Transmission electron microscopy (TEM) and associated electron energy-loss spectroscopy (EELS) study show intergrowth of Ce4+-rich pyrochlore (metamict) and Ce3+-rich pyrochlore (partially metamict) in a Ce-rich pyrochlore from a rare earth element (REE) ore deposit of Inner Mongolia, Northern China. The partially metamict material is Ba-free and dominated by Ce3+. However, the metamict material is Ba-bearing and dominated by Ce3+,. The Ce4+-rich pyrochlore may result from radiation damage by alpha decay that also causes oxidation of Fe 2+ in titanite, and the interaction with a Ba-bearing oxidizing fluid. The oxidation of Ce3+ in the primary pyrochlore is accompanied by in the loss of REE, Ca, and Pb, a daughter product of U via alpha decay, during the alteration. However, most REE were incorporated in the alteration product, the Ce4+-rich pyrochlore. Based on EDS and EELS analyses, the chemical formulae of the partially metamict Ce3+-rich pyrochlore and metamict Ce4+-rich pyroeblore can be written as: (Ca, Ce3+, U, Pb) 2(Ti, Nb)2O7−x(OH)x, and (Ba, Ca, Ce4+, U)2 (Ti, Nb)2O7−y(OH)y, respectively. Ce is the most abundant element among all REE. It is proposed that the alteration takes place in solid-state with oxidizing fluid as a catalyst. The alteration kinetics is controlled by diffusion processes of aqueous species in metamict pyrochlore.

Zirconolite, CaZrTi2O7, has been proposed as an immobilization phase for the disposition of excess weapons Pu and other actinides (e.g., 235U). Due to actinide incorporation, zirconolite is expected to sustain α-decay event damage and become aperiodic (=metamict) over time. The leaching behavior of metamict zirconolite is, therefore, of interest. Because groundwater in a variety of geologic settings contains up to saturation concentrations of silicic acid, H4SiO4, silica-saturated solutions were used. Natural, metamict (> 1026 α-decay events per mn3) zirconolite grains, nominally (Ca, Th)ZrTi2O7 (U.S. National Museum sample #B20392, Walawada, Sri Lanka) were leached in two separate silica-saturated solutions at 150 °C for 60 days. Surface features and alteration products were examined using scanning electron microscopy (SEM) and quantitative energy dispersive X-ray spectroscopy (EDS). Secondary electron (SEI) images of the surfaces of the leached grains from both experiments revealed pores, probably due to the accumulation of He-bubbles from α-decay events, of approximately 4% as estimated by contrast enhanced gray-scale analysis of digital images. SEI of the zirconolite surface before leaching showed a smooth surface. The pores not only increase the surface area of the metamict zirconolite, but also act as nucleation sites for alteration phase growth. One experiment was conducted in a silica-saturated solution containing approximately 100 ppm P as measured by atomic absorption spectroscopy (AAS). The main alteration phase was euhedral, monoclinic cheralite, (Th, Ca, Ce)(P, Si)O4 (monazite group). The second experiment was conducted in the absence of P. The main alteration phase was subhedral cubic thorianite, ThO2.

Corrosion of zirconolite in a closed system at elevated temperature and pressure for various fluid compositions has been investigated by several analytical techniques. Together with previous studies, the results indicate a weak corrosion below 250° C at 5 MPa, Above 400° C, secondary TiO2 phases crystallize in acidic and neutral fluids on the corroded surfaces and in near-surface pores, At 550° C and above, zirconolite starts to be replaced by perovskite and “calzirtite” in basic fluids. Our results indicate that zirconolite is reasonably durable in acidic. neutral and basic fluids up to 500° C and 50 MPa which corresponds to deep borehole conditions.

Synroc specimens doped individually with ˜ 1 wt% of Np or Pu have been studied, after leaching for >2,500 days, using α-spectroscopy and SEM. In the last leaching period, the leachant was replaced on either a daily or monthly basis. When the leachant was replaced on a daily basis the release rate of Np was similar to that measured when the samples were first leached, whereas for the monthly replacement the release rates of Np or Pu were about a factor of 20 to 30 lower than that for daily replacement. These findings agree generally with the results obtained from surface examination of the samples which showed that the thickness of actinidedepleted surface layer of anatase was 0.3 μm for monthly replacement but only 0.1 μm when the leachant was replaced frequently. Overall, the results suggest that release of actinides from Synroc are controlled by solubility limiting effects at the surface area to volume ratios employed in MCC-1 tests.

Four titanate ceramics were characterized and tested. These ceramics were similar to those proposed for Pu disposition, with Ce as a surrogate for Pu. The baseline ceramic contained Ti, U, Ca, Hf, Gd, and Ce, and was made up of only four crystalline phases: pyrochlore, zirconolite, rutile, and brannerite. The three other ceramics contained different amounts of impurities that are expected in the feed. Impurities are defined as any element other than Ti, U, Ca, Hf, Gd, and Ce. The ceramics that contained impurities contained different phases than the baseline. The addition of impurities led to the absence of brannerite and the presence of amorphous silicate, Ca-Al-Ti, and perovskite phases. The results from 3 day, 90°C MCC-I tests with impurity ceramics were significantly different than the results from tests with the baseline ceramic. Overall, the addition of impurities to these titanate ceramics altered the phase assemblages, which in turn, affected the corrosion behavior.

This paper focuses on the reaction of glass-bonded zeolite and sodalite ceramic waste forms, naturally occurring sodalite, and their respective alteration phases that formed after reaction in a Vapor Hydration Test (VHT). Scanning electron microscope (SEM) analysis of the glass-bonded zeolite and sodalite samples indicate that NaCI is the major alteration phase present, along with KCI, Na-Al-Si, and Na-K-Ca-Al-Si phases. The order of precipitation of salt crystals from ceramic waste form occurred in the sequence NaCl ⇒ KC1 ⇒ CaCl2 and MgCl2. We also found aluminum silicate phases containing Ca, Mg, K-Na-Ca, K-Mg, Na-K-Mg, K-Ca-Na-Mg, as well as CaCO3, SiO2. The ratio of Si/Al in the alteration phases was noted to increase from approximately 1.5 to 5 during the progressive alteration of the samples. This trend indicates a relative increase in the contribution of silicon from the relatively Si-rich glass regions of the samples over time. Phases enriched in Cs, Ba and rare earth elements (RE) were also observed on the altered samples. Alteration phases initially appeared along the micro-fractures during the early stages of alteration. The number of micro-fractures, density of alteration phase cover, and grain size of alteration phases increased with reaction time and temperature.

The Defense Waste Processing Facility (DWPF), at the Savannah River Site (SRS), is processing and immobilizing the radioactive high level waste sludge at SRS into a durable borosilicate glass for final geological disposal. The DWPF is currently processing the second, million gallon batch of radioactive sludge. This second batch is primarily from Tank 42. Each time a new batch of radioactive sludge is to be processed by the DWPF, the process flowsheet is to be tested and demonstrated to ensure an acceptable melter feed and glass can be made. This demonstration was completed in the Shielded Cells Facility in the Savannah River Technology Center at SRS.

This paper presents the processing and offgas data, and compositional analyses obtained during the preparation of a melter feed for this demonstration. A second paper in this conference describes the properties of the glass produced from this feed. The demonstration used Tank 42 sludge slurry and the DWPF process control strategy for blending the sludge slurry with Frit 200 to make an acceptable melter feed. To prepare feed for the melter, the flowsheet requires that the radioactive sludge slurry be treated with nitric and formic acid to adjust rheology and remove mercury. During this step, hydrogen is formed from the decomposition of the formic acid. The acidified sludge slurry is then mixed with the prescribed amount of glass forming frit and evaporated to the proper weight percent solids to prepare feed to the melter. During this step hydrogen is also formed. Results indicate that the H2 generation rate is below the DWPF safety limits and an acceptable melter feed was produced.

The Defense Waste Processing Facility (DWPF) at the Savannah River Site (SRS) is currently processing and immobilizing the radioactive high level waste sludge at SRS into a durable borosilicate glass for final geological disposal. The DWPF has recently finished processing the first radioactive sludge batch, and is ready for the second batch ofradioactive sludge. The second batch is primarily sludge from Tank 42. Before processing this batch in the DWPF, the DWPF process flowsheet has to be demonstrated with a sample of Tank 42 sludge to ensure that an acceptable melter feed and glass can be made. This demonstration was recently completed in the Shielded Cells Facility at SRS. An earlier paper in these proceedings described the sludge composition and processes necessary for producing an acceptable melter feed [1]. This paper describes the preparation and characterization of the glass from that demonstration. Results substantiate that Tank 42 sludge after mixing with the proper amount of glass forming frit (Frit 200) can be processed to make an acceptable glass.

Mixtures of simulated Hanford low-activity aqueous waste and glass precursor additives were heat treated at various temperatures up to 1000°C and characterized using X-ray diffraction, X-ray fluorescence spectroscopy, scanning electron microscopy, gas chromatography mass spectroscopy, and inductively coupled plasma – mass spectroscopy. It was found after heating the simulant mixture up to 700°C, the majority of the sodium nitrite present in the original waste reacted with the silica and perhaps boron species in the additives to form C– and N-containing gases, oxygen gas, and lower melting phases which contributed to the observed batch volume expansion beginning at these temperatures. Elemental analyses of minor components in the simulant mixture showed that I, Cl, and S exhibited significant losses during heating, presumably due to volatilization or entrainment. For the case of I, nearly 90% on an elemental basis was lost from the simulant mixture after processing to 1000°C, whereas Re exhibited less than 40% loss.

The liquidus temperature of a waste glass (the temperature at which the system glass + crystalline material is in equilibrium with the amorphous glass of the same composition) often is a measure of the solubility of a specific component in the waste in the base glass. A new approach toward predicting the liquidus temperatures of waste glasses has been developed based on solubility. The approach predicts liquidus temperature as well as more complicated expressions developed by other researchers. At the same time, simple process control limits can be easily derived from this model, which is difficult to do with other approaches. While the approach is not applicable to all waste glass components, it can be used for important species such as nickel-iron spinel, chromium oxide, and probably plutonium dioxide. The approach also has led to two new techniques for determining liquidus of waste glass systems.

The Savannah River Technology Center (SRTC) is developing an immobilization process for graphite fines residues generated during nuclear materials production activities at the Rocky Flats Environmental Technology Site (Rocky Flats). The continued storage of this material has been identified as an item of concern. The residue was generated during the cleaning of graphite casting molds and potentially contains reactive plutonium metal. The average residue composition is 73 wt% graphite, 15 wt% calcium fluoride (CaF2), and 12 wt% plutonium oxide (PuO2 ). Approximately 950 kg of this material are currently stored at Rocky Flats.

The strategy of the immobilization process is to microencapsulate the residue by mixing with a sodium borosilicate (NBS) glass frit and heating at nominally 700°C. The resulting waste form would be sent to the Waste Isolation Pilot Plant (WIPP) for disposal. Since the PuO2 concentration in the residue averages 12 wt%, the immobilization process was required to meet the intent of safeguards termination criteria by limiting plutonium recoverability based on a test developed by Rocky Flats. The test required a plutonium recovery of less than 4 g/kg of waste form when a sample was leached using a nitric acid/CaF2 dissolution flowsheet.

Immobilization experiments were performed using simulated graphite fines with cerium oxide (CeO2) as a surrogate for PuO2 and with actual graphite fines residues. Small-scale surrogate experiments demonstrated that a 4:1 frit to residue ratio was adequate to prevent recovery of greater than 4 g/kg of cerium from simulated waste forms. Additional experiments investigated the impact of varying concentrations of CaF2 and the temperature/heating time cycle on the cerium recovery. Optimal processing conditions developed during these experiments were subsequently demonstrated at full-scale with surrogate materials and on a smaller scale using actual graphite fines.

In general, the recovery of cerium from the full-scale waste forms was higher than for smaller scale experiments. The presence of CaF2 also caused a dramatic increase in cerium recovery not seen in the small-scale experiments. However, the results from experiments with actual graphite fines were encouraging. A 4:1 frit to residue ratio, a temperature of 700°C, and a 2 hr heating time produced waste forms with plutonium recoveries of 4±1 g/kg. With an increase in the frit to residue ratio, waste forms fabricated at this scale should meet the Rocky Flats product specification. The scale-up of the waste form fabrication process to nominally 3 kg is expected to require a 5:1 to 6:1 frit to residue ratio and maintaining the waste form centerline temperature at 700°C for 2 hr.

The parameters of processing of large-scale (50-100 kg) radwaste-containing blocks were studied. Batches consisted of exothermic mixture, oxidizer, mineral additive, and radwaste surrogate. Such mixtures provide process temperature up to ˜2400 K. In the present work exothermic mixtures composed of Ca-Fe silicide, zirconium, silicon, and boron were intermixed with oxidizer (mixture of KMnO4 and V2O5), natural zircon concentrate as mineral additive, and cerium oxide (actinide surrogate). Tests were carried out in a bench-scale plant. Burning of the reaction mixture was remotely initiated from electric supply unit. Melt formation started in the nearbottom area sustained upwards frontally, The melt temperature ranged between 1600 and 1900 K. The melt formation rate during the tests ranged between 0.7 and 2.5 kg/s. Melt cooling after reaction completion was kept for about 15-18 hours. Volatile aerosols were essentially retained in the bulk. The initial volume of the mixture was reduced after reaction by factors of 2 to 3. The densities of the blocks were ranged between 3.2 and 3.9 kg/dm3.

Samples of the materials obtained were examined by optical microscopy, X-ray diffraction, and scanning electron microscopy. The product consisted of glass-like body with metallic spheres (˜0.5-1.5 mm) distributed within. The base was composed of glass (˜60-80 vol.% of total) and crystalline phases (mainly zirconia). Monoclinic baddeleyite was found to be the major crystalline phase. The minor phase was tetragonal zirconia. Cerium (actinide surrogate) was partitioned among the glass and zirconia. The Ce2O3 concentration in glass ranged between ˜0.4 wt.% and ˜0.8 wt.% whereas its content in zirconia was found to be lower (up to ˜0.3 wt %). The metallic constituent consisted of vanadium and Fe-V-Si alloy whose average composition wasFe7O2V19.3Mn1.1 Si9.4.

Vitrification has become the international method of choice for the immobilization of highlevel nuclear wastes (HLW) from both commercial and defense fuel reprocessing activities. However, there are clear economic incentives for "next-generation" melters that might achieve increased waste loadings and higher throughput rates. In addition, BNFL is investigating potential immobilization processes for projected high-zirconium waste streams that would be generated from advanced nuclear fuel reprocessing processes. One approach towards addressing both of these challenges is waste vitrification at temperatures above the limits that are imposed on all currently operating HLW melters (around 1100 -1200°C). A unique high-temperature joule-heated melter that has been operating at the Vitreous State Laboratory for several years was used to conduct a matrix of tests designed to address these issues. The tests used simulants of two reprocessing waste streams: (i) A blend of "Magnox" and "Oxide" wastes, similar to those currently being vitrified at BNFL's Sellafield site; and (ii) The projected high-zirconium wastes. Each type of waste was investigated both as a wet slurry feed using chemical additives and as a dry calcined waste feed mixed with glass frit. Three target glass compositions were designed for these tests in order to permit processing over a wide range of temperatures. Data were collected on processing rates, off-gas composition, glass product composition and product quality, as well as on general processing behavior. In all cases, glass production rates increased rapidly with processing temperature, typically doubling for every 100°C, and decreased with increasing water content of the feed. The glass production rates per unit melt surface area were significantly higher than those typically observed for hot-wall induction melters operating under comparable conditions. Interestingly, off-gas emissions were much greater for the slurry feeds than for the comparable solid feeds. The resulting glass products were characterized for melt viscosity and electrical conductivity, leachability by both soxhlet and PCT procedures, glass transition temperature, and the amounts and types of any secondary phases that were present.

Nonisothermal kinetics of spinel crystallization in a high-level waste (HLW) glass was predicted using Mehl-Avrami-Johnson-Kolmogorov equation coefficients from isothermal data. The volume fraction of spinel was determined as a function of time, temperature, and cooling rate. The results were verified experimentally. Also predicted was the spatial distribution of spinel in a HLW glass canister. Finally, a parameter study was performed, and an empirical equation was proposed relating the final spinel volume fraction in glass to dimensionless numbers for cooling rate, phase equilibrium, and crystallization kinetics.

The development of glass materials for long-term storage of high-level waste implies determining the glass thermal stability and notably assessing the risk of devitrification. Previous studies of the French nuclear waste glass have identified the crystalline phases and the domains in which they form, and have shown that devitrification is minimal. Modeling the long-term crystallization behavior requires an investigation of the thermodynamic and kinetic mechanisms liable to induce crystallization during cooling. The first step in this approach is to determine the nucleation and growth curves for each of the component phases. Crystals including CaMoO4, CeO2 and ZnCr2O4 were identified in glasses heat-treated at temperatures between 630°C and 1110°C, and the time and thermal dependence of the CaMoO4 morphology were evaluated. The nucleation and growth kinetics of these phases were determined by optical microscopy and SEM, and the impact of impurities was addressed by studying two glasses, with and without platinoid elements. The results indicated enhanced nucleation kinetics in glass containing platinoid elements. No induction time was observed before permanent nucleation in either of the glasses, and rapid saturation of nucleation kinetics—synonymous with the depletion of active centers of nucleation— was detected after a few hours. Furthermore, similar growth kinetics were observed in both glasses. The nucleation and growth curves coincided for all the phases. Peak values were much higher for nucleation than for growth kinetics, confirming the need for a thorough investigation of the mechanisms occurring in and below the glass transition range, i.e. in the non-equilibrium state.

A wide variety of plutonium-containing wastes have been generated from both defense and commercial nuclear programs. The Advanced Mixed Waste Treatment Project (AMWTP) involves the privatized treatment of 65,000 m3 of mixed TRU and low-level waste that is in storage at DOE's INEEL site. BNFL Inc. leads the commercial team including GTS Duratek, Inc., that has been selected to design, construct, and operate a facility to perform that service. Since vitrification is a potential treatment option, the current study was conducted to investigate issues surrounding the incorporation of the approximately 700 kg of plutonium that is present in the AMWTP waste. A series of glasses were prepared with plutonium (0.5 wt%), with cerium (as a potential surrogate), with both, and with neither. Scanning electron microscopy coupled with energy dispersive x-ray spectroscopy (SEM-EDS) and optical analyses confirmed that both elements were completely dissolved into the as-melted glasses. Heat treatments forced the formation of secondary phases, which were analyzed for Pu and Ce incorporation. Plutonium-containing glasses were spiked with sulfates and remelted to force the formation of a sulfate salt phase, which was analyzed to determine partitioning of Pu and Ce. Coupons of Monofrax K-3 glass contact refractory were exposed to melts of these glasses for six days at 1208°C using both standard and modified versions of ASTM procedure C621-84. Thickness loss, microstructural alteration, and the extent to which Pu and Ce were incorporated into the material were then determined by methods employing SEMEDS\ wavelength dispersive x-ray spectroscopy (WDS). Extreme heat treatments were found to lead to the formation of small amounts (< 0.1 vol%) of either (Pu, Ce)O2, or PuO2, depending upon whether Ce was present or absent, respectively. The Pu level in the sulfate phase was at least a factor often below that of the parent glass, and no tendency to concentrate Pu in the K-3 refractory above the level in the parent glass was found.

The corrosion of the glass-contact refractory Monofrax K-3 in two sodium-rich aluminosilicate melts has been studied at 1208 and 1283°C using a modified ASTM procedure with constant agitation of the melt by air bubbling. The results for the monolithic refractory indicate a fast initial stage involving phase dissolution and transformation and a later “passivated” stage in which the surface of the refractory has been substantially modified. The composition of the stable spinel phase in the altered layer on monolithic coupons of K-3 is almost identical to the equilibrium composition bracketed by the dissolution of powdered K-3 into under-saturated melts on the one hand, and by crystallization of spinels from supersaturated melts on the other. The temperature and melt shear viscosity were found to have significant effects on the rates of K-3 dissolution and transformation.

Inconel-690, a Cr-Ni-Fe-based “superalloy,” has become the material of choice for electrodes in joule-heated waste glass melters and is currently employed in the high-level nuclear waste vitrification systems at West Valley and DWPF, as well as in GTS Duratek's privatized M-Area mixed waste vitrification facility at Savannah River. Future applications of joule-heated vitrification technologies will necessitate an assessment of the limits of performance of this material under more demanding conditions than have been studied previously. In this work, Inconel 690 electrodes were tested in several simulated sodium-rich aluminosilicate waste glasses in wide ranges of AC current density, electrical waveform, temperature, and glass composition.

Because of their high chemical durability and waste loading capacity, iron phosphate glasses are a natural candidate for a nuclear waste disposal medium. We have studied the effects of uranium on the structure of iron phosphate glasses with both neutron and high-energy x-raydiffraction. The results of neutron scattering, which is mostly sensitive to pair correlations involving light atoms i.e. O-O, Fe-O and P-O, indicate the main structural features of the base glass are largely unaffected by the addition of UO2. The nearest-neighborbour P-O, Fe-O and O- O peaks remain at the same position in real space and their intensities scale approximately with concentration. These findings are consistent with earlier results using Raman scattering and EXAFS on the Fe-K edge, where in both cases the spectra remain similar to the base glass. The results of high-energy x-ray scattering, which is sensitive to correlations involving the heavier atoms and thus complements the neutron measurements, are also consistent with the overall picture of uranium occupying interstitial sites in the relatively undisturbed base glass structure. Combining the neutron and x-ray data for a 10 mol% UO2 glass suggests the intriguing possibility of a U6+ uranyl ion configuration although further work is needed to establish the precise local structure and valence state of uranium in these glasses.

A range of compositions of high-zirconia borosilicate glasses were formulated and their structures investigated by a combination of techniques. These compositions have potential applications for high-level nuclear waste storage in combination with advanced reprocessing methods. Raman and Zr EXAFS data were collected for a series of glasses spanning a range of zirconia concentrations. The Raman spectra indicate that Zr acts as a silicate network modifier, where the silicate tetrahedral network depolymerizes as the zirconia content increases. Zr EXAFS analysis indicates that Zr is found in octahedral sites, and to a minor extent, sevencoordinated sites. As the zirconia content increases, the fraction of seven-coordinated Zr-sites increases; this may be the cause of ZrO2 baddeleyite crystallization that was observed in some Zr-rich glasses investigated.