<%BANNER%>

Experimental Determination of the Dry Oxidation Behavior of a Compositional Range of Uranium-Thorium Mixed-Oxide Pellet ...


PAGE 1

EXPERIMENTAL DETERMINATION OF THE DRY OXIDATION BEHAVIOR OF A COMPOSITIONAL RANGE OF URANIUM-THORIUM MIXED-OXIDE PELLET FRAGMENTS By LISA ARGO A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2003

PAGE 2

Copyright 2003 by Lisa Argo

PAGE 3

iii ACKNOWLEDGMENTS This work would not have been completed without the support of many. More than anyone else, I thank my advisors (Ronald Baney and James Tulenko) for their wisdom, patience, understanding, and occasional cracking whip. I would like to acknowledge the contribution of Paul Demkowicz for his expertise and meticulous editing; and Noriko Shibuya for her technical support, as we investigated the ins and outs of many synthesis methods. I thank my sister, Irene, for the late night pep talks and advice. I thank my husband, Paul, for pushing me back into and out of school. This work was funded through a grant from the Nuclear Engineering Research Initiative (NERI) project #99-0153 Advanced Proliferation Resistant, Lower Cost, Uranium-Thorium Dioxide Fuels for Light Water Reactors.

PAGE 4

iv TABLE OF CONTENTS page ACKNOWLEDGMENTS ..........................................................................................................iii TABLE OF CONTENTS...........................................................................................................iv LIST OF TABLES.....................................................................................................................vi LIST OF FIGURES.................................................................................................................viii ABSTRACT..............................................................................................................................xi CHAPTER 1 2 3 4 INTRODUCTION................................................................................................................1 REVIEW OF LITERATURE................................................................................................5 Current Status of UO 2 and ThO 2 Research............................................................................5 Uranium Dioxide............................................................................................................5 Thorium Oxide...............................................................................................................9 Uranium-Thorium Mixed Oxide...................................................................................11 Background of Synthesis Methods......................................................................................14 Kinetic Analysis.................................................................................................................16 MATERIALS AND METHODS ........................................................................................20 Material Synthesis............................................................................................................. .20 Oxalate Co-Precipitation..............................................................................................20 Ammonium Hydroxide Co-Precipitation.......................................................................26 Co-Milled Mixe d Oxides..............................................................................................27 Characterization..................................................................................................................32 X-ray Diffraction (XRD)..............................................................................................32 Elemental Analysis.......................................................................................................35 Particle Morphology .....................................................................................................36 Pellet Density...............................................................................................................37 Dry Oxidation.....................................................................................................................38 Thermogravimetric Analyzer........................................................................................38 Kinetic Analysis...........................................................................................................39 RESULTS AND DISCUSSION..........................................................................................43 Material Synthesis............................................................................................................. .43 Thermogravimetry..............................................................................................................4 5

PAGE 5

v 5 Kinetic Analysis.................................................................................................................54 Mixed Oxide (U 0.236 Th 0.764 O 2 ).......................................................................................55 Mixed Oxide (U 0.368 Th 0.632 O 2 ).......................................................................................61 Mixed Oxide (U 0.50 Th 0.50 O 2 ).........................................................................................66 Pure Uranium Dioxide (UO 2 )........................................................................................71 SUMMARY AND CONCLUSIONS..................................................................................74 APPENDIX A ACTIVATION ENERGIES................................................................................................77 B X-RAY DIFFRAC TION PATTERNS.................................................................................80 C SECONDARY ELECTRON IMAGES OF UNIRRADIATED PELLETS ...........................91 REFERE NCES.........................................................................................................................93 BIOGRAPHICAL SKETCH.....................................................................................................97

PAGE 6

vi LIST OF TABLES Table page 2-1. Defect formation energies for UO 2 ......................................................................................6 2-2. Oxidation reactions of UO 2+x ...............................................................................................7 2-3. Material properties of ThO 2 and UO 2 ..................................................................................9 2-4. Cation and anion formation and migration energies in ThO 2 ..............................................10 2-5. Composition dependence on oxygen partial pressure for UO 2+x and (U,Th)O 2+x .................13 2-6. Rate laws for a simple process A P...............................................................................18 3-1. Evolution of oxalate-synthesized (U,Th)O 2 .......................................................................23 3-2. 2 2 factorial for 20% UO 2 -80% ThO 2 blended oxide.........................................................30 3-3. Co-milled U 0.2 Th 0.8 O 2 pressing conditions 2 2 factorial results..........................................30 3-4. Pellet manufacture conditions ...........................................................................................31 3-5. Solid state theoretical reaction models...............................................................................41 4-1. Calculated (U,Th)O 2 oxidized lattice parameters by three methods....................................43 4-2. ICP-AES and LECO carbon analysis results and calculated metal valence.........................45 4-3. Rate coefficients of isothermal (U 0.236 Th 0.764 )O 2 agreement to diffusion models.................56 4-4. Kinetic results for nonisothermal (U 0.236 Th 0.764 )O 2 agreement to diffusion models..............56 4-5. Rate coefficient of isothermal (U 0.368 Th 0.632 )O 2 agreement to 3D diffusion model...............63 4-6. Kinetic results for nonisothermal (U 0.368 Th 0.632 )O 2 agreement to 3D diffusion model..........63 4-7. Rate coefficient of isothermal (U 0.50 Th 0.50 )O 2 agreement to 3D diffusion model.................68 4-8. Kinetic results for nonisothermal (U 0.500 Th 0.500 )O 2 in 3D diffusion and Avrami-Erofeev...68 5-1. Estimated E and A by model-free and model-fit techniques of (U,Th)O 2 and UO 2 .............76 A-1. Published estimates of U 3 O 7 /U 4 O 9 activation energy of formation....................................77

PAGE 7

vii A-2. Published estimates of U 3 O 8 formation on UO 2 activation energies...................................77 A-3. Published estimates of UO 2 cation and anion diffusion activation energies........................78 A-4. Published estimates of ThO 2 cation and anion diffusion activation energies......................78 A-5. Published estimates of diffusion in (U,Th)O 2 activation energies......................................79 A-6. Kinetic parameters of mixed urania-thoria oxides.............................................................79 B-1. U 0.368 Th 0.632 O 2+x isothermal 400 C air oxidized compared to JCPDS standards..................80 B-2. XRD peak intensities for 450 C isotherm of co-milled 23.6% UO 2 fragments...................81 B-3. XRD peak intensities for 500 C isotherm of co-milled 23.6% UO 2 fragments...................82 B-4. XRD peak intensities for 550 C isotherm of co-milled 23.6% UO 2 fragments...................83 B-5. XRD peak intensities for 400 C isotherm of co-milled 36.8% UO 2 fragments...................84 B-6. XRD peak intensities for 450 C isotherm of co-milled 36.8% UO 2 fragments...................85 B-7. XRD peak intensities for 500 C isotherm of co-milled 36.8% UO 2 fragments...................86 B-8. XRD peak intensities for 375 C isotherm of co-milled 50.0% UO 2 fragments...................87 B-9. XRD peak intensities for 400 C isotherm of co-milled 50.0% UO 2 fragments...................88 B-10. XRD peak intensities for 425 C isotherm of co-milled 50.0% UO 2 fragments.................89 B-11. XRD peak intensities for 450 C isotherm of co-milled 50.0% UO 2 fragments.................90

PAGE 8

viii LIST OF FIGURES Figure page 3-1. Oxalate path co-precipitation............................................................................................ .21 3-2. Optical micrograph of nominal 20% UO 2 calcined oxalate-synthesized powder.................22 3-3. X-Ray Diffraction evolution of U 0.05 Th 0.95 O 2 oxalate synthesized powder..........................24 3-4. Vulcan Muffle furnace used for drying, decomposition, and calcination............................25 3-5. Lindberg high te mperature tube furnace............................................................................25 3-6. Ammonium hydroxide path co-precipitation......................................................................26 3-7. Ammonium hydroxide synthesized (U 0.2 Th 0.8 )O 2+x before (l) and after (r) calcinations.......27 3-8. Optical micrographs of Alfa Aesar UO 2 (l) and ThO 2 (r)...................................................28 3-9. 8000M SPEX Certiprep Mixer/Mill (l) and zirconia mill jar (r) .........................................29 3-10. Pellets prepared using 20% U powder synthesized by the oxalate technique....................31 3-11. Optical micrographs of nominal U 0.2 Th 0.8 O 2+x ..................................................................37 3-12. TA Instruments TGA 2050 thermogravimetric analyzer ...................................................38 4-1. Lattice parameter evolution with respect to UO 2 content....................................................44 4-2. Nonisothermal (U 0.236 Th 0.764 )O 2 oxidation TGA data at heating rates of 1, 3, and 5 C/min.47 4-3. Nonisothermal (U 0.368 Th 0.632 )O 2 oxidation TGA data at heating rates of 1, 3, and 5 C/min.48 4-4. Nonisothermal (U 0.500 Th 0.500 )O 2 oxidation TGA data at heating rates of 1 and 5 C/min......48 4-5. Nonisothermal UO 2 oxidation TGA data at heating rate of 3 C/min...................................49 4-6. Isothermal oxidation TGA data for (U 0.236 Th 0.764 )O 2 fragments (90 250 m)...................50 4-7. Isothermal oxidation TGA data for (U 0.368 Th 0.632 )O 2 fragments (90 250 m)...................50 4-8. Isothermal oxidation TGA data for (U 0.500 Th 0.500 )O 2 fragments (90 250 m)...................51 4-9. Isothermal oxidation TGA data for UO 2 fragments (90 250 m).....................................51

PAGE 9

ix 4-10. Mean uranium valence for isothermally oxidized (U y Th 1-y )O 2 and UO 2 ............................54 4-11. Isotherm at 450 C of (U 0.236 Th 0.764 )O 2 oxidation fit to 2D and 3D Diffusion models.........57 4-12. Isotherm at 500 C of (U 0.236 Th 0.764 )O 2 oxidation fit to 2D and 3D Diffusion models.........57 4-13. Isotherm at 550 C of (U 0.236 Th 0.764 )O 2 oxidation fit to 2D and 3D Diffusion models.........58 4-14. Arrhenius plot of (U 0.236 Th 0.764 )O 2 isotherms fit to 3D Diffusion models..........................58 4-15. Nonisotherm at 1 C/min (U 0.236 Th 0.764 )O 2 Arrhenius plot fit to 2D and 3D diffusion........59 4-16. Nonisotherm at 3 C/min (U 0.236 Th 0.764 )O 2 Arrhenius plot fit to 2D and 3D diffusion........59 4-17. Nonisotherm at 5 C/min (U 0.236 Th 0.764 )O 2 Arrhenius plot fit to 2D and 3D diffusion........60 4-18. Model free (U 0.236 Th 0.764 )O 2 isotherms plotted at = 0.4, 0.5, 0.6....................................60 4-19. Isotherm at 450 C for (U 0.368 Th 0.632 )O 2 oxidation fit to 3D Diffusion...............................63 4-20. Isotherm at 475 C for (U 0.368 Th 0.632 )O 2 oxidation fit to 3D Diffusion...............................64 4-21. Isotherm at 500 C for (U 0.368 Th 0.632 )O 2 oxidation fit to 3D Diffusion...............................64 4-22. Arrhenius plot of (U 0.368 Th 0.632 )O 2 isotherms fit to 3D Diffusion......................................65 4-23. Nonisotherms at 1, 3, and 5 C/min (U 0.368 Th 0.632 )O 2 Arrhenius plot fit.............................65 4-24. Model free (U 0.368 Th 0.632 )O 2 isotherms plotted at = 0.4, 0.5, 0.6....................................66 4-25. Isotherm at 375 C for (U 0.50 Th 0.50 )O 2 oxidation in 3D Diffusion and Avrami-Erofeev.....68 4-26. Isotherm at 400 C for (U 0.50 Th 0.50 )O 2 oxidation in 3D Diffusion and Avrami-Erofeev.....69 4-27. Isotherm at 425 C for (U 0.50 Th 0.50 )O 2 oxidation in 3D Diffusion and Avrami-Erofeev.....69 4-28. Isotherm at 450 C for (U 0.50 Th 0.50 )O 2 oxidation in 3D Diffusion and Avrami-Erofeev.....70 4-29. Arrhenius plot of (U 0.500 Th 0.500 )O 2 isotherms in 3D Diffusion reaction model...................70 4-30. Nonisotherms at 1 and 5 C/min (U 0.500 Th 0.500 )O 2 Arrhenius plot......................................71 4-31. Model free (U 0.500 Th 0.500 )O 2 isotherms plotted at = 0.4, 0.5, 0.6....................................71 4-32. Model free UO 2 isotherms plotted at = 0.4, 0.5, 0.6......................................................73 B-1. XRD pattern for 450 C isotherm of co-milled 23.6% UO 2 fragments................................81 B-2. XRD pattern for 500 C isotherm of co-milled 23.6% UO 2 fragments................................82 B-3. XRD pattern for 550 C isotherm of co-milled 23.6% UO 2 fragments................................83

PAGE 10

B-4. XRD pattern for 400 C isotherm of co-milled 36.8% UO 2 fragments................................84 B-5. XRD pattern for 450 C isotherm of co-milled 36.8% UO 2 fragments................................85 B-6. XRD pattern for 500 C isotherm of co-milled 36.8% UO 2 fragments................................86 B-7. XRD pattern for 375 C isotherm of co-milled 50.0% UO 2 fragments................................87 B-8. XRD pattern for 400 C isotherm of co-milled 50.0% UO 2 fragments................................88 B-9. XRD pattern for 425 C isotherm of co-milled 50.0% UO 2 fragments................................89 B-10. XRD pattern for 450 C isotherm of co-milled 50.0% UO 2 fragments..............................90 C-1. Unpolished U 0.05 Th 0.95 O 2+x pellet surface. Scale bar is 21 microns....................................91 C-2. Unpolished U 0.05 Th 0.95 O 2+x pellet surface. Scale bar is 20 microns....................................91 C-3. Broken surface of U 0.2 Th 0.8 O 2+x pellet...............................................................................92

PAGE 11

xi Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science EXPERIMENTAL DETERMINATION OF THE DRY OXIDATION BEHAVIOR OF A COMPOSITIONAL RANGE OF URANIUM-THORIUM MIXED-OXIDE PELLET FRAGMENTS By Lisa Argo December 2003 Chair: Ronald H. Baney Major Department: Materials Science and Engineering Oxidation of (U y Th 1-y )O 2 (y = 0.236, 0.368, 0.500) solid solutions was investigated using thermal gravimetric analysis and compared with UO 2 The UO 2 and ThO 2 powders were ground, pressed into pellets, and sintered at 1650 C in a reducing atmosphere. Gravimetric oxidation data for all samples, including UO 2 exhibited single-step behavior. This was interpreted as the effect of the low surface-to-volume ratio of the fragments used (specific surface areas were 0.01 to 0.02 m 2 /g) on the relative contributions of surface and bulk oxidation reactions, with bulk reactions dominating in this case compared to powders with higher surface area. Model-fitting kinetic analysis suggested diffusion (y = 0.236, 0.368) and possible nucleation and growth (y = 0.500) as possible mechanisms. Activation energy, E, for isothermal oxidation calculated by model-fitting increased with uranium content, 62.7 17.9 and 171 8 kJ/mol for y = 0.236 and 0.368, respectively. X-ray diffraction patterns and uranium valence calculations do confirm that the ultimate oxidation is inhibited in the (U,Th)O 2 compared to UO 2 Lattice structure remained cubic fluorite and did not undergo any phase transformation. Cation valencies indicated that

PAGE 12

xii uranium does not proceed to its maximum oxidation state. The absence of quantitative particle and grain size data restricted the models in fully describing system complexity.

PAGE 13

1 CHAPTER 1 INTRODUCTION The objective of this project was to determine the long-term stability of ThO 2 /UO 2 high level waste. Radionuclide leaching from spent UO 2 fuel is a major long-term storage concern for radiological materials. Mixed oxide (U,Th)O 2 fuels were considered an alternative because of the natural abundance of ThO 2 nonproliferation benefits, and potentially improved long-term spentfuel storage capability. Oxidation behavior of (U,Th)O 2 solid solutions was measured as a function of the oxide composition in order to determine the relative advantage these materials possess over conventional UO 2 Uranium dioxide oxidizes to a multitude of naturally occurring off-stoichiometric phases, (i.e., U 3 O 7 U 4 O 9 and U 3 O 8 ). These phase changes induce pellet cracking (due to a 36% volume increase associated with the UO 2 U 3 O 8 transformation) and generally facilitate radionuclide release from the crystal lattice. Additionally, the higher oxidation states, U(V) and U(VI), form water-soluble species. Release of hazardous radionuclides from the spent-fuel into the surrounding environment, therefore, may occur through surface oxidation or catastrophic fracture during long-term storage. Since the life expectancy of uranium and thorium based high-level radioactive waste greatly exceeds any reasonable length of time to obtain experimental data, predictions of waste behavior rely upon accelerat ed tests and/or models based on an understanding of system behavior. Half-lives reported for uranium-238, -235, and thorium-232 are 4.46 10 9 years, 7.04 10 8 years, and 1.4 10 10 years, respectively [Win03]. For instance, after gamma irradiation of CANDU polycrystalline UO 2 pellets in oxygenrich/-free and moisture-rich/-free atmospheres, Sunder and Miller [Sun96] observed U(VI) by xray photoelectron spectroscopy (XPS) and multiple uranium oxide phases (UO 2 U 3 O 7 U 3 O 8

PAGE 14

2 U 8 O 19 UO 3 U 16 O 37 etc.) by x-ray diffraction (XRD). Equivalent to gamma fields associated with 10to 20-year-old used CANDU fuel [Sun96], irradiated pellets did not reach temperatures in excess of 150 C during Sunder and Millers 2-day experiments. Uranium oxidation thus was restricted to surface crystalline growth. Catastrophic pellet fracture was not observed. Serrano et al. [Ser01] evaluated oxidized LWR UO 2 spent-fuel pieces in wet leaching experiments. Initial uranium fractional release was clearly related to oxidation level (starting O/M) of the fuel. Fission products identified in the leachate solution were plutonium-239, molybdenum-98, cesium-137, strontium-90, and technetium-99, where all but Pu-239 had release rates greater than for uranium. Serrano et al. [Ser01] attributed the increased fractional release rates to mobility along grain boundaries; and higher solubility of some fission products (like Tc or Cs). Shoesmith [Sho00] identified unoxidized UO 2 as a slow dissolving semiconducting oxide with the major rate-controlling process as surface ionic species formation by surface charge transfer or alteration. Most (> 90%) fission products and actinides generated in-reactor, in fact, are retained in the UO 2 fuel matrix and expected to be released as the fuel degrades in storage [Sho00]. The sensitivity of UO 2 solubility when oxidized, therefore, makes fuel dissolution and radionuclide release dependent on repository redox conditions (i.e., oxidant supply) [Sho00]. Currently, various waste forms and packages encapsulate, stabilize, shield, and otherwise prevent or minimize high-level radioactive waste release while in storage at geologic repositories or interim sites. Two waste-form systems, borosilicate glass and Synroc, are currently used; although glass is still the preferred method for civil high-level radioactive waste. Both systems immobilize radioactive actinides within a material matrix. The liquid properties of glass permit the glass matrix to accommodate impurities nearly indiscriminately. Synroc, which is in developmental stages in the U.S. for military waste and surplus military plutonium storage, comprises titanate minerals (such as titanium dioxide, hollandite, zirconolite, and perovskite). Synroc, like vitrified glass, is globally recognized for its chemical durability and resistance to leaching at high temperatures.

PAGE 15

3 In the same vein as glass vitrification and Synroc, mixed (U,Th)O 2 oxide solid solutions are pursued in this study as a material barrier specifically to uranium release. Thorium oxide, like the actinide oxides used for nuclear-reactor fuels, has a fluorite-type cubic lattice structure. Cations are arranged in a face-centered cubic close-packing sublattice. Anions occupy all tetrahedral interstitial positions, forming a simple cubic sublattice, with cations occupying one half the interstitial sites, and the remainder are vacant. Large octahedral interstitial holes remain where interstitial ions can be introduced to form a hyper-stoichiometric oxide. Diffusion characteristics of the fluorite structure indicate that anionic mobility (diffusion) is greater than cationic and ionic self-diffusion, because both anions and cations are sensitive to deviations from stoichiometry [And83]. Deviations from ThO 2 stoichiometry, however, are hindered. Unlike UO 2 thorium in ThO 2 is present in its maximum stable oxidation state, Th(IV), and cannot both accommodate excess anions and maintain charge neutrality. In fact, ThO 2 oxidation is largely independent of oxygen partial pressure [And76, And79, Haw68]. Additionally, thorium oxide is structurally and atomically similar to uranium oxide, unlike borosilicate glass and titanate minerals. The stability and physical properties of the thorium oxide matrix present an alternative to pure UO 2 fuel in-reactor and may potentially replace other waste-form systems in minimizing radionuclide release in storage. Because ThO 2 is in its maximum oxidation state, thorium oxide is stronger and more durable than UO 2 in various ways. Solubility of ThO 2 over a wide range of aqueous solutions is extremely low as compared to UO 2 under reducing conditions [Tay96]. The ThO 2 (like UO 2 ) is not susceptible to radiation-induced phase transformation to an amorphous state [Tay96]. Grain growth, a major cause of fission product release and governed by cation diffusion, is expected to be similar or lower than UO 2 Since thorium oxide is a better thermal conductor, with a higher melting point and slower cation diffusion [And83, Tay96], it is expected to run cooler and undergo less grain growth for a given power rating and fuel geometry [Tay96].

PAGE 16

4 Once a waste-form system has been used, the spent-fuel is further contained and packaged in systems (such as the use of lead canisters and other shielding or absorbent materials immediately surrounding an individual waste container). Despite the development of long-lived nuclear waste containers, the final barrier to radionuclide release is probably the waste-form system. Wet and dry waste vault/repository conditions are expected. For instance, according to the Nuclear Waste Management Organization, Canada initially stores spent nuclear fuel in waterfilled pools called Irradiated Fuel Bays; and 7 to 10 years later transfers it to dry storage facilities. Assessing overall repository performance therefore demands an understanding of potential fuel degradation in wet and dry conditions. The scope of this research consists of the manufacture and dry oxidative behavior of urania-thoria oxide solid solutions under accelerated dry-storage conditions as compared with pure UO 2 Kinetic analysis is used to assess mixed (U,Th)O 2 oxide behavior versus UO 2 Various processing methods were considered, and a significant portion of this study was dedicated to evaluating and selecting a synthesis technique that would yield a homogenous product. Our study did not attempt to assess the in-reactor performance of urania-thoria fuels.

PAGE 17

5 CHAPTER 2 REVIEW OF LITERATURE Current Status of UO 2 and ThO 2 Research Uranium Dioxide Natural uranium is primarily composed of two isotopes (U-238 and the fissionable U235). The fissionable isotope characterizes the economical value of nuclear fuel. The Oakridge National Laboratory Review website [Gab93] reported that nuclear fusion generates approximately 2 10 9 kWh/ton and coal combustion 6150 kWh/ton. From a mass to powergenerated standpoint, uranium is undoubtedly superior. In generating this power, however, the radioactive and toxic byproducts generated dictate a complex waste system to prevent harm to the surrounding environment. The final waste form, after mechanical barriers have degraded, will ultimately be a material system that must immobilize mobile species, such as uranium. Defect structure. Uranium dioxide has a cubic fluorite-type lattice structure, like other actinide oxides such as ThO 2 and PuO 2 In this crystal structure, diffusion characteristics favor anion diffusivity more than cation diffusivity. Point defects considered for diffusion in fluoritecubic oxides are the oxygen Frenkel pair, cation Frenkel pair, and Schottky trio. Written in Kroger-Vink notation, defect-formation energies for stoichiometric UO 2 are shown in Table 2-1 suggesting that the most dominant defect is the oxygen Frenkel pair [And83]. Ando and Oishi [And83] reported that stoichiometric UO 2 and ThO 2 showed similar diffusion characteristics with activation energies ranging between 200 and 275 kJ/mol, which they interpreted to be intrinsic diffusion. The similarity in stoichiometric UO 2 and ThO 2 would further support that anion diffusivity is closely tied to the simple cubic anion sublattice, almost

PAGE 18

6 regardless of cation species. Uranium self-diffusion in stoichiometric UO 2 in turn, is closely related to the face-centered cubic cation sublattice. Table 2-1. Defect formation energies for UO 2 Defect Kroger-Vink Notation Energy (kJ/mol) Oxygen Frenkel pair ]][[//iOOV 482 Uranium Frenkel pair ]][[4/4iUUV 1784 Schottky trio 2/4]][[OUVV 993 Hyperstoichiometry. In the case of hyperstoichiometric uranium dioxide, anion and cation mobility are further enhanced. Qualitatively, it is known that using oxygen-rich materials facilitate sintering of UO 2 thereby enhancing cation diffusion. Ando and Oishi [And83] reported published activation energies 90 to 100 kJ/mol for oxygen self-diffusion in UO 2+x that appeared to be independent of the degree change of x. Hawkins and Alcock [Haw68] measured cation tracer-concentration depth profiles by alpha-ray spectrometry in single and polycrystalline hyperstoichiometric (UO 2+x x = 0.01, 0.03, 0.10, 0.15) uranium oxide. They observed that volume contributions dominated near surface, whereas grain boundary contributions to diffusion in polycrystalline samples were virtually identical to the measured profile at large penetrations. Log plots of U diffusion coefficient as a function of off-stoichiometry revealed orders of magnitude increase in the cation diffusion coefficient with small departures from stoichiometry, regardless of single or polycrystalline sample [Haw68]. Insufficient data prevented Hawkins and Alcock from ascertaining the mechanism of cation diffusion in UO 2+x Oxidation behavior. The oxidation of UO 2 is commonly observed as a two-step process [Bla58, Ban68, McE97a, McE97b]. Initially, excess oxygen is interstitially accommodated into the cubic fluorite-type UO 2 structure [And83], resulting in an oxygen hyperstoichiometry and ultimately distorting the cubic lattice. Given a sufficient supply of oxygen and thermal energy to the system, the O/U ratio increases until the cubic matrix no longer supports the excess anions. The material then undergoes a phase transformation to U 3 O 7 /U 4 O 9 that Willis [Wil87] denotes as

PAGE 19

7 an ordered superlattice structure, and subsequently transforms to orthorhombic U 3 O 8 Boase and Vandergraaf [Boa77] observed that tetragonal U 3 O 7 /U 4 O 9 remained the stable phase at temperatures below 250 C, whereas U 3 O 8 was the stable end product at higher temperatures. Table 2-2 lists the three competing reactions that are suggested as the principal mechanisms that occur for UO 2 oxidation [Boa77]. The transformation of UO 2 to U 3 O 7 occurs initially and is a surface reaction that proceeds at the solid/gas interface, with a limiting thickness of ~22 nm. Boase and Vandergraaf [Boa77] reported that the oxidation of U 3 O 7 to U 3 O 8 at 235C was significantly slower than the UO 2 to U 3 O 8 reaction. The two bulk reactions that form U 3 O 8 therefore occur simultaneously and in competition with each other, particularly in high surface area powders. Blackburn et al. [Bla58] measured the rate of UO 2 oxidation to determine the reaction mechanism. They observed that the weight gain versus time curve for the UO 2 to UO 2.33 reaction did not indicate any apparent change in mechanism. Below 300 C, a surface layer of single phase -U 3 O 7 was immediately formed and further oxidation was controlled by oxygen diffusion through this phase, such that the oxidation rate was inversely proportional to oxide thickness. The second oxidation stage, which has been shown to be a nucleation and growth process, [Bla58] exhibited an induction period, a gradually increasing reaction rate, and finally a region in which the rate slows as the final composition approaches UO 2.67 Table 2-2. Oxidation reactions of UO 2+x Oxidation reaction Comments (1) 6UO 2 + O 2 2U 3 O 7 Fast; surface only (2) 2U 3 O 7 + O 2 2U 3 O 8 Slow at temperatures < 250 C; surface only in coarse powders (3) 3UO 2 + O 2 U 3 O 8 Faster than reaction (2); occurs in the bulk only Surface area dependence. The individual contributions of these competing reactions (#2 and #3 from Table 2-2) in the UO 2 system are dependent on the specific surface area of the sample, and separating them is often difficult. Because both surface and bulk reactions occur simultaneously, the surface to volume ratio of the sample has a significant bearing on the relative

PAGE 20

8 contributions of the various oxidation mechanisms. In fact, Dobrov et al. [Dob98] cited work that showed the surface oxygen exchange is a relatively slow process and is the rate-limiting step, not solid-state diffusion, of fuel oxidation in thin samples where surface contribution is greater than bulk. Experiments reported using powders with specific surface areas ranging from 0.61 to 19 m 2 /g [Ban68] indicated a direct relationship between the final O/U ratios of the oxidized materials and the specific surface area of the samples. Taylor et al. [ Tay98 ] studied the rate of U 3 O 8 formation as a function of pellet surface roughness at 250 C and noted that reaction times increased by a factor of four with increasing roughness, corresponding to an increase in the nucleation and growth rate constant, by two orders of magnitude. A nonlinear relationship between and surface roughness was observed such that was independent of surface roughness with very fine (< 1 m) or coarse (18-100 m) polishing agent particle size. For 1-18 m particle size polishing media, values increase with media particle size. Taylor et al. attributed this correlation of with roughness to a combination of at least three factors: increasing macroscopic surface density of nucleation sites, increasing microscopic surface density of nucleation sites, and preferred oxidation in the <111> direction. The 36% volume expansion is largely accommodated in the <111> UO 2 direction, which becomes the <001> direction ( c axis) of orthorhombic U 3 O 8 Activation energies. A wide range of experimentally determined activation energy values has been reported for the oxidation of UO 2 The results of previous UO 2 oxidation studies have been summarized in the literature [McE97a-b]. See Appendix A. Activation energies for the formation of U 3 O 7 /U 4 O 9 on unirradiated UO 2 fuel and spent LWR fuel have been reported to range from 90-120 kJ/mol [McE97a]. Values for the oxidation of UO 2 fuel and spent-fuel to U 3 O 8 determined with a variety of experimental techniques, exhibit a much larger range, from 48-172 kJ/mol [McE97b]. The broad range in experimental values for UO 2 U 3 O 8 oxidation highlights the complexity of the oxidation process for this material, such as grain boundary diffusion, off-stoichiometry (O/M), oxidizing conditions, and/or in-reactor conditions.

PAGE 21

9 Thorium Oxide A naturally abundant material, thorium oxide has found use in a variety of products from portable gas mantles to catalyzing petroleum cracking. It is also an identified fertile material used for producing nuclear fuel. Although not fissile, thorium-233, like uranium-238, absorbs slow neutrons to produce fissile uranium-233. The comparison with uranium reveals a host of properties that can be exploited in the waste end of the nuclear fuel cycle. Thorium oxide possesses a molecular weight, cation radius, and lattice structure similar or identical to UO 2 Not surprisingly, ThO 2 forms solid solutions with UO 2 at all U/Th ratios. Table 2-3 summarizes some material properties of ThO 2 and UO 2 Table 2-3. Material properties of ThO 2 and UO 2 Material property ThO 2 UO 2 Metal electron configuration [Rn] 6d 2 7s 2 [Rn] 5f 3 6d 1 7s 2 Metal atomic number 90 92 Metal atomic weight 238.0289 g/mol 232.0381 g/mol Metal atomic radius (neutral) 1.8 1.75 Cation radius 1.18 1.14 Melting point 3390 C 2827 C Density + 9.986 g/m 3 10.977 g/m 3 Lattice structure + Fm3m (225) Fm3m (225) Lattice parameter + 5.600 5.467 Lattice energy 10397 kJ/mol 10644 kJ/mol Radioisotope half-life 1.4 10 10 y (Th 232 ) 4.46 10 9 y (U 238 ) radii for coordination number (CN) = 8 and valence = 4+ [Ric92] + JCPDS Diffraction data 04-0556 and 41-1422, respectively Defect structure. Transport phenomena within the thorium oxide lattice are important to predict mobility behavior of inclusions, such as uranium and fission products generated inreactor. As with other fluorite-structured materials, anion Frenkel defects (oxygen vacancyinterstitial pair) are thought to be the predominant intrinsic disorder, with cation vacancies and holes as minor defects [And76, Col83]. Calculated effective formation and migration energies for intrinsic disorder reported by Colbourn and Mackrodt [Col83] further support thorium oxide stability over uranium dioxide, as summarized in Table 2-4, where cation interstitials appear to be far too energetic to play any significant role in the cation disorder of thorium oxide.

PAGE 22

10 Table 2-4. Cation and anion formation and migration energies in ThO 2 ThO 2 (eV) Effective formation energies for intrinsic disorder Anion vacancy ~3.00 Anion interstitial ~3.00 Cation vacancy 5.93 Cation interstitial 13.87 Anion migration Vacancy 0.78 Direct exchange 5.80 Direct interstitial 3.27 Interstitialcy mechanism 0.92 Direct interstitial migration of O i 4.28 Cation migration Vacancy 7.04 Divacancy 5.36 Trivacancy 6.35 Oxidation behavior. Results published by Ando et al. [And76, And79] of single crystal 99.99% purity ThO 2 indicate that oxygen diffusion kinetics are rate limited by surface exchange at temperatures 973 C to 1593 C with respect to time. Concavity was observed at initial and final times when fractional O 18 uptake is plotted as a function of diffusion time. However, oxygen self-diffusion, with respect to temperature dependence, did not appear to be influenced by specimen size and pre-annealing. The temperature dependence of oxygen self-diffusion split into a high (> 1100 C) and low (< 1100 C) region, which Ando et al. [And76] attributed to intrinsic and extrinsic oxygen diffusion, respectively. Further work by Ando and Oishi [And79] show ThO 2 ionic conduction primarily comes from oxygen self-diffusion and is independent of oxygen partial pressure. Impurities in thoria. Experimental investigations of the presence of impurities in the thorium oxide lattice are largely concerned with electrical conductivity of solid solutions to shed light on the defect structure. Colbourn and Mackrodt [Col83] reported vacancy binding energies for diand tri-valent cation impurities, including Th 3+ which give rise to anion vacancies for ionic conduction in the thorium oxide lattice. Di-valent impurities gave rise to greater substitution and vacancy binding energies than tri-valent; and decreased in magnitude with decreasing dopant radii. [Col83] As would be expected, smaller impurities and, particularly,

PAGE 23

11 valency close to 4+ ease inclusions into the lattice. It stands to reason that U 4+ with its similar radii to Th 4+ and any of the lanthanide or actinide elements, particularly those with 4+ valency, would be energetically favorable impurities accommodated in the thorium oxide lattice. Uranium-Thorium Mixed Oxide The oxidation behavior of (U,Th)O 2 solid solutions has been studied previously. Within a composition range of 24% to 90% moles of uranium dioxide, (U,Th)O 2 phases are always single phase cubic fluorite type when oxidized at temperatures below 200 C [And54]. In high uranium concentration (> 95% mol) unoxidized pellets at room temperature, lattice parameters increased linearly with decreasing uranium content, reflecting Vegards law that implies atomic volume is conserved regardless of local lattice distortions [Tsu98]. Chandramouli et al. [Cha98] noted the cubic fluorite structure was maintained irrespective of uranium content of unoxidized (U,Th)O 2 solid solutions. Oxidation of compositions containing less than 50% mol UO 2 yielded only fluorite phases under any conditions. Mixed oxides containing greater than 50% mol UO 2 however, formed second non-cubic phases, hexagonal or orthorhombic U 3 O 8 [And54, Cha98]. X-ray diffraction patterns indicated the presence of a U 3 O 8 second phase, which disappears upon reduction heating at 1300 C, results in uranium depletion from the solid solution [Cha98]. Anderson et al. [And54] observed that mixed oxides containing less than 78% UO 2 were stable in air up to 1400 C, except for 66% UO 2 which lost its cubic structure in high-pressure oxygen conditions. Otherwise, high uranium compositions, such as U 0.9 Th 0.1 O 2 evidenced x-ray pattern broadening (indicative of unit cell contraction), breakdown of cubic lattice symmetry when oxidized below 200 C, phase separation above 200 C oxidation, and a final orthorhombic U 3 O 8 like structure at 600 C [And54]. Anderson et al. deduced that for UO 2 concentrations from 15 to 78% mol, excess oxygen entered interstitial sites in the fluorite lattice based on changes in unitcell density, which increased with oxidation in agreement with theoretical calculations. Unit cell

PAGE 24

12 parameters, however, demonstrated a parabolic relationship with uranium valence, such that a minimum was observed near U valence = 5. Anderson et al. interpreted this behavior as a competition between initial ionic attraction and volumetric accommodation of excess anions. Aronson and Clayton [Aro60] examined (U,Th)O 2 compositions synthesized by ammonium hydroxide co-precipitation, with UO 2 contents ranging from 30 to 90% mol and x-ray diffraction patterns confirmed a single phase fluorite-type structure present at all compositions. Following controlled oxidation and annealing at 800 C, it was observed that the added oxygen was accommodated without destroying the fluorite structure. The activation energy of (U,Th)O 2 oxidation has been studied using thermogravimetric techniques [Ant00]. All of the samples studied exhibited single-step weight-gain curves. For low uranium content (U y Th 1-y )O 2 powders (y = 0.15 and 0.30), the average activation energy for oxidation was calculated to be 45 kJ/mol. Higher uranium content powders (y = 0.72 and 0.77) had average activation energies of 74 kJ/mol. The higher activation energy for higher uranium content powders was attributed to the phase separation that occurs when these materials are oxidized at high temperatures. Oxidation potential. Ugajin [Uga82] prepared (U,Th)O 2 samples with 5%, 10%, and 20% mol UO 2 to measure the oxygen potential of non-stoichiometric mixed oxides as compared to UO 2+x at 1000 C to 1200 C in CO 2 /CO atmosphere. Oxygen potentials of mixed oxides, like UO 2 increased with increasing O/M at all temperatures and compositions. Assuming Th valence is constant at 4+, Ugajin observed oxygen potentials increased with increasing Th content and U valency more so than UO 2+x at greater deviations from stoichiometry. Supplemented by Anthonysamy et al. [Ant97] and Arima et al. [Ari00] measurements of 1%, 3%, 5%, 54% and 90% mol UO 2 the compositional range of (U,Th)O 2+x oxygen potential data dependence upon uranium valence and the U/(U+Th) ratio was further affirmed.

PAGE 25

13 Matsui and Naitos [Mat85] work with 20% and 40% mol UO 2 at 1000 C to 1100 C in H 2 /CO 2 atmosphere further interpreted the differences between (U,Th)O 2+x and UO 2+x oxygen potential behavior according to the dependence of x on oxygen partial pressure expressed as x P(O 2 ) 1/n Assuming a complex defect (2:2:2) model with two different interstitial oxygen types (O i a and O i b ) and one oxygen vacancy (V O ), Matsui and Naito showed UO 2+x to have three different dependencies of x upon P(O 2 ) and four regions for (U,Th)O 2+x summarized in Table 2-5 [Mat85]. Table 2-5. Composition dependence on oxygen partial pressure for UO 2+x and (U,Th)O 2+x Phase Composition n for x P(O 2 ) 1/n Defect Model UO 2+x x < 0.003 0.003 < x < 0.006 x > 0.006 2 12 2 Neutral defect {2O i a O i b 2V O } {2O i a O i b 2V O } 5 {2(Oi aOi b2VO)} (U,Th)O 2+x x < 0.001 0.001 < x < 0.003 0.003 < x < 0.008 x > 0.008 2 4 12 4 Neutral defect {2O i a O i b 2V O } {2O i a O i b 2V O } {2O i a O i b 2V O } {2(Oi aOi b2VO)} Polycrystalline diffusion. Furuya [Fur68] measured cation ( 237 U) diffusion in ThO 2 and UO 2 -ThO 2 polycrystalline pellets with average grain sizes over 60 m over temperatures ranges 1800 C to 2000 C and 1800 C to 2300 C, respectively. As expected, lattice diffusion made a significant contribution near surface, resulting in a non-linear concentration profile. Deeper penetration regions show 237 U concentration varied linearly with distance, characteristic of grain boundary diffusion. Excluding the section nearest to the free surface, Furuya [Fur68] calculated lattice and grain-boundary diffusion contributions. Comparison of the associated activation energies showed that lattice diffusion through (U,Th)O 2 (360 kJ/mol) is greater than ThO 2 (320 kJ/mol). Furuya deduced the smaller lattice spacing in (U,Th)O 2 caused the potential barrier at the saddle point to increase compared with ThO 2 Both values were greater than values reported in nominally stoichiometric UO 2 (304 kJ/mol), with implications of the influence of offstoichiometry on cation diffusion. Activation energies for grain boundary diffusion (269 and 201

PAGE 26

14 kJ/mol), as expected, were significantly lower than lattice diffusion and also saw the same relationship between (U,Th)O 2 and ThO 2 respectively. Ando et al. [And85] measured uranium oxide evaporation mechanism and ratecontrolling step for polycrystalline Th 0.90 U 0.10 O 2.05 and Th 0.75 U 0.25 O 2.13 after heating at 1650 C in flowing air for 32 h. UO 2 and ThO 2 concentrations varied to a depth of approximately 20 m and 80 m from the surface in the 10% UO 2 and 25% UO 2 samples, respectively. The concentration profiles showed that uranium oxide preferentially vaporizes. Pores and cracks were also observed after evaporation annealing, with the cracks and voids decreasing with increasing depth. Cation diffusion, further enhanced by grain boundaries, was identified as the rate-controlling mechanism. Background of Synthesis Methods The powder-pellet route involves generation and handling of fine powder or particles (<1 m) and is hence associated with the problem of radiotoxic dust hazard and fire hazard (applicable for carbide and nitride powders). Further, the fine powders are not free flowing and pose problems in remote and automated fabrication. Microhomogeneity of fissile species in mixed oxide is not fully obtained since the oxide powders are mechanically mixed. The alternative sol-gel type processes use dust-free, free-flowing and coarse (100 to 2000 m) particles as starting materials for pellet making. The choice of uranium oxide as a nuclear fuel is because of its high melting point (2828 20 C), corrosion resistance to radiation damage, and irradiation stability. The physico-chemical characteristics of UO 2 such as density, surface area, pore structure, grain size, crystallite size, sphericity, and oxygen to uranium ratio, depend mainly on the preparation method [Abd90]. Solgel wet methods are typically preferred over traditional dry powder milling for producing spherical particles. The sol-gel processes have been developed and applied at Oak Ridge National Laboratory (USA), AERE (United Kingdom), KEMA (Netherlands), CNEN (Italy),

PAGE 27

15 KFA Julich (FRG), and Tokai, Japan for high density thorium oxide and/or uranium oxide microspheres [Abd90]. The oxalate precipitation method is the primary process for commercially produced thorium oxide. White et al. [Whi81] investigated precipitation temperature, agitation method, and digestion time to refine powder sinterability and density without milling for ThO 2 and 25% UO 2 ThO 2 A 10 C temperature, mechanical stirring, and 15 min digestion yielded the most sinterable powder with 96% theoretical densities (TD) without milling. Temperature was identified to have the most effect upon particle morphology, surface area, crysta llite size, and sinterability. The lower (10 C) digestion temperature, however, yielded cubic particles less than 1 m in size. At 70 C, the particles were square platelets varying in size from 1 to 3 m along the square edge. Longer digestion times rounded off the edges and generally produced more uniform particle sizes Ganguly et al. [Gan86] examined thorium oxide and uranium oxide microspheres by well-established Societa Nationale Matanodoti (SNA) and Kernforschungsanlage Julich (KFA) external gelation and KFA and Keuring Electrotechnische Materialen Arnkem (KEMA) internal gelation processes. Pellets made by each of the processes, however, presented relatively poor densities and general behavior. Particle size, density, and crushing strength influenced pressing characteristics. Specific surface area, crystallite size and additives determined sintering behavior. Ganguly et al. noted that adding a sintering aid (CaO) and pore forming additive (carbon black) improved pelletization and sintering ( 94% TD). Chandramouli et al. [Cha98] observed that surface area of mixed oxide powders were influenced by composition and calcinations method. As U content increased, surface area of conventionally calcined powders decreased. Microwave heating also resulted in low surface area powders with large crystallite sizes. Solid solutions with 15% mol U content calcined by a graphite coupling agent yielded surface areas nearly twice that of conventional furnace calcined

PAGE 28

16 powders. Residual carbon content, also, was greater in microwave-calcined powders than those heated by conventional furnace. Kinetic Analysis The Arrhenius approach is used to interpret the rate dependence of thermal decomposition of solids on temperature. Predicting how quickly the solid state (U,Th)O 2 system approaches equilibrium does not necessarily require full understanding of the complicating features of real systems. Simple approximate theoretical kinetic models can be used to interpret experimental data and make predictions of (U,Th)O 2 behavior at the end of the nuclear fuel cycle. Additionally, knowledge of UO 2 and ThO 2 defect structures and oxidation mechanisms can lay the groundwork for fully revealing the mechanisms that predominate (U,Th)O 2 oxidation (i.e., diffusion and/or nucleation and growth). Kinetic reaction models. Basic solid state reaction models, f( ), are largely based on simplified geometrical assumptions for reaction particles, such as geometry of reaction interface. Contracting area and volume models assume its basis in the reaction interface and corresponding spatial movement. The Johnson-Mehl-Avrami-Erofeev-Kolmogorov and Prout-Tompkins kinetic reaction models describe nucleation and growth processes. Diffusion controlled reaction models, which have largely been developed from gas-solid interactions, assume simplified geometry such as spherical or cylindrical particle shapes and ball-and-stick lattice structures. Arrhenius equation. Generally, homogeneous kinetics is assumed to behave according to the simple differential kinetic rate and Arrhenius Equations 2-1 and 2-2, respectively The activation energy, E, is the energy barrier or threshold that must be overcome to enable the bond redistribution steps required to convert reactants into products. The pre-exponential term, or frequency factor, A, is a measure of the frequency of occurrence of the homogeneous reaction situation. This is typically envisioned as including the vibration frequency in the reaction coordinate. From a reaction dynamics treatment [Gal02], the activation energy is identified as the difference between the energy of the molecules undergoing reaction and the overall average

PAGE 29

17 energy. Degree of conversion or fraction reacted, rate coefficient, and kinetic model are expressed as (where 0 1), k(T), and f(), respectively. (2-1) )()( fTkdtdb= (2-2) =RTEATkexp)( The hypothesis is that the reaction involves only an active part of all reactant molecules that, according to the Maxwell-Boltzmann distribution law, is an exponential function of temperature [Lvo01]. Galwey and Brown [Gal02] expressed that the Maxwell-Boltzmann equation, which is applicable to homogeneous gaseous systems and a starting point for theoretical explanation of Arrhenius behavior in homogeneous reactions, does not adequately represent the energy distribution of the immobilized reactants of solid-state processes. Mechanisms, such as three-body collisions or linked sequence of steps, which are unique to the constrained mobility of solid-state systems, oppose a single contributing step during the entire reaction as a realistic predictor of decomposition behavior. Additionally, reactant energies in liquid or solution are associated with individual molecules, whereas band theory describes the energy distribution within crystals. Garn [Lvo01] noted that no discrete activated states can exist during the decomposition of a solid. Subsequently, the Arrhenius kinetic parameters (A and E) have lost their physical meaning. Because of the spatial constraints of solid-state processes, there are no collisions of freely moving reactant molecules as defined by the frequency factor for homogeneous reactions. Energy transfer through vibrational interactions within the crystal, also, occurs so fast that no substantial deviations from the average energy can take place [Lvo01]. Limitations. The assumption that kinetic parameters are intrinsic constants, which uniquely characterize a given solid-state process, however, is rather controversial and often leads to misunderstandings since solid-state processes typically exhibit complex kinetics [Mal01]. Although mathematical descriptions can be determined for most solid-state processes to optimize

PAGE 30

18 desired variables, the underlying mechanism is not necessarily easily obtained [Mal01]. Unlike liquid and gaseous systems, reactants in solid-state processes are spatially constrained and place additional complications on the system. Further difficulties that add to the complexity of solid-state processes are inhomogeneous reactant distribution irregular shapes, polydispersity, shielding and overlapping of reacting particles, or preferred orientations [Mal01]. Empirical rate laws developed to address the anomalous reaction orders (kinetic exponents) in the rate law as a result of these complexities are shown in Table 2-6. Table 2-6. Rate laws for a simple process A P Reaction in solids )1( b=kdtd 1 st order )1( b=kdtd with autocatalysis Empirical rate laws nkdtd)1( b= 1 st order NMkdtd)1( b= with autocatalysis Fractional conversion (0 1), such that [A] = 1and [P] = Solid-state decomposition reaction rates are influenced by numerous factors that could inhibit determination of kinetic parameters. Although it is typically assumed that sample temperature is equal to that of the furnace, [Lvo01] Smith and Topleys work where single crystals in vacuum measured temperatures 4 to 8 K lower than the furnace. Lvov also previously developed a program that theoretically computed the layer-by-layer temperature distribution in powder samples decomposing in vacuum and foreign gases. Depending upon the total number of layers, n, temperature deviations can differ significantly for powders and single crystals. For instance, the temperature of the central layer of a Mg(OH) 2 sample with n = 1000, 10,000 and furnace temperature of 500 K or 600 K is actually 427 and 387 K, respectively [Lvo01]. Single crystal Mg(OH) 2 in vacuum with temperatures 550, 600, and 663 K is expected to have surface temperatures of 549.6, 593 and 628 K, respectively [Lvo01]. Correcting kinetic parameters for

PAGE 31

19 this self-cooling effect is naturally simpler for single crystals. This effect and number of layers, otherwise, must be determined through measurement of grain size, grain number, powder mass, crucible geometry, environmental conditions, and heating rate. With these corrections, it is possible to explain the differences between kinetic parameters obtained under different conditions. Other than the self-cooling effect, other reasons for scatter of reported values of A and E may be differences between mathematical modes by different authors and difference in the kinetic parameters for the nucleation and steady-state stages of decomposition [Lvo01]. Lvov [Lvo01] reviewed a physical approach for interpreting thermal decomposition of solids that is based on the Hertz-Langmuir prediction of proportional dependence of the evaporation rate on the equilibrium partial pressure of the vapor which, in its turn, depends exponentially on temperature. Most solid-state decomposition reactions proceed under conditions far from equilibrium. The traditional chemical approach tackles this deviation from equilibrium by connecting it to the theoretically unpredicted energy barrier, activation energy. In other words, non-equilibrium decomposition into equilibrium products [Lvo01]. The physical approach, however, contends that the reason for the deviation lies in the decomposition of the reactant into primary non-equilibrium gaseous species different from those at equilibrium. Lvov [Lvo01] asserts this approach permits quantitative interpretation of the mechanism of nucleation and the energy source supporting decomposition, retardation in the presence of gaseous products, low vaporization coefficients, thermal stability, effect of selfcooling, the Topley-Smith effect, and kinetic compensation effect.

PAGE 32

20 CHAPTER 3 MATERIALS AND METHODS Material Synthesis Three methods of powder fabrication were used, two of which employed sol-gel type approaches for obtaining a solid solution. The sol-gel type processes yield a co-precipitate from an aqueous solution of uranium and thorium nitrate salts followed by complexation with either an oxalate or hydroxide group. The final method engaged in this study adhered to traditional comilled powder techniques. Wet chemistry methods of material synthesis were initially pursued for its known homogeneity advantages over powder-mixing techniques and minimization of dust hazard. Both uranium and thorium are alpha emitters. The hazard, therefore, to be considered for workers are dust particulates becoming airborne and entering the human body. Safety gear, such as protective gloves, eyewear, and clothing, are necessary. Wet synthesis methods, however, present a reduced hazard since the process is largely contained in liquid form prior to heat treatment. Additionally, particle morphology of sol-gel type synthesis is largely spherical; presenting a free-flowing powder that is ideal when filling dies for pellet making. In contrast, mixing by traditional powder co-milling methods occurs entirely in powder form and generates platelet-like particles with less desirable flow characteristics. The details of each material fabrication techniques are presented in the ensuing sections Oxalate Co-Precipitation Since UO 2 and ThO 2 are iso-structural, both have face-centered cubic CaF 2 -type lattices, and have similar thermodynamic properties. Their fabrication processes therefore are nearly identical. Aqueous oxalic acid is utilized as the complexation agent to form the mixed oxide. A

PAGE 33

21 flow chart schematic in Figure 3-1 provides a brief synopsis of the steps entailed. Uranium oxynitrate hexahydrate (UO 2 (NO 3 ) 2 6H 2 O) and thorium nitrate hydrate (Th(NO 3 ) 4 xH 2 O) are supplied by Alfa Aesar. Analytical grade oxalic acid ((COOH) 2 2H 2 O) and sodium formaldehyde sulfoxylate dihydrate (HOCH 2 SO 2 Na 2H 2 O), also purchased from Alfa Aesar, are the complexing and reducing agents, respectively. A 115-volt motorized three-blade stirrer supplied the agitation for the co-precipitation step. Dissolve U -nitrate ( s ) in D.I. water. Add Th-nitrate ( s ) to retained filtrate. Reduce U ( VI ) to U ( IV ) with NaFS*. Di g est ~24 h. Dry in air, 120C for 5 h. Decompose in air, 350C for 5 h. Calcine in air, 900C for 24 h. Reduce in 5%H-Ar, 800C for 10 h. Filter p reci p itate. Cop reci p itate with excess Oxalic Acid ( a q) Se p arate & discard excess NaFS ( s ) Figure 3-1. Oxalate path co-precipitation. *NaFS is sodium formaldehyde sulfoxalate. The desired molar ratio of uranium to thorium determines the starting quantities of solid nitrate salts. Uranium (IV) nitrate solution is prepared by dissolving uranium (VI) oxynitrate hexahydrate in a sufficient volume of demineralized water to yield an approximately 1M concentration. Since thorium (IV) nitrate is later added in solid form, additional water is added to the solution to achieve a thorium concentration of 1M. The nitrate solution is then reduced via addition of a six-fold excess of the reducing agent at room temperature and covered for 24 h. The solution changes colors from yellow to orange to dark green. After digestion, the excess reducing agent solid is removed by vacuum-assisted filtration with Whatman #42 ashless filter paper and thorium nitrate hydrate is dissolved into the nitrate solution.

PAGE 34

22 Figure 3-2. Optical micrograph of nominal 20% UO 2 calcined oxalate-synthesized powder mounted with a collodion/amyl acetate solution at 50x magnification To form the co-precipitate, a six-fold excess of the oxalic acid complexing agent, dissolved in demineralized water to a 1M concentration, is added via one of two dropwise methods to the nitrate solution under ~600 rpm mechanical agitation. The two assemblies, buret and dropping funnel, were employed for oxalic acid addition to yield a fine particle size and spherical morphology. The buret formed a somewhat finer particle size than the dropping funnel due to the smaller aperture of a buret, which produced a smaller drop size. Because the agitation speed was held constant, drop size was the primary means of controlling particle size. A coarser powder, however, was desired to reduce airborne dust hazard and assist sieve particle size classification. Powders fabricated by either method demonstrated the desired free-flowing characteristics for pellet making. An optical micrograph (Figure 3-2) of nominal 20%UO 2 80%ThO 2 calcined powder demonstrates the spherical morphology observed with this method. Following the complexation of U and Th to the oxalate group, the precipitate is separated by vacuum-assisted filtration through #42 Whatman paper and dried in air at room temperature until it releases easily from the paper. The filtrate liquid is clear, an indicator that most of the uranium did not remain in solution. X-Ray Fluorescence performed by co-researcher Shibuya of the liquid filtrate confirmed this assertion. The precipitate cake, which is white-colored, is

PAGE 35

23 transferred to a glazed ceramic crucible. Eliminating the residual moisture and undesired organics (i.e., carbon) subsequently requires a series of drying, decomposition, and calcination steps carried out in air, followed by a final reduction heating in flowing 5%H 2 -Ar. Powder color following reduction that was performed in a tube furnace on an Alumina boat varied from tan to dark orange. Because of the multiple oxidation states available to uranium, which readily oxidizes, the final reduction step is needed to return it to U(IV) and face-centered cubic lattice structure. Since the powder will later be formed into pellet geometry and sintered, reduction is carried out at 800C to prevent particles from sintering. The molecular evolution of the coprecipitate is depicted in Table 3-1 where U/Th is 1 and X-Ray Diffraction pattern in Figure 3-3 for nominal U/Th = 5/95. Table 3-1. Evolution of oxalate-synthesized (U,Th)O 2 Gas T(C) Equation Filter air 25 U(NO 3 ) 4 Th(NO 3 ) 4 + 4H 2 C 2 O 4 [U(C 2 O 4 ) 2 Th(C 2 O 4 ) 2 ] nH 2 O + 8HNO 3 Dry air 120 [U(C 2 O 4 ) 2 Th(C 2 O 4 ) 2 ] xH 2 O [U(C 2 O 4 ) 2 Th(C 2 O 4 ) 2 ] + nH 2 O Decompose air 350 [U(C 2 O 4 ) 2 Th(C 2 O 4 ) 2 ] U(CO 3 ) 2 Th(CO 3 ) 2 + 4CO Calcine air 900 U(CO 3 ) 2 Th(CO 3 ) 2 UO 2+x ThO 2 + 4CO 2 Reduce 5%H-Ar 800 UO 2+x ThO 2 + nH 2 UO 2 ThO 2 + nH 2 O Drying steps carried out in air were performed in a Vulcan 3-550 Muffle Furnace. The final reduction drying at 800C took place in either a CM Rapid Temp or Lindberg/Blue 54434C tube furnace with Mullite and, later, 99.8% Alumina open-ended tubes supplied by Coors. Two sets of custom-made stainless steel caps were installed to control gas atmosphere during firing. The initial design was suspected of not maintaining an appropriate seal between the steel and ceramic, presenting carbon content issues for preliminary synthesis attempts. The latter design incorporated a high temperature Viton polymer gasket situated between the stainless steel cap and ceramic tube. Additionally, a vacuum set-up was assembled whereby the tube atmosphere could be evacuated and backfilled with the desired gas, thereby preventing the material from encountering an oxidative atmosphere. The tube atmosphere is evacuated and backfilled with 5%H-Ar three times before initiating the desired firing regime. Figures 3-4 and 3-5 show each furnace.

PAGE 36

24 Figure 3-3. X-Ray Diffraction evolution of U 0.05 Th 0.95 O 2 oxalate synthesized powder.

PAGE 37

25 Figure 3-4. Vulcan Muffle furnace used for drying, decomposition, and calcination Figure 3-5. Lindberg high temperature tube furnace (right) with alumina tube and stainless steel cap (above)

PAGE 38

26 Ammonium Hydroxide Co-Precipitation Another synthesis path used was derived from a modification of the well-documented Ammonium Diuranate (ADU) process, which is used in commercial processing of UO 2 fuel. The purposes of attempting this synthesis path did not extend beyond establishing a preliminary qualitative comparison with the oxalate path and traditional blending. Consequently, only a nominal 20% UO 2 composition of moderate quantity (5 grams) was processed. The quantities and specifications of the subsequent procedure were adapted from parameters made available by Dr. E. Lahoda of Westinghouse, Inc. A brief schematic of the procedure follows in Figure 3-6. Add Th-nitrate ( s ) Filter p reci p itate. Add HF such that F/U = 4. Pour nitrate solution in N H 4OH ( N H 3/U > 30 ) bath Dissolve U -nitrate ( s ) in D.I. water. Calcine in 5%H-Ar, 630C for 4 h. Figure 3-6. Ammonium hydroxide path co-precipitation Starting constituents were uranium oxynitrate hexahydrate (UO 2 (NO 3 ) 2 b6H 2 O) and thorium nitrate hydrate (Th(NO 3 ) 4 bxH 2 O) supplied by Alfa Aesar, and 49% diluted hydrofluoric acid (HF) and 29% diluted ammonium hydroxide (NH 4 OH). Desired molar quantities of uranium (IV) and thorium (IV) nitrate were determined. In demineralized water, uranium (IV) nitrate was dissolved to yield a 0.19 M solution. A molar excess of hydrofluoric acid was added to the uranium (IV) nitrate solution by pipette to raise the molar ratio F/U to 4. No color change or other apparent reaction observed. At this stage, thorium (IV) nitrate was stirred into the solution until dissolved. The solution took on a transparent green-yellow hue. A 30-fold molar excess of ammonium hydroxide aqueous bath was prepared separately. Without agitation, the nitrate salt

PAGE 39

27 solution was poured directly into the ammonium hydroxide bath. A bright yellow colored precipitate immediately formed. The precipitate was removed from the liquid by vacuum-assisted filtration through a Whatman #42 ashless paper and rinsed with demineralized water. The precipitate had an orange juice pulp-like appearance, as shown in Figure 3-7 at left, which was starkly different from the white oxalate precipitate. Additionally, the volume of material for this 5-gram batch appeared more voluminous than oxalate-synthesized batches. Following filtration, the yellow cake was air dried until it easily released from the filter paper. Figure 3-7. Ammonium hydroxide synthesized (U 0.2 Th 0.8 )O 2+x before (l) and after (r) calcinations Since the hydroxide path did not introduce organic substituents, the heat treatment scheme was decidedly simpler than for the oxalate path. In a flowing 5%H-Ar reduction atmosphere, the yellow cake was calcined at 630C for 4 hours. A significant volume reduction and color change from yellow to nearly black occurred during calcinations as shown in Figure 3-7 at right. X-ray diffraction patterns, taken prior to and following calcinations, revealed the cubic fluorite structure. Silicon was used as an external standard for XRD measurements of hydroxide-synthesized material before and after calcinations. Co-Milled Mixed Oxides Uranium dioxide and (U y Th 1-y )O 2 solid solutions with y = 0.236, 0.368, and 0.500 were prepared by mixing appropriate amounts of UO 2 and ThO 2 powders supplied by Siemens (SM) and Alfa Aesar (AA), respectively. Thorium and uranium oxide powders supplied by Alfa Aesar are manufactured to a below 325 mesh (< 45m) and 50 mesh (< 300m) particle size

PAGE 40

28 distributions, respectively. Optical micrographs of these powders affixed with a 1:7 collodion and amyl acetate solution are shown in Figure 3-8. From the micrographs, it is apparent that the ThO 2 agglomerates easily, forming large clumps, suggesting a platelet-like morphology. Uranium (IV) oxide particles from AA, on the other hand, presented spherical morphology and demonstrated good flow. Figure 3-8. Optical micrographs of Alfa Aesar UO 2 (l) and ThO 2 (r) Co-milling. Preliminary co-milling trials were conducted with high-density polyethylene (HDPE) jars, zirconia ball media, and a laboratory roller mill assembly. Milled powder became imbedded in the HDPE jar walls and clung to ball media. Cleaning jars and media was tedious and difficult. The material lost in the cleaning process, also, directly has an effect on the final U/Th composition. This set-up, therefore, was abandoned in favor of a small batch high impact mixer. Powders were weighed and batch milled for 60 minutes in a zirconium oxide jar with two 10-mm yttria-stabilized zirconia (YSZ) milling media using a high energy 8000M SPEX Certiprep Mixer/Mill. The mixer employs a torsional figure eight motion for pulverizing and mixing. Zirconia ceramic vial, cap, and ball media, as supplied by the vendor, accommodates loads up to 20 mL for mixing. No zirconia contamination from the container or milling media was found in the milled powders. To prevent cross contamination, grinding vials were cleaned

PAGE 41

29 with dilute nitric acid between each batch. Figure 3-9 displays the mixer and mill jar used in this study. Figure 3-9. 8000M SPEX Certiprep Mixer/Mill (l) and zirconia mill jar (r) Preliminary runs showed that the high impact velocity of the grinding balls caused powder to pack onto the vial walls. It was necessary, therefore, to mill in 15-minute intervals, scraping vial walls between each interval. After observing this, there was concern the particle size difference between starting powders inhibited effective blending if finer particles tended to pack more easily than coarse particles along vial walls. Because of the particle size and morphology difference between Alfa Aesar UO 2 and ThO 2 a 22 factorial experiment was initially performed to determine effect of fine (< 63 m) versus unsieved milled UO 2 and high (100 MPa) versus low (50 MPa) compaction pressure on sintered pellet density. Pre-milling as-received UO 2 (AA) spherical particles resulted in a platelet-like morphology similar to as-received ThO 2 (AA). A 30-gram batch of UO 2 (AA) was milled four 15-minutes consecutive intervals before sieving through a 230 mesh (63 m) to achieve a fine powder batch. It was observed that powder packing along vial walls became progressively hard-packed with each successive milling/blending intervals. Four 15-gram batches of 20% UO 2 80% ThO 2 blended for 60 minutes were processed according to the parameters of the experimental matrix shows in Table 3-2. Each batch yielded similar (within 2%) green and sinter densities, indicating little

PAGE 42

30 difference between using fine or coarse UO 2 (AA) and 50 MPa or 1000 MPa uniaxial compaction pressure. Table 3-2. 2 2 factorial for 20% UO 2 -80% ThO 2 blended oxide 50 MPa 100 MPa Unsieved UO 2 < 63 m UO 2 Green and sintered densities were measured three times each with calipers to determine volume. The pellets were sintered at 1650 C in a flowing 5%H-Ar atmosphere. Pellet densities were measured after 5 and 15 hour dwells at 1650 C. Those results are summarized in Table 3-3. Table 3-3. Co-milled U 0.2 Th 0.8 O 2 pressing conditions 2 2 factorial results Pellet Description Green Density Sintered Density (5 h) Sintered Density (15 h) Unsieved 50 MPa 60.2% 85.8% 97.2% Unsieved 100 MPa 61.7% 85.7% 98.7% Sieved 50MPa 61.2% 87.8% 98.1% Sieved 100 MPa 63.8% 86.7% 98.7% From the above results, the pressing parameters (100 MPa, sieved) were selected to make pellets for dry oxidation study. P. Demkowicz confirmed by the water immersion technique that sintered densities for the dry oxidation pellets > 98% theoretical density. Those pellets, subsequently, were fragmented by mortar and pestle and classified by ASTM sieves. The 90 to 250 m particle size range was selected. Demkowicz measured specific surface area of fragments in the 90 to 250 m range using the multi-point Brunauer-Emmett-Teller (BET) method in a Quantachrome Autosorb 1, with krypton as the adsorbate gas, and found all batches to be within 0.01 0.02 m 2 /g. Sample quantities were between 2 and 10-g, dried under vacuum at 120 C for approximately 12 hours prior to analysis. Pellet pressing. The milled ThO 2 -UO 2 powders were initially sieved through 100-mesh (< 149 m) to remove coarse particles before compaction. The powders were compacted into pellets in a 13-mm diameter stainless steel die using a single action laboratory press at 100 MPa. Pure UO 2 pellets were pressed using pressures of 200 MPa. Stearic acid was used as a lubricant on the die, anvil, and punch sidewalls. The dimensions of the pressed pellets were measured with

PAGE 43

31 calipers and the green densities, obtained by geometrical calculations, were 56% to 64% theoretical density (%TD). Pellets were arranged on a powder mound of matching composition to minimize any possible reactivity with the ceramic setter. The pressed co-milled pellets (Figure 3-10 shows oxalate synthesized pellets) were sintered at 1650C for 20 hours in a 5%H 2 -Ar gas mixture using a tube furnace fitted with stainless steel end caps. Sintered densities were obtained by immersion methods in deionized water and were between 95% to 98% theoretical density. Those conditions are summarized in Table 3-4. Table 3-4. Pellet manufacture conditions Pellet Description Pressure (MPa) Green Density (%TD) Sinter Density (%TD) 23.6% UO 2 76.4% ThO 2 100 63.5 0.2 96.8 0.4 36.8% UO 2 63.2% ThO 2 100 62.6 0.2 96.2 0.2 50.0% UO 2 50.0% ThO 2 100 61.4 0.1 95.3 0.3 100% UO 2 200 49 96.7 Figure 3-10. Pellets prepared using 20% U powder synthesized by the oxalate technique. Powders compacted at 200MPa (left) and 300MPa (right) and sintered at 1600C for 5 hours in Ar-5%H atmosphere Pellets were manually crushed in a zirconium container with an alumina pestle. Fragments were sieved using 60-mesh (250 m) and 170-mesh (90 m) ASTM screens. Generally, about 51% to 61% of the powder was retained on the 170-mesh sieve and about 35% to 42% was collected below 170-mesh. The portions of the crushed pellets below 170-mesh (< 90 m) were analyzed using x-ray diffraction. The XRD patterns indicated that the samples were

PAGE 44

32 complete solid solutions, with no evidence of secondary phases. The powder distribution retained on the 170 mesh and 60 mesh were not measured by XRD to confirm lattice structure or composition. For materials with significantly different hardness, fragmentation may inadvertently separate the hard and soft materials. More brittle materials may fragment into smaller particles whereas the harder materials remain large, altering the powder composition. Brinell and Vickers hardness reported for uranium (2400 MN/m 2 and 1960 MN/m 2 ) and thorium (400 MN/m 2 and 350 MN/m 2 ) by webelements.com indicate that U is the tougher material. It is likely the fine particle distributions may have a lower U/Th ratio than coarser distributions. Again, only the powder retained below 60-mesh/above 170-mesh further characterized. It is unknown whether the fragmentation process may have yielded a high U and low U powder composition. The specific surface area of the crushed pellets was measured using the BrunauerEmmett-Teller (BET) method with krypton as the adsorbate gas. The 90 to 250 m powder samples for surface area analysis were rinsed in demineralized water and dried in a vacuum prior to BET analysis to remove all fine particulates. Measured surface areas for all 90 to 250 m pellet fragments at all compositions were 0.01 to 0.02 m 2 /g. Characterization X-ray Diffraction (XRD) Crystal structure indexing, chemical analysis, and lattice parameter calculations were determined from continuous x-ray diffraction scans on a XRD Philips APD 3720 Diffractometer with a Cu anode target, where K 1 wavelength is 1.54056 Generator voltage and current were 40 kV and 20 mA, respectively. With a step size of 0.02 a range from 10 to 139 was scanned for each sample, with an external Si standard. Lattice parameters were calculated by a graphical and numerical least squares methods and angular separation reported by Popovic [Pop73, Pop85].

PAGE 45

33 Popovic method. Unit cell dimensions were calculated from the angular separation between two Bragg reflections on the basis that shifts occur in the same direction from resultant systematic aberrations. The separation (t = n 2 n 1 ), therefore, includes the difference between systematic aberrations at the two positions, which is generally smaller than the resultant aberration at either position and removes the necessity of absolute measurements. For a cubic structure, lattice parameter a derived from the separation t between two diffraction lines, n 1 and n 2 is given by Equations 3-1 through 3-5. (3-1) t t 2212si n 4cosBBa= where (3-2, 3-3) 2111222221112NNBNNB=+= and (3-4, 3-5) 2,1222=++=ilkhNiiii To avoid temperature variations, the second reflection should be chosen within 10 to 20 from the first reflection [Pop73]. The highest possible diffraction angles are used because of the sine nature of Bragg law, = 2dsinn, where n values near 90 change slowly and, therefore, yield greater accuracy. The sensitivity of this method is reported as (3-6) t n n t si n coscos21f=faa Method of least squares. The lattice parameter, a, of a cubic structure is directly proportional to the d-spacing according to the relationship (3-7) 222lkhda++= where h, k, l are the miller indices defining any particular set of planes. From measurement of the Bragg angle, n, interplanar d-spacing can be determined according to Bragg law, = 2dsinn, and a calculated. Since the term sinn appears in Bragg law, precision in d or a depends on precision in sinn, not the measured n. Because of the nature of sinn, values of sinn change slowly with n near 90. A very accurate sinn value, therefore, can be obtained from a measurement of n

PAGE 46

34 that may not be particularly precise, provided that n is near 90. In other words, the diffracted beam angle is more sensitive to changes in plane spacing when n is large. Obtaining measurements near 2n = 180 are impossible. Since the calculated lattice parameter values approach the true value as 2n nears 180, the true value for a can be obtained by extrapolating a plot of measured values against 2n to 180. The sinn function, however, yields a nonlinear curve, which is difficult to accurately extrapolate. Depending upon the kind of camera employed, n can be inserted into certain functions that present linear curves for extrapolation of lattice parameter. For a diffractometer, that function is either cos 2 n or cos 2 n/sinn depending upon the predominant source or error [Cul78, p.360]. Systematic errors from using a flat, instead of curved, specimen and absorption in the specimen are attributed to the cos 2 n function. Displacement from the diffractometer axis, which is typically the largest single source of error, is attributed to cos 2 n/ sinn. A lattice parameter value is then obtained by linear graphic extrapolation of the a vs. cos 2 n or cos 2 n/ sinn plot such that n = 90. Cohens method. Similar to the least squares method, Cohens method [Cul78, p.363] applies the least squares technique to the observed sin 2 n values directly, instead of to a vs. cos 2 n or cos 2 n/ sinn plots. Squaring, taking the logarithm, and differentiating the Bragg law yield the relationship below. (3-8) ddf=f2si n sin22nn The fd/d term is replaced with an extrapolation function (-D/R)(cos 2 n/sinn), where D is specimen displacement and R diffractometer radius, to account for diffractometer error [Cul78, p.359]. The difference, fsin 2 n, between the true, as defined by unit cell geometry, and observed sin 2 n values is set equal to the above equation. Simultaneously solving Equation 3-9 for the set of Bragg lines yields the true lattice parameter, a o at some constant wavelength For ease in calculation, the terms are separated into Equations 3-10 through 3-13.

PAGE 47

35 (3-9) nnnnsincos2sin)(4)(sin22222222RDlkhaobso=f=++ (3-10) 224oaC= (3-11) 222lkh++= (3-12) 1012RDA= (3-13) nntsincos102= The experimental values of sin 2 n(obs), and t for each back-reflection inserted into the appropriate terms leave only C and A to be solved. Employing the numeric method of least squares detailed by Cullity [Cul78, p.364], normal Equations 3-14 and 3-15 for cubic systems are simultaneously solved to obtain C and A. (3-14, 3-15) +=+=2222sinsinttnttnACAC Elemental Analysis Inductively coupled plasma Auger electron spectrometry (ICP-AES). Measurements obtained from ICP-AES were not performed by the author, but by James Jerden, Ph.D. at Argonne National Laboratory. The general technique and results obtained from Dr. Jerden are reported because of their importance in determining the synthesis process of samples to undergo thermogravimetric analysis. ICP-AES measures the mass spectrum of a sample, typically, from lithium (Z = 3) to uranium (Z = 92), yielding semi-quantitative, quantitative, elemental and isotopic information for each element. It is a destructive technique where the aqueous sample is nebulized into an aerosol and swept into the ionizing plasma, a high-temperature, atmospheric pressure, and partially ionized gas. The ions generated are carried into a detector for analysis by Auger electron spectrometry. Detection limits are typically in the sub-ppb range for most elements.

PAGE 48

36 Carbon analysis. The LECO WC-200 Carbon Analyzer is a unit specifically designed for tungsten carbide application that provides high precision measurement of carbon content in metals, ceramics, and other inorganic materials. A high frequency induction furnace combusts samples, with accelerators (~ 1 gram each of iron and copper chips) in a quartz crucible, in a pure oxygen environment. The carbon-bearing elements are reduced, releasing the carbon, which immediately binds with the oxygen to form CO and CO 2 These gases are then carried into a molecular trap and then released into infrared (IR) cells. As CO 2 passes through the IR cell, it absorbs IR energy at a precise wavelength within the IR spectrum, preventing it from reaching the IR detector. After passing all IR energy through a narrow bandpass filter to ensure the signal can only be attributed to CO 2 the concentration is detected as a reduction in the energy level at the detector. A five-place balance and burned-off crucibles and lids were handled only with clean tongs to minimize additive error contributions. Particle Morphology An exhaustive classification of particle size demands complex sampling and numerous optical micrographs to yield a statistically significant portrayal of size and distribution. Light optic techniques were used to present a qualitative comparison of particle size and morphology processed via co-precipitation, blending, and commercial methods shown in Figure 3-11. The spherical particles of oxalate synthesized and commercial UO 2 aided pellet making in terms of flowing material into the die and during pressing. Particles did not stick to die walls, unlike the blended (U,Th)O 2 and commercial ThO 2 powders. Since wet synthesized (U,Th)O 2 powders were abandoned for dry oxidation study, this sticking behavior was addressed by using a lubricant, stearic acid, to aid pellet release from the die. These were not necessary to form successful pellets with the oxalate synthesized powders.

PAGE 49

37 a) 20 U S S 1100 b ) 20 % U B l e nde d c) UO2 Commercial (10x) d) ThO2 Commercial Figure 3-11. Optical micrographs of nominal U 0.2 Th 0.8 O 2+x made by oxalate synthesis (a), blended (b), and commercial UO 2 (c) and ThO 2 (d) powders Pellet Density Green pellet densities were measured with a Mettler AB-104S balance and calipers. Volumes were calculated under the assumption that pellets formed perfect cylinders. Densities of sintered pellets, however, were measured by immersion in water. First, a dry weight (A) was taken. Then, the pellet is immersed in water and the difference between the weight wet and immersed pellet is the buoyancy (P). The density is calculated based on Equation 3-16. (3-16) waterPArrb=

PAGE 50

38 effect upon oxide decomposition. 050 thermogravimetric analyzer. chosen based on the results funs Dry Oxidation Thermogravimetric Analyzer Nonisothermal and isothermal dry oxidation experiments were conducted on the crushed pellets (90 to 250 m powder) in a TA Instruments TGA 2050 thermal gravimetric analyzer, as shown in Figure 3-12, with air flowing at 90 cm 3 /min through the sample chamber. Approximately 8 to 20 mg of the 90 to 250 m pellet fragments were heated to 900C at rates of 1, 3, and 5C/min while monitoring the weight changes. The 90 to 250 m fragments were used in order to provide a narrow particle size distribution. Restricting samples to the same size distribution was a means to maintain relatively similar surface to volume ratios, normalizing its Figure 3-12. TA Instruments TGA 2 The appropriate temperatures for isothermal experiments were rom the nonisothermal gravitational analysis for each composition. The isothermal rwere performed using the same gas flow used for the nonisothermal experiments. The length of the isothermal experiments varied for each composition and temperature, as measurements were

PAGE 51

39 e re also characterized with x-ray diffraction after oxidation. All (U,Th)O2+x samples se ed data from nonisothermal and isothermal dry oxidation of mixed oxides ere su n s e he model-free technique, on the other hand, bypasses assumptions of a specific reaction odel. terminated after the sample exhibited no further weight change. The ramp-up time (to) to reach the designated temperature was between 8 and 15 minutes for all experiments. This period is neglected in the presentation of figures and all calculations. The measured weight gain data ar converted to fraction reacted, versus time plots for purposes of calculating kinetic parameters. Isotherms for (U0.368Th0.632)O2+x were repeated in triplicate to determine the statistical significance of measured data Samples we were found to be the cubic fluorite structure, demonstrating that there were no bulk ph a changes during oxidation for any of the solid solution compositions. Kinetic Analysis The measur wbjected to standard model-fitting techniques and a model-free method reported by Vyazovkin and Wight [Vya99]. Standard methods estimate kinetic parameters and reactio mechanisms based on theoretical reaction models. Arrhenius parameters from isothermal and nonisothermal data, however, often disagree because of the differing nature of a constant versu variable heating profile. Calculations derived from isothermal data, therefore, are traditionally considered more reliable because the temperature variable is held constant, reducing the number of parameters simultaneously determined by fitting the data using a particular reaction model. Consequently, the macroscopic nature of thermoanalytical techniques does not always elucidate system complexity, such as overlapping mechanism. Additionally, measured data may fit more than one hypothetical reaction model, demonstrating the danger of force-fitting data to inadequat models. T mActivation energies instead are determined as a function of the degree of conversion ( ) and/or temperature. Instead of yielding a global composite activation energy for the system, the

PAGE 52

40 of isothermal data entails a ompard based ecause d on general kinetic Equation 3-17 and revious activation energy dependence on is capable of revealing process complexities. Vyazovkin and Wight [Vya99] reported the model-free method capable of producing consistent kinetic information from both isothermal and nonisothermal data Model-fitting method. Traditional kinetic analysis cison of measured values with theoretical reaction models. These models are deriveon the geometry of interface initiation and advance and/or diffusion processes occurring in the solid. Natural limitations are inherent in translating experimental data into pertinent kinetic parameters based on the models. It is often the case that factors important to yielding an important conclusion regarding kinetic processes may not be experimentally accessible. Bof restrictions experienced in quantitatively characterizing particle morphology in this study and the assumption that pellet fragments were not spherical, developing a particle size dependent termwas hindered. It was anticipated that restricting the particle size and, consequently, surface area to a common range in this study would normalize the surface area influence and allow for comparisons between U/Th compositions in this study. Solid-state decomposition kinetic analysis is base ply mentioned Arrhenius Equation 2-2, where is the fraction reacted, f() is the reactionmodel, k(T) is the rate coefficient (also denoted k), t is time, A is the frequency factor, E is the activation energy, R is the gas constant, and T is the temperature. (3-17) tTkfb=)()( =RTEATkexp)( (2-2) Theoretical relationships, f(), that have found the greatest application in solid state inetic as. he knalysis are summarized in Table 3-5 [Bam80]. If experimental data is plotted as f() vt, the linearity of the plot is an indication of how well the data agrees with a particular reaction model. Care must be taken in accurately defining the final yield, = 1.0, to prevent distorting t

PAGE 53

41 3-5. Solid state theoretical reaction models Reaction model f() Reaction model f() vs. t curve. Rate coefficients, subsequently, are obtained from the linear slope of the f() vs. t plot. Table Acceleratory rate equation Decelera tory rate equations Power law 1/n (Based on diffusi on mechanisms) () uations -)ln(1-) + )]1/2 (1-)1/3]2 ln(1-)]1/3 )2/3 (1-) respect to ) -) 2 Exponential law ln One-dimensional diffusion Sigmoid rate eq Two-dimensional diffusion (1 [-ln(1Three-dimensional diffusion [1 Avrami-Erofeev [Ginstling-Brounshtein [1 (2/3)] (1[-ln(1-)] 1/4 (Based on geometric models) Prout-Tompkins ln[/(1)] Contracting area 1 (1-)1/2 1/3 Contracting volume 1 (Based on order with First order -ln(1-) -1 Second order (1 Third order (1-)-2 The equation below is obtained by taking the natural logarithm of the Arrhenius relationive a (3-18) ship [Ban68]. From Equation 3-18, it is clear that a plot of ln[k(T)] vs. 1/T will gstraight line with slope equal to E/R and an intercept equal to lnA. E1 ATRTkln)(ln+= Model-free method. The model-free method was applied solely to isothermal rate data this s (3-19) intudy. Under isothermal conditions, the reaction model is assumed to be independent of the heating rate. By rearranging Equation 3-17, substituting into Equation 2-2, and taking the natural logarithm, Equation 3-19 is obtained, where t,i is the reaction time for a selected fractionreacted, for a given isotherm, i. +=)(lnln,fARTEti A ln(t) vs. 1/T plot at a particular value is constructed from the isothermal oxidation ata obt dained at several temperatures. The linear slope of the ln(t) vs. 1/T plot yields the value E/R, at the fraction reacted, without making any assumptions about the reaction model, f().

PAGE 54

42 Because the intercept term in Equation 3-19 is a function of both A and f( ), the frequency factor (A) cannot be determined by this method without identifying f( ). The benefits of this method, however, are such that reaction complexities may be revealed by a dependence of E on instead of simply yielding an overall value for the reaction. Consequently, there is a heavy reliance upon meaningful definition of final yield where = 1.0.

PAGE 55

43 CHAPTER 4 RESULTS AND DISCUSSION Material Synthesis X-ray diffraction. Bulk homogeneity of the mixed oxides was confirmed by XRD to be cubic fluorite in structure at all U/Th compositions before and after air oxidation. A Philips APD 3720 Diffractometer measured pellet fragments with silicon as a standard. Measured diffraction powder patterns were compared with JCPDS standards to qualitatively confirm Bragg angles (2 n ) and identify corresponding miller indices. See Appendix B for details of the diffraction pattern measured for U 0.368 Th 0.632 O 2+x pellet fragments isothermally oxidized at 400 C in flowing air. All U/Th compositions qualitatively presented the same cubic fluorite diffraction pattern. Three methods of calculating unit cell parameters were used: method of least squares (LSQ), Cohens method [Cul78], and Popovics method [Pop73, Pop85]. Starting compositions of all unoxidized material batches are assumed stoichiometric. Unit cell volume or lattice parameter decreases with respect to increasing percent UO 2 which is consistent with published findings as presented in Figure 4-1. Table 4-1 summarizes the lattice parameters as calculated by the three methods for isothermally oxidized (U,Th)O 2 pellet fragments. As more of the smaller uranium cation is introduced into the thorium oxide matrix, it is apparent that the lattice shrinks to accommodate without changing the cubic fluorite structure. Table 4-1. Calculated (U,Th)O 2 oxidized lattice parameters by three methods Description LSQ (cos 2 n ) LSQ (cos 2 n /sin n ) Cohens Popovic U 0.236 Th 0.764 O 2 5.568 0.003 5.564 0.002 5.563 0.001 5.564 0.005 U 0.368 Th 0.632 O 2 5.542 0.002 5.543 0.001 5.543 0.001 5.541 0.003 U 0.500 Th 0.500 O 2 5.523 0.004 5.522 0.003 5.523 0.002 5.525 0.002 LSQ: least squares method uses two functions for diffractometer measured patterns.

PAGE 56

44 Figure 4-1. Lattice parameter evolution with respect to UO 2 content. Dashed lines indicate the lattice parameters for reference oxides ThO 2 and UO 2 Linear trendline show that results from this study are in agreement with Anderson et al. (1954) Inductively coupled plasma Auger electron spectrometer. Samples of (U,Th)O 2 synthesized by the oxalate and ammonium hydroxide pathways were measured by James Jerden, Ph.D. at Argonne National Laboratory to obtain weight percent uranium and thorium. Assuming whatever remained was oxygen, those results, as summarized in Table 4-2 as weight percent, it was apparent something other than oxygen was also present. As a consequence, carbon analysis was pursued to determine whether the excess could be attributed to residual carbon from the synthesis process. 5.575.545.525.465.485.55.525.545.565.585.65.6215%25%35%45%55%65%75%85%95%% UO2Lattice Parameter (A) Anderson (1954) this study UO2ThO2 Carbon analyzer. A tungsten carbide standard (0.26 0.02 g) was measured five times with a standard deviation of 0.03% by weight carbon. An empty crucible established a blank measured carbon concentration of 0.00385% by weight. Absolute and relative uncertainty were calculated to be {(0.03) 2 + (0.03) 2 } = 0.04 and (0.04/measurement 100%), respectively [Har91]. Oxalate and ADU synthesized (U,Th)O 2 samples with nominal percent U content of 20%, 35%, and 50% measured results are summarized in Table 4-2 as weight percent. Since it was supposed the excess identified by ICP-MS was residual carbon from the oxalate process, the ADU synthesized nominal 20% UO 2 sample was not measured by this instrument.

PAGE 57

45 From the above results, it is apparent the elemental analysis techniques used were unable to fully identify the chemical composition of wet synthesized (U,Th)O 2 materials. Ando et al. [And76, And84] also noted the presence of impurities for ThO 2 synthesis by sol-gel and arc fusion techniques. Without a full chemical identification, the possible influence of unknown impurities could possibly inhibit evaluation of the thorium oxide matrix effect upon uranium oxidation. As a consequence, wet chemistry synthesis techniques were abandoned in favor of traditional powder process methods to avoid introduction of impurities associated with the synthesis path. Table 4-2. ICP-AES and LECO carbon analysis results and calculated metal valence of wet synthesized (U y Th 1-y )O 2+x indicate impurities must be present. Sample (Nominal composition) U wt % Th wt % C wt % Balance* y x V M ADU (U 0.20 Th 0.80 )O 2+x 15.8 66.6 -17.6 0.192 1.11 6.22 Oxalate (U 0.20 Th 0.80 )O 2+x 11.6 70.9 0.01 17.5 0.141 1.09 6.18 Oxalate (U 0.35 Th 0.65 )O 2+x 21.3 54.1 0.17 24.4 0.283 2.72 9.44 Oxalate (U 0.50 Th 0.50 )O 2+x 32.7 38.0 0.19 29.1 0.463 4.03 12.06 Assumes balance is only oxygen. Metal valences (V M ) calculated from x are impossible; knowing that V Th maximum is +4 and V U is 6. Thermogravimetry Nonisothermal. Thermal gravimetric analysis curves for air oxidation of UO 2 and (U,Th)O 2 samples under nonisothermal conditions are shown in Figures 4-2 through 4-5 as percent weight gain vs. temperature. All of the data exhibited single-step behavior, including the data for UO 2 (Figure 4-5). This is consistent with observations of air oxidation of (U y Th 1-y )O 2 materials (y = 0.15, 0.30, 0.72, 0.77) performed by Anthonysamy et al. [Ant00] While gravimetric oxidation curves for UO 2 often exhibit a two-step behavior corresponding to oxidation from UO 2 to U 3 O 7 to U 3 O 8 the absence of a well defined two-step curve for UO 2 in this study is most likely the result of the low specific surface area of the samples (0.01 to 0.02 m 2 /g). With the lower surface area of coarser particle size distributions, the fast surface transformation of UO 2 to U 3 O 7 contributes less to the overall weight gain because the reaction interface decreases. The first step, therefore, diminishes with decreasing surface area. The slower bulk

PAGE 58

46 transformations of UO 2 to U 3 O 8 and U 3 O 7 to U 3 O 8 however, are less dependent upon surface area. The second step remains despite particle size distribution. Thermogravimetric measurements of higher surface area UO 2 powders (1 to 3 m 2 /g) resulted in the expected two-step weight gain curve. Thus, the low surface to volume ratio of the 90 to 250 mm pellet fragments impacted the results such that bulk reactions had a much large effect on the overall oxidation process than the contributions from surface reactions. Figures 4-2 to 4-4 show the mixed (U,Th)O 2 oxides and Figure 4-5 UO 2 fragments under nonisothermal air oxidation at heating rates of 1, 3, 5 C/min as indicated. A weight decrease of (U,Th)O 2 samples at higher temperatures (Figures 4-2 to 4-4) was observed. The UO 2 isothermal sample, Figure 4-5, however, did not present this same behavior. The mixed (U,Th)O 2 oxide weight loss at the higher temperature region was constant and sustained for > 20 minutes. Since the TGA was located on an isolation table and weight loss was not sudden, it was not supposed that fragments were somehow ejected from the pan from an external source. To eliminate the presence of an unknown gas, the pellet fragments, furnace chamber, and gas lines were vacuum purged and backfilled with high purity Argon gas three times at room temperature. A final vacuum purge was backfilled with air prior to beginning each run. These same measures were taken with the nonisothermal oxidation of UO 2 pellet fragments. Those results are presented in Figures 4-2 to 4-5. For the UO 2 sample, the absence of a steady weight decrease following initial weight gain further suggests an external source is unlikely. Further, without measurements beyond 900 C, it is unknown whether the mixed oxide weight would eventually reach a steady state. Knowing the lattice structure evolution from start to finish may confirm what mechanisms influence oxidation. Though not accomplished in this study, XRD measurements of (U,Th)O 2 fragments at various stages of nonisothermal oxidation, particularly at the peak maximum and T = 900 C, may suggest a possible explanation. Possibly, the thoria lattice

PAGE 59

47 initially accommodates excess anions in the large octahedral interstitial sites, with charge neutrality maintained by increasing U valency. Successively more anions, noting peak weight gains of U 0.236 Th 0.464 O 2 (0.76% 0.77%), U 0.368 Th 0.632 O 2 (1.16 1.26%), and U 0.500 Th 0.500 O 2 (1.59 1.65%), are accommodated with increasing U content. Eventually, the lattice may become supersaturated, holding the maximum number of anions with U at its highest valency, yielding a peak maximum. Perhaps, with increasing thermal energy, the structural and energetic strain of maintaining this supersaturated lattice causes the steady ejection of excess anions until a stable, unstrained configuration is achieved. This may account for the decreasing weight. Nonisothermal measurements of finer particle size distributions may also yield surface area dependent or other mechanisms that would clarify the peak maximum and subsequent decreasing weight in the (U,Th)O 2 systems. Like UO 2 there may be fast and slow competing mechanisms. A fast mechanism may dominate the weight gain at lower temperatures. At the higher temperatures, this fast mechanism may have reached completion and the slow mechanism becomes the rate-limiting step. -0.1%0.0%0.1%0.2%0.3%0.4%0.5%0.6%0.7%0.8%1002003004005006007008009001000Temperature (Celsius)Weight gain (%)1 C/min5 C/min3 C/minU0.236Th0.764O2 Figure 4-2. Nonisothermal (U 0.236 Th 0.764 )O 2 oxidation TGA data at heating rates of 1, 3, and 5C/min.

PAGE 60

48 -0.2%0.0%0.2%0.4%0.6%0.8%1.0%1.2%1.4%100200300400500600700800900Temperature (Celsius)Weight gain (%)1 C/min5 C/min3 C/min`U0.368Th0.632O2 Figure 4-3. Nonisothermal (U 0.368 Th 0.632 )O 2 oxidation TGA data at heating rates of 1, 3, and 5C/min. -0.2%0.0%0.2%0.4%0.6%0.8%1.0%1.2%1.4%1.6%1.8%100200300400500600700800900Temperature (Celsius)Weight gain (%)1 C/min5 C/minU0.50Th0.50O2 Figure 4-4. Nonisothermal (U 0.500 Th 0.500 )O 2 oxidation TGA data at heating rates of 1 and 5C/min.

PAGE 61

49 0.0%0.5%1.0%1.5%2.0%2.5%3.0%3.5%4.0%4.5%300350400450500Temperature (Celsius)Weight gain (%)UO23 C/min Figure 4-5. Nonisothermal UO 2 oxidation TGA data at heating rate of 3C/min. Isothermal. The appropriate temperatures for isothermal gravitational analysis experiments were chosen based on the nonisothermal results in Figures 4-2 to 4-5. Generally, temperatures were chosen to encompass the region of the curves where the majority of the weight gain occurred. The results from isothermal gravitational analysis at the selected temperatures is presented as fraction reacted, versus time, t t o (Figures 4-6 to 4-9) where t o is the time required to reach the desired isothermal temperature (between 8 minutes). Some oxidative weight gain did occur during this ramp-up period. Because the time required to achieve complete reaction for the (U,Th)O 2 samples was large (usually more than 1000 minutes), the weight gain during the same ramp-up period was negligible compared to the overall weight gain. Isothermal oxidation of the UO 2 samples, however, was completed in 20 to 120 minutes, depending on the temperature. Contributions to overall weight gain during the ramp period for UO 2 therefore, were often significant and resulted in errors in the calculation of the true degree of completion, This, in turn, affected the vs. t t o curve shapes in Figures 4-6 to 4-9. Consequently, the curve distortion hindered reaction model identification by model-fitting methods as well as accurate measurement by the -dependent model-free method.

PAGE 62

50 At successively higher isothermal temperatures, (U,Th)O 2 rate of oxidation was expected to increase, such that the slope systematically became steeper before reaching steady state. This did not occur in U 0.5 Th 0.5 O 2 (Figure 4-8) and UO 2 (Figure 4-9). Additionally, the instrument was jarred during the 375C run of UO 2 which accounts for the discontinuity at 60 minutes in Figure 4-9. With the exception of U 0.368 Th 0.632 O 2 (Figure 4-7), which was measured in triplicate, only single runs at all isotherms were performed. Slope irregularities, therefore, may be a consequence of experimental scatter. 00.20.40.60.811.20500100015002000250030003500t-to, minfraction reacted, U0.236Th0.764O2550C450C500C Figure 4-6. Isothermal oxidation TGA data for (U 0.236 Th 0.764 )O 2 fragments (90 250 m). Percent weight gain data is converted to fraction reacted, 00.20.40.60.811.202505007501000125015001750t-to, minfraction reacted, 475C450C500C`U0.368Th0.632O2 Figure 4-7. Isothermal oxidation TGA data for (U 0.368 Th 0.632 )O 2 fragments (90 250 m). Percent weight gain data is converted to fraction reacted,

PAGE 63

51 igure 4-8. Isothermal oxidation TGA data for (U0.500Th0.500)O2 fragments (90 250 m). Percent weight gain data is converted to fraction reacted, Figure 4-9. Isotrcent weight gain 00.20.40.60.811.2020040060080010001200t-to, minfraction reacted, U0.50Th0.50O2425C375C400C450C F hermal oxidation TGA data for UO2 fragments (90 250 m). Pe data converted to fraction reacted, 0.00.20.40.60.81.002040608010012t-to, minfraction reacted, 425C375C400C450CUO2 1.20

PAGE 64

52 Cation valence. Based on the final weight gain of each sample after isothermal air xidatioo (4-1) on, the mean uranium valence was calculated. For the calculation, thorium is assumed tmaintain a valence of +4. The weight gain is entirely attributed to excess oxygen, which is used to determine the valence of uranium cations. The initial O/U ratio of the samples was assumed tobe 2. The mean uranium valence (VU) was calculated from the weight gain data using Equations 4-1 and 4-2, where the variable x and y are taken from the formula (UyTh1-y)O2+x, m1 is the initial mass of the sample, fm is the weight change after oxidation, and the v alu e in p a ren t heses in Equation 4-2 is the formula weight of the starting material, ((UyTh1-y)O2+x). x2 yVU4+= +f=16)99.504.264(1ymmx (4-2) Results are plotted in Figure 4-10. Dashed lines indicate the mean uranium valence for the refeons. tain rt the view that ThO2 inhibits uranium from oxidizing further by inhibitin09 to rence oxides U3O7 and U3O8. As indicated in the figure, none of the mixed oxides reached the same degr e e o f oxidation as UO2, based on the mean uranium valence calculatiKnowing the oxidized lattices are still cubic fluorite and uranium valences are increasing, it stands to reason that the lattices are accommodating sufficient excess oxygen anions to maincharge neutrality with the U cations. So, instead of a phase transformation, anions may be packedinto interstitial sites near uranium cations, since thorium valency is a constant +4. Assuming a homogeneous distribution of U and Th cations, oxygen accommodation may be a predominantlybulk mechanism. Without confirmation of the surface composition, a surface-dependent reaction cannot be rejected. This would suppo g the supply of oxygen with the decreased anion mobility in the thorium matrix. Published activation energies (Appendix A) of intrinsic oxygen anion diffusion in ThO2 (2275 kJ/mol) and UO2 (237 to 273 kJ/mol) indicate that stoichiometric UO2 values are c o mparable

PAGE 65

53 ic lower the hO2 n microscope images taken of pellet surfaces by J. Jerden at ANL (Appen rom ments sses contributions. to those for ThO2. Uranium dioxide, however, is typically hyper-stoichiometric because of the ease in which it in corporates excess oxygen into the lattice structure. Activation energies reported for UO2+x are 87 to 124 kJ/mol, which is significantly lower than both stoichiometr ThO2 and UO2, indicating the greater ease in which O diffuses into the lattice. The broad range and variety of sample preparation and measurement conditions may account for the breadth of published values reported. It stands to reason that activation energies calculated for mixed (U,Th)O2 oxides would fall between those reported for stoichiometric ThO2 and hypersto i chiometric UO2+x. The presence of U cations in the ThO2 is expected to threshold for oxygen anion mobility and accommodation in the lat t ice. It is anticipated the T matrix, on the other hand, will limit the anion supply such that cations are unable to attain U maximum valence state. Secondary electro dix C) confirmed the pellets are polycrystalline. Jerden reported grain sizes ranging f 4 to 20 m on unbroken surfaces of 5% UO2 and 20% UO2 co-milled pellets. An image of the broken surface of a 20% UO2 co-milled pell e t shows curio u s gas bubble features. No images were obtained of pellet fragments. Despite no images of the pellet fragments, it is obvious that grain boundaries are present and would have an extrinsic effect upon oxygen mobility. In fact, anions may assemble more easily in the grain boundaries than interstitial sites. Anion migration along grain boundaries, consequently, may yield lower oxidation activation energies than for single crystal systems. Undoubtedly, activation energies calculated for this polycrystalline system would encompass intrinsic and extrinsic mechanisms. Without quantitative measure of pellet fragment grain sizes and boundaries, an extrinsic term cannot be developed for the model-fitting kinetic analysis. The gas bubble features, also, indicate other extrinsic proce may also be at work. Consequently, the models used in this research are limited without more complete characterization of the pellet fragments to account for the intrinsic and extrinsic

PAGE 66

54 Figure 4-10. Me2. Dashed li8, U3O7, and UO sothermal rate data were inserted into the reaction models, f(), from Table 3-5. Linear behavior of f()-time plots would sugget with the theoretical model. lthougence on n, such l ndent re reliable for calculating kinetic an uranium valence for isothermally oxidized (UyTh1-y)O2 and UOnes indicate the mean uranium valence for the reference oxides U3O2. 44.24.44.64.855.2300350400450500550600Temperature (Celsius)U valence y = 0.236 y = 0.368 y = 0.500 y = 1.0 UO2U3O7 5.45.6 U3O8 Kinetic Analysis I st overall agreemen Ah model fitting is the conventional technique for kinetic analysis, the published theoretical reaction models do not necessarily account for all variables that may have influthe proposed mechanisms. For instance, systems that exhibit a surface-dependent reactioas the UO2 U3O7 fast transformation, require an additional term to allow for particle size and shape variations. It has already been noted that surface area has an impact on the measured signafor UO2 o x idatio n. The (U,Th)O2 mixed oxide compositions may likely demonstrate a similar consequence. The reaction models used in this study, however, do not take into account particle size distribution (i.e., surface area). Considering the absence of particle morphology data to develop a particle geometry and/or surface area term, conclusions drawn from the reaction models in agreement with the measured data are restricted. Since the temperature is held constant for the isothermal runs, time is the only depevariable. Isothermal data, therefore, are typically assumed mo

PAGE 67

55 paramet s of conversion, measured isothermal rate data of to reaction models, f( ), summarized in Table 3-5. The twoimensi nd f( ) = a atures greater than 500 C (Figures 4-12 through 413). Ta ed to rate coefficient (k) by the lation othermal (E = 62.1 kJ/mol) and ers than nonisothermal data, where time and temperature vary simultaneously. In the subsequent figures, linear agreement to 2D diffusion, 3D diffusion, and/or Avrami-Erofeev reaction models is shown for isothermal data. Rate coefficients (k) are obtained from the slope of the best-fit linear trendlines. Mixed Oxide (U0.236Th0.764O2) Transformed into degree (U0.236Th0.764)O2 wer e in sert ed in donal and three-dimensional diffusion models, where f( ) = (1)ln(1) + a [1 (1) 1/3] 2, r e spectively, displayed agreement. Reaction models were plotted against t-to, which truncates the initial to ramp-up stage. Model-fitting suggests two dimensional diffusion 450 C (Figure 4-11), followed b y shift to three dimensional d i ffusion at temper ble 4-3 summarizes the rate coefficients and linear fit of both 2D and 3D diffusion reaction models, which were used to construct an Arrhenius plot (Figure 4-27) with E and ln(A) calculated to be 62.1 17.9 kJ/mol and 1.71 2.87, respectively. Linear agreement of nonisothermal rate data was also tested with the diffusion reaction models identified by isothermal data. Measured data were convert re k = f( )/t, according to Equations 2-2, 3-17, and 3-18. Arrhenius log(k) vs 1/T plots were constructed with f( ) for 2D and 3D diffusion models shown in Figures 4-15 to 4-17 for heating rates 1, 3, and 5 C/min to 900 C. Both reaction models displayed linear behavior with high correlation values in the region of interest. It was not discernable which diffusion model would be appropriate. Like the isothermal data, slopes and intercepts from linear fit yield corresponding E and A values shown in Table 4-4. The model-free technique was used to compare with values calculated by model-fitting particularly since there is poor agreement between is

PAGE 68

56 onisoth (= 0.4, e and 94.1 7.7 kJ/mol for fraction reacted 0.4, 0.5, and 0.6 respectively. These values re mor l (U0.236Th0.764)O2 agreement to diffusion models 2D Diffusion 3D Diffusion nermal (E = 92 to 94 kJ/mol) results. The model-free method was applied solely to the isothermal rate data in this study, as shown in Figure 4-18. Reaction times at selected 0.5, 0.6) were arranged into a ln(t) vs. 1/T plot based on the relationship described in Equation 3-19. Frequency factors cannot be determined by this method without identifying f( ). Since neither the 2D or 3D diffusion reaction models could be discerned by isothermal and nonisothermal model fitting, frequency factor calculations were not performed by the model-fre method. Activation energies calculated from the slope yield values of 79.2 6.9 kJ/mol, 87.6 7.6 kJ/mol ae in agreement with those obtained by nonisothermal model-fitting, supporting diffusion as the primary mechanism of oxidation. Table 4-3. Rate coefficients of isotherma Isothermal temperature k (min-1) R2 k (min-1) R2 450 C 3 10-4 0.9969 2.0 10-4 0.9330 500 C 3 10-4 0.8823 2.9 10-4 0.987 1 50 C 8 10 7.1 10 9807 5 -4 0.8365 -4 0. T4. Kinetic results for noniso (U0.236T agreemfusion m2ion iffusion able 4therm a l h0.764)O2 ent to d i f odels D Diffus 3D D Nonisothermal ln(A) (min-1) E (kJ/mol) R2 ln(A) (min-1) E (kJ/mol) R2 8 1 C/min 4.63 81.2 0. 9888 5.48 94.4 0.997 3 C/min 4.36 0.9946 5.27 0.9982 0. 0.1 77.7 92.2 5 C/min 4.83 81.6 9953 5.06 92.2 997

PAGE 69

57 Figure 4-11. Isfusion models. Li Figure 4-12. Isfusion models. Li otherm at 450C of (U0.236Th0.764)O2 oxidation fit to 2D and 3D Difnear correlation values (R2) are 0.9969 and 0.933, respectively 2D Diffusion3D Diffusion00.20.40.60.805001000150020002500300035004000t-tof() 450C11.2 otherm at 500C of (U0.236Th0.764)O2 oxidation fit to 2D and 3D Difnear correlation values (R2) are 0.8823 and 0.9871, respectively. 3D Diffusion00.20.40.60.8050010001500200025003000t-tof() 500C11.22D Diffusion

PAGE 70

58 550C0.811.22D Diffusion Figure 4-13. Isfusion models. Li Figure 4-14. Arrhenius plot of (U0.236Th0.764)O2 isotherms fit to 3D Diffusion models. Rate constants (k) were determined by fitting gravimetric oxidation data with the 3D diffusion reaction model otherm at 550C of (U0.236Th0.764)O2 oxidation fit to 2D and 3D Difnear correlation values (R2) are 0.8365 and 0.9807, respectively. 3D Diffusion00.20.40.6020040060080010001200t-tof() -9-8.6-8.2-7.8-7.41.21.251.31.351.41/T x 103 (K-1)ln(k)U0.236Th0.764O2 -7

PAGE 71

59 igure 4-15. Nonisotherm at 1C/min (U0.236Th0.764)O2 Arrhenius plot fit to 2D and 3D diffusion models. Linear trendlines shown. Figure 4-16. Nod 3D diffusion m 1.E-051.E-041.E-031.E-02log k2D Diffusion3D Diffusion1 C/min to 900C 1.E-060.801.001.201.401.601.80-3-1 1/T x 10 (K) F nisotherm at 3C/min (U0.236Th0.764)O2 Arrhenius plot fit to 2D anodels. Linear trendlines shown. 1.0E-061.0E-051.0E-041.0E-030.800.901.001.101.201.301.401.501.601/T x 10-3 (K-1)log k3D Diffusion3 C/min to 900C 1.0E-022D Diffusion

PAGE 72

60 igure 4-17. Nonisotherm at 5C/min (U0.236Th0.764)O2 Arrhenius plot fit to 2D and 3D diffusion models. Linear trendlines shown. Figure 4-18. Moope of the li 1.0E-051.0E-041.0E-031.0E-02log k2D Diffusion3D Diffusion 1.0E-060.800.901.001.105 C/min to 900C 1.201.301.401.501.601/T x 10-3 (K-1) F del free (U0.236Th0.764)O2 isotherms plotted at = 0.4, 0.5, 0.6. Slnear trendlines gives E/R. -6.5-6-5.5-5-4.5-41.21.251.31.351/T x103 (K-1)-ln(t) U0.236Th0.764O2 -3.5-31.4 = 0.6 = 0.4 = 0.5

PAGE 73

61 ixed Oxide (U0.368Th0.632O2) The same protocol as described in the previous section was applied to (U0.368Th0.632)O2, which was isothermally measured in triplicate at 450 C, 475 C, and 500 C. Unlike the other (U,Th)O2 mixed oxides and UO2, (U0.368Th0.632)O2 isothermal runs were done in triplicate to establish statistical significance and error analysis for the calculated kinetic parameters. Isothermal rate data were inserted into reaction models from Table 3-5. Those that showed nonlinear behavior were rejected, leaving only the 3D diffusion model. Figures 4-19 to 4-21 show the three sets of isothermal rate data per temperature fit to the 3D diffusion reaction model and the associated linear trendlines. Rate coefficients (k) obtained from the slopes of the f( ) vs. l v ctor are calculated. One 475 C isothermal run (labeled iso475 in Figure 4-20), however, was iscarded because, after reaching a maximum weight gain, the sample began to inexplicable lose eight. Although this run is included in Figure 4-20, it is not incorporated into any kinetic scatter was observed in the measured signal at increasingly long well ti 1/lot ion e uld M time plots are compiled into an(k)s. 1/T Arrhenius curve. Activation energy and frequency fa d w analysis methods A ddi tion a lly dmes (i.e., greater than 1000 min). The rate coefficients obtained from the linear slope are detailed in Table 4-5 and assembled into an Arrhenius ln(k) vs.T p (Figure 4-22) to determine the activation energy and frequency factor for isothermal data. The slope of the linear trendline yielded activat energy and frequency factor of 171 8 kJ/mol and 20.8 1.3, respectively. These values ar significantly greater than those determined for isothermal model-fit (U0.236Th0.764)O2. This wo seem to indicate that increasing U content results in an increased threshold to oxidation. The higher frequency factor also suggests there are more events occurring in the higher U content mixed oxide. This, however, is contradictory to the premise that increasingly uranium-rich oxides more readily oxidize than thorium-rich compositions. That (U0.236Th0.764)O2 isothermal data agreed with 2D and 3D diffusion models, whereas only 3D diffusion was identified for

PAGE 74

62 ) and ion model, the plots generated in this study cannot erify w y n l n. t able ee sults. 17 (U0.368Th0.632)O2, may allude to a possible explanation. If diffusion occurs at the surface (2D bulk (3D), it is possible that one takes place more readily than the other. Without a particle and/or grain size term included in the react vhether activation energies calculated reflect whether either system is dominated b surface or bulk reactions or by temperature. The sensitivity to changes in solid-gas and/or graingrain surfaces is also unknown without an interface-dependent term in the reaction model. Linear agreement of nonisothermal rate data to the 3D diffusion reaction model is show in Figure 4-23. Activation energy and frequency factor were calculate d fr om the sl o pes and intercepts of linear trendlines. They are summarized in Table 4-6. As with (U0.236Th0.764)O2, nonisothermal results are significantly different from those for isothermal runs. Nonisotherma and isothermal data appear to support 3D diffusion as the most likely mechanism of oxidatio Values obtained are similar to those for (U0.236Th0.764)O2. Again, the a bse nc e of te r ms to accoun for particle size and grain sizes casts a shadow on these results. The addition of time as a vari bri ngs m ore co m plexity to the calculated results than the solely temperature-dependent isothermal measurements. Model-free plots of (U0.368Th0.632)O2 isotherms yielded linear behavior within the investigated temperature range. There was no indication either by model-fitting or model-fr methods of a multi-step process. This, of course, does not eliminate the possibility of overlapping mechanisms, but is not suggested in these findings. Figure 4-24 summarizes the model-free re Activation energies calculated from the slopes yield values of 103 20 kJ/mol, 105 kJ/mol, and 111 16 kJ/mol for degree of conversion 0.4, 0.5, and 0.6 respecti vel y. Unl ik e (U0.236Th0.764)O2, nonisothermal model-fit and model-free calculated activation energies are similar. The isothermally model-fit values, however, are significantly higher. This, again,

PAGE 75

63 450Cso450so450aiso450b0.10.20.340.560.78025050075010001250t-to, minf() appears to be the nature associated with model-fitting isothermal versus nonisothermal rate data. Both techniques, however, do not disclose a multi-step process. Table 4-5. Rate coefficient of isothermal (U0.368Th0.632)O2 agreement to 3D diffusion model Isothermal temperature k (min-1) R2 4.8310-4 0.932 475C 1.1710-3 0.965 450C 3.9010-4 4.2710-4 0.894 0.984 9.3510-4 0.974 500C 2.7210-3 2.9810-3 0.989 0.989 2.5310-3 0.983 3D Diffusion (min-1) 2 3C/min 6.80 92.5 0.999 5C/min 8.63 108.2 0.999 Table 4-6. Kinetic results for nonisothermal (U0.368Th0.632)O2 agreement to 3D diffusion model Nonisothermal ln(A) E (kJ/mol) R 1C/min 6.48 92.3 0.999 1 i i 0. 0. 0. 0.9 0 15001750 or (U36 Figure 4-19. Isotherm at 450C f0.8Th0.632)O2 oxidation fit to 3D Diffusion reaction model. Three runs measured

PAGE 76

64 Figure 420. Isotherm at 475C for (U0.368Th0.632)O2 oxidation fit to 3D Diffusion reaction model. Run labeled iso475 is later discarded. Figure 4-21. Iseaction model. Th iso475biso475aiso4750.20.30.40.50.60.70.80.9f()475C 00.1 0100200300400500600700800t-to, min otherm at 500C for (U0.368Th0.632)O2 oxidation fit to 3D Diffusion rree runs measured 500Ciso500aiso500biso500c00.10.20.30.40.50.60.70.8050100150200250300350t-to, minf()` 0.9

PAGE 77

65 igure 4-22. Arrhenius plot of (U0.368Th0.632)O2 isotherms fit to 3D Diffusion reaction model. Rate constants (k) were determined by fitting gravimetric oxidation data with the 3D diffusion reaction model Figure 4-23. Not to 3D -7.5-7-6.5-6-5.5ln(k)U0.368Th0.632O2 -81.281.30 1.321.341.361.381.401/T x 103 (K-1) F nisotherms at 1, 3, and 5C/min (UTh)O Arrhenius plot fi -16-14-12-10-8-6-411.11.21.31.41.51.61.71.81.91/T x 10-3 (K-1)ln(k)1 C/min5 C/min3 C/min 0.3680.6322diffusion model.

PAGE 78

66 Figure 4-24. Moope of the Mixed Oxide (U0.50Th0.50O2) Model-fitting of (U0.500Th0.500)O2 air oxidized pellet fragments introduces another reaction model possibility, Avrami-Erofeev. With regard to the raw isothermal data (Figure 4-8), oxidation is more than 50% complete within the first 500 minutes, which may contribute to nonlinear behavior at dwell times < 200 min. However, linear agreement is observed at times greater than 200 min at all isotherms. The validity of using either the reaction model to determine k, E, and A kinetic parameters for the overall process becomes questionable as to its usefulness. Selecting an appropriate model by global linear fit yields little insight into the oxidative mechanism when the bulk of the process is complete in the ill-fit region. Figures 4-25 to 4-28 present the linear fit of isothermal measurements to both Avrami-Erofeev and 3D Diffusion odels. Figure 4-29 summarizes the rate coefficients according to the 3D diffusion reaction model in an Arrhenius plot of ln(k) vs 1/T. When compared to Arrhenius plots of (U0.236Th0.764)havior of (U0.500Th0.500)O2 fit to the 3D diffusion reaction model is inconsistent with the lower U content mixed oxides. Table 4-7 lists the rate coefficients and linear correlation to the 3D diffusion model. del free (U0.368Th0.632)O2 isotherms plotted at = 0.4, 0.5, 0.6. Sl -5.25-4.75-4.25-3.75-3.25-2.75-2.251.281.31.321.341.361.381.41/T x 103 (K-1)-ln(t) = 0.4 = 0.5 = 0.6U0.368Th0.632O2 linear trendlines gives E/R. m O2 and (U0.368Th0.632)O2, the be

PAGE 79

67 ure 4ith a ss ar nd ble 4-8 3D ermal and nonisothermal rate data onform r. Considering the UO2 system is well known for being a complex Although the lower tem pe ratu re s exhibit a high degree of linearity, both models fall apart at higher temperatures. This is particularly apparent in the Arrhenius ln(k) vs. 1/T plot, Fig 29, of the 3D diffusion reaction model. The plot, in fact, yields a poorly fit linear trendline w positive slope. Calculating activation energy and frequency would, undoubtedly, be meaningle for model fit isothermal rate data. Attempts to fit nonisothermal data to a known reaction model also met with similarly inconclusive results as shown in Figure 4-30. At the slower heating rate of 1 C/min, line agreement was observed in both the Avrami-Erofeev and 3D Diffusion models within the temperature range of interest. The higher heating rate, 5 C/min, on the other hand, did not demonstrate linear behavior. This incongruity may be a result of a lag between reaction rate a heating rate, where the temperature rises faster than the material is able to react. Ta sum m ari zes kinetic par ame ter s ca lculated from nonisoth erm al mod el fitting of rate data to the diffusion and Avrami-Erofeev models. The inability to discern whether (U0.500Th0.500)O2 isoth c to known reaction models presented the model-free technique as a desirable method for calculating activation energies. A plot of the isothermal rate data by the model-free method is shown in Figure 4-31. Again, degree of fraction reacted are selected at = 0.4, 0.5, and 0.6. Similar to the (U0.236Th0.764)O2 and (U0.368Th0.632)O2 model-free Arrhenius plots, there is no suggestion of multiple step behavio multiple step oxidation process, the simple linear behavior would suggest the thorium oxide matrix as a stabilizing influence upon UO2 oxidation. Activation energies calculated from the slopes yield values of 85.2 6.5 kJ/mol, 85.6 5.4 kJ/mol, and 86.2 5.4 kJ/mol for degree of conversion 0.4, 0.5, and 0.6 respectively. Considering that isothermal and nonisothermal model-fitting techniques did not yield a

PAGE 80

68 3D diffusion Avrami-E C 610-4 0.9858 110-3 conclusive reaction model, it is not surprising that model-free activation energies are not in agreement with model-fit results. rofeev Table 4-7. Rate coefficient of isothermal (U0.50Th0.50)O2 agreement to 3D diffusion model Isothermal temperature k (min-1) R2 k (min-1) R2 400C 110-3 0.9581 3.310-3 0.9648 450C 6100.7843 2.210-3 0.776 Erofeev models Noniosthermal ln(A) -1E R2 3750.9821 425C 610-4 0.8528 210-3 0.7495 -4 3 Table 4-8. Kinetic results for nonisothermal (U0.500Th0.500)O2 in 3D diffusion and Avrami-3D Diffusion Avrami-Erofeev (min) (kJ/mol) ln(A) (min-1) E (kJ/mol) R2 1C/min 20.12 166.7 0.9894 8.16 90.3 0.9951 375C1.2 5C/min 8.26 99.7 0.8722 11.91 116.2 0.9735 Figure 4-25C for (U0.50Th0.50)O2 oxidation in 3D Diffusion and Avrami-s. Linear th models are show Isotherm at 375Erofeev model o 3D DiffusAvrami-Erofe'ev00.20.411.4020040060080010001200t-tf( 0.8) ion 0.6 trendline for bo n.

PAGE 81

69 igure 4-26. Isotherm at 400C for (U0.50Th0.50)O2 oxidation in 3D Diffusion and Avrami-Erofeev models. Linear trendline for both models are shown. Figure 4-27. Isvrami-Er 400CAvrami-Erofe'ev3D Diffusion00.511.522.53f() 010020030 0 4 00500600700800t-to F otherm at 425C for (U0.50Th0.50)O2 oxidation in 3D Diffusion and Aofeev models. Linear trendline for both models are shown. 3D Diffusion00.511.522.5020040060080010001200t-tof() 425CAvrami-Erofe'ev33.54

PAGE 82

70 Figure 4-28 IsovramiFigure 4-29. A model. Rate coith the 3D di therm at 450C for (U0.50Th0.50)O2 oxidation in 3D Diffusion and A 450CAvrami-Erofe'ev3D Diffusion00.511.522.533.5402004006008001000t-tof() Erofeev models. Linear tre nd lin e f o r both models are shown. rrhenius plot of (U0.500Th0.500)O2 isotherms in 3D Diffusion reactionnstants (k) were determined by fitting gravimetric oxidation data wffusion reaction model -7.5-7.4-7.3-7.21.351.381.411.441.471.51.531.561/T x 103 (K-1)ln(k -7.1-7-6.9-6.8)U0.5Th0.5O

PAGE 83

71 3D DiffusionAvrami-Erofe'evAvrami-Erofe'ev3D Diffusion-12-11-10-9-8-7-6-5-411.11.21.31.41.51.61.71/T x 10-3 (K-1)ln k5 C/min1 C/min Figure 4-31. Model free (U0.500Th0.500)O2 isotherms plotted at = 0.4, 0.5, 0.6. Slope of the linear trendlines gives E/R. Pure Uranium Dioxide (UO2) Figure 4-30. Nonisotherms at 1 a nd 5 C /mi n ( U 0.500Th0.500)O2 Arrhenius plot fit to Avrami-Erofeev and 3D diffusion models. Linear trendlines shown. -5-4.5-4-3.5-3-2.5-21.361.381.41.421.441.461.481.51.521.541.561/T x103 (K-1)-ln(t) = 0.4 = 0.5 = 0.6U0.5Th0.5O2

PAGE 84

72 Kinetic analysis of UO2 pellet fragments encountered similar results to (U0.500Th0.500)O2 material. Linear agreement to the various reaction models was not observed for isothermal and nonisothermal data. Considering UO2 is known to have a multi-step oxidation, this is not unexpected. As McEachern [McE97a] reported, UO2 powders oxidize by parabolic kinetics, which indicate the reaction is diffusion controlled, to form U3O7 at low temperatures. Diffusion through a discrete layer of the product oxide is the limiting reaction. Two mechanisms, concentration-gradient and discrete-layer, are generally accepted as consistent with the diffusioncontrolled kinetics for UO2 powder oxidation to U3O7. Literature yields activation energy to96 In ter38at nucleation and rowth mechanisms have been consistently applied to this step. Again, two models are commonly applied in literature, Johnson-Mehl and Avrami-Erofeev equations. Also, it has been proven correct that there are at least two different activation energies (at different temperature ranges) for U3O8 formation, with a change in oxidation behavior around 300 350 C [McE97b]. It is not unlikely, therefore, to expect that same degree of complexity at temperatures greater than 350 C, as examined in this study. Considering the nucleation and growth complexity of U3O8 formation, weight gain data is susceptible to interference from U3O7/U4O9 formation and linear growth rates are not necessarily applicable. McEachern et al. [McE97b] developed a twodimensional nucleation and growth model for the specific case of U3O8 formation on UO2 pellets, analogous to three-dimensional nucleation and growth models, resulting in a calculated activation energy of 146 Since oxidation of UO2 was pursued as a control for comparison purposes, not necessarily identifying oxidation mechanisms, the model-free method was used to determine E. However, knowing the UO fragments have a low surface-to-bulk ratio, it is likely the plot mainly reflects the bulk contribution to oxidation and not U3O7 formation. estimates of the U3O7 formation on UO2 powders be kJ/mol [McE97a]. ms of U O formation, McEachern et al. [McE97b] reported th g 10 kJ/mol. 2

PAGE 85

73 trendlines gives E/R. Activation energies cal c ulated for = 0.5, 0.7, and 0.9 across the 375 45 0C temperature range are 116 29 kJ/mol, 104 23 kJ/mol, and 100 20 kJ/mol respectively. -3.5-2-1.5-0.50 = 0.5 = 0.9 These are within the range of values reported in literature for pellet fragments. See Appendix A. Figure 4-32. Model free UO2 isotherms plotted at = 0.4, 0.5, 0.6. The slope of the linear -41.361.381.41.421.441.461.481.51.521.541.563 -1 2 -4.5-3-2.5-11/T x 10(K)-ln(t) UO = 0.7

PAGE 86

74 CHAPTER 5 SUMMARY AND CONCLUSIONS Two of the compositions analyzed ((U0.236Th0.764)O2 and (U0.368Th0.632)O2) exhibited satisfactory agreement to any of the basic theoretical reaction models, absent any additional terms to account for particle size or grain size. As a consequence, results calculated with these basic models restrict conclusions that can be drawn (i.e., surface or bulk-dependence). The (U0.500Th0.500)O2 and UO2 compositions did not demonstrate agreement with any of the basic theoretical reaction models. Knowing the two-step nature of UO2, the absence of a particle size dependent term prohibited the identification of an appropriate reaction model. Nonisothermal calculations, additionally, were further hindered by the simultaneous time and temperature dependence o r quantifying particle characteristics would be necessary to render more precise kinetic parameters. For the low U compositions, the 3D Diffusion model provided the best fit for isothermal rate data. (U0.236Th0.764)O2 fragments, however, did present agreement to 2D diffusion at 450 C. Activation energy values of isothermal rate data calculated using the 3D diffusion reaction model were 62.1 17.9 kJ/mol and 171 8 kJ/mol, for the (U0.236Th0.764)O2 and (U0.368Th0.632)O2 samples, respectively Frequency factors (lnA) were 1.71 2.87 min-1and 20.8 1.3 min-1 for increasing U content, respectively. These values are contrary to the hypothesis that uranium-rich compositions would more readily oxidize than thorium-rich ones. Based on the known cubic fluorite structure, it seems likely that the lattice at interstitial sites accommodates the excess anions without having to undergo phase transformations. Additionally, assuming pellet fragments are polycrystalline, excess anions may also accumulate or migrate along grain f measurement. Furthe

PAGE 87

75 boundaries. This incongruity is likely a result of the limitations attributed to the missing particle size and/or grain size terms. Without fully characterizing the pellet fragments, the basic reaction model can not adequately account for possible complexities, such as multiple mechanisms, simultaneous mechanism, surface-depend dary migration. The kinetic analysis in is study does preliminarily suggest that (U,Th)O2 oxidation occurs by diffusion. X-ray diffraction and uranium valence calculations additionally assert that thorium oxide inhibits uranium from reaching its highest valence state. The higher uranium (U0.500Th0.500)O2 fragments did not indicate agreement with diffusioncontrolled models. In fact, the nucleation and growth Avrami-Erofeev reaction model appeared to fit (U0.500Th0.500)O2 rate data better than 3D diffusion in nonisothermally measured fragments. Considering the reaction model agreement alternated between 3D diffusion and Avrami-Erofeev, there is likely more than one mechanism in simultaneous action. In general, nonisothermal rate data, particularly (U0.236Th0.764)O2 and (U0.368Th0.632)O2, did not discriminately identify one model over another, despite exhibiting linear agreement with models identified isothermally. Consequently, little meaning could be drawn from nonisothermal data in this study. Like the nonisothermal data, the relatively untried model-free method reported in 1999 by Vyazovkin and Wight was tentatively used for comparison. There was no correlation noted between (U,Th)O2 composition and model-free calculated activation energy. Table 20 summarizes values obtained by all three methods. Gravimetric analysis of UO2 was also subjected to the same limitations that befell mixed (U,Th)O2 oxides. Single-step nonisothermal gravitational analysis curves were observed for UO2 and all (U,Th)O2 samples. The absence of two-step curves in the case of UO2 is attributed to the large pellet fragments (90 250 m) used, which have a low surface-to-volume ratio compared with much finer powders. It is unclear from this study whether the single-step behavior observed for the solid solutions was also due to the size of the fragments used, or if such behavior is characteristic of (U,Th)O2 materials, regardless of the particle size. ency, grain boun th

PAGE 88

76 ower uranium concentrations where diffusion dominat f midto high U content mixed oxides m ke erstitial sites and vacancies in the thorium-rich lattice. Table 5-1. Estimated E and A by model-free and model-fit techniques of (U,Th)O2 and UO2 The results of this study highlight the complexities in the UO2 and (U,Th)O2 oxidation processes, as well as difficulties in performing kinetic analyses of this nature. None of the (U,Th)O2 samples reached the same degree of uranium oxidation as pure UO2, indicating a stabilizing effect by thorium in the solid solution. It is clear that diffusion is the primary mechanism of oxidation with possible nucleation and growth mechanisms observed at increasingly greater uranium content. This, perhaps, may explain the incongruity between activation energies and uranium content. At the l es, activation energy in crea se s w ith in creasing U/Th. At U/Th = 0.5, however, it is no longer clear that diffusion is the primary mechanism of oxidation. There is a suggestion that nucleati on a nd grow th mechanisms are beginning to play a significant role in the uranium oxidation. Activation energies calculated, subsequently, do not reveal the system complexity suggested by the model-fit method. Again, further investigation o ay reveal the cause behind the lowered activation ene rgy as c o m pared t o lo we r U /T h oxides. This study, however, does put forward suggestions as to possible mechanisms for mixed (U,Th)O2 dry oxidation. X-ray diffraction confirmed that mixed oxide lattice structures remain cubic fluorite during oxidation. Assuming weight gain is solely a result of excess oxygen anions, uranium valency does n o t proceed to the same extent as oxidation of pure UO2. Also, much li UO2, excess anions are likely accommodated at int Model free Isothermal Model Fit No n isothermal Model fit (kJ/mol) (kJ/mol) (min-1) (kJ/mol) (min-1) Composition E E ln(A) E ln(A) (U0.236Th0.764)O2 87 8 62.7 17.9 1.71 2.87 92.9 1.3 5.3 0.2 (U0.368Th0.632)O2 106 31 171 8 20.8 1.3 97.7 9.1 7.3 1.2 UO2 107 42 ----(U0.500Th0.500)O2 86 6 --103 18* 10 3* *Avrami-Erofeev model, otherwise 3D diffusion

PAGE 89

77 ACTIVATION ENERGIES Table A-1. Published estimates of U3O7/U4O9 activation energy of formation APPENDIX A Eact (kJ/mol) Sample T ( C) Ref. 104* UO2 powder 131 164.5 Anderson et al. (1955) 102* UO2 powder 161 350 Aronson et al. (1957) 120 8 UO2 powder 143 211 Walker (1965) 113 17 Spent LWR fuel 175 225 Woodley et al. (1988, 1989) 90.8 UO2 powder and pellets 125 280 [Bla58] 100 Spent LWR fuel 175 195 Einziger et al. (1992) Activation energy re-calculated from investigators original data using the discrete-layer kinetic model Eact (kJ/mol) Sample T ( C) Method Cited Ref. (rather than the concentration-gradient model) [McE97a]. Table A-2. Published estimates of U3O8 formation on UO2 activation energies 146 UO2 powder 278-325 Gravimetric Aronson (1961) 127.6 UO2 powder 315-360 Gravimetric Saito (1975) 2966) 134.7 UO2 powder 312-352 Gravimetric Walker (1965) .5 UO2 pellets 279-361 Gravimetric Walker (1965) 170.2 AGR pellet fragments 200-300 Gravimetric Tucker (1987) 4 124-139 CANDU pellets 200-300 XR D Taylor (1992) 163 UO2 powder 200-350 Gravimetric [Boa77] 63 CANDU fuel element 300-350 Progression of [Boa77] 143 UO2 pellet fragments 250-350 Gravimetric You (1992) 9 UO2 pellet fragments 350-400 Gravimetric You (1992) 94.5 used LWR fuel 300-400 Gravimetric You (1992) 140 Unirradiated CANDU fragments 175400 Gravimetric Hastings (1986) 6) 194 Used LWR fr in nergy was obse to vary as action of oxessure. The value of 120 kJ/molds to tn prior er form post-s rioispn activatgy of 1 cE ~100* UO2 microspheres 300-450 Gravimetric Ohashi (1987) 161.5 UO powder 365-400 DTA Landspersky (1 110 8 AGR pellet fragments 300-550 Gravimetric Tucker (1987) 102+ LWR Pellets 200-250 Gravimetric White (1983) 170 CANDU pellets 330-350 Gravimetric [Boa77] 67 CANDU pellets 350-450 Gravimetric [Boa77] 172 CANDU fuel element 250-300 Progression of oxidation front [Boa77] oxidation front 10 120 Used CANDU fragments 175-400 Gr avimetric Hstings (198 a agments 250-360 Visual exam ation Einziger (1984) Activation e+ rved fun ygen pr correspon he oxidatio to powd ation. The pallatn io ped dlayed a ion ener 60 kJ/mol [M 97b].

PAGE 90

78 Table A-3. Published estimates of UO2 cation and anion diffusion activation energies Eact(kJ/mol) Sample T( C) Method Ref Cation Diffusion 372 U in UO2.01 1400 1650 -energy spectrometry [Haw68] 398 U in UO2.03 1400 1600 -energy spectrometry [Haw68] 439 U in UO2.1 1400 1650 ectrometry [Haw68] 339 U in UO2.15 1350 ectrometry [Haw68] 304 237U in UO2+x ntribution [Fur68] 438 237U in UO2 1300 1600 Cited Lindner and Schmitz [Fur68] [And83] ion -energy sp 1450 -energy sp Lattice diffusion co 411 U in (S) UO2 Cited Reimann & Lundy Anon Diffusi 273 O in UO2 550 780 sk [And83] 600 1500 le [And83] 1250 ari [And83] sk [And83] sk [And83] nt [And83] nt [And83] ontamin et al. [And83] 92 O in UO Cited Contamin et al. [And83] [And83] 2+x Selonitung (1978 100.3.5 Cff reitung (1 Cff ay (1970 Cffited Bayoglu (1 Sen 1975) Cited Au el ern and Belle 237 O in UO2 Cited B amin 248 O in UO2 780 Cited M n & Cont 124 O in UO2.004 Cited Au ern & Belle 124 O in UO2.063 Cited Au ern & Belle 89 89 O in UO2.006 O in UO Cited Co Cited Co amin et al. amin et al. 2.02090 O in UO2.10 Cited Contamin et al. [And83] 92 O in UO2.12 Cited C 2.1697 O in UO2.08 Cited Murh et al. c 99.6 O in UO f diffusi cited Bre ) [Ari00] 1 10 O in UO2+x hemical di usion, cited B 978) [Ari00] 119.2 O in UO2+x hemical di usion, cited L ) [Ari00] 86.6 O n UO2+x i hemical di usion, c 984) [Ari00] 96.7 O i n UO2+x lf diffusio cited Murch ( [Ari00] (P) ptallin Tabl Pub tes of ThO2 cation aniation enEact(l) T( C) olycrys e; (S) single crystal e A4 lished estima d anion d ffusion activ ergies kJ/mo Sample Method Ref Catiiffusi on D on 2 1600 2100 p 68] 46 2045 g 83] 628 238Th in (S) ThO Cited Matz [And83] 3 00 2000 ff n 8] 201 237U in (P) ThO 1800 2000 Grain boun o [Fur68] Aniosion 47 Th in (P) ThO 2Th in (S) ThO2 18 -energy s ectrometry [Haw 626 Cited Kin ke [And 2 237U in (P) ThO2 18 20 Lattice di usion contributio dary diffusion [Fur6 2 contributi n n Diffu 219 18O in (P) ThO 2 18O in (S) ThO 1099 1644 1099 1644 tr tr 1646 is e o nge [And76] rds et al. [And83] 238 O in ThO2 2100 2800 Cited Lam [Fre80] (P) polycrystalline; (S) single crystal mass spec ometry [And84] 209 2 18O in (S) ThO 845 mass spec ometry [And84] 209 2 gas-solid otope exchang (intrinsic c ntribution) [And76] 73.6 18O in (S) ThO 845 1646 gas-solid isotope excha 2(extrinsic contribution) 275 O in ThO2 900 1500 Cited Edwa

PAGE 91

79 S C ethod Table A-5. Published estimates of diff u sion in (U,Th)O2 activation energies Ea ct( kJ/mol) ample T( ) M Ref 360 237U in (P) UO2-ThO2 200 Lattice diffusion contribution ur68] 1800 3 [F 269 23 7U in (P ) U O 2-Th 2 iffusion 9O O2+10 n 8O2+x 940 10 7O2+10 6O2+x 1 cited Matsui (1985) 0.70O2+x 4 cited 9O0.50O2+x 1 cited 1O O2+x 245 4 n, cited 10O0.O2+x 245 4 ion, cited O2 1800 300 Grain boundary d contribution [Fur68] 9.6 in U0.01Th 0. 99 x 940 40 Chemical diffusio [Ari00] 8.3 in U0.03Th 0.97 O 40 Chemical diffusion [Ari00] 1.0 in U0.05Th 0. 95 O x 940 40 Chemical diffusion [Ari00] 6.8 in U0.20Th0. 8 0O 1009 100 Chemical diffusion [Ari00] 10O 7.1 in U0.30Th 245 55 Chemical diffusion a (1990) Furuy [Ari00] 3.1 in U0.40Th 1009 100 Chemical diffusion, Matsui (1985) [Ari00] 12.1 in U0.50Th 0. 50 55 Chemical diffusio Furuya (1990) [Ari00] 8.0 in U0.70Th30 55 Chemical diffus Furuya (1990) [Ari00] Table Kiners of mixe uraniaSam te Activion ene (/mol) A-6 tic para me te d -thoria oxides ple Heatin g r a (K/m in ) at rgy Pre-exponential factor kJ (log A) (/min) (U0.Th )O2 0.5 45 0.3 0.747 0.006 150.85 42 0.3 31 0.2 Isotherms 51 1 2.6 0.4 (U )O 51 0.6 1 0.693 0.004 5 0.471 0.002 0.300.70 2 Th 0. 5 1.820 0.095 1 49 0.5 1.800 0.090 5 49 0.4 1.350 0.073 Isotherms 45 1 2.9 0.3 (U0.Th )O2 0.5 81 0.4 5.860 0.910 720.28 1 4.750 0.760 2 70 0.6 1.890 0.035 6.5 0.6 (U0.723)O2 90 0.6 0.140 70 0.6 5 46 0.4 0.728 0.012 Isotherms 91 1 7Th0. 0.5 7.250 5 66 0.8 0 0.160 Isotherms 82 1 7.2 1.4 [Ant 1 79 0.4 5.860 0.160 2 67 0.8 4.760 0.190 4.60 00]

PAGE 92

80 IX Y DOTable B-1. U0.368Th0.6322+xormal 400 C air oxi to JCPDS standards 2nmeInk 2 nUO2 hkl APPEND B X RA IFFRACTI N PATTERNS O is the dized compared as t* hl 2n 2 5U Int hkl nThO2 Int hkl 2 Int 27.830 51(100) 111 27.791 100 111 2 111 28.281 100 111 7.581 100 32.245 16( 3 741 200 31) 20 0 32 18 6 40 200 1.962 35 200 32. 50 46.270 21(41) 220 46.172 60 220 4 220 49.968 50 220 5.827 58 54.823( 5 754 311 70 45) 31 1 54 76 9 55 311 4.312 64 311 55. 45 57.525 5(10) 222 57.411 12 222 5 222 58.428 8 222 6.984 11 67.510 3( 6 594 400 6) 40 0 67 36 6 8 400 6.821 8 400 68. 10 74.555 9(18) 331 74.403 18 331 7 331 75.775 20 331 3.794 26 7 6.815 6(12) 420 76.659 14 420 76.010 17 420 78.121 15 420 85.820 5(10) 422 85.562 14 422 84.804 20 422 87.314 15 422 92.465 6(12) 511 92.185 14 511 91.316 19 511 103.650 1(2) 440 102.279 6 440 1 4(8) 10 10.600 531 9.069 18 531 113.010 2(4) 111.37 0 600 7 8 60 1 ) 620 121.07 0 23 .01 0 2(4 3 14 62 131.345 2(4) 533 129.02 3 6 9 53 134.440 1(2) 622 131.92 2 2 9 62 Si hkl 2n Int 2 0 111 28.466 8.375 1 0 100 111 47.235 30 220 47.343 55 220 56.060 14 311 56.170 30 311 69.120 2 .194 400 400 69 6 7 331 76.450 6.305 5 11 331 87.955 6 422 88.115 12 422 94.830 2 511 95.048 6 511 106.595 2 440 106.839 3 440 113.995 2 4.230 531 531 11 7 1 620 127.728 27 .42 0 2 8 620 136.745 1 533 137.124 3 533 *he relative intensities dgnated in parre calculated entified U,Th oxide angles, nt including Si, to determe whether ine consistent w ted for U0.25Th0.75O2, ThO2, a UO cubic oxidesandard md reported val shown T o esi in entheses a tensities ar based on the id ith those repor nd 2 Silicon st easured an ues are also

PAGE 93

81 -1. pato 450C isotherm ofo-m.6%O2 ts nod StSiz ,.54 StTim25 sec Generaor Volte:Tuberent -2. XR peakr 450 isof co-milled6%O2ragments n l Figure B XRD tern f r c illed 23 U fragm en A e: Cu Scan ep e: 0.020 : 1 056, 1. 54439 Scan ep e: 0. t ag 40 V r Cu : 20 Table B D inten sities fo C therm o 23. U f 2 d Int h k 27.7 6 11 1 3.2 11 1 1 32.17 2802 01667 01778 176.49 1.2443 4 4 2 0 85.47 1.1351 3 4 2 2 92.015 1.0707 2 5 1 1 103.06 0.9838 1 4 4 0 109.945 0.9407 2 5 3 1 112.4 0.9269 1 6 0 0 122.235 0.8797 1 6 2 0 .7 5 2 0 46.115 .9 8 2 2 54.66 .6 8 3 1 57.31 1.6063 2 2 2 2 67.255 1.3909 1 4 0 0 74 245 1.2763 3 3 3 1 0 1 0 1 3. 1. 29 56.05 1.6394 13 3 1 1 69.035 1.3593 3 4 0 0 76.29 1.2471 5 0 100 200 300 20 30 405 60 70 80 10 110 C 1Si 400Si 1Si420 1Sih 440SiU,Th60 1Si 62h 53 111U, 200U,T 220U,Th U,Th 2U, 0U,Th 33 Th U,Th422U,T 531U, U,Th620U,T 3U,ThU, 622 400ount 5006007008001009001201301402ns111Si220Si31332Si51530Si3SiThh31122Th401U,42h511U,T400T0h53Th 130.455.8483 5 3 3 133.395.8387 6 2 2 (Si) 28.375 1428 100 1 1 1 47.22 9233 2 2 0 3 3 1 87.945 1.1094 5 4 2 2 94.85 1.046 2 5 1 1 106.5750.9609 1 4 4 0 113.9650.9186 2 5 3 1 127.4350.8591 1 6 2 0 136.78 0.8285 1 5 3 3

PAGE 94

82 01002003004005006007008001020304050607080901001101201301402nCounts111Si220Si311Si400Si331Si422Si511Si440Si531Si620Si533Si111U,Th200U,Th220U,Th311U,Th222U,Th400U,Th331U,Th420U,Th422U,Th511U,Th400U,Th531U,Th600U,Th620U,Th533U,Th622U,Th Figure B-2. XRD pattern for 500C isotherm of co-milled 23.6% UO 2 fragments Anode: Cu Scan Step Size: 0.020 : 1.54056, 1.54439 Scan Step Time: 0.25 sec Generator Voltage: 40V Tube Current: 20 Table B-3. XRD peak intensities for 500C isotherm of co-milled 23.6% UO2 fragments 2n d Int h k l 27.775 3.2093 13 1 1 1 32.195 2.7781 4 2 0 0 46.155 1.9651 7 2 2 0 54.705 1.6765 6 3 1 1 57.325 1.6059 2 2 2 2 67.24 1.3912 1 4 0 0 74.27 1.2759 3 3 3 1 76.54 1.2437 4 4 2 0 85.44 1.1354 2 4 2 2 92.015 1.0707 2 5 1 1 103.11 0.9835 1 4 4 0 110.04 0.9401 2 5 3 1 112.38 0.9271 1 6 0 0 122.27 0.8796 1 6 2 0 130.41 0.8485 1 5 3 3 133.44 0.8386 1 6 2 2 Si 28.38 3.1422 100 1 1 1 47.235 1.9227 21 2 2 0 56.045 1.6395 7 3 1 1 69.025 1.3595 1 4 0 0 76.305 1.2469 4 3 3 1 87.945 1.1094 3 4 2 2 94.86 1.0459 2 5 1 1 106.59 0.9608 1 4 4 0 113.99 0.9185 1 5 3 1 127.405 0.8592 1 6 2 0 136.76 0.8286 0 5 3 3

PAGE 95

83 igure B-3. XRD pattern for 550C isotherm of co-milled 23.6% UO2 fragments e9 Scan Step Time: 0.25 sec : 40V ube Current: 20 Table B-4. XRD peak intensities for 550C isotherm of co-milled 23.6% UO2 fragments 01002003004005006007008001020304050607080901001101201301402nCounts111Si220Si311Si400Si331Si422Si511Si440Si531Si620Si533Si111U,Th200U,Th220U,Th311U,Th222U,Th400U,Th331U,Th420U,Th422U,Th511U,Th400U,Th531U,Th600U,Th620U,Th533U,Th622U,Th F Anod: Cu Scan Step Size: 0.020 : 1.54056, 1.5443 Generator Voltage T 2n d Int h k l 27.795 3.207 14 1 1 1 32.215 2.7764 4 2 0 0 46.125 1.9663 7 2 2 0 54.69 1.6769 7 3 1 1 57.37 1.6048 2 2 2 2 67.25 1.391 1 4 0 0 74.265 1.276 4 3 3 1 76.525 1.2439 4 4 2 0 85.46 1.1352 2 4 2 2 92.055 1.0703 2 5 1 1 103.135 0.9833 1 4 4 0 110.01 0.9403 2 5 3 1 112.37 0.9271 1 6 0 0 122.2 0.8799 1 6 2 0 130.435 0.8484 1 5 3 3 133.305 0.839 0 6 2 2 (Si) 8.38 1 1 1 47.23 1.9229 21 2 2 0 56.05 1.6394 8 3 1 1 69.065 1.3588 2 4 0 0 76.29 1.2471 4 3 3 1 87.955 1.1093 4 4 2 2 94.855 1.046 2 5 1 1 106.61 0.9607 1 4 4 0 113.995 0.9185 1 5 3 1 127.495 0.8589 1 6 2 0 136.79 0.8285 0 5 3 3 2 3.1422 100

PAGE 96

84 Figure B4. XRD pattern for 400C isotherm of co-milled 36.8% UO2 fragments 9 Scan Step Time: 0.25 sec : 40V ube Current: 20 able B-5. XRD peak intensities for 400C isotherm of co-milled 36.8% UO2 fragments 01002003004005006007008001020304050607080901001101201301402nCounts111Si220Si311Si400Si331Si422Si511Si440Si531Si620Si533Si111U,Th200U,Th220U,Th311U,Th222U,Th400U,Th331U,Th420U,Th422U,Th511U,Th400U,Th531U,Th600U,Th620U,Th533U,Th622U,Th A n od e : Cu Scan Step Size: 0.020 : 1.54056, 1.5443 Generator Voltage T T 2n d Int h k l 27.83 3.2031 51 1 1 1 32.245 2.7739 16 2 0 0 46.27 1.9605 21 2 2 0 54.87 1.6718 23 3 1 1 57.525 1.6008 5 2 2 2 67.51 1.3863 3 4 0 0 74.555 1.2718 9 3 3 1 76.815 1.2399 6 4 2 0 85.82 1.1314 5 4 2 2 92.465 1.0666 6 5 1 1 103.65 0.9798 1 4 4 0 110.6 0.9369 4 5 3 1 113.01 0.9237 2 6 0 0 123.01 0.8765 2 6 2 0 131.345 0.8453 2 5 3 3 134.44 0.8354 1 6 2 2 Si 28.375 3.1428 100 1 1 1 47.235 1.9227 30 2 2 0 56.06 1.6391 14 3 1 1 69.12 1.3579 2 4 0 0 76.305 1.2469 5 3 3 1 87.955 1.1093 6 4 2 2 94.83 1.0462 2 5 1 1 106.595 0.9608 1 4 4 0 113.995 0.9185 2 5 3 1 127.42 0.8591 2 6 2 0 136.7450.8286 1 5 3 3

PAGE 97

85 igure B-5. XRD pattern for 450C isotherm of co-milled 36.8% UO2 fragments node: Cu Scan Step Size: 0.020 c Table B XRD peak intensities for 450C isotherm of co-milled 36.8% UO2 fragments Ihkl 01002003004005006007008001020304050607080901001101201301402nCounts111Si220Si311Si400Si331Si422Si511Si440Si531Si620Si533Si111U,Th200U,Th220U,Th311U,Th222U,Th400U,Th331U,Th420U,Th422U,Th511U,Th400U,Th531U,Th600U,Th620U,Th533U,Th622U,Th F A : 1.54056, 1.54439 Scan Step Time: 0.25 se G e ne ra tor Voltage: 40V Tube Current: 20 -6. 2n d nt 27.85 3.2008 78 1 1 1 32.27 2.7718 25 2 0 0 45 2 2 0 41 3 1 1 9 2 2 2 4 4 0 0 12 3 3 1 10 4 2 0 7 4 2 2 9 5 1 1 3 4 4 0 8 5 3 1 5 6 0 0 4 6 2 0 3 5 3 3 2 6 2 2 46.29 1.9597 54.88 1.6715 57.535 1.6006 67.52 1.3861 74.515 1.2724 76.905 1.2387 85.785 1.1317 92.45 1.0668 103.625 0.98 110.575 0.9371 112.95 0.924 123.015 0.8764 131.25 0.8457 134.45 0.8354 Si 28.405 3.1395 100 1 1 1 25 2 2 0 11 3 1 1 1 4 0 0 5 3 3 1 5 4 2 2 2 5 1 1 1 4 4 0 2 5 3 1 1 6 2 0 1 5 3 3 47.255 1.9219 56.075 1.6387 69.1 1.3582 76.325 1.2466 87.98 1.1091 94.875 1.0458 106.615 0.9606 113.995 0.9185 127.395 0.8592 136.74 0.8286

PAGE 98

86 2 d Int h k l 01002003004005006007008001020304050607080901001101201301402nCounts111Si220Si311Si400Si331Si422Si511Si440Si531Si620Si533Si111U,Th200U,Th220U,Th311U,Th222U,Th400U,Th331U,Th420U,Th422U,Th511U,Th400U,T531U,Th600U,T620U,Th533U,Th622U,Th Figure B-6. XRD pattern for 500C isotherm of co-milled 36.8% UO 2 fragments Anode: Cu Scan Step Size: 0.020 : 1.54056, 1.54439 Scan Step Time: 0.25 sec Generator Voltage: 40V Tube Current: 20 Table B-7. XRD peak intensities for 500C isotherm of co-milled 36.8% UO2 fragments n 273. .85 2008 24 1 1 1 2.285 2.7705 14 2 0 0 46.27 1.9605 14 2 2 0 54.85 1.6724 12 3 1 1 57.53 1.6007 3 2 2 2 67.55 1.3856 2 4 0 0 74.53 1.2721 6 3 3 1 6.885 1.2389 4 4 2 0 85.8 1.1316 3 4 2 2 92.46 1.0667 4 5 1 1 03.65 0.9798 1 4 4 0 10.515 .9374 3 5 3 1 12.975 0.9239 2 6 0 0 22.995 0.8765 2 6 2 0 131.33 0.8454 1 5 3 3 34.36 0.8357 1 6 2 2 8.405 3.1395 10 1 1 1 255 38 2 2 0 6.065 1.639 15 3 1 1 9.085 1.3585 4 4 0 0 6.325 1.2466 7 3 3 1 7.95 1.1094 7 4 2 2 94.86 1.0459 3 5 1 1 06.59 0.9608 2 4 4 0 13.965 0.9186 3 5 3 1 27.395 0.8592 2 6 2 0 136.76 0.8286 1 5 3 3 3 7 1 10 1 1 1 Si 20 47. 1.9219 5 6 7 8 1 1 1

PAGE 99

87 Figure B7. XRD pattern for 375C isotherm of co-milled 50.0% UO2 fragments 9 Scan Step Time: 0.25 sec : 40V ube Current: 20 able B-8. XRD peak intensities for 375C isotherm of co-milled 50.0% UO2 fragments 01002003004005006007008001020304050607080901001101201301402nCounts111Si220Si311Si400Si331Si422S511S440S531Si620S533Si111U,Th200U,T220U,T311U,T222U,Th400U,T331U,T420U,T422U,T511U,T400U,T531U,T600U,T620U,T533U,T622U,T A n od e : Cu Scan Step Size: 0.020 : 1.54056, 1.5443 Generator Voltage T T 2n d Int h k l 27.92 3.1929 5 1 1 1 32.405 2.7605 2 2 0 0 46.47 1.9525 3 2 2 0 55.16 1.6637 4 3 1 1 56.2 1.6354 4 2 2 2 67.78 1.3814 1 4 0 0 74.915 1.2665 2 3 3 1 76.52 1.2439 2 4 2 0 86.26 1.1267 1 4 2 2 92.805 1.0636 1 5 1 1 104.37 0.9750 0 4 4 0 111.205 0.9335 1 5 3 1 113.985 0.9185 2 6 0 0 123.7 0.8736 0 6 2 0 128.005 0.8570 0 5 3 3 132.465 0.8417 0 6 2 2 (Si) 28.37 3.1433 100 113.9850.9185 2 5 3 1 127.41 0.8592 1 6 2 0 136.7 0.8287 1 5 3 3 1 1 1 47.22 1.9233 26 2 2 0 56.04 1.6397 9 3 1 1 69.04 1.3593 2 4 0 0 76.3 1.2470 4 3 3 1 87.945 1.1094 4 4 2 2 94.84 1.0461 2 5 1 1 106.635 0.9605 1 4 4 0

PAGE 100

88 igure B-8. XRD pattern for 400C isotherm of co-milled 50.0% UO2 fragments node: Cu Scan Step Size: 0.020 c Table B XRD peak intensities for 400C isotherm of co-milled 50.0% UO2 fragments 2n d Int h k l 01002003004005006007008001020304050607080901001101201301402nCounts111Si220Si311Si400Si331Si422S511S440S531Si620S533Si111U,Th200U,T220U,T311U,T222U,Th400U,T331U,T420U,T422U,T511U,T400U,T531U,T600U,T620U,T533U,T622U,T F A : 1.54056, 1.54439 Scan Step Time: 0.25 se G e ne ra tor Voltage: 40V Tube Current: 20 -9. 27 3.1879 .965 11 1 1 1 32.415 2.7597 5 2 0 0 463525 1.9504 7 2 2 0 55.105 1.6652 9 3 1 1 57.845 1.5927 2 2 2 2 67.83 1.3805 1 4 0 0 74.865 1.2673 2 3 3 1 77.17 1.2351 2 4 2 0 86.235 1.1270 2 4 2 2 92.865 1.0631 2 5 1 1 104.265 0.9757 1 4 4 0 111.15 0.9338 2 5 3 1 114.005 0.9184 2 6 0 0 123.765 0.8734 1 6 2 0 132.31 0.8422 1 5 3 3 28.38 3.1422 100 5 1.6394 11 11 1 1 1 47.235 1.9227 28 2 2 0 6.05 3 1 1 69.05 1.3591 2 4 0 0 76.3 1.2470 5 3 3 1 87.945 1.1094 6 4 2 2 94.855 1.0460 3 5 1 06.595 0.9607 2 4 4 0 114.005 0.9184 2 5 3 1 27.415 0.8592 1 6 2 0 136.765 0.8286 1 5 3 3 135.355 0.8327 1 6 2 2 (Si) 1

PAGE 101

89 01002003004005006007008001020304050607080901001101201301402nCounts111Si220Si311Si400Si331Si422S511S440S531Si620S533Si111U,Th200U,T220U,T311U,T222U,Th400U,T331U,T420U,T422U,T511U,T400U,T531U,T600U,T620U,T533U,T622U,T Figure B-9. XRD pattern for 425C isotherm of co-milled 50.0% UO 2 fragments Anode: Cu Scan Step Size: 0.020 : 1.54056, 1.54439 Scan Step Time: 0.25 sec Generator Voltage: 40V Tube Current: 20 Table B-10. XRD peak intensities for 425C isotherm of co-milled 50.0% UO2 fragments 2n d Int h k l 28.045 3.1790 13 1 1 1 32.43 2.7585 5 2 0 0 46.515 1.9507 5 2 2 0 55.165 1.6636 8 3 1 1 57.81 1.5936 2 2 2 2 67.87 1.3798 1 4 0 0 74.87 1.2672 2 3 3 1 77.18 1.2349 2 4 2 0 86.22 1.1271 2 4 2 2 92.885 1.0629 2 5 1 1 104.145 0.9765 1 4 4 0 111.2 0.9335 2 5 3 1 113.615 0.9205 1 6 0 0 123.835 0.8731 1 6 2 0 2.29 0 5 3 3 35.195 0.8332 0 6 2 2 (Si) 28.405 3.1395 100 1 1 1 47.25 1.9221 28 2 2 0 56.07 1.6389 8 3 1 1 69.085 1.3585 2 4 0 0 76.325 1.2466 4 3 3 1 87.96 1.1093 4 4 2 2 94.87 1.0459 2 5 1 1 106.635 0.9605 1 4 4 0 113.975 0.9186 2 5 3 1 1350.8422 1127.4150.8592 1 6 2 0 136.79 0.8285 0 5 3 3

PAGE 102

90 Figure B10. XRD pattern for 450C isotherm of co-milled 50.0% UO2 fragments e9 Scan Step Time: 0.25 sec : 40V ube Current: 20 able B-11. XRD peak intensities for 450C isotherm of co-milled 50.0% UO2 fragments 01002003004005006007008001020304050607080901001101201301402nCounts111Si220Si311Si400Si331Si422S511S440S531Si620Si533Si111U,Th200U,T220U,T311U,T222U,Th400U,T331U,Th420U,T422U,T511U,T400U,T531U,T600U,T620U,T533U,T622U,T Anod: Cu Scan Step Size: 0.020 : 1.54056, 1.5443 Generator Voltage T T 2n d Int h k l 27.965 3.1879 18 1 1 1 32.39 2.7618 10 2 0 0 46.49 1.9517 13 2 2 0 55.105 1.6652 12 3 1 1 57.765 1.5947 3 2 2 2 67.815 1.3808 3 4 0 0 74.885 1.2670 4 3 3 1 77.18 1.2349 3 4 2 0 86.2 1.1273 3 4 2 2 92.91 1.0627 3 5 1 1 104.16 0.9764 1 4 4 0 111.205 0.9335 2 5 3 1 113.985 0.9185 3 6 0 0 123.725 0.8735 1 6 2 0 132.14 0.8427 1 5 3 3 5.35 1 6 2 2 (Si) 28.37 3.1433 100 1 1 1 47.225 1.9231 21 2 2 0 56.205 1.6352 5 3 1 1 59.13 1.3577 1 4 0 0 76.3 1.2470 4 3 3 1 87.95 1.1094 4 4 2 2 94.875 1.0458 2 5 1 1 106.76 0.9597 0 4 4 0 113.985 0.9185 3 5 3 1 127.405 0.8592 1 6 2 0 1350.8327 5 3 3

PAGE 103

PENDIX C AGES OF UNIRRADIATED PELLETS James Jerden: ANL-E) 91 APSECONDARY ELECTRON IM(Microscopy by Figure C-1. Unpolished U0.05Th0.95O2+x pellet surface. Scale bar is 21 microns. Grain size generally ranges from ~4 to ~20 microns. White dots on surface are dust particles introduced during SEM sample prep. [PD020222-2] igure C-2. Unpolished U0.05Th0.95O2+x pelleurfaccaas Gzea am ~20 W d ue dpal dg Sme r020222F t s e. S le b r i 20 m icrons. rain si e g ner lly r nges fro ~4 to m icrons. hite ots on s rface ar ust rtic es in troduced urin EM sa pl p ep. [PD 2]

PAGE 104

92 Figure C-3. Broken surface of U0.2Th0.8O2+x pellet. Splitting was accomplished by a single blow within a heavy iron mortar-pestle. Note curious gas bubble looking features these are pervasive on The unbroken pellet surface of this sample looksFigures 42. [20BL0202-4] ns broken surfaces of this pellet. similar to Scale bar is 27 micro

PAGE 105

93 EFERENCES Abd90 hemical characteristics of uranium oxide microspheres Isotopenpraxis, 26 11 (1990) 524-529. nd54 J.S. Anderson, D.N. Edgington, L.E.J. Roberts, and E. Wait, The oxides of uranium. Part IV. The system UO2-ThO2-O, Journal of the Chemical Society (1954) 33243331. And76 K. And a, Self-diffusion of oxygen in single crystal thorium oxide, Journal of Chemical Physics 65 7 (1976) 2751-2755. And79 K. Ando, and Y. Oishi, Oxygen self-diffusion and conduction mechanism of thoria, Journal of Nuclear Science and Technology 16, 3 (1979) 214-220. And83 K. Ando, and Y. Oishi, Diffusion characteristics of actinide oxides review article, Journal of Nuclear Science and Technology 20, 12 (1983) 973-982. And84 K. Ando, Y. Oishi, T. Tsuji, and Y. Hidaka, Grain-boundary depressed diffusion of oxygen ions in ThO2, Journal of Nuclear Materials 120 (1984) 99-101. And85 K. Ando, Y. Ikeda, S. Morita, R. Watanabe, K. Shiba, and M. Handa, Preferential vaporization of uranium oxide in the polycrystalline thoria-urania solid solution, Journal of Nuclear Materials, 136 (1985) 186-191. Ant97 S. Anthonysamy, K. Nagarajan, and P.R. Vasudeva Rao, Studies on the oxygen potentials of (UyTh1-y)O2+x solid solutions, Journal of Nuclear Materials 247 (1997) 273-276. Ant00 S. Anthonysamy, K. Joseph, T. Gnanasekaran, and P.R. Vasudeva Rao, Studies on the kinetics of oxidation of urania-thoria solid solutions in air, Journal of Nuclear Materials 280 (2000) 25-32. Ari00 T. Arima, K. Kitano, K. Takemura, and H. Furuya, Chemical diffusion of oxygen in thoria-urania solid solution, Thermochimica Acta 344 (2000) 37-45. Aro60 S. Aronson, and J.C. Clayton, Thermodynamic properties of nonstoichiometric urania-thoria solid solutions, Journal of Chemical Physics 32 3 (1960) 749-754. Bam80 C.H. Bamford and C.F.H. Tipper (Eds.), Chemical Kinetics. Reactions in the Solid RA.S. Abdel-Halim, Physico-c produced by internal gelation, A o, Y. Oishi, and Y. Hidak State New York, Elsevier, (1980) pp. 68-83. Ban68 M.J. Bannister, The storage behavior of uranium dioxide powders review article, Journal of Nuclear Materials, 26 (1968) 174-184.

PAGE 106

94 Bla58 P.E. Blackburn, J. Weissbart, and E.A. Gulbransen, Oxidation of uranium dioxide, Journal of Physical Chemistry 62 (1958) 902-908. Boa77 D.G. Boase, and T.T. Vandergraaf, The Canadian spent-fuel storage canister: some material aspects, Nuclear Technology 32 (1977) 60-71. Cha98 V. Chandramouli, S. Anthon deva Rao, R. Divakar, and D. Sundararaman, Microwave synthesis of solid solutions of urania nad thoria a ol83 E.A. Colbourn, and W.C. Mackrodt, The calculated defect structure of thoria, Cul78 ullity, Elements of X-ray Diffraction ysamy, P.R. Vasu comparative study, Journal of Nuclear Materials 254 (1998) 55-64. C Journal of Nuclear Materials, 118 (1983) 50-59. B.D. C 2nd edition, Massachusetts, AddisionWesley Publishing Co. (1978). Dob98 and H. Manenc, Kinetics of UO2 oxidation in steam atmosphere, Journal of Nuclear re80 R. Freer, Bibliography self-diffusion and impurity diffusion in oxides, Journal of ur68 H. Furuya, Lattice and grain-boundary diffusion of uranium in ThO2 and ThO2-UO2 ab93 A. Gabbard, Coal combustion: nuclear resource or danger, Oak Ridge National al02 A.K. Galwey and M.E. Brown, Application of the Arrhenius equation to solid state Gan86 y, H. Langen, E. Zimmer, E.R. Merz, Sol-gel microsphere pelletization process for fabrication of high-density ThO2-2% UO2 fuel for advanced pressurized Har91 ical Analysis B.V. Dobrov, V.V. Likhanskii, V.D. Ozrin, A.A. Solodov, M.P. Dissane Materials 255 (1998) 59-66. F Materials Science 15 (1980) 803-824. F solid solution, Journal of Nuclear Materials 26 (1968) 123-128. G Laboratory Review 26, 3 & 4, (1993) http://www.ornl.gov/info/ornlreview/rev2634/text/colmain.html Accessed 10/2003. G kinetics: can this be justified? Thermochimica Acta 386 (2002) 91-98. C. Gangul heavy water reactors, Nuclear Technology 73 (1986) 84-95. D.C. Harris, Quantitative Chem 3rd edition, New York, (1991). 2-122. and a Acta 373 (2001) 97-124. Mal01 etic analysis of solid-state processes, Journal of the Materials Research Society 16, 6 (2001) 1862-1871. Mat85 2+x and (Th1-yUy)O2+x, Journal of Nuclear Materials 132 (1985) 212-221. Haw68R.J. Hawkins, and C.B. Alcock, A study of cation diffusion in UO2+x and ThO2 using -ray spectrometry, Journal of Nuclear Materials 26 (1968) 11 Lvo01 B.V. Lvov, Review. The physical approach to the interpretation of the kinetics mechanisms of thermal decomposition of solids: the state of the art, Thermochimic J. Ma lek, T. Mitsuhashi, and J.M. Criado, Kin T. Matsui, and K. Naito, Oxygen potentials of UO

PAGE 107

95 McE97a of U3O7 formation on UO2, Journal of Nuclear Materials, 245 (1997) 238-247. McE97b and D.D. Wood, Determination of the activation energy for the formation of U3O7 on UO2, Journal Pop73 fraction lines, Journal of Applied Crystallography 6 (1973) 122-128. Pop85 alloys by x-ray diffraction, Crystal Research and Technology 20, 4 (1985) 552-555. Ric92 ic Engineering R.J. McEachern, A review of kinetic data on the rate R.J. McEachern, J.W. Choi, M. Kolar, W. Long, P. Taylor of Nuclear Materials 249, (1997) 58-69. S. Popovic, Unit-cell dimension measurements from pairs of x-ray dif S. Popovic, Unit-cell parameter measurements of D.W. Richerson, Modern Ceram New York, Marcel Dekker, Inc. (1992). Ser01 ce of on the dissolution behavior of high burn-up LWR spent-fuel, Journal of Nuclear Materials 294 (2001) 339-343. Sho00 ses under waste disposal conditions review, Journal of Nuclear Materials 282 (2000) 1-31. Sun96 uel by air in gamma radiation at 150 C, Journal of Nuclear Materials 231 (1996) 121-131. Tay96 waste forms for direct disposal, Nuclear Technology 116 (1996) 222-230. Tay98 specimen roughness on the rate of formation of U3O8 on UO2 in air at 250 C, Journal of Tsu98 cations on lattice constants of (MyU1-y)O2.00 (M=Pu, Th, La) at low doped cation concentrations, Materials 96 (1981) 305-313. J.A. Serrano, J.P. Glatz, E.H. Toscano, J. Barrero, and D. Papaioannou, Influen low-temperature air oxidation D.W. Shoesmith, Fuel corrosion proces S. Sunder, and N.H. Miller, Oxidation of CANDU uranium oxide f P. Taylor, W.H. Hocking, L.H. Johnso n, R.J. McEachern, and S. Sunder, A comparison of (Th,P u)O2 and UO2 fuels as P. Taylor, R.J. McEachern, D.C. Doern, and D.D. Wood, The influence of Nuclear Materials 256 (1998) 213-217. T. Tsuji, M. Iwashita, T. Yamashita, and K. Ohuchi, Effect of Journal of A lloys and Compounds 271-273 (1998) 391-394. Uga82 M. Ugajin, Oxygen potentials of (Th,U)O2+x solid solutions, Journal of Nucle a r Materials 110 (1982) 140-146. Vya99 S. Vyazovkin and C.A. Wight, Model-free and model-fitting approaches to kinetic analysis of isothermal and nonisothermal data, Thermochimica Acta 340/341 (1999) 53Whi81 G.D. White, L.A. Bray, and P.E. Hart, Optimization of thorium oxalate precipitation conditions relative to derived oxide sinterability, Journal of Nuclear

PAGE 108

96 il87 B.T.M. Willis, Crystallographic studies of anion-excess uranium oxides, Journal of in03 M. Winter, Webelements periodic table (professional edition), University of ed W Chemical Society, Faraday Transactions 2 83 (1987) 1073-1081. W Sheffield and WebElements Ltd, UK, (2003) www.webelements.com Access 10/2003.

PAGE 109

97 BIOGRAPHICAL SKETCH Lisa Argo was born in Gainesville, Florida where she attended the University of Florida. While an undergraduate student, the author was active in the American Ceramics Society student branch as Correspondence Secretary and the 1999 Engineering Fair Planning Committee as F ood Coordinator. After graduating with her Bachelor of Science degree in Materials Science and Engineering in 1999, she worked as an Application Engineer at Elan Technology in Midway, Georgia. The following year, the author married and returned to Gainesville, Florida where she worked as Research Associate for James Tulenko in the UF Department of Nuclear and Radiological Engineering and co-advised by Dr. Ronald Baney in the Department of Materials Science and Engineering. Four months later, the author enrolled in the University of Florida towards completion of a Master of Science in Materials Science and Engineering under Dr. Ronald Baney.


Permanent Link: http://ufdc.ufl.edu/UFE0001521/00001

Material Information

Title: Experimental Determination of the Dry Oxidation Behavior of a Compositional Range of Uranium-Thorium Mixed-Oxide Pellet Fragments
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0001521:00001

Permanent Link: http://ufdc.ufl.edu/UFE0001521/00001

Material Information

Title: Experimental Determination of the Dry Oxidation Behavior of a Compositional Range of Uranium-Thorium Mixed-Oxide Pellet Fragments
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0001521:00001


This item has the following downloads:


Full Text












EXPERIMENTAL DETERMINATION OF THE DRY OXIDATION
BEHAVIOR OF A COMPOSITIONAL RANGE OF
URANIUM-THORIUM MIXED-OXIDE PELLET FRAGMENTS









By

LISA ARGO


A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE

UNIVERSITY OF FLORIDA


2003



























Copyright 2003

by

Lisa Argo
















ACKNOWLEDGMENTS

This work would not have been completed without the support of many. More than

anyone else, I thank my advisors (Ronald Baney and James Tulenko) for their wisdom, patience,

understanding, and occasional cracking whip. I would like to acknowledge the contribution of

Paul Demkowicz for his expertise and meticulous editing; and Noriko Shibuya for her technical

support, as we investigated the ins and outs of many synthesis methods. I thank my sister, Irene,

for the late night pep talks and advice. I thank my husband, Paul, for pushing me back into and

out of school.

This work was funded through a grant from the Nuclear Engineering Research Initiative

(NERI) project #99-0153 Advanced Proliferation Resistant, Lower Cost, Uranium-Thorium

Dioxide Fuels for Light Water Reactors.

















TABLE OF CONTENTS

page

A C K N O W L E D G M EN T S ................................................................................. ....................

T A B L E O F C O N T EN T S ................................................................... .. ................................. iv

L IS T O F T A B L E S .........................................................................................v i

L IST O F F IG U R E S ............................................................................... ..................... viii

A B S T R A C T ...............................................................................................................x i

CHAPTER

1 IN T R O D U C T IO N ................................................................... ................................ 1

2 REV IEW OF LITERA TURE ............................................................... .................... 5

Current Status of UO2 and ThO2 Research ............................................................. .......... 5
U ran iu m D iox id e ...................................................................................... ........ .... 5
T horium O xide ..................................... .................................................... 9
Uranium-Thorium Mixed Oxide ............. ..... .................................. 11
Background of Synthesis M methods .............. .................................. ............................. 14
K inetic A analysis ........................................................ ......... ...... 16

3 MATERIALS AND METHODS ............... .....................................20

M material S y nth esis .............. ..... ............ ................. ............................................. 2 0
Oxalate Co-Precipitation .......... ....................... ... ................................................... 20
Ammonium Hydroxide Co-Precipitation.................... ...... ......................... 26
C o-M killed M ixed O xides ......... ................................................................................ 27
Characterization...................................... ............ .. .. .. ................... 32
X -ray D iffraction (X RD ) ................................... ................. ......... ............... 32
Elemental Analysis ................................... ............................ .... ......... 35
Particle M orphology .............................................................. ...................... 36
Pellet D density ................................... .................. ...... ........... 37
Dry Oxidation ................................................ 38
Therm ogravim etric A nalyzer ........................................................... ............ .. ...... 38
Kinetic Analysis ...................................................................... ......... 39

4 RESULTS AND DISCUSSION........................ ....... ............................... 43

M material S ynth esis ....................................................................... .................................. 4 3
T h erm og rav im etry .................................................................. .................................... 4 5










Kinetic Analysis ..................... .. ............ ........... ............ 54
M ixed Oxide (Uo 236Th0 7640 2) ..................... ......................................... ............... 55
M ixed Oxide (Uo 368Th0 6320 2) ..................... ......................................... ............... 61
M ixed O xide (U o 5oT ho 500 2) ........................................................................ 66
Pure U ranium D ioxide (U 2)............................................................... ............... 71

5 SUMMARY AND CONCLUSIONS ......................................................... ............... 74

APPENDIX

A A CTIV A TION EN ER G IES .................................... ................................... .................... .... 77

B X-RAY DIFFRACTION PATTERNS .................................... .................. ............... 80

C SECONDARY ELECTRON IMAGES OF UNIRRADIATED PELLETS ........................ 91

R E F E R E N C E S ................................ ........................................ ................ 9 3

BIOGRA PH ICAL SKETCH ................................................................... ........................... 97








































v
















LIST OF TABLES


Table page

2-1. Defect formation energies for U02 ............................................................... .... 6

2-2. Oxidation reactions of U 2+ ........................................................ .............................. 7

2-3. M material properties of ThO 2 and UO2 ........................................... ........................... 9

2-4. Cation and anion formation and migration energies in ThO2 ............................................ 10

2-5. Composition dependence on oxygen partial pressure for U02+x and (U,Th)02+ .............. 13

2-6. Rate laws for a simple process A P .................................................................. 18

3-1. Evolution of oxalate-synthesized (U,Th)02. ............................................................. 23

3-2. 2x2 factorial for 20% U02-80% ThO2 blended oxide................................. ............ 30

3-3. Co-milled U 2Tho 802 pressing conditions 2x2 factorial results....................................... 30

3-4. Pellet m manufacture conditions ................................................ ............................... 31

3-5. Solid state theoretical reaction m odels ...................................................... .............. 41

4-1. Calculated (U,Th)02 oxidized lattice parameters by three methods................................43

4-2. ICP-AES and LECO carbon analysis results and calculated metal valence.........................45

4-3. Rate coefficients of isothermal (Uo 236Th0 764)02 agreement to diffusion models ................. 56

4-4. Kinetic results for nonisothermal (Uo 236Th0 764)02 agreement to diffusion models.............. 56

4-5. Rate coefficient of isothermal (Uo 368Th0 632)02 agreement to 3D diffusion model .............. 63

4-6. Kinetic results for nonisothermal (Uo 368Th0 632)02 agreement to 3D diffusion model.......... 63

4-7. Rate coefficient of isothermal (Uo 50Tho 50)02 agreement to 3D diffusion model ............. 68

4-8. Kinetic results for nonisothermal (Uo 500Tho 500)02 in 3D diffusion and Avrami-Erofe'ev ... 68

5-1. Estimated E and A by model-free and model-fit techniques of (U,Th)O2 and UO2 .............76

A-1. Published estimates of U307/U409 activation energy of formation ..... ......................... 77










A-2. Published estimates of U308 formation on U02 activation energies.............................. 77

A-3. Published estimates of U02 cation and anion diffusion activation energies........................ 78

A-4. Published estimates of ThO2 cation and anion diffusion activation energies................... 78

A-5. Published estimates of diffusion in (U,Th)02 activation energies.................................. 79

A-6. Kinetic parameters of mixed urania-thoria oxides .................................. 79

B-1. Uo 368Th0 63202+x isothermal 4000C air oxidized compared to JCPDS standards................ 80

B-2. XRD peak intensities for 4500C isotherm of co-milled 23.6% U02 fragments ............... 81

B-3. XRD peak intensities for 5000C isotherm of co-milled 23.6% U02 fragments ............... 82

B-4. XRD peak intensities for 550C isotherm of co-milled 23.6% U02 fragments ............... 83

B-5. XRD peak intensities for 4000C isotherm of co-milled 36.8% U02 fragments ............... 84

B-6. XRD peak intensities for 4500C isotherm of co-milled 36.8% U02 fragments ............... 85

B-7. XRD peak intensities for 5000C isotherm of co-milled 36.8% U02 fragments ............... 86

B-8. XRD peak intensities for 3750C isotherm of co-milled 50.0% U02 fragments ................. 87

B-9. XRD peak intensities for 4000C isotherm of co-milled 50.0% U02 fragments................. 88

B-10. XRD peak intensities for 4250C isotherm of co-milled 50.0% U02 fragments ............. 89

B-11. XRD peak intensities for 4500C isotherm of co-milled 50.0% U02 fragments ............... 90
















LIST OF FIGURES


Figure page

3-1. O xalate path co-precipitation................................................................................. 2 1

3-2. Optical micrograph of nominal 20% UO2 calcined oxalate-synthesized powder ............... 22

3-3. X-Ray Diffraction evolution of Uo 05Tho 9502 oxalate synthesized powder.......................... 24

3-4. Vulcan Muffle furnace used for drying, decomposition, and calcination ............................ 25

3-5. Lindberg high temperature tube furnace ................................................. ...... .........25

3-6. Ammonium hydroxide path co-precipitation...... ............ .. ......... ............... 26

3-7. Ammonium hydroxide synthesized (Uo 2Tho s)02+x before (1) and after (r) calcinations....... 27

3-8. Optical micrographs of Alfa Aesar UO2 (1) and ThO2 (r) ................................................ 28

3-9. 8000M SPEX Certiprep Mixer/Mill (1) and zirconia mill jar (r) ............... ...................29

3-10. Pellets prepared using 20% U powder synthesized by the oxalate technique ................... 31

3-11. Optical micrographs of nominal U 2Th 802+ ............................................. ..................37

3-12. TA Instruments TGA 2050 thermogravimetric analyzer................................................ 38

4-1. Lattice parameter evolution with respect to U02 content............................................ 44

4-2. Nonisothermal (Uo 236Th0 764)02 oxidation TGA data at heating rates of 1, 3, and 5C/min. 47

4-3. Nonisothermal (Uo 368Tho 632)02 oxidation TGA data at heating rates of 1, 3, and 5C/min. 48

4-4. Nonisothermal (Uo 5ooTho 500)02 oxidation TGA data at heating rates of 1 and 5C/min......48

4-5. Nonisothermal U02 oxidation TGA data at heating rate of 30C/min ............................... 49

4-6. Isothermal oxidation TGA data for (Uo 236Tho 764)02 fragments (90 250 pm) ................. 50

4-7. Isothermal oxidation TGA data for (Uo 368Tho 632)02 fragments (90 250 pm) ................. 50

4-8. Isothermal oxidation TGA data for (Uo 5ooTho 500)02 fragments (90 250 pm) ................... 51

4-9. Isothermal oxidation TGA data for U02 fragments (90 250 pm).............................. 51










4-10. Mean uranium valence for isothermally oxidized (UyThly)02 and U02............................54

4-11. Isotherm at 4500C of (Uo 236Tho 764)02 oxidation fit to 2D and 3D Diffusion models......... 57

4-12. Isotherm at 5000C of (Uo 236Th0 764)02 oxidation fit to 2D and 3D Diffusion models......... 57

4-13. Isotherm at 5500C of (Uo 236Tho 764)02 oxidation fit to 2D and 3D Diffusion models......... 58

4-14. Arrhenius plot of (Uo 236Tho 764)02 isotherms fit to 3D Diffusion models .......................... 58

4-15. Nonisotherm at lC/min (Uo 236Tho 764)02 Arrhenius plot fit to 2D and 3D diffusion ........ 59

4-16. Nonisotherm at 30C/min (Uo 236Tho 764)02 Arrhenius plot fit to 2D and 3D diffusion ........ 59

4-17. Nonisotherm at 50C/min (Uo 236Tho 764)02 Arrhenius plot fit to 2D and 3D diffusion ........ 60

4-18. Model free (Uo 236Th0 764)02 isotherms plotted at a = 0.4, 0.5, 0.6 ................................ 60

4-19. Isotherm at 4500C for (Uo 368Tho 632)02 oxidation fit to 3D Diffusion .............................. 63

4-20. Isotherm at 4750C for (Uo 368Tho 632)02 oxidation fit to 3D Diffusion .............................. 64

4-21. Isotherm at 5000C for (Uo 368Tho 632)02 oxidation fit to 3D Diffusion .............................. 64

4-22. Arrhenius plot of (Uo 368Tho 632)02 isotherms fit to 3D Diffusion .................................. 65

4-23. Nonisotherms at 1, 3, and 50C/min (Uo 368Tho 632)02 Arrhenius plot fit ............................. 65

4-24. Model free (Uo 368Tho 632)02 isotherms plotted at a = 0.4, 0.5, 0.6 ........................ 66

4-25. Isotherm at 3750C for (Uo 5oTho 50)02 oxidation in 3D Diffusion and Avrami-Erofe'ev..... 68

4-26. Isotherm at 4000C for (Uo 50Tho 50)02 oxidation in 3D Diffusion and Avrami-Erofe'ev..... 69

4-27. Isotherm at 4250C for (Uo 50Tho 50)02 oxidation in 3D Diffusion and Avrami-Erofe'ev..... 69

4-28. Isotherm at 4500C for (Uo 50Tho 50)02 oxidation in 3D Diffusion and Avrami-Erofe'ev..... 70

4-29. Arrhenius plot of (Uo 5ooTho 500)02 isotherms in 3D Diffusion reaction model................... 70

4-30. Nonisotherms at 1 and 5C/min (Uo 5ooTho 500)02 Arrhenius plot ..................................71

4-31. Model free (Uo 5ooTho 500)02 isotherms plotted at a = 0.4, 0.5, 0.6 ............... .......... 71

4-32. Model free U02 isotherms plotted at a = 0.4, 0.5, 0.6....... .........................................73

B-1. XRD pattern for 4500C isotherm of co-milled 23.6% U02 fragments.............................. 81

B-2. XRD pattern for 5000C isotherm of co-milled 23.6% U02 fragments.............................. 82

B-3. XRD pattern for 5500C isotherm of co-milled 23.6% U02 fragments.............................. 83










B-4. XRD pattern for 4000C isotherm of co-milled 36.8% U02 fragments.............................. 84

B-5. XRD pattern for 4500C isotherm of co-milled 36.8% U02 fragments.............................. 85

B-6. XRD pattern for 5000C isotherm of co-milled 36.8% U02 fragments.............................. 86

B-7. XRD pattern for 3750C isotherm of co-milled 50.0% U02 fragments.............................. 87

B-8. XRD pattern for 4000C isotherm of co-milled 50.0% U02 fragments.............................. 88

B-9. XRD pattern for 4250C isotherm of co-milled 50.0% U02 fragments.............................. 89

B-10. XRD pattern for 4500C isotherm of co-milled 50.0% U02 fragments.............................. 90

C-1. Unpolished Uo 05Tho 9502+x pellet surface. Scale bar is 21 microns................................ 91

C-2. Unpolished Uo 05Tho 9502+x pellet surface. Scale bar is 20 microns................................ 91

C-3. Broken surface of Uo 2Tho 802+x pellet ..................................................................... 92
















Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science

EXPERIMENTAL DETERMINATION OF THE DRY OXIDATION BEHAVIOR OF A
COMPOSITIONAL RANGE OF URANIUM-THORIUM MIXED-OXIDE PELLET
FRAGMENTS

By

Lisa Argo

December 2003

Chair: Ronald H. Baney
Major Department: Materials Science and Engineering

Oxidation of (UThiy)O2 (y = 0.236, 0.368, 0.500) solid solutions was investigated using

thermal gravimetric analysis and compared with U02. The UO2 and ThO2 powders were ground,

pressed into pellets, and sintered at 16500C in a reducing atmosphere. Gravimetric oxidation data

for all samples, including UO2, exhibited single-step behavior. This was interpreted as the effect

of the low surface-to-volume ratio of the fragments used (specific surface areas were 0.01 to 0.02

m2/g) on the relative contributions of surface and bulk oxidation reactions, with bulk reactions

dominating in this case compared to powders with higher surface area. Model-fitting kinetic

analysis suggested diffusion (y = 0.236, 0.368) and possible nucleation and growth (y = 0.500) as

possible mechanisms. Activation energy, E, for isothermal oxidation calculated by model-fitting

increased with uranium content, 62.7 17.9 and 171 8 kJ/mol for y = 0.236 and 0.368,

respectively. X-ray diffraction patterns and uranium valence calculations do confirm that the

ultimate oxidation is inhibited in the (U,Th)O2 compared to U02. Lattice structure remained

cubic fluorite and did not undergo any phase transformation. Cation valencies indicated that









uranium does not proceed to its maximum oxidation state. The absence of quantitative particle

and grain size data restricted the models in fully describing system complexity.
















CHAPTER 1
INTRODUCTION

The objective of this project was to determine the long-term stability of ThO2/U02 high

level waste. Radionuclide leaching from spent U02 fuel is a major long-term storage concern for

radiological materials. Mixed oxide (U,Th)02 fuels were considered an alternative because of the

natural abundance of ThO2, nonproliferation benefits, and potentially improved long-term spent-

fuel storage capability. Oxidation behavior of (U,Th)O2 solid solutions was measured as a

function of the oxide composition in order to determine the relative advantage these materials

possess over conventional UO2.

Uranium dioxide oxidizes to a multitude of naturally occurring off-stoichiometric phases,

(i.e., U307, U409, and U308). These phase changes induce pellet cracking (due to a 36% volume

increase associated with the UO2 U308 transformation) and generally facilitate radionuclide

release from the crystal lattice. Additionally, the higher oxidation states, U(V) and U(VI), form

water-soluble species. Release of hazardous radionuclides from the spent-fuel into the

surrounding environment, therefore, may occur through surface oxidation or catastrophic fracture

during long-term storage. Since the life expectancy of uranium and thorium based high-level

radioactive waste greatly exceeds any reasonable length of time to obtain experimental data,

predictions of waste behavior rely upon accelerated tests and/or models based on an

understanding of system behavior. Half-lives reported for uranium-238, -235, and thorium-232

are 4.46x109 years, 7.04x108 years, and 1.4x101 years, respectively [Win03].

For instance, after gamma irradiation of CANDU polycrystalline U02 pellets in oxygen-

rich/-free and moisture-rich/-free atmospheres, Sunder and Miller [Sun96] observed U(VI) by x-

ray photoelectron spectroscopy (XPS) and multiple uranium oxide phases (U02, U307, U308,










U8019, U03, U16037, etc.) by x-ray diffraction (XRD). Equivalent to gamma fields associated

with 10- to 20-year-old used CANDU fuel [Sun96], irradiated pellets did not reach temperatures

in excess of 1500C during Sunder and Miller's 2-day experiments. Uranium oxidation thus was

restricted to surface crystalline growth. Catastrophic pellet fracture was not observed. Serrano et

al. [SerOl] evaluated oxidized LWR U02 spent-fuel pieces in wet leaching experiments. Initial

uranium fractional release was clearly related to oxidation level (starting O/M) of the fuel.

Fission products identified in the leachate solution were plutonium-239, molybdenum-98,

cesium-137, strontium-90, and technetium-99, where all but Pu-239 had release rates greater than

for uranium. Serrano et al. [SerO1] attributed the increased fractional release rates to mobility

along grain boundaries; and higher solubility of some fission products (like Tc or Cs). Shoesmith

[ShoOO] identified unoxidized U02 as a "slow dissolving semiconducting oxide" with the major

rate-controlling process as surface ionic species formation by surface charge transfer or alteration.

Most (> 90%) fission products and actinides generated in-reactor, in fact, are retained in the U02

fuel matrix and expected to be released as the fuel degrades in storage [ShoOO]. The sensitivity of

UO2 solubility when oxidized, therefore, makes fuel dissolution and radionuclide release

dependent on repository redox conditions (i.e., oxidant supply) [ShoOO].

Currently, various waste forms and packages encapsulate, stabilize, shield, and otherwise

prevent or minimize high-level radioactive waste release while in storage at geologic repositories

or interim sites. Two waste-form systems, borosilicate glass and Synroc, are currently used;

although glass is still the preferred method for civil high-level radioactive waste. Both systems

immobilize radioactive actinides within a material matrix. The liquid properties of glass permit

the glass matrix to accommodate impurities nearly indiscriminately. Synroc, which is in

developmental stages in the U.S. for military waste and surplus military plutonium storage,

comprises titanate minerals (such as titanium dioxide, hollandite, zirconolite, and perovskite).

Synroc, like vitrified glass, is globally recognized for its chemical durability and resistance to

leaching at high temperatures.










In the same vein as glass vitrification and Synroc, mixed (U,Th)02 oxide solid solutions

are pursued in this study as a material "barrier" specifically to uranium release. Thorium oxide,

like the actinide oxides used for nuclear-reactor fuels, has a fluorite-type cubic lattice structure.

Cations are arranged in a face-centered cubic close-packing sublattice. Anions occupy all

tetrahedral interstitial positions, forming a simple cubic sublattice, with cations occupying one

half the interstitial sites, and the remainder are vacant. Large octahedral interstitial holes remain

where interstitial ions can be introduced to form a hyper-stoichiometric oxide. Diffusion

characteristics of the fluorite structure indicate that anionic mobility (diffusion) is greater than

cationic and ionic self-diffusion, because both anions and cations are sensitive to deviations from

stoichiometry [And83]. Deviations from ThO2 stoichiometry, however, are hindered. Unlike

UO2, thorium in ThO2 is present in its maximum stable oxidation state, Th(IV), and cannot both

accommodate excess anions and maintain charge neutrality. In fact, ThO2 oxidation is largely

independent of oxygen partial pressure [And76, And79, Haw68]. Additionally, thorium oxide is

structurally and atomically similar to uranium oxide, unlike borosilicate glass and titanate

minerals.

The stability and physical properties of the thorium oxide matrix present an alternative to

pure U02 fuel in-reactor and may potentially replace other waste-form systems in minimizing

radionuclide release in storage. Because ThO2 is in its maximum oxidation state, thorium oxide

is stronger and more durable than U02 in various ways. Solubility of ThO2 over a wide range of

aqueous solutions is extremely low as compared to UO2 under reducing conditions [Tay96]. The

ThO2 (like UO2) is not susceptible to radiation-induced phase transformation to an amorphous

state [Tay96]. Grain growth, a major cause of fission product release and governed by cation

diffusion, is expected to be similar or lower than UO2. Since thorium oxide is a better thermal

conductor, with a higher melting point and slower cation diffusion [And83, Tay96], it is expected

to run cooler and undergo less grain growth for a given power rating and fuel geometry [Tay96].










Once a waste-form system has been used, the spent-fuel is further contained and

packaged in systems (such as the use of lead canisters and other shielding or absorbent materials

immediately surrounding an individual waste container). Despite the development of long-lived

nuclear waste containers, the final barrier to radionuclide release is probably the waste-form

system. Wet and dry waste vault/repository conditions are expected. For instance, according to

the Nuclear Waste Management Organization, Canada initially stores spent nuclear fuel in water-

filled pools called Irradiated Fuel Bays; and 7 to 10 years later transfers it to dry storage facilities.

Assessing overall repository performance therefore demands an understanding of potential fuel

degradation in wet and dry conditions.

The scope of this research consists of the manufacture and dry oxidative behavior of

urania-thoria oxide solid solutions under accelerated dry-storage conditions as compared with

pure U02. Kinetic analysis is used to assess mixed (U,Th)O2 oxide behavior versus UO2.

Various processing methods were considered, and a significant portion of this study was

dedicated to evaluating and selecting a synthesis technique that would yield a homogenous

product. Our study did not attempt to assess the in-reactor performance of urania-thoria fuels.
















CHAPTER 2
REVIEW OF LITERATURE

Current Status of UO2 and ThO2 Research

Uranium Dioxide

Natural uranium is primarily composed of two isotopes (U-238 and the fissionable U-

235). The fissionable isotope characterizes the economical value of nuclear fuel. The Oakridge

National Laboratory Review website [Gab93] reported that nuclear fusion generates

approximately 2x109 kWh/ton and coal combustion 6150 kWh/ton. From a mass to power-

generated standpoint, uranium is undoubtedly superior. In generating this power, however, the

radioactive and toxic byproducts generated dictate a complex waste system to prevent harm to the

surrounding environment. The final waste form, after mechanical barriers have degraded, will

ultimately be a material system that must immobilize mobile species, such as uranium.

Defect structure. Uranium dioxide has a cubic fluorite-type lattice structure, like other

actinide oxides such as ThO2 and PuO2. In this crystal structure, diffusion characteristics favor

anion diffusivity more than cation diffusivity. Point defects considered for diffusion in fluorite-

cubic oxides are the oxygen Frenkel pair, cation Frenkel pair, and Schottky trio. Written in

Kroger-Vink notation, defect-formation energies for stoichiometric UO2 are shown in Table 2-1

suggesting that the most dominant defect is the oxygen Frenkel pair [And83].

Ando and Oishi [And83] reported that stoichiometric UO2 and ThO2 showed similar

diffusion characteristics with activation energies ranging between 200 and 275 kJ/mol, which

they interpreted to be intrinsic diffusion. The similarity in stoichiometric UO2 and ThO2 would

further support that anion diffusivity is closely tied to the simple cubic anion sublattice, almost









regardless of cation species. Uranium self-diffusion in stoichiometric U02, in turn, is closely

related to the face-centered cubic cation sublattice.

Table 2-1. Defect formation energies for UO2
Defect Kroger-Vink Notation Energy (kJ/mol)
Oxygen Frenkel pair [VJ ] [O, ] 482
Uranium Frenkel pair [V 4] [U4' ] 1784
Schottky trio [V 4][Vo']2 993

Hyperstoichiometry. In the case of hyperstoichiometric uranium dioxide, anion and

cation mobility are further enhanced. Qualitatively, it is known that using oxygen-rich materials

facilitate sintering of U02, thereby enhancing cation diffusion. Ando and Oishi [And83] reported

published activation energies 90 to 100 kJ/mol for oxygen self-diffusion in UO2+x that appeared to

be independent of the degree change ofx.

Hawkins and Alcock [Haw68] measured cation tracer-concentration depth profiles by

alpha-ray spectrometry in single and polycrystalline hyperstoichiometric (UO2+x, x = 0.01, 0.03,

0.10, 0.15) uranium oxide. They observed that volume contributions dominated near surface,

whereas grain boundary contributions to diffusion in polycrystalline samples were virtually

identical to the measured profile at large penetrations. Log plots of U diffusion coefficient as a

function of off-stoichiometry revealed orders of magnitude increase in the cation diffusion

coefficient with small departures from stoichiometry, regardless of single or polycrystalline

sample [Haw68]. Insufficient data prevented Hawkins and Alcock from ascertaining the

mechanism of cation diffusion in UO2+x.

Oxidation behavior. The oxidation of U02 is commonly observed as a two-step process

[Bla58, Ban68, McE97a, McE97b]. Initially, excess oxygen is interstitially accommodated into

the cubic fluorite-type UO2 structure [And83], resulting in an oxygen hyperstoichiometry and

ultimately distorting the cubic lattice. Given a sufficient supply of oxygen and thermal energy to

the system, the O/U ratio increases until the cubic matrix no longer supports the excess anions.

The material then undergoes a phase transformation to U307/U409 that Willis [Wil87] denotes as










an ordered superlattice structure, and subsequently transforms to orthorhombic U308. Boase and

Vandergraaf [Boa77] observed that tetragonal U307/U409 remained the stable phase at

temperatures below 250C, whereas U3sOwas the stable end product at higher temperatures.

Table 2-2 lists the three competing reactions that are suggested as the principal mechanisms that

occur for UO2 oxidation [Boa77].

The transformation of U02 to U307 occurs initially and is a surface reaction that proceeds

at the solid/gas interface, with a limiting thickness of -22 nm. Boase and Vandergraaf [Boa77]

reported that the oxidation of U307 to U308 at 235C was significantly slower than the U02 to

U308 reaction. The two bulk reactions that form U308 therefore occur simultaneously and in

competition with each other, particularly in high surface area powders.

Blackburn et al. [Bla58] measured the rate of UO2 oxidation to determine the reaction

mechanism. They observed that the weight gain versus time curve for the U02 to U02 33 reaction

did not indicate any apparent change in mechanism. Below 3000C, a surface layer of single

phase y-U307 was immediately formed and further oxidation was controlled by oxygen diffusion

through this phase, such that the oxidation rate was inversely proportional to oxide thickness.

The second oxidation stage, which has been shown to be a nucleation and growth process,

[Bla58] exhibited an induction period, a gradually increasing reaction rate, and finally a region in

which the rate slows as the final composition approaches UO2 67.

Table 2-2. Oxidation reactions of UO2+
Oxidation reaction Comments
(1) 6U02 + 02 2U307 Fast; surface only
(2) 2U307 + 02 2U308 Slow at temperatures < 2500C; surface only in coarse powders
(3) 3U02 + 02 -) U308 Faster than reaction (2); occurs in the bulk only


Surface area dependence. The individual contributions of these competing reactions

(#2 and #3 from Table 2-2) in the U02 system are dependent on the specific surface area of the

sample, and separating them is often difficult. Because both surface and bulk reactions occur

simultaneously, the surface to volume ratio of the sample has a significant bearing on the relative










contributions of the various oxidation mechanisms. In fact, Dobrov et al. [Dob98] cited work that

showed the surface oxygen exchange is a relatively slow process and is the rate-limiting step, not

solid-state diffusion, of fuel oxidation in thin samples where surface contribution is greater than

bulk. Experiments reported using powders with specific surface areas ranging from 0.61 to

19 m2/g [Ban68] indicated a direct relationship between the final O/U ratios of the oxidized

materials and the specific surface area of the samples. Taylor et al. [Tay98] studied the rate of

U308 formation as a function of pellet surface roughness at 2500C and noted that reaction times

increased by a factor of four with increasing roughness, corresponding to an increase in the

nucleation and growth rate constant, K, by two orders of magnitude. A nonlinear relationship

between K and surface roughness was observed such that K was independent of surface roughness

with very fine (< 1 p.m) or coarse (18-100 pm) polishing agent particle size. For 1-18 p.m particle

size polishing media, K values increase with media particle size. Taylor et al. attributed this

correlation of K with roughness to a combination of at least three factors: increasing macroscopic

surface density of nucleation sites, increasing microscopic surface density of nucleation sites, and

preferred oxidation in the <111> direction. The 36% volume expansion is largely accommodated

in the <111> UO2 direction, which becomes the <001> direction (c axis) of orthorhombic U308.

Activation energies. A wide range of experimentally determined activation energy

values has been reported for the oxidation of U02. The results of previous UO2 oxidation studies

have been summarized in the literature [McE97a-b]. See Appendix A. Activation energies for

the formation of U307/U409 on unirradiated UO2 fuel and spent LWR fuel have been reported to

range from 90-120 kJ/mol [McE97a]. Values for the oxidation of U02 fuel and spent-fuel to

U308, determined with a variety of experimental techniques, exhibit a much larger range, from

48-172 kJ/mol [McE97b]. The broad range in experimental values for UO2 -> U308 oxidation

highlights the complexity of the oxidation process for this material, such as grain boundary

diffusion, off-stoichiometry (O/M), oxidizing conditions, and/or in-reactor conditions.










Thorium Oxide

A naturally abundant material, thorium oxide has found use in a variety of products from

portable gas mantles to catalyzing petroleum cracking. It is also an identified fertile material used

for producing nuclear fuel. Although not fissile, thorium-233, like uranium-238, absorbs slow

neutrons to produce fissile uranium-233. The comparison with uranium reveals a host of

properties that can be exploited in the waste end of the nuclear fuel cycle.

Thorium oxide possesses a molecular weight, cation radius, and lattice structure similar

or identical to U02. Not surprisingly, ThO2 forms solid solutions with U02 at all U/Th ratios.

Table 2-3 summarizes some material properties of ThO2 and UO2.

Table 2-3. Material properties of ThO2 and U02
Material property ThO2 U02
Metal electron configuration [Rn] (-'s2 [Rn] 5f36d17s2
Metal atomic number 90 92
Metal atomic weight 238.0289 g/mol 232.0381 g/mol
Metal atomic radius (neutral) 1.8 A 1.75 A
Cation radius* 1.18 A 1.14 A
Melting point 33900C 28270C
Density+ 9.986 g/m3 10.977 g/m3
Lattice structure Fm3m (225) Fm3m (225)
Lattice parameter+ 5.600 A 5.467 A
Lattice energy 10397 kJ/mol 10644 kJ/mol
Radioisotope half-life 1.4 x 10'0 y (Th232) 4.46 x 109 y (U238)
* radii for coordination number (CN) = 8 and valence = 4+ [Ric92]
+JCPDS Diffraction data 04-0556 and 41-1422, respectively

Defect structure. Transport phenomena within the thorium oxide lattice are important to

predict mobility behavior of inclusions, such as uranium and fission products generated in-

reactor. As with other fluorite-structured materials, anion Frenkel defects (oxygen vacancy-

interstitial pair) are thought to be the predominant intrinsic disorder, with cation vacancies and

holes as minor defects [And76, Col83]. Calculated effective formation and migration energies

for intrinsic disorder reported by Colbourn and Mackrodt [Col83] further support thorium oxide

stability over uranium dioxide, as summarized in Table 2-4, where cation interstitials appear to be

far too energetic to play any significant role in the cation disorder of thorium oxide.










Table 2-4. Cation and anion formation and migration energies in ThO2
ThO2 (eV)
Effective formation energies for intrinsic disorder
Anion vacancy -3.00
Anion interstitial -3.00
Cation vacancy 5.93
Cation interstitial 13.87
Anion migration
Vacancy 0.78
Direct exchange 5.80
Direct interstitial 3.27
Interstitialcy mechanism 0.92
Direct interstitial migration of O1 4.28
Cation migration
Vacancy 7.04
Divacancy 5.36
Trivacancy 6.35

Oxidation behavior. Results published by Ando et al. [And76, And79] of single crystal

99.99% purity ThO2 indicate that oxygen diffusion kinetics are rate limited by surface exchange

at temperatures 9730C to 15930C with respect to time. Concavity was observed at initial and

final times when fractional 018 uptake is plotted as a function of diffusion time. However,

oxygen self-diffusion, with respect to temperature dependence, did not appear to be influenced by

specimen size and pre-annealing. The temperature dependence of oxygen self-diffusion split into

a high (> 11000C) and low (< 11000C) region, which Ando et al. [And76] attributed to intrinsic

and extrinsic oxygen diffusion, respectively. Further work by Ando and Oishi [And79] show

ThO2 ionic conduction primarily comes from oxygen self-diffusion and is independent of oxygen

partial pressure.

Impurities in thoria. Experimental investigations of the presence of impurities in the

thorium oxide lattice are largely concerned with electrical conductivity of solid solutions to shed

light on the defect structure. Colboum and Mackrodt [Col83] reported vacancy binding energies

for di- and tri-valent cation impurities, including Th3+, which give rise to anion vacancies for

ionic conduction in the thorium oxide lattice. Di-valent impurities gave rise to greater

substitution and vacancy binding energies than tri-valent; and decreased in magnitude with

decreasing dopant radii. [Col83] As would be expected, smaller impurities and, particularly,










valency close to 4+ ease inclusions into the lattice. It stands to reason that U4+, with its similar

radii to Th4+, and any of the lanthanide or actinide elements, particularly those with 4+ valency,

would be energetically favorable impurities accommodated in the thorium oxide lattice.

Uranium-Thorium Mixed Oxide

The oxidation behavior of (U,Th)O2 solid solutions has been studied previously. Within

a composition range of 24% to 90% moles of uranium dioxide, (U,Th)O2 phases are always

single phase cubic fluorite type when oxidized at temperatures below 2000C [And54]. In high

uranium concentration (> 95% mol) unoxidized pellets at room temperature, lattice parameters

increased linearly with decreasing uranium content, reflecting Vegard's law that implies atomic

volume is conserved regardless of local lattice distortions [Tsu98]. Chandramouli et al. [Cha98]

noted the cubic fluorite structure was maintained irrespective of uranium content of unoxidized

(U,Th)O2 solid solutions.

Oxidation of compositions containing less than 50% mol UO2 yielded only fluorite

phases under any conditions. Mixed oxides containing greater than 50% mol UO2, however,

formed second non-cubic phases, hexagonal or orthorhombic U308 [And54, Cha98]. X-ray

diffraction patterns indicated the presence of a U308 second phase, which disappears upon

reduction heating at 13000C, results in uranium depletion from the solid solution [Cha98].

Anderson et al. [And54] observed that mixed oxides containing less than 78% UO2 were stable in

air up to 14000C, except for 66% UO2 which lost its cubic structure in high-pressure oxygen

conditions. Otherwise, high uranium compositions, such as Uo 9Tho 102, evidenced x-ray pattern

broadening (indicative of unit cell contraction), breakdown of cubic lattice symmetry when

oxidized below 2000C, phase separation above 2000C oxidation, and a final orthorhombic U308-

like structure at 6000C [And54]. Anderson et al. deduced that for UO2 concentrations from 15 to

78% mol, excess oxygen entered interstitial sites in the fluorite lattice based on changes in unit-

cell density, which increased with oxidation in agreement with theoretical calculations. Unit cell









parameters, however, demonstrated a parabolic relationship with uranium valence, such that a

minimum was observed near U valence = 5. Anderson et al. interpreted this behavior as a

competition between initial ionic attraction and volumetric accommodation of excess anions.

Aronson and Clayton [Aro60] examined (U,Th)O2 compositions synthesized by

ammonium hydroxide co-precipitation, with UO2 contents ranging from 30 to 90% mol and x-ray

diffraction patterns confirmed a single phase fluorite-type structure present at all compositions.

Following controlled oxidation and annealing at 8000C, it was observed that the added oxygen

was accommodated without destroying the fluorite structure.

The activation energy of (U,Th)O2 oxidation has been studied using thermogravimetric

techniques [Ant00]. All of the samples studied exhibited single-step weight-gain curves. For low

uranium content (UyThliy)O2 powders (y = 0.15 and 0.30), the average activation energy for

oxidation was calculated to be 45 kJ/mol. Higher uranium content powders (y = 0.72 and 0.77)

had average activation energies of 74 kJ/mol. The higher activation energy for higher uranium

content powders was attributed to the phase separation that occurs when these materials are

oxidized at high temperatures.

Oxidation potential. Ugajin [Uga82] prepared (U,Th)02 samples with 5%, 10%, and

20% mol U02 to measure the oxygen potential of non-stoichiometric mixed oxides as compared

to UO2+x at 10000C to 12000C in C02/CO atmosphere. Oxygen potentials of mixed oxides, like

UO2, increased with increasing O/M at all temperatures and compositions. Assuming Th valence

is constant at 4+, Ugajin observed oxygen potentials increased with increasing Th content and U

valency more so than UO2+x at greater deviations from stoichiometry. Supplemented by

Anthonysamy et al. [Ant97] and Arima et al. [AriOO] measurements of 1%, 3%, 5%, 54% and

90% mol U02, the compositional range of (U,Th)O2+x oxygen potential data dependence upon

uranium valence and the U/(U+Th) ratio was further affirmed.










Matsui and Naito's [Mat85] work with 20% and 40% mol U02 at 10000C to 11000C in

H2/C02 atmosphere further interpreted the differences between (U,Th)02+x and U02+x oxygen

potential behavior according to the dependence ofx on oxygen partial pressure expressed as x c

P(O2)1/n. Assuming a complex defect (2:2:2) model with two different interstitial oxygen types

(O1 and Ob) and one oxygen vacancy (Vo), Matsui and Naito showed UO2+x to have three

different dependencies ofx upon P(02) and four regions for (U,Th)02+x summarized in Table 2-5

[Mat85].

Table 2-5. Composition dependence on oxygen partial pressure for UO2+x and (U,Th)02+x
Phase Composition n for x o P(O2)/n Defect Model
U02+x x < 0.003 2 Neutral defect {20laOlb2Vo}
0.003 x > 0.006 2 {2(OlaOb2Vo)}'
(U,Th)2+x x < 0.001 2 Neutral defect {20laOlb2Vo}
0.001 < x < 0.003 4 {20laO b2Vo}'
0.003 < x < 0.008 12 {20laO b2Vo}'
x > 0.008 4 {2(OlaOb2Vo)}"'

Polycrystalline diffusion. Furuya [Fur68] measured cation (237U) diffusion in ThO2 and

U02-ThO2 polycrystalline pellets with average grain sizes over 60 pm over temperatures ranges

18000C to 20000C and 18000C to 23000C, respectively. As expected, lattice diffusion made a

significant contribution near surface, resulting in a non-linear concentration profile. Deeper

penetration regions show 237U concentration varied linearly with distance, characteristic of grain

boundary diffusion. Excluding the section nearest to the free surface, Furuya [Fur68] calculated

lattice and grain-boundary diffusion contributions. Comparison of the associated activation

energies showed that lattice diffusion through (U,Th)02 (360 kJ/mol) is greater than ThO2 (320

kJ/mol). Furuya deduced the smaller lattice spacing in (U,Th)02 caused the potential barrier at

the saddle point to increase compared with ThO2. Both values were greater than values reported

in nominally stoichiometric U02 (304 kJ/mol), with implications of the influence of off-

stoichiometry on cation diffusion. Activation energies for grain boundary diffusion (269 and 201









kJ/mol), as expected, were significantly lower than lattice diffusion and also saw the same

relationship between (U,Th)O2 and ThO2, respectively.

Ando et al. [And85] measured uranium oxide evaporation mechanism and rate-

controlling step for polycrystalline Th 9oUo 1002 05 and Tho 75Uo 2502 13 after heating at 16500C in

flowing air for 32 h. UO2 and ThO2 concentrations varied to a depth of approximately 20 pm and

80 p.m from the surface in the 10% UO2 and 25% UO2 samples, respectively. The concentration

profiles showed that uranium oxide preferentially vaporizes. Pores and cracks were also observed

after evaporation annealing, with the cracks and voids decreasing with increasing depth. Cation

diffusion, further enhanced by grain boundaries, was identified as the rate-controlling mechanism.

Background of Synthesis Methods

The powder-pellet route involves generation and handling of fine powder or particles

(<1 pm) and is hence associated with the problem of radiotoxic dust hazard and fire hazard

(applicable for carbide and nitride powders). Further, the fine powders are not free flowing and

pose problems in remote and automated fabrication. Microhomogeneity of fissile species in

mixed oxide is not fully obtained since the oxide powders are mechanically mixed. The

alternative sol-gel type processes use dust-free, free-flowing and coarse (100 to 2000p.m)

particles as starting materials for pellet making.

The choice of uranium oxide as a nuclear fuel is because of its high melting point (2828 +

200C), corrosion resistance to radiation damage, and irradiation stability. The physico-chemical

characteristics of U02, such as density, surface area, pore structure, grain size, crystallite size,

sphericity, and oxygen to uranium ratio, depend mainly on the preparation method [Abd90]. Sol-

gel "wet" methods are typically preferred over traditional dry powder milling for producing

spherical particles. The sol-gel processes have been developed and applied at Oak Ridge

National Laboratory (USA), AERE (United Kingdom), KEMA (Netherlands), CNEN (Italy),









KFA Julich (FRG), and Tokai, Japan for high density thorium oxide and/or uranium oxide

microspheres [Abd90].

The oxalate precipitation method is the primary process for commercially produced

thorium oxide. White et al. [Whi81] investigated precipitation temperature, agitation method, and

digestion time to refine powder sinterability and density without milling for ThO2 and 25% UO2-

ThO2. A 100C temperature, mechanical stirring, and 15 min digestion yielded the most sinterable

powder with 96% theoretical densities (TD) without milling. Temperature was identified to have

the most effect upon particle morphology, surface area, crystallite size, and sinterability. The

lower (100C) digestion temperature, however, yielded cubic particles less than 1 pm in size. At

700C, the particles were square platelets varying in size from 1 to 3 [pm along the square edge.

Longer digestion times rounded off the edges and generally produced more uniform particle sizes

Ganguly et al. [Gan86] examined thorium oxide and uranium oxide microspheres by

well-established Societa Nationale Matanodoti (SNA) and Kernforschungsanlage Julich (KFA)

external gelation and KFA and Keuring Electrotechnische Materialen Arnkem (KEMA) internal

gelation processes. Pellets made by each of the processes, however, presented relatively poor

densities and general behavior. Particle size, density, and crushing strength influenced pressing

characteristics. Specific surface area, crystallite size and additives determined sintering behavior.

Ganguly et al. noted that adding a sintering aid (CaO) and pore forming additive (carbon black)

improved pelletization and sintering (2 94% TD).

Chandramouli et al. [Cha98] observed that surface area of mixed oxide powders were

influenced by composition and calcinations method. As U content increased, surface area of

conventionally calcined powders decreased. Microwave heating also resulted in low surface area

powders with large crystallite sizes. Solid solutions with 15% mol U content calcined by a

graphite coupling agent yielded surface areas nearly twice that of conventional furnace calcined










powders. Residual carbon content, also, was greater in microwave-calcined powders than those

heated by conventional furnace.

Kinetic Analysis

The Arrhenius approach is used to interpret the rate dependence of thermal

decomposition of solids on temperature. Predicting how quickly the solid state (U,Th)02 system

approaches equilibrium does not necessarily require full understanding of the complicating

features of real systems. Simple approximate theoretical kinetic models can be used to interpret

experimental data and make predictions of (U,Th)02 behavior at the end of the nuclear fuel cycle.

Additionally, knowledge of U02 and ThO2 defect structures and oxidation mechanisms can lay

the groundwork for fully revealing the mechanisms that predominate (U,Th)02 oxidation (i.e.,

diffusion and/or nucleation and growth).

Kinetic reaction models. Basic solid state reaction models, f(a), are largely based on

simplified geometrical assumptions for reaction particles, such as geometry of reaction interface.

Contracting area and volume models assume its basis in the reaction interface and corresponding

spatial movement. The Johnson-Mehl-Avrami-Erofe'ev-Kolmogorov and Prout-Tompkins

kinetic reaction models describe nucleation and growth processes. Diffusion controlled reaction

models, which have largely been developed from gas-solid interactions, assume simplified

geometry such as spherical or cylindrical particle shapes and ball-and-stick lattice structures.

Arrhenius equation. Generally, homogeneous kinetics is assumed to behave according

to the simple differential kinetic rate and Arrhenius Equations 2-1 and 2-2, respectively. The

activation energy, E, is the energy barrier or threshold that must be overcome to enable the bond

redistribution steps required to convert reactants into products. The pre-exponential term, or

frequency factor, A, is a measure of the frequency of occurrence of the homogeneous reaction

situation. This is typically envisioned as including the vibration frequency in the reaction co-

ordinate. From a reaction dynamics treatment [Gal02], the activation energy is identified as the

difference between the energy of the molecules undergoing reaction and the overall average









energy. Degree of conversion or fraction reacted, rate coefficient, and kinetic model are

expressed as a (where 0 < a < 1), k(T), and f(a), respectively.

da
(2-1) = k(T). f(a)
dt


(2-2) k(T)= Aexp- -


The hypothesis is that the reaction involves only an "active part" of all reactant molecules

that, according to the Maxwell-Boltzmann distribution law, is an exponential function of

temperature [LvoO 1]. Galwey and Brown [Gal02] expressed that the Maxwell-Boltzmann

equation, which is applicable to homogeneous gaseous systems and a starting point for theoretical

explanation of Arrhenius behavior in homogeneous reactions, does not adequately represent the

energy distribution of the immobilized reactants of solid-state processes. Mechanisms, such as

three-body collisions or linked sequence of steps, which are unique to the constrained mobility of

solid-state systems, oppose a single contributing step during the entire reaction as a realistic

predictor of decomposition behavior. Additionally, reactant energies in liquid or solution are

associated with individual molecules, whereas band theory describes the energy distribution

within crystals. Garn [LvoO 1] noted that no discrete activated states can exist during the

decomposition of a solid. Subsequently, the Arrhenius kinetic parameters (A and E) have lost

their physical meaning. Because of the spatial constraints of solid-state processes, there are no

collisions of freely moving reactant molecules as defined by the frequency factor for

homogeneous reactions. Energy transfer through vibrational interactions within the crystal, also,

occurs so fast that no substantial deviations from the average energy can take place [LvoO ].

Limitations. The assumption that kinetic parameters are intrinsic constants, which

uniquely characterize a given solid-state process, however, is rather controversial and often leads

to misunderstandings since solid-state processes typically exhibit complex kinetics [MalO1].

Although mathematical descriptions can be determined for most solid-state processes to optimize









desired variables, the underlying mechanism is not necessarily easily obtained [Mal01]. Unlike

liquid and gaseous systems, reactants in solid-state processes are spatially constrained and place

additional complications on the system. Further difficulties that add to the complexity of solid-

state processes are inhomogeneous reactant distribution irregular shapes, polydispersity, shielding

and overlapping of reacting particles, or preferred orientations [MalO ]. Empirical rate laws

developed to address the anomalous reaction orders (kinetic exponents) in the rate law as a result

of these complexities are shown in Table 2-6.

Table 2-6. Rate laws for a simple process A P
Reaction in solids
da
= k (1- a) 1st order
dt
da
= k a(1 a) with autocatalysis
dt
Empirical rate laws
da
= k (1 a)" 1t order
dt
dac
= k a" (1 a)N with autocatalysis
dt
Fractional conversion a (0 < a < 1), such that [A] = 1-a and [P] = a.

Solid-state decomposition reaction rates are influenced by numerous factors that could

inhibit determination of kinetic parameters. Although it is typically assumed that sample

temperature is equal to that of the furnace, [Lvo 01] Smith and Topley's work where single

crystals in vacuum measured temperatures 4 to 8 K lower than the furnace. L'vov also previously

developed a program that theoretically computed the layer-by-layer temperature distribution in

powder samples decomposing in vacuum and foreign gases. Depending upon the total number of

layers, n, temperature deviations can differ significantly for powders and single crystals. For

instance, the temperature of the central layer of a Mg(OH)2 sample with n = 1000, 10,000 and

furnace temperature of 500 K or 600 K is actually 427 and 387 K, respectively [Lvo01]. Single

crystal Mg(OH)2 in vacuum with temperatures 550, 600, and 663 K is expected to have surface

temperatures of 549.6, 593 and 628 K, respectively [Lvo01]. Correcting kinetic parameters for










this self-cooling effect is naturally simpler for single crystals. This effect and number of layers,

otherwise, must be determined through measurement of grain size, grain number, powder mass,

crucible geometry, environmental conditions, and heating rate. With these corrections, it is

possible to explain the differences between kinetic parameters obtained under different

conditions. Other than the self-cooling effect, other reasons for scatter of reported values of A

and E may be differences between mathematical modes by different authors and difference in the

kinetic parameters for the nucleation and steady-state stages of decomposition [LvoO 1].

L'vov [LvoO1] reviewed a "physical" approach for interpreting thermal decomposition of

solids that is based on the Hertz-Langmuir prediction of proportional dependence of the

evaporation rate on the equilibrium partial pressure of the vapor which, in its turn, depends

exponentially on temperature. Most solid-state decomposition reactions proceed under conditions

far from equilibrium. The traditional "chemical" approach tackles this deviation from

equilibrium by connecting it to the theoretically unpredicted energy barrier, activation energy. In

other words, "non-equilibrium decomposition into equilibrium products" [Lvo01]. The

"physical" approach, however, contends that the reason for the deviation lies in the

decomposition of the reactant into primary non-equilibrium gaseous species different from those

at equilibrium. L'vov [LvoO 1] asserts this approach permits quantitative interpretation of the

mechanism of nucleation and the energy source supporting decomposition, retardation in the

presence of gaseous products, low vaporization coefficients, thermal stability, effect of self-

cooling, the Topley-Smith effect, and kinetic compensation effect.
















CHAPTER 3
MATERIALS AND METHODS

Material Synthesis

Three methods of powder fabrication were used, two of which employed sol-gel type

approaches for obtaining a solid solution. The sol-gel type processes yield a co-precipitate from

an aqueous solution of uranium and thorium nitrate salts followed by complexation with either an

oxalate or hydroxide group. The final method engaged in this study adhered to traditional co-

milled powder techniques.

Wet chemistry methods of material synthesis were initially pursued for its known

homogeneity advantages over powder-mixing techniques and minimization of dust hazard. Both

uranium and thorium are alpha emitters. The hazard, therefore, to be considered for workers are

dust particulates becoming airborne and entering the human body. Safety gear, such as protective

gloves, eyewear, and clothing, are necessary. Wet synthesis methods, however, present a reduced

hazard since the process is largely contained in liquid form prior to heat treatment. Additionally,

particle morphology of sol-gel type synthesis is largely spherical; presenting a free-flowing

powder that is ideal when filling dies for pellet making. In contrast, mixing by traditional

powder co-milling methods occurs entirely in powder form and generates platelet-like particles

with less desirable flow characteristics. The details of each material fabrication techniques are

presented in the ensuing sections

Oxalate Co-Precipitation

Since U02 and ThO2 are iso-structural, both have face-centered cubic CaFz-type lattices,

and have similar thermodynamic properties. Their fabrication processes therefore are nearly

identical. Aqueous oxalic acid is utilized as the complexation agent to form the mixed oxide. A










flow chart schematic in Figure 3-1 provides a brief synopsis of the steps entailed.

Uranium oxynitrate hexahydrate (U02(N03)2 6H20) and thorium nitrate hydrate

(Th(N03)4 2 xHzO) are supplied by Alfa Aesar. Analytical grade oxalic acid ((COOH)2 2H20)

and sodium formaldehyde sulfoxylate dihydrate (HOCH2SO2Na 2H20), also purchased from

Alfa Aesar, are the completing and reducing agents, respectively. A 115-volt motorized three-

blade stirrer supplied the agitation for the co-precipitation step.


Dissolve U-nitrate (s) in D.I. water.

Reduce U(VI) to U(IV) with NaFS*. Digest -24 h.

Separate & discard excess NaFS (s).

Add Th-nitrate (s) to retained filtrate.
4 --
Co-precipitate with excess Oxalic Acid (aq).

Filter precipitate.







Figure 3-1. Oxalate path co-precipitation. *NaFS is sodium formaldehyde sulfoxalate.

The desired molar ratio of uranium to thorium determines the starting quantities of solid

nitrate salts. Uranium (IV) nitrate solution is prepared by dissolving uranium (VI) oxynitrate

hexahydrate in a sufficient volume of demineralized water to yield an approximately 1M

concentration. Since thorium (IV) nitrate is later added in solid form, additional water is added to

the solution to achieve a thorium concentration of 1M. The nitrate solution is then reduced via

addition of a six-fold excess of the reducing agent at room temperature and covered for 24 h. The

solution changes colors from yellow to orange to dark green. After digestion, the excess reducing

agent solid is removed by vacuum-assisted filtration with Whatman #42 ashless filter paper and

thorium nitrate hydrate is dissolved into the nitrate solution.


























Figure 3-2. Optical micrograph of nominal 20% U02 calcined oxalate-synthesized powder
mounted with a collodion/amyl acetate solution at 50x magnification


To form the co-precipitate, a six-fold excess of the oxalic acid completing agent,

dissolved in demineralized water to a 1M concentration, is added via one of two dropwise

methods to the nitrate solution under -600 rpm mechanical agitation. The two assemblies, buret

and dropping funnel, were employed for oxalic acid addition to yield a fine particle size and

spherical morphology. The buret formed a somewhat finer particle size than the dropping funnel

due to the smaller aperture of a buret, which produced a smaller drop size. Because the agitation

speed was held constant, drop size was the primary means of controlling particle size. A coarser

powder, however, was desired to reduce airborne dust hazard and assist sieve particle size

classification. Powders fabricated by either method demonstrated the desired free-flowing

characteristics for pellet making. An optical micrograph (Figure 3-2) of nominal 20%UO2 -

80%ThO2 calcined powder demonstrates the spherical morphology observed with this method.

Following the complexation of U and Th to the oxalate group, the precipitate is separated

by vacuum-assisted filtration through #42 Whatman paper and dried in air at room temperature

until it releases easily from the paper. The filtrate liquid is clear, an indicator that most of the

uranium did not remain in solution. X-Ray Fluorescence performed by co-researcher Shibuya of

the liquid filtrate confirmed this assertion. The precipitate cake, which is white-colored, is










transferred to a glazed ceramic crucible. Eliminating the residual moisture and undesired

organic (i.e., carbon) subsequently requires a series of drying, decomposition, and calcination

steps carried out in air, followed by a final reduction heating in flowing 5%H2-Ar. Powder color

following reduction that was performed in a tube furnace on an Alumina boat varied from tan to

dark orange. Because of the multiple oxidation states available to uranium, which readily

oxidizes, the final reduction step is needed to return it to U(IV) and face-centered cubic lattice

structure. Since the powder will later be formed into pellet geometry and sintered, reduction is

carried out at 800C to prevent particles from sintering. The molecular evolution of the co-

precipitate is depicted in Table 3-1 where U/Th is 1 and X-Ray Diffraction pattern in Figure 3-3

for nominal U/Th = 5/95.

Table 3-1. Evolution of oxalate-synthesized (U,Th)02.
Gas T(C) Equation
Filter air 25 U(N03)4-Th(NO3)4 + 4H2C204 [U(C204)2-Th(C204)2] nHO2 + 8HNO3
Dry air 120 [U(C204)2-Th(C204)2] xH20 [U(C204)2-Th(C204)2] + nH20
Decompose air 350 [U(C204)2-Th(C204)2] -> U(C03)2-Th(CO3)2 + 4CO
Calcine air 900 U(C03)2-Th(C03)2 U02+x-ThO2 + 4C02
Reduce 5%/H-Ar 800 U02+-ThO2 + nH2 U02-ThO2 + nHO2

Drying steps carried out in air were performed in a Vulcan 3-550 Muffle Furnace. The

final reduction drying at 800C took place in either a CM Rapid Temp or Lindberg/Blue 54434C

tube furnace with Mullite and, later, 99.8% Alumina open-ended tubes supplied by Coors. Two

sets of custom-made stainless steel caps were installed to control gas atmosphere during firing.

The initial design was suspected of not maintaining an appropriate seal between the steel and

ceramic, presenting carbon content issues for preliminary synthesis attempts. The latter design

incorporated a high temperature Viton polymer gasket situated between the stainless steel cap and

ceramic tube. Additionally, a vacuum set-up was assembled whereby the tube atmosphere could

be evacuated and backfilled with the desired gas, thereby preventing the material from

encountering an oxidative atmosphere. The tube atmosphere is evacuated and backfilled with

5%oH-Ar three times before initiating the desired firing regime. Figures 3-4 and 3-5 show each

furnace.


























Reduced (800C/1Oh)














Calcined (900C/24h)









Decomposed (350C/5h)





Dried (120C/5h)






Filtered


5 25 45 65 85 105 125
2Theta


Figure 3-3. X-Ray Diffraction evolution of Uo o5Tho 9502 oxalate synthesized powder.







































Figure 3-4. Vulcan Muffle furnace used for drying, decomposition, and calcination


Figure 3-5. Lindberg high temperature tube furnace (right) with alumina tube and stainless steel
cap (above)


II'










Ammonium Hydroxide Co-Precipitation

Another synthesis path used was derived from a modification of the well-documented

Ammonium Diuranate (ADU) process, which is used in commercial processing of U02 fuel. The

purposes of attempting this synthesis path did not extend beyond establishing a preliminary

qualitative comparison with the oxalate path and traditional blending. Consequently, only a

nominal 20% UO2 composition of moderate quantity (5 grams) was processed. The quantities

and specifications of the subsequent procedure were adapted from parameters made available by

Dr. E. Lahoda of Westinghouse, Inc. A brief schematic of the procedure follows in Figure 3-6.


Dissolve U-nitrate (s) in D.I. water.

Add HF such that F/U = 4.

Add Th-nitrate (s).

Pour nitrate solution in NH4OH
(NH3/U > 30) bath

Filter precipitate.

Calcine in 5%/H-Ar, 6300C for 4 h.


Figure 3-6. Ammonium hydroxide path co-precipitation

Starting constituents were uranium oxynitrate hexahydrate (UO2(NO3)2-6H20) and

thorium nitrate hydrate (Th(NO3)4-xH20) supplied by Alfa Aesar, and 49% diluted hydrofluoric

acid (HF) and 29% diluted ammonium hydroxide (NH40H). Desired molar quantities of uranium

(IV) and thorium (IV) nitrate were determined. In demineralized water, uranium (IV) nitrate was

dissolved to yield a 0.19 M solution. A molar excess of hydrofluoric acid was added to the

uranium (IV) nitrate solution by pipette to raise the molar ratio F/U to 4. No color change or

other apparent reaction observed. At this stage, thorium (IV) nitrate was stirred into the solution

until dissolved. The solution took on a transparent green-yellow hue. A 30-fold molar excess of

ammonium hydroxide aqueous bath was prepared separately. Without agitation, the nitrate salt









solution was poured directly into the ammonium hydroxide bath. A bright yellow colored

precipitate immediately formed. The precipitate was removed from the liquid by vacuum-

assisted filtration through a Whatman #42 ashless paper and rinsed with demineralized water.

The precipitate had an "orange juice" pulp-like appearance, as shown in Figure 3-7 at left, which

was starkly different from the white oxalate precipitate. Additionally, the volume of material for

this 5-gram batch appeared more voluminous than oxalate-synthesized batches. Following

filtration, the yellow cake was air dried until it easily released from the filter paper.

'Ak .... ..






S ,..; .



Figure 3-7. Ammonium hydroxide synthesized (Uo 2Tho s)02 before (1) and after (r) calcinations

Since the hydroxide path did not introduce organic substituents, the heat treatment

scheme was decidedly simpler than for the oxalate path. In a flowing 5%H-Ar reduction

atmosphere, the yellow cake was calcined at 630C for 4 hours. A significant volume reduction

and color change from yellow to nearly black occurred during calcinations as shown in Figure 3-7

at right. X-ray diffraction patterns, taken prior to and following calcinations, revealed the cubic

fluorite structure. Silicon was used as an external standard for XRD measurements of hydroxide-

synthesized material before and after calcinations.

Co-Milled Mixed Oxides

Uranium dioxide and (UyThi-y)02 solid solutions with y = 0.236, 0.368, and 0.500 were

prepared by mixing appropriate amounts of U02 and ThO2 powders supplied by Siemens (SM)

and Alfa Aesar (AA), respectively. Thorium and uranium oxide powders supplied by Alfa Aesar

are manufactured to a below 325 mesh (< 45pm) and 50 mesh (< 300pm) particle size










distributions, respectively. Optical micrographs of these powders affixed with a 1:7 collodion

and amyl acetate solution are shown in Figure 3-8. From the micrographs, it is apparent that the

ThO2 agglomerates easily, forming large clumps, suggesting a platelet-like morphology.

Uranium (IV) oxide particles from AA, on the other hand, presented spherical morphology and

demonstrated good flow.















Figure 3-8. Optical micrographs ofAlfa Aesar U02 (1) and ThO2 (r)

Co-milling. Preliminary co-milling trials were conducted with high-density polyethylene

(HDPE) jars, zirconia ball media, and a laboratory roller mill assembly. Milled powder became

imbedded in the HDPE jar walls and clung to ball media. Cleaning jars and media was tedious

and difficult. The material lost in the cleaning process, also, directly has an effect on the final

U/Th composition. This set-up, therefore, was abandoned in favor of a small batch high impact

mixer.

Powders were weighed and batch milled for 60 minutes in a zirconium oxide jar with two

10-mm yttria-stabilized zirconia (YSZ) milling media using a high energy 8000M SPEX

Certiprep Mixer/Mill. The mixer employs a torsional "figure eight" motion for pulverizing and

mixing. Zirconia ceramic vial, cap, and ball media, as supplied by the vendor, accommodates

loads up to 20 mL for mixing. No zirconia contamination from the container or milling media

was found in the milled powders. To prevent cross contamination, grinding vials were cleaned










with dilute nitric acid between each batch. Figure 3-9 displays the mixer and mill jar used in this

study.


I'












Figure 3-9. 8000M SPEX Certiprep Mixer/Mill (1) and zirconia mill jar (r)


Preliminary runs showed that the high impact velocity of the grinding balls caused

powder to pack onto the vial walls. It was necessary, therefore, to mill in 15-minute intervals,

scraping vial walls between each interval. After observing this, there was concern the particle

size difference between starting powders inhibited effective blending if finer particles tended to

pack more easily than coarse particles along vial walls. Because of the particle size and

morphology difference between Alfa Aesar UO2 and ThO2, a 2x2 factorial experiment was

initially performed to determine effect of fine (< 63 pm) versus unsieved milled UO2 and high

(100 MPa) versus low (50 MPa) compaction pressure on sintered pellet density. Pre-milling as-

received UO2 (AA) spherical particles resulted in a platelet-like morphology similar to as-

received ThO2 (AA). A 30-gram batch of U02 (AA) was milled four 15-minutes consecutive

intervals before sieving through a 230 mesh (63 pm) to achieve a fine powder batch. It was

observed that powder packing along vial walls became progressively hard-packed with each

successive milling/blending intervals. Four 15-gram batches of 20% UO2 80% ThO2 blended

for 60 minutes were processed according to the parameters of the experimental matrix shows in

Table 3-2. Each batch yielded similar (within 2%) green and sinter densities, indicating little










difference between using fine or coarse U02 (AA) and 50 MPa or 1000 MPa uniaxial compaction

pressure.

Table 3-2. 2x2 factorial for 20% UO2-80% ThO2 blended oxide
50 MPa 100 MPa
Unsieved UO2 x x
< 63 uin UO2 x x

Green and sintered densities were measured three times each with calipers to determine

volume. The pellets were sintered at 16500C in a flowing 5%H-Ar atmosphere. Pellet densities

were measured after 5 and 15 hour dwells at 1650C. Those results are summarized in Table 3-3.

Table 3-3. Co-milled U0 2Tho 802 pressing conditions 2x2 factorial results
Pellet Description Green Density Sintered Density (5 h) Sintered Density (15 h)
Unsieved 50 MPa 60.2% 85.8% 97.2%
Unsieved 100 MPa 61.7% 85.7% 98.7%
Sieved 50MPa 61.2% 87.8% 98.1%
Sieved 100 MPa 63.8% 86.7% 98.7%


From the above results, the pressing parameters (100 MPa, sieved) were selected to make

pellets for dry oxidation study. P. Demkowicz confirmed by the water immersion technique that

sintered densities for the dry oxidation pellets > 98% theoretical density. Those pellets,

subsequently, were fragmented by mortar and pestle and classified by ASTM sieves. The 90 to

250 |pm particle size range was selected. Demkowicz measured specific surface area of

fragments in the 90 to 250 pm range using the multi-point Brunauer-Emmett-Teller (BET)

method in a Quantachrome Autosorb 1, with krypton as the adsorbate gas, and found all batches

to be within 0.01 0.02 m2/g. Sample quantities were between 2 and 10-g, dried under vacuum

at 1200C for approximately 12 hours prior to analysis.

Pellet pressing. The milled ThO2-U02 powders were initially sieved through 100-mesh

(< 149 ipm) to remove coarse particles before compaction. The powders were compacted into

pellets in a 13-mm diameter stainless steel die using a single action laboratory press at 100 MPa.

Pure U02 pellets were pressed using pressures of 200 MPa. Stearic acid was used as a lubricant

on the die, anvil, and punch sidewalls. The dimensions of the pressed pellets were measured with









calipers and the green densities, obtained by geometrical calculations, were 56% to 64%

theoretical density (%TD). Pellets were arranged on a powder mound of matching composition

to minimize any possible reactivity with the ceramic setter. The pressed co-milled pellets (Figure

3-10 shows oxalate synthesized pellets) were sintered at 16500C for 20 hours in a 5%H2-Ar gas

mixture using a tube furnace fitted with stainless steel end caps. Sintered densities were obtained

by immersion methods in deionized water and were between 95% to 98% theoretical density.

Those conditions are summarized in Table 3-4.

Table 3-4. Pellet manufacture conditions
Pellet Description Pressure Green Density Sinter Density
(MPa) (%TD) (%TD)
23.6% UO2- 76.4% ThO2 100 63.5 0.2 96.8 0.4
36.8% UO2- 63.2% ThO2 100 62.6 0.2 96.2 0.2
50.0% UO2 50.0% ThO2 100 61.4 0.1 95.3 0.3
100% U02 200 49 96.7










S.





Figure 3-10. Pellets prepared using 20% U powder synthesized by the oxalate technique.
Powders compacted at 200MPa (left) and 300MPa (right) and sintered at 16000C for 5
hours in Ar-5%H atmosphere

Pellets were manually crushed in a zirconium container with an alumina pestle.

Fragments were sieved using 60-mesh (250 |pm) and 170-mesh (90 pm) ASTM screens.

Generally, about 51% to 61% of the powder was retained on the 170-mesh sieve and about 35%

to 42% was collected below 170-mesh. The portions of the crushed pellets below 170-mesh (<

90 pm) were analyzed using x-ray diffraction. The XRD patterns indicated that the samples were










complete solid solutions, with no evidence of secondary phases. The powder distribution retained

on the 170 mesh and 60 mesh were not measured by XRD to confirm lattice structure or

composition. For materials with significantly different hardness, fragmentation may

inadvertently separate the "hard" and "soft" materials. More brittle materials may fragment into

smaller particles whereas the harder materials remain large, altering the powder composition.

Brinell and Vickers hardness reported for uranium (2400 MN/m2 and 1960 MN/m2) and thorium

(400 MN/m2 and 350 MN/m2) by webelements.com indicate that U is the "tougher" material. It

is likely the fine particle distributions may have a lower U/Th ratio than coarser distributions.

Again, only the powder retained below 60-mesh/above 170-mesh further characterized. It is

unknown whether the fragmentation process may have yielded a high U and low U powder

composition.

The specific surface area of the crushed pellets was measured using the Brunauer-

Emmett-Teller (BET) method with krypton as the adsorbate gas. The 90 to 250 pm powder

samples for surface area analysis were rinsed in demineralized water and dried in a vacuum prior

to BET analysis to remove all fine particulates. Measured surface areas for all 90 to 250 pLm

pellet fragments at all compositions were 0.01 to 0.02 m2/g.

Characterization

X-ray Diffraction (XRD)

Crystal structure indexing, chemical analysis, and lattice parameter calculations were

determined from continuous x-ray diffraction scans on a XRD Philips APD 3720 Diffractometer

with a Cu anode target, where Kox wavelength is 1.54056 A. Generator voltage and current were

40 kV and 20 mA, respectively. With a step size of 0.020, a range from 100 to 1390 was scanned

for each sample, with an external Si standard. Lattice parameters were calculated by a graphical

and numerical least squares methods and angular separation reported by Popovic [Pop73, Pop85].









Popovic method. Unit cell dimensions were calculated from the angular separation

between two Bragg reflections on the basis that shifts occur in the same direction from resultant

systematic aberrations. The separation (6 = 02 01), therefore, includes the difference between

systematic aberrations at the two positions, which is generally smaller than the resultant

aberration at either position and removes the necessity of absolute measurements. For a cubic

structure, lattice parameter a derived from the separation 6 between two diffraction lines, 01 and

02, is given by Equations 3-1 through 3-5.

2 B -B2 cos
(3-1) a 1 where
4sin2 8

B, NJ+N2 2 N =h2+k2+l2
(3-2, 3-3) and (3-4, 3-5)
B = 2 NN, i= 1,2

To avoid temperature variations, the second reflection should be chosen within 100 to 200 from

the first reflection [Pop73]. The highest possible diffraction angles are used because of the sine

nature of Bragg law, X = 2dsin0, where 0 values near 900 change slowly and, therefore, yield

greater accuracy. The sensitivity of this method is reported as

Aa AcosO, cosO2
(3-6)
a sin 6

Method of least squares. The lattice parameter, a, of a cubic structure is directly

proportional to the d-spacing according to the relationship

(3-7) a= dh2 +k2+2

where h, k, 1 are the miller indices defining any particular set of planes. From measurement of

the Bragg angle, 0, interplanar d-spacing can be determined according to Bragg law, X = 2dsin0,

and a calculated. Since the term sin0 appears in Bragg law, precision in d or a depends on

precision in sin0, not the measured 0. Because of the nature of sin0, values of sin0 change slowly

with 0 near 900. A very accurate sin0 value, therefore, can be obtained from a measurement of 0










that may not be particularly precise, provided that 0 is near 900. In other words, the diffracted

beam angle is more sensitive to changes in plane spacing when 0 is large.

Obtaining measurements near 20 = 1800 are impossible. Since the calculated lattice

parameter values approach the true value as 20 nears 1800, the true value for a can be obtained by

extrapolating a plot of measured values against 20 to 1800. The sin0 function, however, yields a

nonlinear curve, which is difficult to accurately extrapolate. Depending upon the kind of camera

employed, 0 can be inserted into certain functions that present linear curves for extrapolation of

lattice parameter. For a diffractometer, that function is either cos20 or cos20/sin0 depending upon

the predominant source or error [Cul78, p.360]. Systematic errors from using a flat, instead of

curved, specimen and absorption in the specimen are attributed to the cos20 function.

Displacement from the diffractometer axis, which is typically the largest single source of error, is

attributed to cos20/ sine. A lattice parameter value is then obtained by linear graphic

extrapolation of the a vs. cos20 or cos20/ sine plot such that 0 = 90.

Cohen's method. Similar to the least squares method, Cohen's method [Cul78, p.363]

applies the least squares technique to the observed sin20 values directly, instead of to a vs. cos20

or cos20/ sine plots. Squaring, taking the logarithm, and differentiating the Bragg law yield the

relationship below.

A sin20 = -2 Ad
(3-8) =sn -2
sin2 0 d

The Ad/d term is replaced with an extrapolation function (-D/R)(cos20/sin0), where D is specimen

displacement and R diffractometer radius, to account for diffractometer error [Cul78, p.359]. The

difference, Asin20, between the true, as defined by unit cell geometry, and observed sin20 values

is set equal to the above equation. Simultaneously solving Equation 3-9 for the set of Bragg lines

yields the true lattice parameter, ao, at some constant wavelength k. For ease in calculation, the

terms are separated into Equations 3-10 through 3-13.










(3-9) sin2 (obs) (h2 +k2 +2) = Asin2 = 2D cos2 0sin
4a2 R


(3-10) C -
4a2

(3-11) a= h2 +k2 + 2

2D 1
(3-12) A =2
R 10

(3-13) = 10cos2 Osin8

The experimental values of sin20(obs), a and 6 for each back-reflection inserted into the

appropriate terms leave only C and A to be solved. Employing the numeric method of least

squares detailed by Cullity [Cul78, p.364], normal Equations 3-14 and 3-15 for cubic systems are

simultaneously solved to obtain C and A.

Zasin2 0= C a2 + AZ aS
(3-14, 3-15) si
sin2 0 = C- aS +A-'S2

Elemental Analysis

Inductively coupled plasma Auger electron spectrometry (ICP-AES). Measurements

obtained from ICP-AES were not performed by the author, but by James Jerden, Ph.D. at

Argonne National Laboratory. The general technique and results obtained from Dr. Jerden are

reported because of their importance in determining the synthesis process of samples to undergo

thermogravimetric analysis.

ICP-AES measures the mass spectrum of a sample, typically, from lithium (Z = 3) to

uranium (Z = 92), yielding semi-quantitative, quantitative, elemental and isotopic information for

each element. It is a destructive technique where the aqueous sample is nebulized into an aerosol

and swept into the ionizing plasma, a high-temperature, atmospheric pressure, and partially

ionized gas. The ions generated are carried into a detector for analysis by Auger electron

spectrometry. Detection limits are typically in the sub-ppb range for most elements.










Carbon analysis. The LECO WC-200 Carbon Analyzer is a unit specifically designed

for tungsten carbide application that provides high precision measurement of carbon content in

metals, ceramics, and other inorganic materials. A high frequency induction furnace combusts

samples, with accelerators (- 1 gram each of iron and copper chips) in a quartz crucible, in a pure

oxygen environment. The carbon-bearing elements are reduced, releasing the carbon, which

immediately binds with the oxygen to form CO and CO2. These gases are then carried into a

molecular trap and then released into infrared (IR) cells. As CO2 passes through the IR cell, it

absorbs IR energy at a precise wavelength within the IR spectrum, preventing it from reaching the

IR detector. After passing all IR energy through a narrow bandpass filter to ensure the signal can

only be attributed to CO2, the concentration is detected as a reduction in the energy level at the

detector. A five-place balance and burned-off crucibles and lids were handled only with clean

tongs to minimize additive error contributions.

Particle Morphology

An exhaustive classification of particle size demands complex sampling and numerous

optical micrographs to yield a statistically significant portrayal of size and distribution. Light

optic techniques were used to present a qualitative comparison of particle size and morphology

processed via co-precipitation, blending, and commercial methods shown in Figure 3-11.

The spherical particles of oxalate synthesized and commercial U02 aided pellet making

in terms of flowing material into the die and during pressing. Particles did not stick to die walls,

unlike the blended (U,Th)O2 and commercial ThO2 powders. Since wet synthesized (U,Th)O2

powders were abandoned for dry oxidation study, this sticking behavior was addressed by using a

lubricant, stearic acid, to aid pellet release from the die. These were not necessary to form

successful pellets with the oxalate synthesized powders.





















a) 201JSS-1100


c) UO2 Commercial (lOx)


d) ThO2 Commercial


Figure 3-11. Optical micrographs of nominal Uo 2Tho 802+x made by oxalate synthesis (a),
blended (b), and commercial U02 (c) and ThO2 (d) powders


Pellet Density

Green pellet densities were measured with a Mettler AB-104S balance and calipers.

Volumes were calculated under the assumption that pellets formed perfect cylinders. Densities of

sintered pellets, however, were measured by immersion in water. First, a dry weight (A) was

taken. Then, the pellet is immersed in water and the difference between the weight wet and

immersed pellet is the buoyancy (P). The density is calculated based on Equation 3-16.

A
(3-16) P = Pwater
P










Dry Oxidation

Thermogravimetric Analyzer

Nonisothermal and isothermal dry oxidation experiments were conducted on the

crushed pellets (90 to 250 p.m powder) in a TA Instruments TGA 2050 thermal gravimetric

analyzer, as shown in Figure 3-12, with air flowing at 90 cm3/min through the sample chamber.

Approximately 8 to 20 mg of the 90 to 250 p.m pellet fragments were heated to 9000C at rates of

1, 3, and 5C/min while monitoring the weight changes. The 90 to 250 p.m fragments were used

in order to provide a narrow particle size distribution. Restricting samples to the same size

distribution was a means to maintain relatively similar surface to volume ratios, normalizing its

effect upon oxide decomposition.























Figure 3-12. TA Instruments TGA 2050 thermogravimetric analyzer.

The appropriate temperatures for isothermal experiments were chosen based on the

results from the nonisothermal gravitational analysis for each composition. The isothermal runs

were performed using the same gas flow used for the nonisothermal experiments. The length of

the isothermal experiments varied for each composition and temperature, as measurements were










terminated after the sample exhibited no further weight change. The ramp-up time (to) to reach

the designated temperature was between 8 and 15 minutes for all experiments. This period is

neglected in the presentation of figures and all calculations. The measured weight gain data are

converted to fraction reacted, a, versus time plots for purposes of calculating kinetic parameters.

Isotherms for (Uo 368Th0 632)02+x were repeated in triplicate to determine the statistical significance

of measured data.

Samples were also characterized with x-ray diffraction after oxidation. All (U,Th)02+x

samples were found to be the cubic fluorite structure, demonstrating that there were no bulk phase

changes during oxidation for any of the solid solution compositions.

Kinetic Analysis

The measured data from nonisothermal and isothermal dry oxidation of mixed oxides

were subjected to standard model-fitting techniques and a model-free method reported by

Vyazovkin and Wight [Vya99]. Standard methods estimate kinetic parameters and reaction

mechanisms based on theoretical reaction models. Arrhenius parameters from isothermal and

nonisothermal data, however, often disagree because of the differing nature of a constant versus

variable heating profile. Calculations derived from isothermal data, therefore, are traditionally

considered more reliable because the temperature variable is held constant, reducing the number

of parameters simultaneously determined by fitting the data using a particular reaction model.

Consequently, the macroscopic nature ofthermoanalytical techniques does not always elucidate

system complexity, such as overlapping mechanism. Additionally, measured data may fit more

than one hypothetical reaction model, demonstrating the danger of force-fitting data to inadequate

models.

The model-free technique, on the other hand, bypasses assumptions of a specific reaction

model. Activation energies instead are determined as a function of the degree of conversion (a)

and/or temperature. Instead of yielding a global composite activation energy for the system, the









activation energy dependence on a is capable of revealing process complexities. Vyazovkin and

Wight [Vya99] reported the model-free method capable of producing consistent kinetic

information from both isothermal and nonisothermal data

Model-fitting method. Traditional kinetic analysis of isothermal data entails a

comparison of measured values with theoretical reaction models. These models are derived based

on the geometry of interface initiation and advance and/or diffusion processes occurring in the

solid. Natural limitations are inherent in translating experimental data into pertinent kinetic

parameters based on the models. It is often the case that factors important to yielding an

important conclusion regarding kinetic processes may not be experimentally accessible. Because

of restrictions experienced in quantitatively characterizing particle morphology in this study and

the assumption that pellet fragments were not spherical, developing a particle size dependent term

was hindered. It was anticipated that restricting the particle size and, consequently, surface area

to a common range in this study would normalize the surface area influence and allow for

comparisons between U/Th compositions in this study.

Solid-state decomposition kinetic analysis is based on general kinetic Equation 3-17 and

previously mentioned Arrhenius Equation 2-2, where a is the fraction reacted, f(a) is the reaction

model, k(T) is the rate coefficient (also denoted k), t is time, A is the frequency factor, E is the

activation energy, R is the gas constant, and T is the temperature.

(3-17) f(a) = k(T) t


(2-2) k(T) = Aexp-


Theoretical relationships, f(a), that have found the greatest application in solid state

kinetic analysis are summarized in Table 3-5 [Bam80]. If experimental data is plotted as f(a) vs.

t, the linearity of the plot is an indication of how well the data agrees with a particular reaction

model. Care must be taken in accurately defining the final yield, a = 1.0, to prevent distorting the










a vs. t curve. Rate coefficients, subsequently, are obtained from the linear slope of the f(a) vs. t

plot.

Table 3-5. Solid state theoretical reaction models
Reaction model f(a) Reaction model f(a)
Acceleratory rate equation Deceleratory rate equations
Power law a(1n (Based on diffusion mechanisms)
Exponential law ln(a) One-dimensional diffusion a2
Sigmoid rate equations Two-dimensional diffusion (1-a)ln(1-a) + a
[-ln(1-a)]1/2 Three-dimensional diffusion [1 (1-a)1/3]2
Avrami-Erofe'ev [-ln(l-a)]1/3 Ginstling-Brounshtein [1 (2a/3)] (1-a)2/3
[-ln(1-(a)]1/4 (Based on geometric models)
Prout-Tompkins ln[a/(1-a)] Contracting area 1 (-)1/2
Contracting volume 1 (1-a)1/3
(Based on order with respect to a)
First order -ln(l-a)
Second order (1-a)-1
Third order (1-a)-2


The equation below is obtained by taking the natural logarithm of the Arrhenius

relationship [Ban68]. From Equation 3-18, it is clear that a plot of ln[k(T)] vs. 1/T will give a

straight line with slope equal to -E/R and an intercept equal to InA.


(3-18) Ink(T)=L- E I +lnA


Model-free method. The model-free method was applied solely to isothermal rate data

in this study. Under isothermal conditions, the reaction model is assumed to be independent of

the heating rate. By rearranging Equation 3-17, substituting into Equation 2-2, and taking the

natural logarithm, Equation 3-19 is obtained, where to,, is the reaction time for a selected fraction

reacted, a, for a given isotherm, i.

E A
(3-19) In t,, IR- +ln f l


A -ln(t) vs. 1/T plot at a particular a value is constructed from the isothermal oxidation

data obtained at several temperatures. The linear slope of the -ln(t) vs. 1/T plot yields the value -

E/R, at the fraction reacted, a, without making any assumptions about the reaction model, f(a).






42


Because the intercept term in Equation 3-19 is a function of both A and f(a), the frequency factor

(A) cannot be determined by this method without identifying f(a). The benefits of this method,

however, are such that reaction complexities may be revealed by a dependence of E on a, instead

of simply yielding an overall value for the reaction. Consequently, there is a heavy reliance upon

meaningful definition of final yield where a = 1.0.
















CHAPTER 4
RESULTS AND DISCUSSION

Material Synthesis

X-ray diffraction. Bulk homogeneity of the mixed oxides was confirmed by XRD to be

cubic fluorite in structure at all U/Th compositions before and after air oxidation. A Philips APD

3720 Diffractometer measured pellet fragments with silicon as a standard. Measured diffraction

powder patterns were compared with JCPDS standards to qualitatively confirm Bragg angles (20)

and identify corresponding miller indices. See Appendix B for details of the diffraction pattern

measured for U0 368Th0 63202+x pellet fragments isothermally oxidized at 4000C in flowing air. All

U/Th compositions qualitatively presented the same cubic fluorite diffraction pattern.

Three methods of calculating unit cell parameters were used: method of least squares

(LSQ), Cohen's method [Cul78], and Popovic's method [Pop73, Pop85]. Starting compositions

of all unoxidized material batches are assumed stoichiometric. Unit cell volume or lattice

parameter decreases with respect to increasing percent U02, which is consistent with published

findings as presented in Figure 4-1. Table 4-1 summarizes the lattice parameters as calculated by

the three methods for isothermally oxidized (U,Th)02 pellet fragments. As more of the smaller

uranium cation is introduced into the thorium oxide matrix, it is apparent that the lattice shrinks to

accommodate without changing the cubic fluorite structure.

Table 4-1. Calculated (U,Th)02 oxidized lattice parameters by three methods
Description LSQ (cos2O) LSQ (cos2O/sinO) Cohen's Popovic
Uo 236Th 76402 5.568 0.003 5.564 0.002 5.563 0.001 5.564 0.005
Uo 368Th 63202 5.542 0.002 5.543 0.001 5.543 0.001 5.541 0.003
U 500Th050002 5.523 + 0.004 5.522 + 0.003 5.523 + 0.002 5.525 + 0.002
LSQ: least squares method uses two functions for diffractometer measured patterns.











5.62

---------------------------------------
5.6 _____________________________ ThO2

S5.58 0Anderson (1954)
5.57 *this study
a 5.56
E 5.54
S5.54
S.5.52
a 5.52 -

S5.5

5.48
UO2
--J-------------------D-5----|o
5.46 D
15% 25% 35% 45% 55% 65% 75% 85% 95%
% UO2
Figure 4-1. Lattice parameter evolution with respect to UO2 content. Dashed lines indicate the
lattice parameters for reference oxides ThO2 and UO2. Linear trendline show that
results from this study are in agreement with Anderson et al. (1954)

Inductively coupled plasma Auger electron spectrometer. Samples of (U,Th)02

synthesized by the oxalate and ammonium hydroxide path\\ a% s were measured by James Jerden,

Ph.D. at Argonne National Laboratory to obtain weight percent uranium and thorium. Assuming

whatever remained was oxygen, those results, as summarized in Table 4-2 as weight percent, it

was apparent something other than oxygen was also present. As a consequence, carbon analysis

was pursued to determine whether the excess could be attributed to residual carbon from the

synthesis process.

Carbon analyzer. A tungsten carbide standard (0.26 0.02 g) was measured five times

with a standard deviation of 0.03% by weight carbon. An empty crucible established a blank

measured carbon concentration of 0.00385% by weight. Absolute and relative uncertainty were

calculated to be {(0.03)2 + (0.03)2} = 0.04 and (0.04/measurement x 100%), respectively [Har91].

Oxalate and ADU synthesized (U,Th)02 samples with nominal percent U content of 20%, 35%,

and 50% measured results are summarized in Table 4-2 as weight percent. Since it was supposed

the excess identified by ICP-MS was residual carbon from the oxalate process, the ADU

synthesized nominal 20% UO2 sample was not measured by this instrument.










From the above results, it is apparent the elemental analysis techniques used were unable

to fully identify the chemical composition of wet synthesized (U,Th)02 materials. Ando et al.

[And76, And84] also noted the presence of impurities for ThO2 synthesis by sol-gel and arc

fusion techniques. Without a full chemical identification, the possible influence of unknown

impurities could possibly inhibit evaluation of the thorium oxide matrix effect upon uranium

oxidation. As a consequence, wet chemistry synthesis techniques were abandoned in favor of

traditional powder process methods to avoid introduction of impurities associated with the

synthesis path.

Table 4-2. ICP-AES and LECO carbon analysis results and calculated metal valence of wet
synthesized (UyThi -)O2+x indicate impurities must be present.
Sample (Nominal U wt Th wt C wt Ba *
.. Balance* y x VM
composition) % % %
ADU (Uo 2oTho 8o)O2+x 15.8 66.6 -- 17.6 0.192 1.11 6.22
Oxalate (Uo 20Tho 8)02+x 11.6 70.9 0.01 17.5 0.141 1.09 6.18
Oxalate (Uo 35Th0 65)02+x 21.3 54.1 0.17 24.4 0.283 2.72 9.44
Oxalate (Uo 5Th 5o)O2+x 32.7 38.0 0.19 29.1 0.463 4.03 12.06
Assumes balance is only oxygen. Metal valences (VM) calculated from x are impossible; knowing that
VTh maximum is +4 and Vu is 6.

Thermogravimetry

Nonisothermal. Thermal gravimetric analysis curves for air oxidation of U02

and (U,Th)02 samples under nonisothermal conditions are shown in Figures 4-2 through 4-5 as

percent weight gain vs. temperature. All of the data exhibited single-step behavior, including the

data for U02 (Figure 4-5). This is consistent with observations of air oxidation of (UyThly)O2

materials (y = 0.15, 0.30, 0.72, 0.77) performed by Anthonysamy et al. [AntOO] While

gravimetric oxidation curves for UO2 often exhibit a two-step behavior corresponding to

oxidation from U02 to U307 to U308, the absence of a well defined two-step curve for U02 in this

study is most likely the result of the low specific surface area of the samples (0.01 to 0.02 m2/g).

With the lower surface area of coarser particle size distributions, the fast surface transformation

of U02 to U307 contributes less to the overall weight gain because the reaction interface

decreases. The first "step", therefore, diminishes with decreasing surface area. The slower bulk










transformations of U02 to U308 and U307 to U308, however, are less dependent upon surface

area. The second "step" remains despite particle size distribution. Thermogravimetric

measurements of higher surface area U02 powders (1 to 3 m2/g) resulted in the expected two-step

weight gain curve. Thus, the low surface to volume ratio of the 90 to 250 mm pellet fragments

impacted the results such that bulk reactions had a much large effect on the overall oxidation

process than the contributions from surface reactions. Figures 4-2 to 4-4 show the mixed

(U,Th)02 oxides and Figure 4-5 U02 fragments under nonisothermal air oxidation at heating rates

of 1, 3, 5C/min as indicated.

A weight decrease of (U,Th)02 samples at higher temperatures (Figures 4-2 to 4-4) was

observed. The U02 isothermal sample, Figure 4-5, however, did not present this same behavior.

The mixed (U,Th)02 oxide weight loss at the higher temperature region was constant and

sustained for > 20 minutes. Since the TGA was located on an isolation table and weight loss was

not sudden, it was not supposed that fragments were somehow ejected from the pan from an

external source. To eliminate the presence of an unknown gas, the pellet fragments, furnace

chamber, and gas lines were vacuum purged and backfilled with high purity Argon gas three

times at room temperature. A final vacuum purge was backfilled with air prior to beginning each

run. These same measures were taken with the nonisothermal oxidation of U02 pellet fragments.

Those results are presented in Figures 4-2 to 4-5. For the U02 sample, the absence of a steady

weight decrease following initial weight gain further suggests an external source is unlikely.

Further, without measurements beyond 9000C, it is unknown whether the mixed oxide weight

would eventually reach a steady state.

Knowing the lattice structure evolution from start to finish may confirm what

mechanisms influence oxidation. Though not accomplished in this study, XRD measurements of

(U,Th)02 fragments at various stages of nonisothermal oxidation, particularly at the peak

maximum and T = 9000C, may suggest a possible explanation. Possibly, the thoria lattice











initially accommodates excess anions in the large octahedral interstitial sites, with charge

neutrality maintained by increasing U valency. Successively more anions, noting peak weight

gains of UO 236Th0 46402 (0.76% 0.77%), Uo 368Th0 63202 (1.16 1.26%), and Uo 500Tho 50002 (1.59

- 1.65%), are accommodated with increasing U content. Eventually, the lattice may become

"supersaturated", holding the maximum number of anions with U at its highest valency, yielding

a peak maximum. Perhaps, with increasing thermal energy, the structural and energetic strain of

maintaining this "supersaturated" lattice causes the steady ejection of excess anions until a stable,

unstrained configuration is achieved. This may account for the decreasing weight.

Nonisothermal measurements of finer particle size distributions may also yield surface

area dependent or other mechanisms that would clarify the peak maximum and subsequent

decreasing weight in the (U,Th)02 systems. Like UO2, there may be "fast" and "slow" competing

mechanisms. A "fast" mechanism may dominate the weight gain at lower temperatures. At the

higher temperatures, this "fast" mechanism may have reached completion and the "slow"

mechanism becomes the rate-limiting step.




0 7%

S6% 1 C/min
5 C/min
^' 05%

S04%-

03%
_._ 3 C/min
02% -

01% -

0 0% Uo.236Tho.76402

-0 1%
100 200 300 400 500 600 700 800 900 1000
Temperature (Celsius)


Figure 4-2. Nonisothermal (Uo 236Th0 764)02 oxidation TGA data at heating rates of 1, 3, and
50C/min.


















1 2%

1 0% 1 C/min
5C
08% -

0) 06% -
S// 3 C/n
0 4% -

2% -

00% -

-0 2%
100 200 300 400 500 600
Temperature (Celsius)


800 900


Figure 4-3. Nonisothermal (Uo 368Tho 632)02 oxidation TGA data at heating rates of 1, 3, and
50C/min.


100 200 300 400 500 600 700 800 900
Temperature (Celsius)


Figure 4-4. Nonisothermal (Uo 50oTho 500)02 oxidation TGA data at heating rates of 1 and
50C/min.












4 5%

4 0%

3 5%

30% 3 Cmin

2 5% -
0)
S20% -

15% -

1 0% -

0 5% U02

0 0%
300 350 400 450 500
Temperature (Celsius)


Figure 4-5. Nonisothermal UO2 oxidation TGA data at heating rate of 30C/min.



Isothermal. The appropriate temperatures for isothermal gravitational analysis

experiments were chosen based on the nonisothermal results in Figures 4-2 to 4-5. Generally,

temperatures were chosen to encompass the region of the curves where the majority of the weight

gain occurred. The results from isothermal gravitational analysis at the selected temperatures is

presented as fraction reacted, a, versus time, t to, (Figures 4-6 to 4-9) where to is the time

required to reach the desired isothermal temperature (between 8-15 minutes). Some oxidative

weight gain did occur during this ramp-up period. Because the time required to achieve complete

reaction for the (U,Th)02 samples was large (usually more than 1000 minutes), the weight gain

during the same ramp-up period was negligible compared to the overall weight gain. Isothermal

oxidation of the UO2 samples, however, was completed in 20 to 120 minutes, depending on the

temperature. Contributions to overall weight gain during the ramp period for U02, therefore,

were often significant and resulted in errors in the calculation of the true degree of completion, a.


This, in turn, affected the a vs. t to curve shapes in Figures 4-6 to 4-9. Consequently, the curve

distortion hindered reaction model identification by model-fitting methods as well as accurate

measurement by the a-dependent model-free method.











At successively higher isothermal temperatures, (U,Th)02 rate of oxidation was expected

to increase, such that the slope systematically became steeper before reaching steady state. This

did not occur in Uo 5Tho 502 (Figure 4-8) and U02 (Figure 4-9). Additionally, the instrument was


jarred during the 3750C run of U02, which accounts for the discontinuity at 60 minutes in Figure


4-9. With the exception of Uo 368Th0 63202 (Figure 4-7), which was measured in triplicate, only

single runs at all isotherms were performed. Slope irregularities, therefore, may be a

consequence of experimental scatter.


12


550C0--


450C
06 450C
g0

m04


0 2 Uo.236Tho.76402



0 500 1000 1500 2000 2500 3000 3500
t-to, min

Figure 4-6. Isothermal oxidation TGA data for (Uo 236Th0 764)02 fragments (90 250 pim).
Percent weight gain data is converted to fraction reacted, a.


12


- 08

06
06


S04


02


- 450C
475C


U0368Th0.63202


0 250 500 750 1000 1250 1500 1750
t-to, min

Figure 4-7. Isothermal oxidation TGA data for (Uo 368Th0 632)02 fragments (90 250 pim).
Percent weight gain data is converted to fraction reacted, a.


1 500C
1


















12

425C
1 450C

-608- 375C

0m 400C
2 06


0 04

02
U0.50Tho0.5002

0 200 400 600 800 1000 1200
t-to, min

Figure 4-8. Isothermal oxidation TGA data for (Uo 500Tho 500)02 fragments (90 250 pm).
Percent weight gain data is converted to fraction reacted, a.







12

425C

08 450C

0os 400C

I / / 375C
S06


04 -


02
U02
00
0 20 40 60 80 100 120
t-to, min

Figure 4-9. Isothermal oxidation TGA data for U02 fragments (90 250 [pm). Percent weight
gain data converted to fraction reacted, a.









Cation valence. Based on the final weight gain of each sample after isothermal air

oxidation, the mean uranium valence was calculated. For the calculation, thorium is assumed to

maintain a valence of +4. The weight gain is entirely attributed to excess oxygen, which is used

to determine the valence of uranium cations. The initial O/U ratio of the samples was assumed to

be 2. The mean uranium valence (Vu) was calculated from the weight gain data using Equations

4-1 and 4-2, where the variable x and y are taken from the formula (UyThi y)O2+x, mi is the initial

mass of the sample, Am is the weight change after oxidation, and the value in parentheses in

Equation 4-2 is the formula weight of the starting material, ((UyThiy)O2+x).


(4-1) V, =4+ 2x
y

Am [(264.04 + 5.99y)
(4-2) x=--
mL 16

Results are plotted in Figure 4-10. Dashed lines indicate the mean uranium valence for

the reference oxides U307 and U308. As indicated in the figure, none of the mixed oxides

reached the same degree of oxidation as U02, based on the mean uranium valence calculations.

Knowing the oxidized lattices are still cubic fluorite and uranium valences are increasing, it

stands to reason that the lattices are accommodating sufficient excess oxygen anions to maintain

charge neutrality with the U cations. So, instead of a phase transformation, anions may be packed

into interstitial sites near uranium cations, since thorium valency is a constant +4. Assuming a

homogeneous distribution of U and Th cations, oxygen accommodation may be a predominantly

bulk mechanism. Without confirmation of the surface composition, a surface-dependent reaction

cannot be rejected.

This would support the view that ThO2 inhibits uranium from oxidizing further by

inhibiting the supply of oxygen with the decreased anion mobility in the thorium matrix.

Published activation energies (Appendix A) of intrinsic oxygen anion diffusion in ThO2 (209 to

275 kJ/mol) and U02 (237 to 273 kJ/mol) indicate that stoichiometric UO2 values are comparable










to those for ThO2. Uranium dioxide, however, is typically hyper-stoichiometric because of the

ease in which it incorporates excess oxygen into the lattice structure. Activation energies

reported for UO2+, are 87 to 124 kJ/mol, which is significantly lower than both stoichiometric

ThO2 and U02, indicating the greater ease in which O diffuses into the lattice. The broad range

and variety of sample preparation and measurement conditions may account for the breadth of

published values reported. It stands to reason that activation energies calculated for mixed

(U,Th)02 oxides would fall between those reported for stoichiometric ThO2 and

hyperstoichiometric UO2+x. The presence of U cations in the ThO2 is expected to lower the

threshold for oxygen anion mobility and accommodation in the lattice. It is anticipated the ThO2

matrix, on the other hand, will limit the anion supply such that cations are unable to attain U

maximum valence state.

Secondary electron microscope images taken of pellet surfaces by J. Jerden at ANL

(Appendix C) confirmed the pellets are polycrystalline. Jerden reported grain sizes ranging from

4 to 20 |pm on unbroken surfaces of 5% UO2 and 20% UO2 co-milled pellets. An image of the

broken surface of a 20% UO2 co-milled pellet shows curious "gas bubble" features. No images

were obtained of pellet fragments. Despite no images of the pellet fragments, it is obvious that

grain boundaries are present and would have an extrinsic effect upon oxygen mobility. In fact,

anions may assemble more easily in the grain boundaries than interstitial sites. Anion migration

along grain boundaries, consequently, may yield lower oxidation activation energies than for

single crystal systems. Undoubtedly, activation energies calculated for this polycrystalline

system would encompass intrinsic and extrinsic mechanisms. Without quantitative measurements

of pellet fragment grain sizes and boundaries, an extrinsic term cannot be developed for the

model-fitting kinetic analysis. The "gas bubble" features, also, indicate other extrinsic processes

may also be at work. Consequently, the models used in this research are limited without more

complete characterization of the pellet fragments to account for the intrinsic and extrinsic

contributions.












56

54 U308 A a
520 0
8 8 o
5 8
00
_48
U307 O
> ----------------------------------------
46

44
42 y=0 236 y= 0368

4
UO2 oy=0500 Ay=10

300 350 400 450 500 550 600
Temperature (Celsius)
Figure 4-10. Mean uranium valence for isothermally oxidized (UyThliy)02 and U02. Dashed
lines indicate the mean uranium valence for the reference oxides U308, U307, and
UO2.



Kinetic Analysis

Isothermal rate data were inserted into the reaction models, f(a), from Table 3-5. Linear

behavior of f(a)-time plots would suggest overall agreement with the theoretical model.

Although model fitting is the conventional technique for kinetic analysis, the published

theoretical reaction models do not necessarily account for all variables that may have influence on

the proposed mechanisms. For instance, systems that exhibit a surface-dependent reaction, such

as the UO2 U307 "fast" transformation, require an additional term to allow for particle size and

shape variations. It has already been noted that surface area has an impact on the measured signal

for U02 oxidation. The (U,Th)02 mixed oxide compositions may likely demonstrate a similar

consequence. The reaction models used in this study, however, do not take into account particle

size distribution (i.e., surface area). Considering the absence of particle morphology data to

develop a particle geometry and/or surface area term, conclusions drawn from the reaction

models in agreement with the measured data are restricted.

Since the temperature is held constant for the isothermal runs, time is the only dependent

variable. Isothermal data, therefore, are typically assumed more reliable for calculating kinetic









parameters than nonisothermal data, where time and temperature vary simultaneously. In the

subsequent figures, linear agreement to 2D diffusion, 3D diffusion, and/or Avrami-Erofe'ev

reaction models is shown for isothermal data. Rate coefficients (k) are obtained from the slopes

of the best-fit linear trendlines.

Mixed Oxide (Uo.236Tho.76402)

Transformed into degree of conversion, a, measured isothermal rate data of

(Uo 236Th0 764)02 were inserted into reaction models, f(a), summarized in Table 3-5. The two-

dimensional and three-dimensional diffusion models, where f(a) = (1-a)ln(1-a) + a and f(a) =

[1 (1-a)1/3]2, respectively, displayed agreement. Reaction models were plotted against t-to,

which truncates the initial to ramp-up stage.

Model-fitting suggests two dimensional diffusion < 4500C (Figure 4-11), followed by a

shift to three dimensional diffusion at temperatures greater than 5000C (Figures 4-12 through 4-

13). Table 4-3 summarizes the rate coefficients and linear fit of both 2D and 3D diffusion

reaction models, which were used to construct an Arrhenius plot (Figure 4-27) with E and ln(A)

calculated to be 62.1 17.9 kJ/mol and 1.71 2.87, respectively.

Linear agreement of nonisothermal rate data was also tested with the diffusion reaction

models identified by isothermal data. Measured data were converted to rate coefficient (k) by the

relation k = f(a)/t, according to Equations 2-2, 3-17, and 3-18. Arrhenius log(k) vs 1/T plots

were constructed with f(a) for 2D and 3D diffusion models shown in Figures 4-15 to 4-17 for

heating rates 1, 3, and 5C/min to 9000C. Both reaction models displayed linear behavior with

high correlation values in the region of interest. It was not discernable which diffusion model

would be appropriate. Like the isothermal data, slopes and intercepts from linear fit yield

corresponding E and A values shown in Table 4-4.

The model-free technique was used to compare with values calculated by model-fitting,

particularly since there is poor agreement between isothermal (E = 62.1 kJ/mol) and










nonisothermal (E = 92 to 94 kJ/mol) results. The model-free method was applied solely to the

isothermal rate data in this study, as shown in Figure 4-18. Reaction times at selected a (= 0.4,

0.5, 0.6) were arranged into a -In(t) vs. 1/T plot based on the relationship described in Equation

3-19. Frequency factors cannot be determined by this method without identifying f(a). Since

neither the 2D or 3D diffusion reaction models could be discerned by isothermal and

nonisothermal model fitting, frequency factor calculations were not performed by the model-free

method.

Activation energies calculated from the slope yield values of 79.2 6.9 kJ/mol, 87.6 +

7.6 kJ/mol, and 94.1 7.7 kJ/mol for fraction reacted 0.4, 0.5, and 0.6 respectively. These values

are more in agreement with those obtained by nonisothermal model-fitting, supporting diffusion

as the primary mechanism of oxidation.



Table 4-3. Rate coefficients of isothermal (UO 236Th0 764)02 agreement to diffusion models
2D Diffusion 3D Diffusion
Isothermal temperature k (min-1) R2 k (min-1) R2
4500C 3x10-4 0.9969 2.0x10-4 0.9330
5000C 3x10-4 0.8823 2.9x10-4 0.9871
5500C 8x10-4 0.8365 7.1x10-4 0.9807


Table 4-4. Kinetic results for nonisothermal (UO 236Th0 764)02 agreement to diffusion models
2D Diffusion 3D Diffusion
Nonisothermal ln(A) E (kJ/mol) R2 ln(A) E (kJ/mol) R2
(min-1) (min'1)
lC/min 4.63 81.2 0.9888 5.48 94.4 0.9978
30C/min 4.36 77.7 0.9946 5.27 92.2 0.9982
50C/min 4.83 81.6 0.9953 5.06 92.2 0.9971




















450C
1

2D Diffusion
0.8


0.6 -


0.4
3D Diffusion

0.2


0
0 500 1000 1500 2000 2500 3000 3500 4000
t-to

Figure 4-11. Isotherm at 4500C of (U 236Th0 764)02 oxidation fit to 2D and 3D Diffusion models.
Linear correlation values (R2) are 0.9969 and 0.933, respectively







1.2
500C
1 _2D Diffusion


0.8


,0.6
4--
3D Diffusion
0.4


0.2


0


0 500 1000 1500
t-to


2000 2500 3000


Figure 4-12. Isotherm at 5000C of(U 236Th 764)02 oxidation fit to 2D and 3D Diffusion models.
Linear correlation values (R2) are 0.8823 and 0.9871, respectively.


























S0.6 -
4--

0.4


0.2 -

0
0 200 400 600
t-to


800 1000 1200


Isotherm at 5500C of (Uo 236Th0 764)02 oxidation fit to 2D and 3D Diffusion models.
Linear correlation values (R2) are 0.8365 and 0.9807, respectively.


U0 236Th0 76402


1.2 1.25 1.3 1.35 1.4
1/T x 103 (K1)
Arrhenius plot of (Uo 236Th0 764)02 isotherms fit to 3D Diffusion models. Rate
constants (k) were determined by fitting gravimetric oxidation data with the 3D
diffusion reaction model


Figure 4-13.


Figure 4-14.


















1.E-02


1.E-03



0 1.E-04
0


1 .E-05


2D Diffusion


3D Diffusion


1 C/min to 900C


1.E-06
0.80


1.00 1.20 1.40
1/T x 103 (K1)


1.60


1.80


Figure 4-15. Nonisotherm at lC/min (Uo 236Th0 764)02 Arrhenius plot fit to 2D and 3D diffusion
models. Linear trendlines shown.








1.0E-02

2D Diffusion
1.0E-03

3D Diffusion

D 1.0E-04 "



1.0E-05

3 C/min to 900C

1.0E-06 I I


0.80 0.90 1.00 1.10 1.20 1.30
1/T x 103 (K1)


1.40 1.50 1.60


Figure 4-16. Nonisotherm at 30C/min (Uo 236Th0 764)02 Arrhenius plot fit to 2D and 3D diffusion
models. Linear trendlines shown.





















1.0E-02




1.0E-03




0) 1.0E-04

1.0E-05

1.OE-05


5 C/min to 900C
1.0E-06 00
0.80 0.90 1.00


3D Diffusion


Figure 4-17. Nonisotherm at 5C/min (Uo 236Th0 764)02 Arrhenius plot fit to 2D and 3D diffusion
models. Linear trendlines shown.








-3


Oa=0.4
Da=0.5
Aa=0.6


-35


-4 ..

A-.


-45


-5


55


1.10 1.20 1.30 1.40 1.50 1.60
1/Tx 103 (K-1)


-o


-6 U 236Th0 76402

-6 5


13
1/T x103 (K1)


1 35


Figure 4-18. Model free (Uo 236Th0 764)02 isotherms plotted at o = 0.4, 0.5, 0.6. Slope of the
linear trendlines gives -E/R.










Mixed Oxide (U0.368Th0.63202)

The same protocol as described in the previous section was applied to (Uo 368Th0 632)02,

which was isothermally measured in triplicate at 4500C, 475C, and 5000C. Unlike the other

(U,Th)02 mixed oxides and U02, (Uo 368Th0 632)02 isothermal runs were done in triplicate to

establish statistical significance and error analysis for the calculated kinetic parameters.

Isothermal rate data were inserted into reaction models from Table 3-5. Those that showed

nonlinear behavior were rejected, leaving only the 3D diffusion model. Figures 4-19 to 4-21

show the three sets of isothermal rate data per temperature fit to the 3D diffusion reaction model

and the associated linear trendlines. Rate coefficients (k) obtained from the slopes of the f(a) vs.

time plots are compiled into a In(k) vs. 1/T Arrhenius curve. Activation energy and frequency

factor are calculated. One 4750C isothermal run (labeled "iso475" in Figure 4-20), however, was

discarded because, after reaching a maximum weight gain, the sample began to inexplicable lose

weight. Although this run is included in Figure 4-20, it is not incorporated into any kinetic

analysis methods. Additionally, scatter was observed in the measured signal at increasingly long

dwell times (i.e., greater than 1000 min).

The rate coefficients obtained from the linear slope are detailed in Table 4-5 and

assembled into an Arrhenius In(k) vs. 1/T plot (Figure 4-22) to determine the activation energy

and frequency factor for isothermal data. The slope of the linear trendline yielded activation

energy and frequency factor of 171 8 kJ/mol and 20.8 1.3, respectively. These values are

significantly greater than those determined for isothermal model-fit (Uo 236Th0 764)02. This would

seem to indicate that increasing U content results in an increased threshold to oxidation. The

higher frequency factor also suggests there are more events occurring in the higher U content

mixed oxide. This, however, is contradictory to the premise that increasingly uranium-rich

oxides more readily oxidize than thorium-rich compositions. That (Uo 236Th0 764)02 isothermal

data agreed with 2D and 3D diffusion models, whereas only 3D diffusion was identified for










(Uo 368Th0 632)02, may allude to a possible explanation. If diffusion occurs at the surface (2D) and

bulk (3D), it is possible that one takes place more readily than the other. Without a particle

and/or grain size term included in the reaction model, the plots generated in this study cannot

verify whether activation energies calculated reflect whether either system is dominated by

surface or bulk reactions or by temperature. The sensitivity to changes in solid-gas and/or grain-

grain surfaces is also unknown without an interface-dependent term in the reaction model.

Linear agreement of nonisothermal rate data to the 3D diffusion reaction model is shown

in Figure 4-23. Activation energy and frequency factor were calculated from the slopes and

intercepts of linear trendlines. They are summarized in Table 4-6. As with (Uo 236Th0 764)02,

nonisothermal results are significantly different from those for isothermal runs. Nonisothermal

and isothermal data appear to support 3D diffusion as the most likely mechanism of oxidation.

Values obtained are similar to those for (Uo 236Th0 764)02. Again, the absence of terms to account

for particle size and grain sizes casts a shadow on these results. The addition of time as a variable

brings more complexity to the calculated results than the solely temperature-dependent isothermal

measurements.

Model-free plots of (Uo 368Th0 632)02 isotherms yielded linear behavior within the

investigated temperature range. There was no indication either by model-fitting or model-free

methods of a multi-step process. This, of course, does not eliminate the possibility of overlapping

mechanisms, but is not suggested in these findings. Figure 4-24 summarizes the model-free

results.

Activation energies calculated from the slopes yield values of 103 20 kJ/mol, 105 17

kJ/mol, and 111 16 kJ/mol for degree of conversion 0.4, 0.5, and 0.6 respectively. Unlike

(Uo 236Th0 764)02, nonisothermal model-fit and model-free calculated activation energies are

similar. The isothermally model-fit values, however, are significantly higher. This, again,











appears to be the nature associated with model-fitting isothermal versus nonisothermal rate data.

Both techniques, however, do not disclose a multi-step process.



Table 4-5. Rate coefficient of isothermal (UO 368Th0 632)02 agreement to 3D diffusion model
Isothermal temperature k (min-') R2
4500C 3.90x10-4 0.894
4.83x10-4 0.932
4.27x10-4 0.984
4750C 1.17x10-3 0.965
9.35x10-4 0.974
5000C 2.72x10-3 0.989
2.98x10-3 0.989
2.53x10-3 0.983



Table 4-6. Kinetic results for nonisothermal (UO 368Th0 632)02 agreement to 3D diffusion model
3D Diffusion
Nonisothermal ln(A) E (kJ/mol) R2
(min-1)
lC/min 6.48 92.3 0.999
30C/min 6.80 92.5 0.999
50C/min 8.63 108.2 0.999


1
0.9 450C
0.8
0.7 iso'
0.6

40.5 iso450b
0.4
0.3 iso
0.2
0.1
0
0 250 500 750 1000
t-to, min


1250 1500 1750


Figure 4-19. Isotherm at 4500C for (Uo 368Th0 632)02 oxidation fit to 3D Diffusion reaction model.
Three runs measured.




















0.9

0.8 475C
iso475a
0.7

0.6 iso475

0.5

4 0.4 iso475b

0.3

0.2

0.1

0
0 100 200 300 400 500 600 700 800
t-to, min

Figure 4-20. Isotherm at 4750C for (UO 368Th0 632)02 oxidation fit to 3D Diffusion reaction model.
Run labeled "iso475" is later discarded.







0.9

0.8 500C

0.7 -

0.6 -

0.5 iso500b iso500a
t, iso5OOa
0.4

0.3

0.2
iso500c
0.1

0


0 50 100 150 200
t-to, min


250 300 350


Figure 4-21. Isotherm at 5000C for (Uo 368Th0 632)02 oxidation fit to 3D Diffusion reaction model.
Three runs measured


























-6.5


Figure 4-22.


-8
1.28 1.30 1.32 1.34 1.36 1.38 1.40

1/Tx 103 (K1)

Arrhenius plot of (Uo 368Th0 632)02 isotherms fit to 3D Diffusion reaction model.
Rate constants (k) were determined by fitting gravimetric oxidation data with the
3D diffusion reaction model


-4


-6


-8


-10


-12


-14


5 C/min


3 C/min


1 C/min


1 11 12 13 14 15 16 17 18 1 9
1/T x 10-3 (K1)

Figure 4-23. Nonisotherms at 1, 3, and 50C/min (Uo 368Th0 632)02 Arrhenius plot fit to 3D
diffusion model.


UO 368Th0 63202











-225
0 Oa=0.4
-275 Oa=0.5
S "''--.. Au=0.6
325

-3 75 A --..


-425

-4 75
UO 368Th0 63202 A
-5 25
1 28 13 1 32 1 34 1 36 1 38 14
1/T x 103(K1)
Figure 4-24. Model free (Uo 368Th0 632)02 isotherms plotted at x = 0.4, 0.5, 0.6. Slope of the
linear trendlines gives -E/R.


Mixed Oxide (U0.50Tho.5002)

Model-fitting of (Uo 500Th0 500)02 air oxidized pellet fragments introduces another reaction

model possibility, Avrami-Erofe'ev. With regard to the raw isothermal data (Figure 4-8),

oxidation is more than 50% complete within the first 500 minutes, which may contribute to

nonlinear behavior at dwell times < 200 min. However, linear agreement is observed at times

greater than 200 min at all isotherms. The validity of using either the reaction model to determine

k, E, and A kinetic parameters for the overall process becomes questionable as to its usefulness.

Selecting an appropriate model by global linear fit yields little insight into the oxidative

mechanism when the bulk of the process is complete in the ill-fit region. Figures 4-25 to 4-28

present the linear fit of isothermal measurements to both Avrami-Erofe'ev and 3D Diffusion

models. Figure 4-29 summarizes the rate coefficients according to the 3D diffusion reaction

model in an Arrhenius plot of In(k) vs 1/T. When compared to Arrhenius plots of

(Uo 236Th0 764)02 and (Uo 368Th0 632)02, the behavior of (Uo 500Th0 500)02 fit to the 3D diffusion

reaction model is inconsistent with the lower U content mixed oxides. Table 4-7 lists the rate

coefficients and linear correlation to the 3D diffusion model.










Although the lower temperatures exhibit a high degree of linearity, both models fall apart

at higher temperatures. This is particularly apparent in the Arrhenius ln(k) vs. 1/T plot, Figure 4-

29, of the 3D diffusion reaction model. The plot, in fact, yields a poorly fit linear trendline with a

positive slope. Calculating activation energy and frequency would, undoubtedly, be meaningless

for model fit isothermal rate data.

Attempts to fit nonisothermal data to a known reaction model also met with similarly

inconclusive results as shown in Figure 4-30. At the slower heating rate of lC/min, linear

agreement was observed in both the Avrami-Erofe'ev and 3D Diffusion models within the

temperature range of interest. The higher heating rate, 5C/min, on the other hand, did not

demonstrate linear behavior. This incongruity may be a result of a lag between reaction rate and

heating rate, where the temperature rises faster than the material is able to react. Table 4-8

summarizes kinetic parameters calculated from nonisothermal model fitting of rate data to the 3D

diffusion and Avrami-Erofe'ev models.

The inability to discern whether (Uo 500Tho 500)02 isothermal and nonisothermal rate data

conform to known reaction models presented the model-free technique as a desirable method for

calculating activation energies. A plot of the isothermal rate data by the model-free method is

shown in Figure 4-31. Again, degree of fraction reacted are selected at a = 0.4, 0.5, and 0.6.

Similar to the (Uo 236Th0 764)02 and (Uo 368Th0 632)02 model-free Arrhenius plots, there is no

suggestion of multiple step behavior. Considering the UO2 system is well known for being a

complex multiple step oxidation process, the simple linear behavior would suggest the thorium

oxide matrix as a stabilizing influence upon UO2 oxidation.

Activation energies calculated from the slopes yield values of 85.2 6.5 kJ/mol, 85.6 5.4

kJ/mol, and 86.2 5.4 kJ/mol for degree of conversion 0.4, 0.5, and 0.6 respectively.

Considering that isothermal and nonisothermal model-fitting techniques did not yield a










conclusive reaction model, it is not surprising that model-free activation energies are not in

agreement with model-fit results.


Table 4-7. Rate coefficient of isothermal (Uo 5oTho 50)02 agreement to 3D diffusion model
3D diffusion Avrami-Erofe'ev
Isothermal temperature k (min-1) R2 k (min-1) R2
3750C 6x10-4 0.9858 1x10-3 0.9821
4000C lxl0-3 0.9581 3.3x10-3 0.9648
4250C 6x10-4 0.8528 2x10-3 0.7495
4500C 6x10-4 0.7843 2.2x10-3 0.7763


Table 4-8. Kinetic results for nonisothermal (Uo 5ooTho 500)02 in 3D diffusion and Avrami-
Erofe'ev models
3D Diffusion Avrami-Erofe'ev
Noniosthermal ln(A) E R2 ln(A) E R2
(min1) (kJ/mol) (min1) (kJ/mol)
lC/min 20.12 166.7 0.9894 8.16 90.3 0.9951
50C/min 8.26 99.7 0.8722 11.91 116.2 0.9735


0 200 400 600 800 1000 1200
t-to
Figure 4-25. Isotherm at 3750C for (Uo 50Tho 50)02 oxidation in 3D Diffusion and Avrami-
Erofe'ev models. Linear trendline for both models are shown.






















400C


2.5


2


1.5

1


0.5


0
0 -':::::i -----------

0 100 200 300 400
t-to


Avrami-Erofe'ev


Figure 4-26. Isotherm at 4000C for (Uo 50Tho 50)02 oxidation in 3D Diffusion and Avrami-
Erofe'ev models. Linear trendline for both models are shown.









4


3.5

3
Arami-E
2.5

2 2
4--0
1.5

1 -Z 31

0.5

0 -
0 200 400 600
t-to


800 1000 1200


500 600 700 800


Figure 4-27. Isotherm at 4250C for (Uo 50Th0 50)02 oxidation in 3D Diffusion and Avrami-
Erofe'ev models. Linear trendline for both models are shown.





















3.5--

3

2.5 Avrami-Erofe'ev

2

1.5
1 3D Diffusion

0.5

0
0 200 400 600 800 1000
t-to

Figure 4-28 Isotherm at 4500C for (Uo 50Th0 50)02 oxidation in 3D Diffusion and Avrami-
Erofe'ev models. Linear trendline for both models are shown.







-6.8


-6.9


UosTh0sO


-7.2
--7.2
-7.3


-7.4


Figure 4-29.


7 5 ..
1.35 1.38 1.41 1.44 1.47 1.5 1.53 1.56

1/T x 103 (K1)

Arrhenius plot of (Uo 500Th0 500)02 isotherms in 3D Diffusion reaction model. Rate
constants (k) were determined by fitting gravimetric oxidation data with the 3D
diffusion reaction model








71











-4
5 C/min
-5 -

-6 Avram-Erofe'ev 1 C/min

-7 3D Diffusion
SAvrami-Erofe'ev
-8

-9

-10
3D Diffusion
-11

-12
1 1.1 1.2 1.3 1.4 1.5 1.6 1.7
1/Tx 103 (K1)

Figure 4-30. Nonisotherms at 1 and 50C/min (Uo 5ooTho 500)02 Arrhenius plot fit to Avrami-
Erofe'ev and 3D diffusion models. Linear trendlines shown.






Figure 4-31. Model free (Uo 5ooTho 500)02 isotherms plotted at x = 0.4, 0.5, 0.6. Slope of the
-2 -

A 0.4
-25 Oa=0.5
Sa =0.6
-3 ; -" ,


-35


-4-


-45
Uo5Tho5O2

-5
136 138 14 142 144 146 148 15 152 154 156
1/T x103 (K1)

linear trendlines gives -E/R.


Pure Uranium Dioxide (UO2)









Kinetic analysis of U02 pellet fragments encountered similar results to (Uo 5o0Th0 500)02

material. Linear agreement to the various reaction models was not observed for isothermal and

nonisothermal data. Considering U02 is known to have a multi-step oxidation, this is not

unexpected. As McEachern [McE97a] reported, U02 powders oxidize by parabolic kinetics,

which indicate the reaction is diffusion controlled, to form U307 at low temperatures. Diffusion

through a discrete layer of the product oxide is the limiting reaction. Two mechanisms,

concentration-gradient and discrete-layer, are generally accepted as consistent with the diffusion-

controlled kinetics for U02 powder oxidation to U307. Literature yields activation energy

estimates of the U307 formation on U02 powders to be 96 kJ/mol [McE97a].

In terms of U308 formation, McEachern et al. [McE97b] reported that nucleation and

growth mechanisms have been consistently applied to this step. Again, two models are

commonly applied in literature, Johnson-Mehl and Avrami-Erofe'ev equations. Also, it has been

proven correct that there are at least two different activation energies (at different temperature

ranges) for U308 formation, with a change in oxidation behavior around 300 3500C [McE97b].

It is not unlikely, therefore, to expect that same degree of complexity at temperatures greater than

3500C, as examined in this study. Considering the nucleation and growth complexity of U308

formation, weight gain data is susceptible to interference from U307/U409 formation and linear

growth rates are not necessarily applicable. McEachern et al. [McE97b] developed a two-

dimensional nucleation and growth model for the specific case of U308 formation on U02 pellets,

analogous to three-dimensional nucleation and growth models, resulting in a calculated activation

energy of 146 10 kJ/mol.

Since oxidation of UO2 was pursued as a control for comparison purposes, not

necessarily identifying oxidation mechanisms, the model-free method was used to determine E.

However, knowing the U02 fragments have a low surface-to-bulk ratio, it is likely the plot mainly

reflects the bulk contribution to oxidation and not U307 formation.












Activation energies calculated for a = 0.5, 0.7, and 0.9 across the 375 450C


temperature range are 116 29 kJ/mol, 104 + 23 kJ/mol, and 100 20 kJ/mol respectively.





0 -

-0 5
1 D= 0.7

-15

2

-25 '' '-
-3 A A'" a""--'.'"-

35

-4- UO

-45
136 1 38 1 4 1 42 144 1 46 148 15 152 1 54 156
1/Tx 103 (K1)

These are within the range of values reported in literature for pellet fragments. See Appendix A.


Figure 4-32. Model free U02 isotherms plotted at a = 0.4, 0.5, 0.6. The slope of the linear
trendlines gives -E/R.
















CHAPTER 5
SUMMARY AND CONCLUSIONS

Two of the compositions analyzed ((Uo 236Th0 764)02 and (Uo 368Th0 632)02) exhibited

satisfactory agreement to any of the basic theoretical reaction models, absent any additional terms

to account for particle size or grain size. As a consequence, results calculated with these basic

models restrict conclusions that can be drawn (i.e., surface or bulk-dependence). The

(Uo 500Th0 500)02 and U02 compositions did not demonstrate agreement with any of the basic

theoretical reaction models. Knowing the two-step nature of U02, the absence of a particle size

dependent term prohibited the identification of an appropriate reaction model. Nonisothermal

calculations, additionally, were further hindered by the simultaneous time and temperature

dependence of measurement. Further quantifying particle characteristics would be necessary to

render more precise kinetic parameters.

For the low U compositions, the 3D Diffusion model provided the best fit for isothermal

rate data. (Uo 236Th0 764)02 fragments, however, did present agreement to 2D diffusion at 4500C.

Activation energy values of isothermal rate data calculated using the 3D diffusion reaction model

were 62.1 17.9 kJ/mol and 171 8 kJ/mol, for the (Uo 236Th0 764)02 and (Uo 368Th0 632)02

samples, respectively Frequency factors (InA) were 1.71 2.87 min'and 20.8 1.3 min' for

increasing U content, respectively. These values are contrary to the hypothesis that uranium-rich

compositions would more readily oxidize than thorium-rich ones. Based on the known cubic

fluorite structure, it seems likely that the lattice at interstitial sites accommodates the excess

anions without having to undergo phase transformations. Additionally, assuming pellet

fragments are polycrystalline, excess anions may also accumulate or migrate along grain









boundaries. This incongruity is likely a result of the limitations attributed to the missing particle

size and/or grain size terms. Without fully characterizing the pellet fragments, the basic reaction

model can not adequately account for possible complexities, such as multiple mechanisms,

simultaneous mechanism, surface-dependency, grain boundary migration. The kinetic analysis in

this study does preliminarily suggest that (U,Th)02 oxidation occurs by diffusion. X-ray

diffraction and uranium valence calculations additionally assert that thorium oxide inhibits

uranium from reaching its highest valence state.

The higher uranium (Uo 500Th0 500)02 fragments did not indicate agreement with diffusion-

controlled models. In fact, the nucleation and growth Avrami-Erofe'ev reaction model appeared

to fit (Uo 500Th0 500)02 rate data better than 3D diffusion in nonisothermally measured fragments.

Considering the reaction model agreement alternated between 3D diffusion and Avrami-Erofe'ev,

there is likely more than one mechanism in simultaneous action.

In general, nonisothermal rate data, particularly (Uo 236Th0 764)02 and (Uo 368Th0 632)02, did

not discriminately identify one model over another, despite exhibiting linear agreement with

models identified isothermally. Consequently, little meaning could be drawn from nonisothermal

data in this study. Like the nonisothermal data, the relatively untried model-free method reported

in 1999 by Vyazovkin and Wight was tentatively used for comparison. There was no correlation

noted between (U,Th)02 composition and model-free calculated activation energy. Table 20

summarizes values obtained by all three methods.

Gravimetric analysis of UO2 was also subjected to the same limitations that befell mixed

(U,Th)02 oxides. Single-step nonisothermal gravitational analysis curves were observed for UO2

and all (U,Th)02 samples. The absence of two-step curves in the case of U02 is attributed to the

large pellet fragments (90 250 pm) used, which have a low surface-to-volume ratio compared

with much finer powders. It is unclear from this study whether the single-step behavior observed

for the solid solutions was also due to the size of the fragments used, or if such behavior is

characteristic of (U,Th)02 materials, regardless of the particle size.










The results of this study highlight the complexities in the U02 and (U,Th)02 oxidation

processes, as well as difficulties in performing kinetic analyses of this nature. None of the

(U,Th)02 samples reached the same degree of uranium oxidation as pure U02, indicating a

stabilizing effect by thorium in the solid solution. It is clear that diffusion is the primary

mechanism of oxidation with possible nucleation and growth mechanisms observed at

increasingly greater uranium content. This, perhaps, may explain the incongruity between

activation energies and uranium content. At the lower uranium concentrations where diffusion

dominates, activation energy increases with increasing U/Th. At U/Th = 0.5, however, it is no

longer clear that diffusion is the primary mechanism of oxidation. There is a suggestion that

nucleation and growth mechanisms are beginning to play a significant role in the uranium

oxidation. Activation energies calculated, subsequently, do not reveal the system complexity

suggested by the model-fit method. Again, further investigation of mid- to high U content mixed

oxides may reveal the cause behind the lowered activation energy as compared to lower U/Th

oxides.

This study, however, does put forward suggestions as to possible mechanisms for mixed

(U,Th)02 dry oxidation. X-ray diffraction confirmed that mixed oxide lattice structures remain

cubic fluorite during oxidation. Assuming weight gain is solely a result of excess oxygen anions,

uranium valency does not proceed to the same extent as oxidation of pure U02. Also, much like

UO2, excess anions are likely accommodated at interstitial sites and vacancies in the thorium-rich

lattice.


Table 5-1. Estimated E and A by model-free and model-fit techniques of (U,Th)02 and U02
Model free Isothermal Model Fit Nonisothermal Model fit
Composition E E In(A) E In(A)
(kJ/mol) (kJ/mol) (min-1) (kJ/mol) (min1)
(Uo 236Th0 764)02 87 + 8 62.7+ 17.9 1.71 +2.87 92.9+ 1.3 5.3 0.2
(Uo 368Tho 632)02 106 +31 171 8 20.8 1.3 97.7 9.1 7.3 1.2
(U 500Th0500)02 86 +6 -- 103 + 18* 10 + 3*
U02 107 +42 -- -- -- --
*Avrami-Erofe'ev model, otherwise 3D diffusion

















APPENDIX A
ACTIVATION ENERGIES


Table A-1. Published estimates of U307/U409 activation energy of formation
Eat, (kJ/mol) Sample T (C) Ref.
104* U02 powder 131 -164.5 Anderson et al. (1955)
102* U02 powder 161 350 Aronson et al. (1957)
90.8 U02 powder and pellets 125 280 [Bla58]
120+ 8 UO2 powder 143 -211 Walker (1965)
100 Spent LWR fuel 175 -195 Einziger et al. (1992)
113 + 17 Spent LWR fuel 175 225 Woodley et al. (1988, 1989)
* Activation energy re-calculated from investigators original data using the discrete-layer kinetic model
(rather than the concentration-gradient model) [McE97a].

Table A-2. Published estimates of U308 formation on UO2 activation energies


Sample
U02 powder
U02 powder
U02 microspheres
U02 powder
U02 powder
U02 pellets
AGR pellet fragments
AGR pellet fragments
CANDU pellets
LWR Pellets
U02 powder
CANDU pellets
CANDU pellets
CANDU fuel element


63 CANDU fuel element


T (C) Method
278-325 Gravimetric
315-360 Gravimetric
300-450 Gravimetric
365-400 DTA
312-352 Gravimetric
279-361 Gravimetric
200-300 Gravimetric
300-550 Gravimetric
200-300 XRD
200-250 Gravimetric
200-350 Gravimetric
330-350 Gravimetric
350-450 Gravimetric
250-300 Progression of
oxidation front
300-350 Progression of
oxidation front


Cited Ref.
Aronson (1961)
Saito (1975)
Ohashi (1987)
Landspersky (1966)
Walker (1965)
Walker (1965)
Tucker (1987)
Tucker (1987)
Taylor (1992)
White (1983)
[Boa77]
[Boa77]
[Boa77]
[Boa77]

[Boa77]


143 U02 pellet fragments 250-350 Gravimetric You (1992)
109 U02 pellet fragments 350-400 Gravimetric You (1992)
94.5 used LWR fuel 300-400 Gravimetric You (1992)
140 Unirradiated CANDU fragments 175-400 Gravimetric Hastings (1986)
120 Used CANDU fragments 175-400 Gravimetric Hastings (1986)
194 Used LWR fragments 250-360 Visual examination Einziger (1984)
* Activation energy was observed to vary as a function of oxygen pressure.
+ The value of 120 kJ/mol corresponds to the oxidation prior to powder formation. The post-spallation
period displayed an activation energy of 160 kJ/mol [McE97b].


Ea,t (kJ/mol)
146
127.6
~100*
161.5
134.7
110.5
170.2
48
124-139
102
163
170
67
172










Table A-3. Published estimates of U02 cation and anion diffusion activation energies
Eact(kJ/mol) Sample T(C) Method Ref
Cation Diffusion
372 U in U02 01 1400 1650 a-energy spectrometry [Haw68]
398 U in U0203 1400 1600 a-energy spectrometry [Haw68]
439 U in U021 1400 1650 a-energy spectrometry [Haw68]
339 U in U0215 1350 1450 a-energy spectrometry [Haw68]
304 237U in U02+ Lattice diffusion contribution [Fur68]
438 237U in UO2 1300 1600 Cited Lindner and Schmitz [Fur68]
411 U in (S) U02 Cited Reimann & Lundy [And83]
Anion Diffusion
273 O in U02 550 780 Cited Auskem and Belle [And83]
237 O in U02 600 1500 Cited Belle [And83]
248 O in U02 780 1250 Cited Marin & Contamin [And83]
124 O in U02 004 Cited Auskem & Belle [And83]
124 O in U02 063 Cited Auskem & Belle [And83]
89 O in U02 006 Cited Contamin et al. [And83]
89 O in U02 020 Cited Contamin et al. [And83]
90 O in U02 10 Cited Contamin et al. [And83]
92 O in U0212 Cited Contamin et al. [And83]
92 O in U0216 Cited Contamin et al. [And83]
97 O in U02 08 Cited Murch et al. [And83]
99.6 O in U02+ Self diffusion, cited Breitung (1978) [AriOO]
100.1 103.5 O in U02+ Chemical diffusion, cited Breitung [AriOO]
(1978)
119.2 O in U02+ Chemical diffusion, cited Lay (1970) [AriOO]
86.6 O in U02+x Chemical diffusion, cited Bayoglu [AriOO]
(1984)
96.7 O in U02+xSelf diffusion, cited Murch (1975) [AriOO]
(P) polycrystalline; (S) single crystal


Table A-4. Published estimates of ThO2 cation and anion diffusion activation energies
Eact(kJ/mol) Sample T(C) Method Ref
Cation Diffusion
247 Th in (P) Th2 1600 2100 a-energy spectrometry [Haw68]
626 Th in (S) ThO2 1846 -2045 Cited King [And83]
628 238Th in (S) ThO2 Cited Matzke [And83]
320 237U in (P) ThO2 1800 -2000 Lattice diffusion contribution [Fur68]
201 237U in (P) ThO2 1800 -2000 Grain boundary diffusion [Fur68]
contribution
Anion Diffusion
219 180 in (P) ThO2 1099 1644 mass spectrometry [And84]
209 180 in (S) ThO2 1099 -1644 mass spectrometry [And84]
209 180 in (S) ThO2 845 1646 gas-solid isotope exchange [And76]
(intrinsic contribution)
73.6 180 in (S) ThO2 845 1646 gas-solid isotope exchange [And76]
(extrinsic contribution)
275 O in Th02 900 1500 Cited Edwards et al. [And83]
238 O in ThO2 2100 2800 Cited Lam [Fre80]
(P) polycrystalline; (S) single crystal










Table A-5. Published estimates of diffusion in (U,Th)02 activation energies


E,ct(kJ/mol)
360
269


99.6
88.3
71.0
66.8


Sample
237U in (P) UO2-ThO2
237U in (P) U02-ThO2


O in Uo o Tho 9902+x
O in Uo 03Tho 9702+x
O in Uo osTho 9502+x
O in Uo 20Tho 8002+x


107.1 O in Uo 30Tho 7002+x

93.1 O in Uo 40Tho 5002+x

112.1 O in Uo 5oTho 5o02+x

108.0 O in Uo 7oTh0 3002+x


T(C)
1800 -2300
1800 -2300


Method
Lattice diffusion contribution
Grain boundary diffusion
contribution


940 1040 Chemical diffusion
940 1040 Chemical diffusion
940 1040 Chemical diffusion
1009 1100 Chemical diffusion, cited
Matsui (1985)
245 455 Chemical diffusion, cited
Furuya (1990)
1009 1100 Chemical diffusion, cited
Matsui (1985)
245 455 Chemical diffusion, cited
Furuya (1990)
245 455 Chemical diffusion, cited
Furuya (1990)


Table A-6. Kinetic parameters of mixed urania-thoria oxides
Sample Heating rate Activation energy Pre-exponential factor
(K/min) (kJ/mol) (log A) (/min)
(Uo 15Tho 85)02 0.5 45 0.3 0.747 0.006
1 42 + 0.3 0.693 + 0.004
5 31 + 0.2 0.471 + 0.002
Isotherms 51 + 1 2.6 0.4
(Uo 30Th 70)02 0.5 51 0.6 1.820 + 0.095
1 49 + 0.5 1.800 + 0.090
5 49 + 0.4 1.350 + 0.073
Isotherms 45 1 2.9 0.3
(Uo 72Tho 28)02 0.5 81 0.4 5.860 + 0.910
1 70 + 0.6 4.750 + 0.760
2 70 + 0.6 1.890 + 0.035
5 46 0.4 0.728 0.012
Isotherms 91 + 1 6.5 0.6
(Uo 77Th 23)02 0.5 90 + 0.6 7.250 + 0.140
1 79 0.4 5.860 + 0.160
2 67 0.8 4.760 + 0.190
5 66 0.8 4.600 + 0.160
Isotherms 82 1 7.2 1.4
[AntOO]


Ref
[Fur68]
[Fur68]


[AriOO]
[AriOO]
[AriOO]
[AriOO]

[AriOO]

[AriOO]

[AriOO]

[AriOO]

















APPENDIX B
X-RAY DIFFRACTION PATTERNS


Table B-1. Uo 368Tho 63202+x isothermal 4000C air oxidized compared to JCPDS standards
26meas Int* hkl 2025U Int hkl 20ThO2 Int hkl 20o2 Int hkl
27.830 51(100) 111 27.791 100 111 27.581 100 111 28.281 100 111
32.245 16(31) 200 32.186 40 200 31.962 35 200 32.741 50 200
46.270 21(41) 220 46.172 60 220 45.827 58 220 49.968 50 220
54.870 23(45) 311 54.769 55 311 54.312 64 311 55.754 45 311
57.525 5(10) 222 57.411 12 222 56.984 11 222 58.428 8 222
67.510 3(6) 400 67.366 8 400 66.821 8 400 68.594 10 400
74.555 9(18) 331 74.403 18 331 73.794 26 331 75.775 20 331
76.815 6(12) 420 76.659 14 420 76.010 17 420 78.121 15 420
85.820 5(10) 422 85.562 14 422 84.804 20 422 87.314 15 422
92.465 6(12) 511 92.185 14 511 91.316 19 511
103.650 1(2) 440 102.279 6 440
110.600 4(8) 531 109.069 18 531
113.010 2(4) 600 111.377 8 600
123.010 2(4) 620 121.073 14 620
131.345 2(4) 533 129.026 9 533
134.440 1(2) 622 131.922 9 622
26s, Int hkl
28.375 100 111 28.466 100 111
47.235 30 220 47.343 55 220
56.060 14 311 56.170 30 311
69.120 2 400 69.194 6 400
76.305 5 331 76.450 11 331
87.955 6 422 88.115 12 422
94.830 2 511 95.048 6 511
106.595 2 440 106.839 3 440
113.995 2 531 114.230 7 531
127.420 2 620 127.728 8 620
136.745 1 533 137.124 3 533__
*The relative intensities designated in parentheses are calculated based on the identified U,Th oxide angles,
not including Si, to determine whether intensities are consistent with those reported for Uo 25Tho 7502, ThO2,
and U02 cubic oxides. Silicon standard measured and reported values are also shown










800


700
111s,

600
220s

500


400
0
o
311s,
300
311 JTh

200 111Th 220UTh 331s 422s
S200Uh 420u Th
I 400s,331U h 422U h 511s, 531 U s 533s
100 222UTh511u 620s,


0
10 20 30 40 50 60 70 80 90 100 110 120 130 140
20

Figure B-1. XRD pattern for 450C isotherm of co-milled 23.6% U02 fragments


Anode: Cu
k,1,hI2 : 1.54056, 1.54439
Generator Voltage: 40V
Tube Current: 20


Scan Step Size: 0.0200
Scan Step Time: 0.25 sec


Table B-2. XRD peak intensities for 4500C isotherm of co-milled 23.6% U02 fragments


20 d Int
27.76 3.211 11
32.17 2.7802 5
46.115 1.9667 8
54.66 1.6778 8
57.31 1.6063 2
67.255 1.3909 1
74.245 1.2763 3
76.49 1.2443 4
85.47 1.1351 3
92.015 1.0707 2
103.06 0.9838 1
109.945 0.9407 2
112.4 0.9269 1
122.235 0.8797 1


h k 1
1 1 1
2 0 0
2 2 0
3 1 1
2 2 2
4 0 0
3 3 1
4 2 0
4 2 2
5 1 1
4 4 0
5 3 1
6 0 0
6 2 0
620


130.455 0.8483 1 5 3 3
133.395 0.8387 1 6 2 2
(Si)
28.375 3.1428 100 1 1 1
47.22 1.9233 29 2 2 0
56.05 1.6394 13 3 1 1
69.035 1.3593 3 4 0 0
76.29 1.2471 5 3 3 1
87.945 1.1094 5 4 2 2
94.85 1.046 2 5 1 1
106.575 0.9609 1 4 4 0
113.965 0.9186 2 5 3 1
127.435 0.8591 1 6 2 0
136.78 0.8285 1 5 3 3










800


700
111s,

600


500
220s

400
o

300
331s,
311s, 420uTh
200 111UTh 220U T 311u
331U h 422s,531
201U h 422U h511UTh 531UTh
100 222UTh 4004 0 I s 620s 622UTh
00 ---------------------------------------------------
400u Th s, 4 0U


0
10 20 30 40 50 60 70 80 90 100 110 120 130 140
20

Figure B-2. XRD pattern for 5000C isotherm of co-milled 23.6% U02 fragments


Anode: Cu
k,1,hI2 : 1.54056, 1.54439
Generator Voltage: 40V
Tube Current: 20


Scan Step Size: 0.0200
Scan Step Time: 0.25 sec


Table B-3. XRD peak intensities for 5000C isotherm of co-milled 23.6% U02 fragments


26 d Int
27.775 3.2093 13
32.195 2.7781 4
46.155 1.9651 7
54.705 1.6765 6
57.325 1.6059 2
67.24 1.3912 1
74.27 1.2759 3
76.54 1.2437 4
85.44 1.1354 2
92.015 1.0707 2
103.11 0.9835 1
110.04 0.9401 2
112.38 0.9271 1
122.27 0.8796 1


h k 1
1 1 1
2 0 0
2 2 0
3 1 1
2 2 2
4 0 0
3 3 1
4 2 0
4 2 2
5 1 1
4 4 0
5 3 1
6 0 0
6 2 0
620


130.41 0.8485 1 5 3 3
133.44 0.8386 1 6 2 2
Si
28.38 3.1422 100 1 1 1
47.235 1.9227 21 2 2 0
56.045 1.6395 7 3 1 1
69.025 1.3595 1 4 0 0
76.305 1.2469 4 3 3 1
87.945 1.1094 3 4 2 2
94.86 1.0459 2 5 1 1
106.59 0.9608 1 4 4 0
113.99 0.9185 1 5 3 1
127.405 0.8592 1 6 2 0
136.76 0.8286 0 5 3 3











800


700
111s,

600


500

220s
S400
0

300

311s,
331s,
200 111UTh 220UTh 311u 420U Th
S331U T 42 s,
200UTh 422u h 511u rh 531UTh
100 222U Th 400s 511 ---- 622Urh
400u h


0
10 20 30 40 50 60 70 80 90 100 110 120 130 140
20

Figure B-3. XRD pattern for 550C isotherm of co-milled 23.6% UO2 fragments


Anode: Cu
,1,Ik2 : 1.54056, 1.54439
Generator Voltage: 40V
Tube Current: 20


Scan Step Size: 0.0200
Scan Step Time: 0.25 sec


Table B-4. XRD peak intensities for 550C isotherm of co-milled 23.6% U02 fragments


20 d Int
27.795 3.207 14
32.215 2.7764 4
46.125 1.9663 7
54.69 1.6769 7
57.37 1.6048 2
67.25 1.391 1
74.265 1.276 4
76.525 1.2439 4
85.46 1.1352 2
92.055 1.0703 2
103.135 0.9833 1
110.01 0.9403 2
112.37 0.9271 1
122.2 0.8799 1
130.435 0.8484 1
133.305 0.839 0


h k 1
1 1 1
2 0 0
2 2 0
3 1 1
2 2 2
4 0 0
3 3 1
4 2 0
4 2 2
5 1 1
4 4 0
5 3 1
6 0 0
6 2 0
5 3 3
6 2 2
622


(Si)
28.38 3.1422 100 1 1 1
47.23 1.9229 21 2 2 0
56.05 1.6394 8 3 1 1
69.065 1.3588 2 4 0 0
76.29 1.2471 4 3 3 1
87.955 1.1093 4 4 2 2
94.855 1.046 2 5 1 1
106.61 0.9607 1 4 4 0
113.995 0.9185 1 5 3 1
127.495 0.8589 1 6 2 0
136.79 0.8285 0 5 3 3











800


700


600


500


400
o

300


200


100


0


422UTh 511,


10 20 30 40 50 60 70 80 90 100 110 120 130 140
20


Figure B-4. XRD pattern for 4000C isotherm of co-milled 36.8% U02 fragments


Anode: Cu
,I,1k,2 : 1.54056, 1.54439
Generator Voltage: 40V
Tube Current: 20


Scan Step Size: 0.0200
Scan Step Time: 0.25 sec


Table B-5. XRD peak intensities for 4000C isotherm of co-milled 36.8% U02 fragments


26
27.83
32.245
46.27
54.87
57.525
67.51
74.555
76.815
85.82
92.465
103.65
110.6
113.01
123.01
131.345
134.44


d
3.2031
2.7739
1.9605
1.6718
1.6008
1.3863
1.2718
1.2399
1.1314
1.0666
0.9798
0.9369
0.9237
0.8765
0.8453
0.8354


Si
28.375
47.235
56.06
69.12
76.305
87.955
94.83
106.595
113.995
127.42
136.745


3.1428
1.9227
1.6391
1.3579
1.2469
1.1093
1.0462
0.9608
0.9185
0.8591
0.8286











800


700
111 ,

600

111UTh 311UTh
500 -\220u Th


400
0
220s
300
331uTh 331s
200u4h 0U h I \ 531....
|200Th 4 Th 422u Th 511U Th 531
2002
222UTh 422s, 1600u Th 620 Th 5331Th

400s 440 620 33

0
10 20 30 40 50 60 70 80 90 100 110 120 130 140
28

Figure B-5. XRD pattern for 4500C isotherm of co-milled 36.8% UO2 fragments


Anode: Cu
,I,1o,2 : 1.54056, 1.54439
Generator Voltage: 40V
Tube Current: 20


Scan Step Size: 0.0200
Scan Step Time: 0.25 sec


Table B-6. XRD peak intensities for 4500C isotherm of co-milled 36.8% U02 fragments


20 d
27.85 3.2008
32.27 2.7718
46.29 1.9597
54.88 1.6715
57.535 1.6006
67.52 1.3861
74.515 1.2724
76.905 1.2387
85.785 1.1317
92.45 1.0668
103.625 0.98
110.575 0.9371
112.95 0.924
123.015 0.8764
131.25 0.8457
134.45 0.8354


h k 1
1 1 1
2 0 0
2 2 0
3 1 1
2 2 2
4 0 0
3 3 1
4 2 0
4 2 2
5 1 1
4 4 0
5 3 1
6 0 0
6 2 0
5 3 3
6 2 2
622


Si
28.405 3.1395
47.255 1.9219
56.075 1.6387
69.1 1.3582
76.325 1.2466
87.98 1.1091
94.875 1.0458
106.615 0.9606
113.995 0.9185
127.395 0.8592
136.74 0.8286


1 1 1
2 2 0
3 1 1
4 0 0
3 3 1
4 2 2
5 1 1
4 4 0
5 3 1
6 2 0
5 3 3
533











800


700 111s, 220s


600


500

311s,
400 311u
0 111umT
220U h
300 331s, 422s,
331U Th 420Th
2 UTh
200 511UTh 531U Th
222UTh 400s 422U h 511s, 60UT 620S,
440s, I 622UTh
100 -L 400u 620UT 533Uh533
1000u Tu "


0
10 20 30 40 50 60 70 80 90 100 110 120 130 140
20

Figure B-6. XRD pattern for 5000C isotherm of co-milled 36.8% UO2 fragments


Anode: Cu
o1,,k2 : 1.54056, 1.54439
Generator Voltage: 40V
Tube Current: 20


Scan Step Size: 0.0200
Scan Step Time: 0.25 sec


Table B-7. XRD peak intensities for 5000C isotherm of co-milled 36.8% U02 fragments


26 d Int
27.85 3.2008 24
32.285 2.7705 14
46.27 1.9605 14
54.85 1.6724 12
57.53 1.6007 3
67.55 1.3856 2
74.53 1.2721 6
76.885 1.2389 4
85.8 1.1316 3
92.46 1.0667 4
103.65 0.9798 1
110.515 0.9374 3
112.975 0.9239 2
122.995 0.8765 2
131.33 0.8454 1
134.36 0.8357 1


h k 1
1 1 1
2 0 0
2 2 0
3 1 1
2 2 2
4 0 0
3 3 1
4 2 0
4 2 2
5 1 1
4 4 0
5 3 1
6 0 0
6 2 0
5 3 3
6 2 2
622


Si
28.405 3.1395 100 1 1 1
47.255 1.9219 38 2 2 0
56.065 1.639 15 3 1 1
69.085 1.3585 4 4 0 0
76.325 1.2466 7 3 3 1
87.95 1.1094 7 4 2 2
94.86 1.0459 3 5 1 1
106.59 0.9608 2 4 4 0
113.965 0.9186 3 5 3 1
127.395 0.8592 2 6 2 0
136.76 0.8286 1 5 3 3











800


700
111s,

600
220s,
111UT
500


400
0
o

300
311s,

200
200 331s, 422s 531s,
220 311331 600UT
200U T i( 400s, [ 511S 531u] 622UT
100 U 11222uTh 400T 11420 422 5111 4040\ 1 620 533
t uJ i i- -I .l l l ll UT 400uT I I 620u s 533 ,T53

0
10 20 30 40 50 60 70 80 90 100 110 120 130 140
20


Figure B-7. XRD pattern for 3750C isotherm of co-milled 50.0% U02 fragments


Anode: Cu
1,2I,,, : 1.54056, 1.54439
Generator Voltage: 40V
Tube Current: 20


Scan Step Size: 0.0200
Scan Step Time: 0.25 sec


Table B-8. XRD peak intensities for 3750C isotherm of co-milled 50.0% U02 fragments


20 d Int
27.92 3.1929 5
32.405 2.7605 2
46.47 1.9525 3
55.16 1.6637 4
56.2 1.6354 4
67.78 1.3814 1
74.915 1.2665 2
76.52 1.2439 2
86.26 1.1267 1
92.805 1.0636 1
104.37 0.9750 0
111.205 0.9335 1
113.985 0.9185 2
123.7 0.8736 0


h k 1
1 1 1
2 0 0
2 2 0
3 1 1
2 2 2
4 0 0
3 3 1
4 2 0
4 2 2
5 1 1
4 4 0
5 3 1
6 0 0
6 2 0
620


128.005 0.8570 0 5 3 3
132.465 0.8417 0 6 2 2
(Si)
28.37 3.1433 100 1 1 1
47.22 1.9233 26 2 2 0
56.04 1.6397 9 3 1 1
69.04 1.3593 2 4 0 0
76.3 1.2470 4 3 3 1
87.945 1.1094 4 4 2 2
94.84 1.0461 2 5 1 1
106.635 0.9605 1 4 4 0
113.985 0.9185 2 5 3 1
127.41 0.8592 1 6 2 0
136.7 0.8287 1 5 3 3







88


800


700
111s,


600

220s,
500


400
o


300
311s,
311ul

200 11 UTh 331s, 422s 531s

200U 331uT 511s 600u T
200

400s 10042UT 4221 620s 622u
100 40222UTh 5 440s 620UT 533L s,
0400u5


10 20 30 40 50 60 70 80 90 100 110 120 130 140
20

Figure B-8. XRD pattern for 4000C isotherm of co-milled 50.0% U02 fragments


Anode: Cu Scan Step Size: 0.0200
,Ik2 : 1.54056, 1.54439 Scan Step Time: 0.25 sec
Generator Voltage: 40V
Tube Current: 20


Table B-9. XRD peak intensities for 4000C isotherm of co-milled 50.0% U02 fragments


26 d Int h k 1 (Si)
27.965 3.1879 11 1 1 1 28.38 3.1422 100 1 1 1
32.415 2.7597 5 2 0 0 47.235 1.9227 28 2 2 0
463525 1.9504 7 2 2 0 56.05 1.6394 11 3 1 1
55.105 1.6652 9 3 1 1 69.05 1.3591 2 4 0 0
57.845 1.5927 2 2 2 2 76.3 1.2470 5 3 3 1
67.83 1.3805 1 4 0 0 87.945 1.1094 6 4 2 2
74.865 1.2673 2 3 3 1 94.855 1.0460 3 5 1 1
77.17 1.2351 2 4 2 0 106.595 0.9607 2 4 4 0
86.235 1.1270 2 4 2 2 114.005 0.9184 2 5 3 1
92.865 1.0631 2 5 1 1 127.415 0.8592 1 6 2 0
104.265 0.9757 1 4 4 0 136.765 0.8286 1 5 3 3
111.15 0.9338 2 5 3 1
114.005 0.9184 2 6 0 0
123.765 0.8734 1 6 2 0
132.31 0.8422 1 5 3 3
135.355 0.8327 1 6 2 2