Crystallization of Rhenium Salts in a Simulated Low-Activity Waste Borosilicate Glass Brian J. Riley,* , John S. McCloy,* Ashutosh Goel, Martin Liezers, Michael J. Schweiger, Juan Liu, Carmen P. Rodriguez, and Dong-Sang Kim Pacific Northwest National Laboratory, PO Box 999, Richland, Washington 99352 This study presents the characterization of salt phases that formed on simulated low-activity waste glass melts during a rhenium solubility study. This study with rhenium salts is also applicable to real applications involving radioactive technetium salts. In this synthesis method, oxide glass powder is mixed with the volatile species, vacuum-sealed in a fused quartz ampoule, and then heated in a furnace. This technique restricts the volatile species to the headspace above the melt but still within the sealed ampoule, thus maximizing the concentration of these species that are in contact with the glass. Above the previously determined solubility of Re 7+ in this glass, a molten salt phase segregated to the top of the melt and crystallized into a solid layer. This salt was analyzed with X-ray diffrac- tion, scanning electron microscopy, energy dispersive spectros- copy, as well as wavelength dispersive spectroscopy and was found to be composed of alkali perrhenates (NaReO 4 , KReO 4 ) and alkali sulfates. Similar crystalline inclusions were found in the bulk of some glasses as well. I. Introduction V ITRIFICATION is the planned technology to immobilize the low-activity radioactive waste (LAW) stored at the Hanford site in Washington State. This waste is a byprod- uct of 45 years of plutonium production and currently occu- pies 2.1 9 10 11 m 3 in underground tanks. Within these tanks are elements that span the periodic table, some of which are long-lived isotopes with half-lives (t 1/2 ) lasting for hundreds of thousands to millions of years, most notably technetium-99 ( 99 Tc, t 1/2 = 2.1 9 10 5 yr) and iodine-129 ( 129 I, t 1/2 = 1.6 9 10 7 yr). The Hanford site tanks contain an estimated ~9 9 10 2 TBq (~1500 kg) of 99 Tc and ~1 TBq (~180 kg) of 129 I. 1 Current calculations estimate that >90% of the 99 Tc inventory and 20% of the 129 I inventory in the tanks will be immobilized in the LAW glass. 2 Technetium is a problematic component in glass for sev- eral reasons. Pertechnetate (TcO 4 ) and TcO 2 are the pri- mary forms of Tc present in glass and are at least somewhat soluble in water, although TcO 2 is much less so. 3,4 Thus, if released, the Tc will migrate easily into the ground water. 5,6 Also, Tc-species are volatile so the Tc retention during the vitrification process tends to be low. 3,79 Although it is not preferred, this lost fraction can be captured with off-gas scrubbing and recycled back into the process loop. Technetium volatilization can be higher when preferen- tially segregated into a molten salt phase, if present, and Kim et al. 9 noted that this is more probable when in the presence of sulfates. It is well-known that sulfates are a cause of concern during melting as are other anionic salts (e.g., NO 2 , NO 3 , Cl ,I , CrO 4 2 ) present in LAW. 1014 This is because a layer of salt, sometimes called “gall”, is formed on the melt pool surface. 15 This gall layer is fed by molten inclu- sions of salt that have a lower viscosity and density than the melt and tend to float to the top of the melt where they remain segregated or are reduced (such as SO 4 2 ? SO 2 ) and leave the melt (i.e., volatilize) as a gas. This behavior is a result of these anions being present in depolymerized regions of the melt where they segregate out of the polymer- ized glassy regions during cooling. 1619 Some of the primary limitations to retention are the Tc solubility and volatility in the selected base glass, which will change with glass composi- tion. The ultimate goal of this study was to determine the solubil- ity of Tc in a borosilicate glass for vitrification of Hanford LAW, but rhenium (Re 7+ ) was used as a Tc surrogate to develop the technique. Rhenium is the most commonly selected Tc surrogate for waste glass applications because it is the most similar in chemistry, ionic size, and speciation in glass. 3 The most common form of Tc found in the LAW simu- lants, as well as the liquid and vapor phases during vitrifica- tion, is Tc 7+ , e.g., KTcO 4 , Tc 2 O 7 . 3 Thus, the species selected for this study were KReO 4 and Re 2 O 7 , both containing Re 7+ . The host glass composition selected is based on the AN-105 low-sulfur tank waste, one of the earliest glasses planned to be processed into LAW glass at Hanford and used in recent glass formulation melter studies of Tc retention. 8,20 To avoid some of the known problems with different behavior noted for halide interaction with Re versus Tc, 21,22 fluorine and chlorine were not included in the base glass composition for this study. A process not common to oxide glass processing was intentionally employed in this study to reduce the volatiliza- tion potential of the rhenium salts from the melt during these experiments. This approach, while being drastically different from an actual melter scenario, was taken to avoid the exper- imental artifacts associated with measuring the solubility of volatile components. Here, glass powder was added to a fused quartz ampoule that was subsequently vacuum-sealed prior to heat treatment (Fig. 1). The ampoule provided an enclosure to reduce the loss of volatile species from the glass to the atmosphere during melting where the volatile vapors were still confined in the headspace above the melt. This spe- cial configuration was designed to determine the solubility of rhenium in this glass, as discussed elsewhere. 23,24 Once melted, the glass was air quenched, mounted in resin, cross-sectioned, and characterized. Several different techniques were employed to characterize the glass structure and the crystalline inclusions within the glass. These tech- niques included: X-ray diffraction (XRD), scanning electron C. Jantzen—contributing editor Manuscript No. 32200. Received October 17, 2012; approved February 11, 2013. Manuscript Authored by Battelle Memorial Institute Under Contract Number DE- AC05-76RLO1830 with the US Department of Energy. The US Government retains and the publisher, by accepting this article for publication, acknowledges that the US Government retains a non-exclusive, paid-up, irrevocable, world-wide license to publish or reproduce the published form of this manuscript, or allows others to do so for US Government purposes. *Member, The American Ceramic Society Author to whom correspondence should be addressed. e-mail: brian.riley@pnnl.gov 1150 J. Am. Ceram. Soc., 96 [4] 1150–1157 (2013) DOI: 10.1111/jace.12280 © 2013 The American Ceramic Society J ournal