectrophotometrically at 289 nm (e = 2380 M l cm). UV-visible spectra were obtained using an Ocean-Optics ST2000 spectrometer. X-ray absorption fine structure (XAFS) spectra were acquired at the Stanford Synchrotron Radiation Laboratory (SSRL) at Beamline 4-1 using a Si(220) double crystal monochromator detuned 50% to reduce the higher order harmonic content of the beam. All Tc samples were triply contained inside sealed polyethylene vessels. X-ray absorption fine structure spectra (XAFS) were obtained in the transmission mode at room temperature using Ar filled ionization chambers or in fluorescence yield mode using a multi-pixel Ge-detector system. The spectra were energy calibrated using the first inflection point of the pre-edge peak from the Tc K edge spectrum of an aqueous solution of NH4TcO4 defined as 21044 eV. To determine the Tc K edge charge state shifts, the energies of the Tc K edges at half height were used. EXAFS analysis and radiolysis experiments were carried out as previously described. All operations were carried out in air except as noted. Water was deionized, passed through an activated carbon cartridge to remove organic material and then distilled. Iminodiacetic acid was recrystallized three times from water. All other chemicals were used as received. The Tc(CO)3(H2O)3 stock solution was prepared from TcOCl4(n-Bu4N) by the procedure developed by Alberto then dissolving the reaction product in 0.01M triflic acid. The Tc concentration was determined by liquid scintillation. Solutions for NMR spectroscopy were prepared by addition of 0.10 mL aliquots of the Tc(CO)3(H2O)3 stock solution to 0.90 mL of D2O solutions of 1.1M NaOH with and without 0.11M organic complexant. NMR samples were contained inside a Teflon tube inside a 10 mm screw cap NMR tube. Solutions for XAFS spectroscopy were prepared by addition of 0.20 mL aliquots of the Tc(CO)3(H2O)3 stock solution to 0.80 mL of D2O solutions of 1.1M NaOH with and without 0.11M organic complexant. The Tc(IV) gluconate complex was prepared by reducing a solution of TcO4 (2mM, 1 mL, 2 mmol) in 0.1M potassium gluconate and 1M NaOH with sodium dithionite (2M, 10 mL, 20!mmol). Solutions were sealed under Ar inside 2 mL screw-cap centrifuge tubes, which were placed inside two consecutive heat sealed, heavy walled polyethylene pouches. Pouches were stored under Ar in glass jars sealed with PTFE tape until their spectra were recorded. Results and Discussion Tc(IV) Alkoxide Complexes. As a first step in investigating the behavior of technetium in highly alkaline solutions relevant to high-level waste, solutions of TcO4 in alkaline solution containing organic compounds, including complexants, were irradiated to reduce the TcO4, and the lower-valent technetium products produced were identified. The use of irradiation in these experiments does not imply a similar mechanism for reduction of TcO4 in high-level waste tanks. Both chemical and radiolytic pathways exist for reduction of TcO4 under these conditions, but the radiation-chemical pathway is different from the pathway that is operative here, direct reduction of TcO4 by hydrated electrons from the radiolysis of water. The initial results of the radiolysis experiments showed that none of the carboxylate complexants, citrate, EDTA, or NTA, form stable complexes with Tc(IV) in alkaline solution. Under these conditions, only TcO2•2H2O is produced. However, soluble, lower-valent complexes are produced by the radiolytic reduction of TcO4 in alkaline solution containing the alcohols, ethylene glycol, glyoxylate, and formaldehyde. Although glyoxylate and formaldehyde are aldehydes, they exist as geminal diols in aqueous solution and therefore can act as alkoxide ligands. The EXAFS spectrum and structure of the Tc(IV) glyoxylate complex are shown in Fig.!2, and the structural parameters are given in Table 1. The structure is very similar to that of the well known (H2EDTA)2Tc2(m-O)2 complex with the EDTA ligands replaced by glyoxylate ligands, presumably acting as diolate ligands. Figure 2. EXAFS spectrum and Fourier transform of the Tc(IV) species formed by radiolysis of TcO4 in a solution of 0.1 glyoxylic acid in 1M NaOH; data are shown in gray and the fit in black. The structure of the complex consistent with the EXAFS spectrum is shown on the right. Table 1. Structural parameters of soluble radiolysis product derived from EXAFS. Scattering Path Coordination Number Distance (Å) Debye-Waller Parameter (Å) DE0 (eV) Tc-O 6.7(3) 2.008(3) 0.0058(5) -7.9 Tc-Tc 0.7(1) 2.582(4) 0.003* -7.9 Tc-O-Tc-O 6* 4.06(2) 0.002(3) -7.9 a) Numbers in parenthesis are the standard deviation of the given parameter derived from leastsquares fit to the EXAFS data. The standard deviations do not indicate the accuracy of the numbers; they are an indication of the agreement between the model and the data. In general, coordination numbers have an error of ±25% and bond distances have an error of ±0.5% when compared to data from crystallography. b) Parameters with an asterisk were not allowed to vary during analysis. c) E0 was refined as a global parameter for all scattering paths. The large negative value results from the definition of E0 in EXAFSPAK. d) This scattering path is a 4-legged multiple scattering path between the trans ligands of the technetium coordination sphere. These radiolysis experiments clearly show that soluble Tc(IV) alkoxide complexes can be formed in highly alkaline solution under conditions similar to those found in high-level waste. However, none of the potential ligands examined are present in high-level waste in sufficient concentrations to account for the formation of the soluble non-pertechnetate species. The potential alkoxide ligand present in large quantities in CC waste is gluconate. Moreover, gluconate can act as a tridentate alkoxide ligand (using the hydroxyl groups on carbon atoms 24). The resulting Tc(gluconate)2 complex would presumably be very similar to an analogous complex of Tc(IV) with two tridentate alkoxide ligands described by Anderegg. This complex, Tc[(OCH2)3CN(CH3)]2, is the most hydrolytically stable of the Tc(IV) alkoxide complexes. While most Tc(IV) complexes are stable only above pH 10, Tc[(OCH2)3CN(CH3)]2 is stable towards hydrolysis above pH 4. Consequently, an analogous Tc(IV) gluconate complex would be expected to be quite hydrolytically stable. The colorless Tc(IV) gluconate complex was prepared in situ by reducing TcO4 with dithionate in a solution of 0.1M gluconate and 1M NaOH. The EXAFS spectrum and its Fourier transform of Tc(IV) gluconate are shown in Fig. 3; fit parameters are given in Table 2. The coordination environment of the Tc center is simple: 6 O neighbors at 2.01 Å and 6 C neighbors at 3.37 Å. The bond distances are similar to the aforementioned Tc[(OCH2)3CN(CH3)]2. Although the coordination geometry of the coordinated gluconate ligand cannot be determined directly from the EXAFS data, the similarity between the Tc-O distances in Tc(IV) gluconate and in Tc[(OCH2)3CN(CH3)]2 strongly suggests that the gluconate ligand is coordinated to the Tc center by three hydroxyl groups, as illustrated in Fig. 3, rather than a carboxylate and two hydroxyl