Figure 1. XRD patterns of different single-phase hollandites
This paper reports the on-going studies on melt-processed single-phase hollandites and hollandite-rich multi-phasic ceramic systems for nuclear waste disposal. Oxide and carbonate precursor powders were mixed in stoichiometric proportions and heated to temperatures above 1500°C. Melting and solidification of the mixtures led to crystallization of different phases. High-temperature x-ray diffraction was used to determine the phase formation and assemblage. Scanning electron microscopy was used to reveal the microstructural features. Phase pure hollandite was formed in all single-phase materials except aluminum hollandite. Multi-phase materials were crystallized to different phases such as hollandite, perovskite, zirconolite, and pyrochlore that are distributed randomly as observed in the microstructures.
Reprocessing of spent fuel from nuclear fission reactors to recover uranium (U) and plutonium (Pu) for reuse is a well-known practice. High-level wastes (HLW) produced from this reprocessing contain different radioactive nuclides and other fission products that can pose severe problems for safe treatment and disposal. The traditional method of using silicate glass wasteforms for HLW disposal has some disadvantages such as low solubility of actinides in the glass and high volatility of some HLW elements at glass melting temperatures1. Ceramic materials have been suggested as alternative wasteforms with higher waste loadings.2 Synthetic Rock (SYNROC), titanate-based structure is one of the proposed ceramic hosts for this purpose.3 Different elements in the HLW are partitioned and confined in the crystalline network of different phases like hollandite (BaxAl2xTi8-2xO16), perovskite (CaTiO3), and zirconolite (CaZrTi2O7). Heavy elements like Cs and Ba are partitioned mainly into hollandite phase, zirconolite acts as the major host for actinides and perovskite accommodates most of Sr and rare-earth elements.
Barium hollandites (BaxCsy)[(M)3+2x+yTi4+8-2x-y]O16, where ‘M’ is a trivalent cation and x+y < 2) are reported to have effective confinement of heavier elements like Cs and Rb in their lattices.4,5,6 These structures consist of tunnels running along the z-axis, surrounded by a framework of edge-shared and corner-shared oxygen octahedra. The M3+ and Ti4+ ions are located in the octahedral sites, while the heavier elements like Ba, Cs or Rb occupy the tunnel sites. These tunnel sites are only partially occupied, where Ba2+ and Cs+ ions are in an ordered arrangement.7 Since the ionic radii of these two ions are different, tunnel distortions occur in the lattice and hence, the occupancy ratio Ba2+/Cs+ in the tunnels is determined by the combination of Ti4+ and M3+ in the hollandite. Due to the high volatility of Cs2O and its high reactivity, reducing conditions and hot-pressing have been used for hollandite synthesis.4 In addition to this, the redox conditions during hollandite synthesis affect the conversion of Ti4+ to Ti3+ Partial reduction of Ti4+ during processing alters the structural parameters of hollandites that would affect Cs incorporation.
Several groups8,9 have worked on studying the effects of different trivalent cations on the properties of hollandites intended for Cs confinement. Aubin-Chevaldonnet et al.8 have studied properties of Cs substituted hollandites when different cations, Al3+, Cr3+ and Fe3+, are used. Cs retention in the hollandites, prepared by oxide route and atmospheric sintering, is observed to be the highest (almost 100%) for Fe3+ hollandite. The retention is the least (less than 10%) for Al3+ hollandite and is around 50%, when Cr3+ is used. The Cr3+ hollandites exhibit very high porosity that will affect the chemical durability and mechanical strength of the waste form. Research at Savannah River National Laboratory (SRNL) on the characteristics of different single-phase hollandites prepared by melt-processing9 show that hollandites with Fe3+ had Cs partitioned into parasitic phases like CsTiAlO4, both in atmospheric and reducing conditions. The Cr3+ hollandites do not have any secondary phases and the Cs retention is improved when melting is performed in reducing conditions. Although single-phase hollandite is obtained, the sample has not melted fully.
In the on-going work at Alfred University, the melt-processing, phase evolution, and phase assemblage of single phase-hollandites, with different cations Al3+, Fe3+ and Cr3+ are presented. In addition to these, similar studies on hollandite-rich multi-phase mixtures, from SRNL, are also presented. Table I outlines the composition and notations for different materials discussed below and Table II shows the composition of the multi-phase mixtures from SRNL.
Table I. Different materials in this study
| Chemistry | Short-hand notation |
| Ba1.0Cs0.3Al2.3Ti5.7O16 | Al-SP |
| Ba1.0Cs0.3Fe2.3Ti5.7O16 | Fe-SP |
| Ba1.0Cs0.3Cr2.3Ti5.7O16 | Cr-SP |
| Multi-phase mixtures | Cr-MP |
| Cr/Al/Fe-MP |
Stoichiometric amounts of oxides and carbonates (99.9% purity, Alfa Aesar) were ball-milled together in plastic container, for 30 min., and dried. About 2 – 3 g of the dried powders were placed in alumina crucibles and heated up to 1500°C–1550°C in air, at a heating rate of 5°C/min, held for about 30 min., and cooled to room temperature.
All samples obtained were characterized by x-ray diffraction (XRD) for phase analysis. Microstructures of the polished samples were studied with scanning electron microscope (SEM) and the chemical compositions of different phases were determined with energy dispersion spectrometer (EDS). Phase evolution of hollandite was studied with high-temperature x-ray diffraction (HTXRD). The starting material was heated at the rate of 5°C/min, to simulate the conditions used for melt-processing, and the diffraction data was collected at an interval of 50°C over a temperature range of 800 – 1350°C.
Table II. Composition of multi-phase mixtures*
| Cr/Al/Fe-MP | Cr-MP | |
| wt% of different oxides | ||
| Al2O3 | 1.28 | 0.00 |
| BaO | 12.82 | 12.78 |
| CaO | 1.40 | 1.39 |
| Cr2O3 | 6.36 | 14.57 |
| CdO | 0.11 | 0.11 |
| Ce2O3 | 3.11 | 3.10 |
| Cs2O | 2.88 | 2.87 |
| Eu2O3 | 0.17 | 0.17 |
| Fe2O3 | 6.68 | 0.00 |
| Gd2O3 | 0.16 | 0.16 |
| La2O3 | 1.59 | 1.59 |
| MoO3 | 0.85 | 0.84 |
| Nd2O3 | 5.25 | 5.24 |
| Pr2O3 | 1.46 | 1.45 |
| SeO2 | 0.08 | 0.08 |
| Sm2O3 | 1.08 | 1.07 |
| SnO2 | 0.07 | 0.07 |
| SrO | 0.98 | 0.98 |
| TeO2 | 0.66 | 0.65 |
| TiO2 | 49.37 | 49.24 |
| Y2O3 | 0.63 | 0.63 |
| ZrO2 | 3.00 | 2.99 |
*Ball-mil led and dried mixtures received from SRNL
XRD data different single-phase hollandites are shown in Figure 1. Pure hollandite is formed with no secondary phases in Fe-SP and Cr-SP, However, multiple trials of melting Al-SP show that a secondary phase, Al2TiO5 is always retained in the final reacted material. All peaks in XRD plots of both multi-phase materials are indexed to four predominant phases, hollandites, perovskites, pyrochlores, zirconolite along with traces of unreacted TiO2, as shown in Figure 2. Since a number of these structures are possible between different oxide systems shown in Table 2, indexing a particular peak to a chemical formula was challenging.
Figure 1. XRD patterns of different single-phase hollandites
Figure 2. XRD patterns of multi-phase hollandites
Visual examination of the reacted materials shows that the Cr-SP did not fully melt, although the reaction was complete. Melting is expected to occur at relatively higher temperatures due to the high refractory nature of Cr2O3, and at the same time it helps the formation of hollandite.8 Visual examination of Fe-SP revealed that it melted fully. Fe-SP melt reacted with Al2O3 crucible and could not be separated from the crucible, whereas there was little reaction between Cr-SP and the crucible, and it could be knocked off the crucible surface. However, the multi-phase materials, Cr-MP and Cr/Al/Fe-MP, melted fully. Figure 3 compares the images of these two hollandites formed.
Figure 3. As-melt samples of Fe-SP and Cr-SP
Microstructural features of one of the multi-phase samples (Cr-MP) are shown in Figure 4. Microstructures show the distribution of different phases and porosity. However, limited elemental resolution and contrast of the EDS data could not confirm the chemistry of individual phases present in the samples. Due to the random distribution of very small volumes of different phases, signal detection from an individual phase could not be achieved. Based on the limited EDS data and literature10, different phases in the microstructures are identified as TiO2 (T), Hollandite (H), perovskite (P) and Zirconolite (Z).
Further, Cs could not be detected from wave dispersion spectrometer (WDS) data of any of the melted powders. This was due to melting of the samples in air, which would enhance the volatilization of Cs.
Figure 4. Back-scattered SEM images of polished Cr-MP sample at different magnifications; (a) 500x and (b) 2000x
HTXRD measurements were made on single-phase hollandites over 800 – 1350°C to understand the effects of different trivalent cations on evolution of holladite phase. Figure 5 shows a representative datasets for the phase changes in different systems. In Al-SP, the first hollandite peak was observed on the plot at 1000°C and the intensity of the peak increased further with temperature, as shown. Al2TiO5, which was observed as a secondary phase in all melt trials, appeared only at 1350°C, suggesting that it is a stable high temperature. In Fe-SP, hollandite was not observed until 1250°C, as indicated in Figure 5(b). HTXRD of Cr-SP has interesting features. Hollandite peaks were observed on all plots, starting from 800°C. The final diffraction peaks at 1350°C mostly correspond to hollandite and some unreacted TiO2, indicating that the reaction is almost complete within the used heating cycle. Studying the plots at intermediate temperatures (1000 – 1200°C) showed that barium titanates are formed initially, as indicated in Figure 6, which react at temperatures above 1200°C to form hollandites.
Figure 5. HTXRD plots for (a) Al-SP (b) Fe-SP and (c) Cr-SP
Figure 6. Phase evolution in Cr-SP
Melt-processing and HTXRD studies of different single-phase hollandites showed that Cr3+ stabilized hollandite at lower temperatures and expedited its formation during melting. Mulit-phase samples melted fully under the processing conditions. XRD data confirmed the presence of desired phases - hollandite, perovskite, zirconolite, pyrochlore and titania. SEM data revealed complex microstructures.
The authors acknowledge the support from DOE’s NEUP for our project. We also acknowledge the advice and feedback from Dr. Jim Marra of Savannah River National Laboratory and Dr. John Vienna from Pacific Northwest National Laboratory. SKS acknowledges the support from Kyocera Corporation for the support of the Inamori Professorship.
1W.J. Weber, A. Navrotsky, S. Stefanovsky, E.R. Vance and E. Vernaz, Materials Science of High-Level Nuclear Waste Immobilization, MRS Bulletin, 34, 46–53 (2009).
2S.V. Stefanovsky, S.V. Yudintsev, R. Giere and G.R. Lumpkin, Nuclear waste forms, Energy, Waste and the Environment: A Geochemical perspective (Special Publication 236, Geological Society of London, London, 2004), p. 37.
3A.E. Ringwood, S.E. Kesson, N.G. Ware, W.O. Hibberson and A. Major, Immobilization of High-Level Nuclear reactor Wastes in SYNROC, Nature, 278, 219–223 (1979).
4S.E. Kesson and T.J. White, [BaxCsy][(Ti,Al)3+ 2x+yTi4+8–2x-y]O16 Synroc-Type Hollandites. I. Phase Chemistry, Proc. R. Soc. London. Ser. A, 405, 73–101 (1986).
5R.W. Cheary, A Structural Analysis of Potassium, Rubidium, and Cesium Substitution in Barium Hollandites, Acta Cryst., B43, 28–34 (1987).
6A.Y. Leinekugel-le-Cocq, P. Deniard, S. Jobic, R. Cerny, F. Bart and H. Emerich, Synthesis and Characterization of Hollandite-Tye Material Intended for the Specific Containment of Radioactive Cesium. J. Solid State Chem., 179, 3196–3208 (2006).
7R.W. Cheary, The Immobilization of Cesium in SYNROC, Mater. Sci. For., 27/28, 397–406 (1988).
8V. Aubin-Chevaldonnet, D. Caurant, A. Dannoux, D. Gourier, T. Charpentier, L. Mazerolles and T. Advocat, Preparation and Characterization of (Ba,Cs)(M,Ti)8O16 (M = Al3+, Fe3+, Ga3+, Cr3+, Sc3+, Mg2+) Hollandite Ceramics for Radioactive Cesium Immobilization, J. Nucl. Mater., 366, 137–160 (2007).
9J. Amoroso, J. Marra, S.D. Conradson, M. Tang and K. Brinkman, Melt Processed Single Phase Hollandite Waste Forms for Nuclear Waste Immobilization: Ba1.0Cs0.3A2.3Ti5.7O16; A = Cr, Fe, Al, J. Alloy. Compd., In Press (2013).
10K.S. Brinkman K.M. Fox, J.C. Marra and M. Tang, Crystalline Ceramic Waste Forms: Reference Formulation Report, SRNL Technical Report, FCRD-SWF-2012-000116 (2012).