Measurement of Tc-99 in Savannah River Site High Activity Waste

D.P.DiPrete, C.C. DiPrete, and R. A. Sigg
Westinghouse Savannah River Company
Aiken, SC 29808

Waste cleanup efforts currently underway at the Savannah River Site have created a need to characterize Tc-99 in the various high activity waste matrices currently in Site inventories. The traditional method our laboratory used for analyzing Tc-99 in higher activity matrices was a solvent-solvent extraction method using Aliquat-336 in xylene, which resulted in the problematic generation of mixed waste. In an effort to eliminate the generation of mixed wastes resulting from the Aliquat 336/xylene process, a variety of different separation methodologies have been studied. Eichrom TEVA solid phase extractions using column technology have been employed in a case by case basis over the last several years. More recently, applications using Eichrom TEVA extraction discs and 3M Empore Tc extraction discs have also been explored.

Tc-99 is a pure beta emitter (E_MAX = 0.29 MeV) with a radiological half-life of 2.1x10⁵ years. Production of the isotope was a by-product of the United States Department of Energy’s plutonium production operations at the Savannah River Site (SRS) in South Carolina. Due to its relatively long half-life, as well as its mobility in the environment, Tc-99 is an isotope of particular concern in Site inventories. It has therefore been an issue for waste characterization and disposition.

SRS wastes fall primarily into three categories: High Level Waste (HLW) tank supernate, a salt-cake formation, and a sludge layer on the bottom of the tank. Both the supernate and the more concentrated salt-cake are comprised predominately of Cs-137 in terms of radiological distribution. Supernate Cs-137 activities are typically in the 1 Curie/L range, with Tc levels 4 to 5 orders of magnitude below Cs-137. Other fission products (i.e. Sb-125, Sn-126, the shorter lived Ru-106 etc…) exist at still lower levels, while Sr-90, the actinides, and the lanthanides exist at trace levels. The sludge radiological inventory is primarily made up of hydroxides of Sr, the actinides and the lanthanides with gross activities in the range of 10¹⁰ dpm/g. Cs-137 and isotopes such as Tc-99 exist at substantially lower relative levels in the sludge. Current waste treatment studies can produce material with skewed supernate/sludge radioisotopic distributions, thereby intensifying the challenge to obtain a good separation of the nuclide of interest in the presence of very high levels of interfering nuclides.

The high levels of interfering species in these waste sample matrices pose considerable challenges to the accurate measurement of Tc-99 by radiochemical methods. The laboratory’s traditional radiochemical methods were based on a quaternary amine/xylene extraction. Quaternary amines have been well established over the past five decades as exceptional extractants for Tc-99 in the pertechnetate ion form.¹ The quaternary amine mixture of choice was Aliquat-336, a technical grade (85% purity) mixture of trioctyl and tridecyl methyl ammonium chloride salts. While this method is quite robust, it leads to the generation of mixed waste, which is problematic in terms of both storage and disposal.

Over the past decade, Eichrom Industries made Aliquat-336 commercially available in the form of its solid phase extractant (SPE) TEVA resin. Initially the SPE was available only as a resin. More recently, both 3M and Eichrom made discs containing Tc extracting SPE available. Numerous examples have appeared in the literature for the uses of these products in environmental applications towards the measurement of Tc-99.^2-6 Methods used in these environmental applications were a starting point for the study of new methodologies for Tc-99 determination in high activity waste.

In order to provide a tracer for the Tc-99 work, the Savannah River Technology Center’s (SRTC) Cf-252 neutron-activation-analysis (NAA) facility⁷is utilized to produce Tc-99m. The present SRTC NAA facility holds eight doubly encapsulated Cf-252 pods in a zircalloy source holder near the bottom of a de-ionized light water, 4-meter deep, 1.2-meter diameter tank. The facility has a capacity for 100mg of Cf-252. The current inventory is approximately 55 milligrams of Cf-252, providing a flux of ~4E8 n*cm^-2s^-1.

Tc-99m is generated by the activation of ammonium molybdate (Mo-98) to form Mo-99 (t1/2 = 66h) which rapidly decays to Tc-99m with a 6 hour half-life. The Mo is dissolved in ammonium hydroxide, from which the Tc-99m is extracted with methyl isobutyl ketone (MIBK). The MIBK containing the Tc-99m is evaporated to dryness under a heat lamp and the Tc-99m is re-dissolved in 10mL of 0.5M nitric acid. A 0.1mL aliquot of a 1000 ppm Ru solution is added to the Tc-99m solution to act as a mass carrier, facilitating precise dispension of the Tc-99m tracer solution. Following the production and separation of the Tc-99m tracer solution, a calibrated coaxial HPGe detector is used to quantify the solution’s activity. Typically, charges of Tc-99m generated using this method contain approximately 5E5 dpm Tc-99m.

Liquid scintillation (LS) spectrometry is applied to quantify Tc-99, a pure beta emitter. However, the Tc-99m tracer can interfere with the LS spectrum of Tc-99. The LS protocol used for Tc-99 analyses analyze three beta energy regions. The region from 3 to 22 keV corresponds to lower energy events from species such as Tc-99, or from Pu-241 or tritium. The second region (22 to 330 keV) captures the higher energy Tc-99 events and extends slightly beyond the Tc-99 beta energy endpoint. The third region (3-2000 keV) corresponds to the entire available energy spectrum, with the exception of the luminescence prone first several keV. Tc-99m counting efficiencies for the 3 described regions of this liquid scintillation instrument are ~ 4.4%, 23%, and 39%, respectively. As a result of these appreciable counting efficiencies for the short-lived tracer, care must be taken to ensure several days have elapsed between radiochemical separation of Tc and quantification of Tc-99 using LS counting.

2b. Solvent-Solvent Extraction for Separation of Tc-99 in High Level Supernate Waste Using Aliquat-336

Traditionally, solvent-solvent extraction using Aliquat-336 was used to determine Tc-99 levels in High Activity Waste. Due to the Cs-137 dose rates associated with high activity waste samples, initial sample dilutions are performed in the SRTC Shielded Cell Facility in order to reduce aliquot dose rates to levels more suitable (<10 mRem whole body) for work in the radiological hoods.

Typically, 1mL aliquots of the diluted solution are spiked with Tc-99m tracer. All of the samples and associated QA samples in a batch are adjusted to ~1.0N nitric acid, and approximately 0.25g of Bio-RAD AMP-1 is added to complex Cs-137. The Cs-137 complexed AMP-1 is then removed with a 0.45 micron Teflon syringe filter. Typical Cs-137 reductions using this method are 2-3 orders of magnitude. The treated aliquots are then oxidized with H₂0₂ and NaVO₃, destroying organic species which may complex or reduce Tc. Nitric concentrations are diluted to < 0.1M. At this point, the Tc-99 is extracted into an organic phase using a 1:7 mixture of 30% Aliquat 336/xylene and sample solution. The organic phase is subsequently washed with 10mL 1.0N nitric followed by 10mL of 4M NaOH. An aliquot of the resultant organic phase is added to Ultima Gold AB, analyzed by gamma spectrometry (to quantify Tc-99m recoveries) followed by LS to quantify Tc-99.

SPE discs capable of separating Tc-99 are commercially available through both Eichrom Industries and 3M Corporation. While information is readily available on the extraction agent used in Eichrom TEVA discs, very little technical information is available concerning the 3M Empore discs since the extractant is proprietary. What little information is available on actual disc use is directed primarily at environmental applications.

In an effort to better understand the behavior of the discs, and to determine feasibility of expanding their use from environmental applications to those of higher activity samples, experiments were designed to compare/evaluate the behaviors of the TEVA and the Empore Tc Discs. Throughout the experiments, flow rates for the TEVA product were observed to be on the order of ~0.5L/min while flow rates for the Empore product were observed to be on the order of ~0.13L/min. Some variability in the flow rates of the Empore product were observed.

As the retention of various analytes on Aliquot 336 is fairly well understood, initially an experiment was designed to evaluate the retention of a number of radioisotopes on the Empore discs after subjecting them to various rinse conditions. Each disc was pre-conditioned with 10mL of 0.01N nitric acid. The discs were then loaded with a solution spiked with some Tc-99 as well as radioisotopes contained in an Amersham Mixed Gamma standard (Co-57, Co-60, Sr-85, Y-88, Cd-109, Sn-113, Cs-137, Hg-203, and Am-241). In SRS wastes, Co-60 is of interest for reactor waste matrices, Cs and Sn are of interest in supernate matrices, and Sr, Y, and Am-241 are of interest in the sludge type matrices. Following pre-conditioning and loading of six discs, each was rinsed using different rinsing protocols as defined in Figure 1.

A second experiment was defined to evaluate the retention of Tc-99 on Eichrom and Empore products under differing experimental conditions. A series of discs was pre-conditioned with DI water and loaded with a Tc-99 spiked 10mL DI water load solution. The discs were then subjected in duplicate to a variety of different rinses as defined in Figure 2, and subsequently analyzed by liquid scintillation analysis.

Experiments were designed to mimic the conditions of the traditional solvent-solvent aliquat 336 extraction using an Eichrom TEVA disc, with the addition of some aggressive rinsing steps to evaluate the disc’s performance on HLW supernate matrices. Dilute sample aliquots were spiked with Tc-99m tracer. The samples were adjusted to be ~1.0N nitric acid, and approximately 0.25g of Bio-RAD AMP-1 was added. The AMP-1 was filtered off with a 0.45 micron Teflon syringe filter. The filtrate samples were diluted to 250mL which served to adjust the nitric acid concentration to ~0.003N. The diluted samples were poured through a TEVA filter and the filters were rinsed successively with 50mL, 100mL, 100mL of 0.01N nitric acid, 100mL of 1.0N nitric acid, and finally 100mL of 0.04M NaOH prior to gamma spectroscopy and LS assays.

A second experiment was conducted in which a TEVA column separation preceded a TEVA disc extraction. A 1mL aliquot of diluted supernate was spiked with Tc-99m tracer, and diluted to adjust the nitric acid concentration to 0.2M. The Bio-RAD AMP-1 step was eliminated, and the solutions were added to 0.2M nitric acid pre-conditioned TEVA columns. The columns were rinsed with two successive 10mL 0.1N nitric rinses. The Tc-99 was then eluted from the columns using 2 successive additions of 5mL 9N nitric acid. The eluate was neutralized with NaOH and poured through TEVA Discs pre-conditioned with 0.1N nitric, prior to assay. This experiment was repeated a second time, with the exception that the columns were rinsed with an additional 20mL of 0.1N nitric acid.

Due to the wide variety of sample matrices analyzed in the laboratory, there is often the need to oxidize the Tc present in samples to the extractable pertechnetate ion form. With this in mind, the method combining TEVA columns and TEVA disks previously described was adjusted to include an aggressive oxidation step. Sample aliquots were spiked with Tc-99m tracer, and diluted to volumes of ~20mL maintaining nitric concentrations of < 0.1N. Samples were oxidized for 1 hour at 80ºC with the addition of 5mL of 30 wt% H2O2 and 0.2mL of 20g/L NaVO3 solution. The oxidized solutions were added to pre-conditioned TEVA columns. Each column was then rinsed successively with 10mL, 10mL, and 20mL of 0.1N nitric. Tc-99 was then eluted from the columns with successive 10mL and 5 mL additions of 9N nitric acid. The eluate was adjusted to a nitric concentration of < 0.01N with NaOH, and added to 0.1N nitric acid pre-conditioned TEVA discs. The TEVA discs were subsequently rinsed with 20mL of 0.1 N nitric acid prior to assays.

Due to the inherent risk of cross contamination encountered when working with extremes of sample activities, only virgin materials are used for each SPE disc. The conventional ground glass apparatus for holding filters in place would be prohibitively expensive. Relatively inexpensive Nalgene disposable filter units with the filter papers swapped out with the SPE discs of choice were mounted on a Fisher vacuum manifold.

LS analyses were carried out with a Packard Instruments (PI) 2550 AB, a PI 2750 AB, or a PI 3150 AB equipped with a bismuth germanate gamma-ray guard. Generally, 20ml of Packard Ultima Gold AB liquid scintillation cocktail loaded with the SPE discs in polyethylene vials were assayed the the LS counters. However, in cases where the low-level mode is employed on the liquid scintillation counter, a faster cocktail, such as Ultima Gold, is required.

Gamma analyses were conducted on a Changer Labs gamma sample changer outfitted with a 40% efficient Ortec GEM P-type up-looking coaxial detector. More recently, a NaI well detector has been employed for the Tc-99m analyses. The MCA software used for the gamma measurements was Canberra’s Genie 2K spectroscopy package.

The solvent-solvent extraction methodology for separating Tc-99 from interfering nuclides in SRS high level waste supernate samples is quite robust. Cs-137 decontamination factors greater than 4E6 were observed (Figure 3) with the method. However, there are several problems associated with this method. Tc-99 yields are only in the range of 20% and tetravalent actinides will co-extract with the Tc-99 to a certain extent.⁸ While a caustic scrub is successful in reducing actinide levels, initial clean-up steps may be required for samples high in plutonium or thorium. ALARA is a concern in employing the solvent-solvent extraction method since it requires a large degree of hands-on contact between the analyst and the process. Finally, the method results in the generation of mixed waste, which is problematic in terms of disposition.

SPE membranes were evaluated to determine whether they could alleviate some of the limitations of the solvent-solvent extraction method. Comparisons of the 3M product and the Eichrom product under various rinse conditions illustrated some differences. In every case, the Tc-99 yield was reduced as the nitric acid concentration was increased. The yield-reduction was more severe for the TEVA disc. As rinse conditions became caustic, Tc-99 yields were substantially higher for the TEVA disc. Figure 2 depicts Tc-99 recoveries for the various rinse conditions. The caustic rinses of the Empore discs led to higher quench levels than observed for neutral and acidic rinses. The TEVA discs had chemi-luminescence interferences in the lower energy window of the LS spectra following the caustic rinses. This is likely attributable to the fact they are substantially more absorbent than the Empore discs, introducing more caustic into the LS cocktail, resulting in luminescence. This luminescence interference precluded using the lower energy window for the TEVA LS analysis. If a basic scrub is desired to remove tetravalent actinides from TEVA, this luminescence issue could be avoided by carrying out a subsequent water rinse prior to transferring the disc to the cocktail.

The Empore product decontaminated the suite of evaluated radionuclides most effectively with a dilute NaOH rinse. Decontamination improved as rinse volumes were increased (Figure 1). Although the decontamination levels shown would be adequate for environmental applications, the levels were not adequate for higher activity applications and at this point continued research efforts were focused on TEVA.

The sequence of steps used to test the TEVA products mimicked to a degree the sequence used in the laboratory’s solvent-solvent Aliquat-336 based extraction for HLW supernate matrices with the notable exception of considerably larger volumes of aqueous waste generation in the TEVA evaluations. Briefly, the steps were acidification, Bio-RAD AMP-1 Cs-137 strip, pH adjustment, sample loading on the disc, and sucessive rinses of dilute nitric, slightly more concentrated nitric, and finally a dilute caustic. Cs-137 decontamination factors achieved (Figure 4) fell far short of that required to analyze a HLW supernate matrix. The Aliquat-336 solvent-solvent based experiments clearly demonstrated that cesium is not retained by the extractant. However, the silica support structure of the discs appeared to become contaminated with an intractable Cs-137 load. The laboratory observed similar problems with the newer TEVA discs having extractant bound to a polyester acrylic based support.

Since high nitric acid conditions can elute Tc-99 from Aliquat-336, a combination of an initial concentration on a TEVA column followed by elution and subsequent re-concentration on a TEVA disc was attempted in a series of experiments. Figure 5 depicts Cs-137 decontamination results attained using these combined column and disk processes compared to those achieved using TEVA discs alone. With the slight addition of rinse, the decontamination levels for Cs-137 approached those of the solvent-solvent method. The final methodology used is provided in Section 2e. Yields of the Tc-99m tracers using this method on high level waste supernate tend to be ~70%, roughly triple the solvent-solvent approach. Because the tetravalent actinides are retained in high nitric acid concentration on quaternary amines as nitrido complexes, this method should be readily extendable to SRS high level waste sludge type matrices. The trivalent actinides, lanthanides, and Sr-90 will pass through the initial TEVA column, while the tetravalent actinides will stick to a degree even in dilute nitric acid. However, the tetravalent actinides will remain on the initial column as the Tc-99 is eluted. Initial experiments indicate plutonium decontamination of~4 orders of magnitude are readily achievable. Further improvement is expected with the addition of a caustic rinse step on the column. Some retention of Ru-106 was observed, but the method appears to provide a cleanup of ~3 orders of magnitude. An initial oxidation step at 100 degrees C, which should volatize the ruthenium at the higher oxidation state as suggested in reference 5, may enhance the cleanup. It has been observed that Co-60, which is high in some SRS matrices, will chase Tc-99 through the separation procedure and appear as an interference. To date, the resolution of that issue has been to add several milligrams of an elemental cobalt carrier in the initial steps of the procedure, which isotopically dilutes the Co-60 interference, and eliminates it from the final LSC spectrum.

Figure 1. Radionuclide Decontamination Factors Using 3M
Tc-99 Disks and Various RinsesProtocols

Figure 2. Tc-99 Yields of Eichrom Discs, and 3M Tc Discs
Under Various Rinsing Protocols

Figure 3. Cs-137 Decontamination Factors using an Aliquot 336 Solvent
Solvent Based Tc-99 Extraction Procedure

Figure 4. Cs-137 Decontamination Factors Using Single
TEVA Disc with Various Rinses