<%BANNER%>

Fabrication and Characterization of Silicon Carbide Inert Matrix Fuels Through a Polymer Precursor Route

Permanent Link: http://ufdc.ufl.edu/UFE0042388/00001

Material Information

Title: Fabrication and Characterization of Silicon Carbide Inert Matrix Fuels Through a Polymer Precursor Route
Physical Description: 1 online resource (152 p.)
Language: english
Creator: Shih, Chunghao
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: imf, pip, precursor, sic
Materials Science and Engineering -- Dissertations, Academic -- UF
Genre: Materials Science and Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Plutonium management is a crucial issue for the nuclear industry since it is an unavoidable by-product of the well established uranium fuel cycle. Inert matrix materials are proposed as a matrix material to carry the plutonium fuel so that plutonium can be transmuted in nuclear reactors. Silicon carbide is a promising inert matrix candidate material because of its low neutron absorption cross section, high thermal conductivity and high temperature stability. The drawback of using silicon carbide as a fuel matrix is the high temperature and pressure required for fabrication. Chemical reactions between SiC and plutonium at higher temperatures may also be an issue. In this study, a low temperature process of using crystalline beta silicon carbide particles and a polymer precursor was developed to synthesize SiC inert matrix fuel pellets.The low temperature process is required to prevent the reactions between SiC and the PuO2 fuel material. The effect of the polymer content and the cold pressing pressure on the packing of SiC particles was investigated. The effect of mixing the coarse and fine SiC particles on the density, the microstructure and the pore size distribution was also investigated. It was found that the density and pore size distribution can be controlled by manipulating the SiC size compositions, polymer content and cold pressing pressure. The polymer infiltration and pyrolysis process was also studied. This process can effectively increase the density and close the open pores. The mechanical properties of the fabricated pellets as a function of the processing parameters were measured and compared with the reference mixed oxide fuel. A SiC pellet with a theoretical density of 86 %, a hardness of 5.55 GPa, a fracture strength of 201.0 MPa and a fracture toughness of 1.88 MPa.m1/2 was achieved. The fabricated silicon carbide pellets had better mechanical properties than the reference mixed oxide fuels and other inert matrix candidate materials.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Chunghao Shih.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Baney, Ronald H.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0042388:00001

Permanent Link: http://ufdc.ufl.edu/UFE0042388/00001

Material Information

Title: Fabrication and Characterization of Silicon Carbide Inert Matrix Fuels Through a Polymer Precursor Route
Physical Description: 1 online resource (152 p.)
Language: english
Creator: Shih, Chunghao
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: imf, pip, precursor, sic
Materials Science and Engineering -- Dissertations, Academic -- UF
Genre: Materials Science and Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Plutonium management is a crucial issue for the nuclear industry since it is an unavoidable by-product of the well established uranium fuel cycle. Inert matrix materials are proposed as a matrix material to carry the plutonium fuel so that plutonium can be transmuted in nuclear reactors. Silicon carbide is a promising inert matrix candidate material because of its low neutron absorption cross section, high thermal conductivity and high temperature stability. The drawback of using silicon carbide as a fuel matrix is the high temperature and pressure required for fabrication. Chemical reactions between SiC and plutonium at higher temperatures may also be an issue. In this study, a low temperature process of using crystalline beta silicon carbide particles and a polymer precursor was developed to synthesize SiC inert matrix fuel pellets.The low temperature process is required to prevent the reactions between SiC and the PuO2 fuel material. The effect of the polymer content and the cold pressing pressure on the packing of SiC particles was investigated. The effect of mixing the coarse and fine SiC particles on the density, the microstructure and the pore size distribution was also investigated. It was found that the density and pore size distribution can be controlled by manipulating the SiC size compositions, polymer content and cold pressing pressure. The polymer infiltration and pyrolysis process was also studied. This process can effectively increase the density and close the open pores. The mechanical properties of the fabricated pellets as a function of the processing parameters were measured and compared with the reference mixed oxide fuel. A SiC pellet with a theoretical density of 86 %, a hardness of 5.55 GPa, a fracture strength of 201.0 MPa and a fracture toughness of 1.88 MPa.m1/2 was achieved. The fabricated silicon carbide pellets had better mechanical properties than the reference mixed oxide fuels and other inert matrix candidate materials.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Chunghao Shih.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Baney, Ronald H.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0042388:00001


This item has the following downloads:


Full Text

PAGE 1

FABRICATION AND CHARACTERIZATION OF SILICON CARBIDE INERT MATRIX FUELS THROUGH A POLYMER PRECURSOR ROUTE By CHUNGHAO SHIH A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2010 1

PAGE 2

2010 Chunghao Shih 2

PAGE 3

To my parents 3

PAGE 4

ACKNOWLEDGMENTS I need to thank my advisor, Dr. Ronald H. Baney for his guidance and patience. I would also like to thank the rest of my committee members Prof. El-Shall, Dr Powers, Prof. Nino and Prof. Tulenko for their help and support. I want to thank all my lab-mates for all their support, and everyone in the Particle Engineering Research Center who helped me on my research. I would also like to thank all my friends in Gainesville, FL. I will always remember all the good times we had together in this energetic college town. Finally I would like to acknowledge my funding agency, the U.S. Department of Energy for the financial support to make this research possible. 4

PAGE 5

TABLE OF CONTENTS page ACKNOWLEDGMENTS ..................................................................................................4 LIST OF TABLES ............................................................................................................8 LIST OF FIGURES ..........................................................................................................9 LIST OF ABBREVIATIONS ...........................................................................................14 ABSTRACT ...................................................................................................................15 CHAPTER 1 INTRODUCTION....................................................................................................17 1.1 Challenges and Motivation................................................................................17 1.2 Scientific Approach...........................................................................................20 1.3 Organization of Dissertation..............................................................................21 2 BACKGROUND......................................................................................................24 2.1 Nuclear Energy and Fuel Cycle........................................................................24 2.2 Inert Matrix Fuel................................................................................................25 2.2.1 Inert Matrix Candidate Materials..............................................................25 2.2.2 Radiotoxicity Issues with IMF..................................................................28 2.3 SiC as a Candidate Inert Matrix Material..........................................................29 2.3.1 Low Neutron Absorption Cross Section...................................................29 2.3.2 High Thermal Conductivity of SiC............................................................30 2.3.3 Good Irradiation Behavior of SiC.............................................................32 2.3.4 Good Compatibility of SiC With Coolant (Water).....................................33 2.3.5 Good Compatibility of SiC With Zircaloy Cladding...................................34 2.3.6 Chemical Compatibility of SiC and PuO 2 at High Temperatures.............34 2.3.7 Density Requirements for SiC IMF Pellets..............................................35 2.4 Fabrication of SiC IMF Pellets..........................................................................37 3 PELLET FABRICATION AND CHARACTERIZATION............................................50 3.1 Materials Used..................................................................................................50 3.2 Characterization of the as Received Materials..................................................50 3.2.1 Particle Size Analysis..............................................................................50 3.2.2 Density of the Particles and the Amorphous SiC from the Decomposition of the Polymer Precursor......................................................51 3.2.3 Structure and Thermal Analysis of the Polymer Precursor......................52 3.2.4 Crystallinity Analysis................................................................................54 3.3 Fuel Pellets Synthesis Steps............................................................................55 5

PAGE 6

3.4 Fuel Pellet Characterization..............................................................................56 3.4.1 Density of the Sintered Pellets................................................................56 3.4.2 Pore Size Distribution of the Sintered Pellets..........................................56 3.4.3 Microstructure of the Sintered Pellets......................................................57 4 EFFECT OF POLYMER CONTENT AND COLD PRESSING PRESSURE............67 4.1 Change of the Polymer Precursor During Sintering..........................................67 4.2 Effect of Polymer Content and Pressing Pressure on the Density of the Sintered Pellets...................................................................................................68 4.3 Polymer Precursor as a Lubricant.....................................................................70 4.4 Effect of Polymer Content on Pore Size Distribution.........................................70 4.5 Effect of Cold Pressing Pressure on Pore Size Distribution..............................71 4.6 Summary and Conclusion.................................................................................72 5 EFFECT OF MIXING COARSE AND FINE SILICON CARBIDE PARTICLES.......80 5.1 Effect of Mixing Coarse and Fine Particles on the Sintered Density, Pore Size Distribution and Microstructure....................................................................80 5.1.1 Effect on Density.....................................................................................80 5.1.2 Effect on Pore Size Distribution...............................................................82 5.1.3 Microstructure Characterization...............................................................83 5.2 Utilization of Different Mixing Methods..............................................................84 5.2.1 Different Mixing Procedures....................................................................85 5.2.2 Effect of the Mixing Methods on SiC Particle Size...................................86 5.2.3 Effect of the Mixing Methods on Density.................................................87 5.2.4 Effect of the Mixing Methods on Pore Size Distribution...........................88 5.2.5 Effect of the Mixing Methods on Microstructure.......................................89 5.3 Summary and Conclusion.................................................................................91 6 EFFECT OF POLYMER INFILTRATION AND PYROLYSIS CYCLES.................103 6.1 Introduction.....................................................................................................103 6.2 Experimental Procedure.................................................................................105 6.2.1 PIP with 1 Atmosphere..........................................................................106 6.2.2 PIP with 10 MPa Argon Pressure..........................................................106 6.3 Results and Discussion...................................................................................107 6.3.1 Effect on Density...................................................................................107 6.3.2 Effect on Pore Size Distribution.............................................................108 6.3.3 Effect on Phase Volume Composition...................................................109 6.4 Summary and Conclusion...............................................................................110 7 MECHANICAL PROPERTIES OF THE PELLETS...............................................118 7.1 Introduction.....................................................................................................118 7.2 Experimental Procedures................................................................................122 7.2.1 Vickers Hardness..................................................................................122 7.2.2 Biaxial Fracture Strength and Fracture Toughness...............................122 6

PAGE 7

7.3 Results and Discussion...................................................................................123 7.3.1 Vickers Hardness..................................................................................123 7.3.2 Fracture Strength and Fracture Toughness...........................................126 7.4 Summary and Conclusion...............................................................................130 8 SUMMARY AND FUTURE WORK.......................................................................141 8.1 Summary........................................................................................................141 8.2 Future Work....................................................................................................143 LIST OF REFERENCES.............................................................................................145 BIOGRAPHICAL SKETCH..........................................................................................151 7

PAGE 8

LIST OF TABLES Table page 2-1 Changes in Pu mass and composition after 1500 days of irradiation [5]................38 2-2 Changes in individual Pu isotope densities after 1500 days of irradiation [5].........38 2-3 Examples of inert matrix candidates and their chemical formula [19].....................39 2-4 Some important properties of selected candidate materials for the inert matrix [1].......................................................................................................................39 2-5 Thermal neuron absorption cross section for several inert matrix candidate elements and some actinides [47]......................................................................40 3-1 Measured density of different types of powder and the amorphous SiC................57 7-1 Results of the Vickers hardness measurements and the literature data [81].......131 7-2 Results of the fracture toughness measurements and literature data [81]...........131 7-3 Vickers hardness values of SiC IMF pellets.........................................................132 7-4 Fracture strength of SiC IMF pellets.....................................................................133 7-5 Fracture toughness of SiC IMF pellets.................................................................134 8

PAGE 9

LIST OF FIGURES Figure page 2-1 Nuclear fuel cycle showing extension for once through Pu burning [1]..................41 2-2 Methodology used to progress towards IMF qualification for a once-through irradiation in light water reactor prior to geological disposal [13]........................41 2-3 K-infinity variation with burnup for the Pu-burnable poison unit cells (a) with B, Dy, and Hf as the burnable poison and (b) with Sm, Eu, and Ho. Results for the reference UO 2 and MOX LWR cells are shown for comparison in each case [5]...............................................................................................................42 2-4 Temperature profiles for solid and annular pellet at 400 W/cm liner power. Conditions: fuel composition 90-80 atomic % ZrO 2 7-14% PuO 2 3-6% Er 2 O 3 [42].....................................................................................................................42 2-5 K-infinity vs. full power day burnup graph comparison between ZrO 2 SiC, MgO-ZrO 2 and UO 2 [43]...............................................................................................43 2-6 MAs radioactivity for mixture A and for nuclides having the larger contributions [45].....................................................................................................................44 2-7 MAs radioactivity for mixture B and for nuclides having the larger contributions [45].....................................................................................................................45 2-8 Thermal conductivity of SiC samples with different fabrication routes. Also shown are the thermal conductivity of the neutron-irradiated SiC and the thermal conductivity of a UO 2 pellet [21].............................................................46 2-9 Peak cladding temperature for hot pin for loss of power and loss of primary flow transient [43].......................................................................................................46 2-10 Core average fuel temperature under loss of flow accident [43]..........................47 2-11 The variations of thermal conductivity of SiC IMF specimen as a function of thermal shock temperature and the number of cyclic thermal shock [50]...........47 2-12 Surface relief of two samples bombarded in a 72 MeV iodine beam with an ion dose of 10 15 cm -2 at 1473K. (a) SiC, (b) ZrSiO 4 The SiC sample shows no evidence of the implantations, in contrast to ZrSiO 4 [21]....................................48 2-13 Time dependence of the weight loss for the SSiC and CVD SiC ceramics after erosion testing in water at 663 K [53].................................................................48 9

PAGE 10

2-14 SEM images revealing surface micro-structure of the (a) as polished and corroded SSiC specimens after corrosion testing for (b) 1, (c) 5, and (d) 7 days in water at 663 K [53].................................................................................49 2-15 K-infinity versus burnup for SiC IMF (courtesy of Jiwei Wang)............................49 3-1 Particle size distribution for SiC particles...............................................................58 3-2 SEM images of the as-received particles. (a) 70-100 nm ceria, (b) 0.6m SiC, (c) 1 m SiC, (d) 16.9m SiC............................................................................59 3-3 The densities of the SiC from the decomposition of the polymer precursor as a function of the final sintering temperature [61]....................................................60 3-4 1 H-NMR spectrum of the commercially available polymer precursor SMP-10 and the predicted structure [62]..........................................................................60 3-5 FT-IR spectrum of the polymer precursor, SMP-10 [62].........................................61 3-6 TGA curve of SMP-10 in nitrogen..........................................................................61 3-7 DSC curve of the polymer precursor, SMP-10.......................................................62 3-8 DSC curve of the polymer precursor, SMP-10, between 150 C and 400 C.........62 3-9 XRD profiles of the amorphous SiC and the 0.6 m -SiC particles......................63 3-10 Picture of the 13 mm die......................................................................................63 3-11 Picture of the Carver press..................................................................................64 3-12 Lindberg/Blue Mini-Mite tube furnace...................................................................64 3-13 Sintering temperature profile for SiC IMF pellets..................................................65 3-14 The filling apparatus for the mercury porosimeter................................................65 3-15 The autoscan-60 porosimeter..............................................................................66 4-1 XRD profiles of the sintered SiC pellet (with a green body composition of 90 weight % 0.6 m -SiC particles and 10 weight % SMP-10), the 0.6 m -SiC particles and the amorphous SiC........................................................................74 4-2 Percent theoretical density as a function of polymer content and cold pressing pressure.............................................................................................................75 4-3 SiC particle packing density as a function of polymer content and cold pressing pressure.............................................................................................................75 10

PAGE 11

4-4 The green body volume as a function of the amounts of polymer precursor mixed with 1 gram of 1 micron SiC particles.......................................................76 4-5 Pore size distributions of pellets with 5, 10 and 15 weight % polymer precursor...77 4-6 Schematics showing how geometric arrangements effect the decomposition of the polymer precursor.........................................................................................78 4-7 Pore size distributions of pellets with 70, 200 and 600 MPa cold pressing pressure.............................................................................................................79 5-1 Percent theoretical density as a function of particles size composition and pressing pressure...............................................................................................93 5-2 Pore size distributions and cumulative pore volume of pellets with (a) 100% 0.6 m, (b) 40% 0.6m, 60%16.9 m and (c) 20% 0.6 m and 80% 16.9 m SiC particles.......................................................................................................94 5-3 SEM images (back scattered) of a pellet with 10 weight % polymer precursor and 5 weight % ceria. The SiC particles were composed of 40% fine particles and 60% coarse particles...................................................................................94 5-4 The influence of using the high energy shaker mill on the particle size distributions of the coarse SiC particles..............................................................95 5-5 Density of the sintered pellets as a function of mixing method and cold pressing pressure.............................................................................................................96 5-6 Pore size distributions and cumulative pore volume of pellets mixed with (a) mortar and pestle (b) high energy shaker mill without hexane and (c) high energy shaker mill with hexane as a grinding aid...............................................97 5-7 Optical microscopy images of polished pellets pressed at 200 MPa and mixed with (a), (b) mortar and pestle (c), (d) SPEX without the hexane and (e), (f) SPEX with hexane. The arrow marks the pressing direction of the pellets.........98 5-8 Optical microscopy images of polished pellets pressed at 600 MPa and mixed with (a), (b) mortar and pestle (c), (d) SPEX without the hexane and (e), (f) SPEX with the hexane. The arrow marks the pressing direction of the pellets...99 5-9 Back scattered SEM images of polished pellets pressed at 200 MPa and mixed with (a), (b) mortar and pestle (c), (d) SPEX mill without the hexane and (e), (f) SPEX with the hexane.................................................................................100 5-10 Back scattered SEM images of polished pellets pressed at 600 MPa and mixed with (a), (b) mortar and pestle (c), (d) SPEX without the hexane and (e), (f) SPEX with the hexane...........................................................................101 11

PAGE 12

5-11 Back scattered SEM images of polished pellets with (a): mortar and pestle mixing, 200 MPa (b): mortar and pestle mixing, 600 MPa, (c) SPEX mill without hexane, 200 MPa, (d) SPEX mill without hexane, 600 MPa, (e) SPEX mill with hexane, 200 MPa, (f) SPEX mill with hexane, 600 MPa.....................102 6-1 Changes in apparent mass density and porosity with the increasing number of impregnation and pyrolysis cycles [69].............................................................111 6-2 Schematics of the polymer infiltration system [61]...............................................112 6-3 Helium and argon gas solubility depending on the applied gas pressure [61]......112 6-4 % Theoretical density improvements with PIP cycles with different infiltration pressure...........................................................................................................113 6-5 Pore size distributions and cumulative pore volume of pellets with PIP cycles using 1 atmosphere pressure...........................................................................114 6-6 Pore size distributions and cumulative pore volume of pellets with PIP cycles using 10 MPa argon pressure...........................................................................115 6-7 Phase volume composition of SiC pellets before and after PIP cycles with one atmosphere infiltration pressure.......................................................................116 6-8 Phase volume composition of SiC pellets before and after PIP cycles with 10 MPa argon infiltration pressure.........................................................................117 7-1 Vickers hardness values for the simulated MOX pellet specimens with different densities subjected to thermal shock at different T (in C) [82].......................135 7-2 Irradiated pellets of (A) Y 2 O 3 + UO 2 and (B) MgAl 2 O 4 + UO 2 (arrow indicates the axial direction of pellet) [81]........................................................................135 7-3 The schematic of the pin on three ball set up for measuring biaxial fracture strength and fracture toughness (Courtesy of Nadia Rohbeck)........................136 7-4 Optical microscopy pictures of the Vickers intents from (a): sample H8, (b): sample H9 and (c): sample H10 in Table 7-3...................................................137 7-5 Vickers hardness as a function of % theoretical density......................................137 7-6 Vickers hardness values of porous SiC specimens as a function of % theoretical density.............................................................................................138 7-7 Biaxial fracture strength as a function of % theoretical density............................138 12

PAGE 13

7-8 Optical microscopy images of polished pellets with 1 m SiC particles, with 10 weight% polymer precursor, and were pressed at (a) 70 MPa (sample F1), (b) 200 MPa (sample F3), and (c) 600 MPa (sample F5).......................................139 7-9 Fracture toughness as a function of % theoretical density...................................140 13

PAGE 14

LIST OF ABBREVIATIONS ALI Annual Limited Intake IMF Inert Matrix Fuel K-infinity Neutron multiplication factor assuming an infinitely large reactor hence there is no neutron leakage. It is the average number of neutrons from one fission that cause another fission LWRs Light Water Reactors MAs Minor Actinides MOX Mixed Oxide Fuel PIP Polymer Infiltration and Pyrolysis PWRs Pressurized Water Reactors 14

PAGE 15

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy FABRICATION AND CHARACTERIZATION OF SILICON CARBIDE INERT MATRIX FUELS THROUGH A POLYMER PRECURSOR ROUTE By Chunghao Shih December 2010 Chair: Ronald H. Baney Major: Materials Science and Engineering Plutonium management is a crucial issue for the nuclear industry since it is an unavoidable by-product of the well established uranium fuel cycle. Inert matrix materials are proposed as a matrix material to carry the plutonium fuel so that plutonium can be transmuted in nuclear reactors. Silicon carbide is a promising inert matrix candidate material because of its low neutron absorption cross section, high thermal conductivity and high temperature stability. The drawback of using silicon carbide as a fuel matrix is the high temperature and pressure required for fabrication. Chemical reactions between SiC and plutonium at higher temperatures may also be an issue. In this study, a low temperature process of using crystalline beta silicon carbide (-SiC) particles and a polymer precursor was developed to synthesize SiC inert matrix fuel pellets.The low temperature process is required to prevent the reactions between SiC and the PuO 2 fuel material. The effect of the polymer content and the cold pressing pressure on the packing of SiC particles was investigated. The effect of mixing the coarse and fine SiC particles on the density, the microstructure and the pore size distribution was also investigated. It was found that the density and pore size distribution can be controlled by manipulating the SiC size compositions, polymer 15

PAGE 16

content and cold pressing pressure. The polymer infiltration and pyrolysis process was also studied. This process can effectively increase the density and close the open pores. The mechanical properties of the fabricated pellets as a function of the processing parameters were measured and compared with the reference mixed oxide fuel. A SiC pellet with a theoretical density of 86 %, a hardness of 5.55 GPa, a fracture strength of 201.0 MPa and a fracture toughness of 1.88 MPam 1/2 was achieved. The fabricated silicon carbide pellets had better mechanical properties than the reference mixed oxide fuels and other inert matrix candidate materials. 16

PAGE 17

CHAPTER 1 INTRODUCTION 1.1 Challenges and Motivation Plutonium (Pu) management is a critical issue in the future of nuclear energy since it is an unavoidable by-product of the established and commercialized uranium fuel cycle [1]. Approximately 70 tons of plutonium are generated per year in commercial nuclear reactors around the world. Plutonium is present at an average concentration of 1 weight % in the UO 2 matrix of spent fuels [1]. The spent fuel is stored on-site at the 439 nuclear power plants in 30 different countries [2]. In addition to the plutonium generated in commercial nuclear power plants, there are100 tons of plutonium to be disposed of derived from the dismantling of nuclear warheads in the U.S. and Russia between 1996 and 2005 [1]. The problem of having excess Pu inventory started to gain more attention in the early 90s [3, 4]. The principal causes of the excess inventory are the delay of the large-scale commercialization of fast reactors, the continued domination of the nuclear energy field by light water reactors (LWRs) operating with uranium fuel in a once-through fuel cycle and the dismantling of large numbers of nuclear weapons [5]. The presence of the abundant Pu inventory causes related proliferation concerns since nuclear weapon can be constructed with Pu [6]. The actual amount of Pu metal needed in order to build a nuclear bomb is related to the isotope composition of Pu. With weapon grade Pu (> 93% 239 Pu) only about 3 kilograms of Pu metal are needed [6]. To eliminate proliferation concerns, the excess Pu should be appropriately disposed of. Besides Pu, the spent fuel annually discharged also contains about 6 tons of Am (mostly 241 Am), about 4.8 tons of 237 Np and 0.2 tons of 244 Cm ten years after removal 17

PAGE 18

from the reactor [1]. These minor actinides are a major concern for environmental safety because of their radiotoxicity, decaying heat generation and very long half lives (> one million years). There are three strategies for the disposal of Pu and other minor actinides. The first approach is the direct disposal of the spent nuclear fuel into a repository. This strategy requires a number of repositories to be built and monitored. Meanwhile, it increases the nuclear proliferation risks. The long half-lives of the actinides also pose a challenge to longterm disposal. The materials used for the immobilization of the nuclear waste should be chemically inert and eradiation-tolerant for thousands of years [7]. The second approach is the use of nuclear reactors or particle accelerator driven systems to transmute the radioactive actinides into nonradioactive nuclei. The transmutation of actinides in nuclear reactors is an effective and economic solution to reduce the radioactive nuclear waste and produce electricity. The third approach is the disposal of the Pu in a form or place that renders the Pu inaccessible [3]. Some options might include shooting the Pu into outer space or into the sun. Also, placing it underground around a nuclear device and detonating that device has been discussed. Until today, Pu has been burned in light water reactors (LWRs) in a fuel type called mixed uranium-plutonium oxide (MOX) [8-11]. The MOX fuel is a blend of UO 2 and PuO 2 It contains 238 U, which is a fertile isotope that converts into the fissile isotope 239 Pu by the absorption of a neutron in nuclear reactors. As a result, while burning (transmuting) the initial Pu in the MOX fuel, new Pu is bred. Hence, this process is not efficient in transmuting vast amounts of Pu and does not allow a fast reduction of the stocked Pu inventory. A more efficient way is to replace the UO 2 with a neutron 18

PAGE 19

transparent matrix which contains no fertile materials. Thus, the concept of an inert matrix fuel (IMF) has been proposed [12-18]. IMF is a type of nuclear fuel that consists of a neutron-transparent matrix and a fissile fuel that can be dissolved in the matrix or incorporated as macroscopic inclusions. It replaces the uranium needed in MOX fuel [19]. The key difference between the IMF and the MOX fuel is the replacement of fertile 238 U by a neutron-transparent matrix, which eliminates plutonium breeding as a result of neutron capture [2]. Inert matrix materials are selected for optimum neutronic and operation performance based on the following criteria [20]: Low neutron absorption cross section High melting point High thermal conductivity Good irradiation stability Good chemical stability with the cladding material and coolant The properties of silicon carbide (SiC) make it a very promising candidate as an inert matrix material. Both silicon and carbon have a small thermal neutron absorption cross section (0.0035 barn for carbon, 0.171 barn for silicon). In addition, SiC possesses a very high decomposition temperature of approximately 2700C. The thermal conductivity of -SiC is 30 W/m-K at 1000 C [21], which is about 10 times higher than a MOX fuel. SiC also demonstrates good irradiation stability and good chemical stability to air and hot water [21]. The major drawback (up to this day) of utilizing SiC as an inert matrix material have been the high temperature and pressure required for sintering [22-25]. Another negative aspect have been the chemical reactivity of SiC with PuO 2 at the high processing temperatures [26, 27]. 19

PAGE 20

The use of the pre-ceramic polymer precursors to produce SiC has been known for a long time [28-37]. This process enables the synthesis of SiC at lower temperatures. The synthesis of SiC IMF at temperatures below 1200 C is desirable in order to avoid potential chemical reactions between SiC and PuO 2 In addition, the loss of minor actinides due to their high vapor pressure at higher temperatures can be avoided. The pores inside SiC IMFs associated with these low temperature processing routes can also function as reservoirs for fission gasses. This work focuses on the fabrication of SiC IMF at a temperature of 1050 C with the use of the SiC powder and a SiC polymer precursor. The fabricated fuel was characterized in terms of its density, microstructure, pore size distribution and mechanical properties. 1.2 Scientific Approach In order to fabricate SiC IMF pellets at lower temperatures, crystalline -SiC powder, CeO 2 powder (PuO 2 surrogate [38, 39]) and a liquid polymer precursor were used. SiC, CeO 2 powder and the polymer precursor were mixed to form a slurry. The slurry was uniaxially cold pressed in a die to form a green body. The polymer precursor behaved as a binder to bind the powder together in the green body. The green body was then sintered in a tube furnace to form a SiC IMF pellet. During sintering, the polymer precursor decomposed to amorphous SiC. The amorphous SiC functioned as a matrix phase to bind the SiC particles together to form a pellet. This work focused on establishing the relationships between the process parameters and the properties of the sintered pellets, which include the density, the pore size distribution, the microstructure 20

PAGE 21

and the mechanical properties. Understanding the relationship is important in order to optimize the design and fabrication of the IMF pellets. Initially, various polymer contents and various cold pressing pressures were examined to achieve a systematic understanding of how these processing parameters affect the sintered density and the pore size distribution of the pellets. The ability to further increase the density by using mixtures of the coarse and fine SiC particles was then examined and the optimum coarse and fine SiC particles volume ratio was determined. This optimum volume ratio was used to determine the effect of different mixing methods on the density, the pore size distribution, and the microstructure of the pellets. After the interrelationships between the process parameters and the sintered density has been developed, the pellet with the highest density was chosen to undergo polymer infiltration and pyrolysis cycles to further increase the density and close the open pores. The mechanical properties of the pellets fabricated with different process parameters were measured. The interrelationship between the process parameters and the mechanical properties was established. 1.3 Organization of Dissertation In Chapter 1, the problem and motivation of this research, which is about the excess plutonium inventory and the concept of utilizing the inert matrix fuel to transmute plutonium is addressed. A brief background of topics that are related to this study is given in Chapter 2. The concept of nuclear energy and nuclear fuel cycle is firstly given. Following this is the discussion about the inert matrix fuel, which includes the IMF concept, the literature 21

PAGE 22

search on several inert matrix candidate materials and some radiotoxicity issues with IMF. The concept of using SiC as an inert matrix material is reviewed. Some properties that make SiC a promising inert matrix candidate are discussed, which include low neutron absorption cross section, high thermal conductivity, good irradiation behavior, good compatibility with the coolant and good compatibility with the cladding. The chemical compatibility of SiC and PuO 2 and the density requirement for SiC IMF pellets are discussed briefly. This chapter ends with the literature review of the fabrication of SiC IMF pellets. In Chapter 3, the materials used in this study and the characterization of the as received materials are described. This chapter continues with the description of the fuel pellets fabrication steps and the characterization techniques used to characterize the fabricated pellets. In Chapter 4, the effect of polymer content and cold pressing pressure on the density and the pore size distribution of the sintered pellets was discussed. Chapter 5 begins with the discussion of the effect of mixing coarse and fine SiC particles on the sintered density and the pore size distribution. The chapter continues with the discussion of the utilization of different mixing methods. Different mixing methods are described first and the effect of the mixing methods on the density, the pore size distribution and the microstructure are discussed. Chapter 6 begins with the introduction of the polymer infiltration and pyrolysis process. This process is then used to further increase the density of the sintered pellets and to close the open pores of the pellets. The effectiveness of the polymer infiltration and pyrolysis process is discussed. 22

PAGE 23

In chapter 7, an introduction to the mechanical properties of nuclear fuels is firstly given. The experimental procedures to measure the Vickers hardness, biaxial fracture strength and fracture toughness are described. This chapter ends with the discussion of the resulted mechanical properties of the pellets, which are correlated with the fabrication parameters. The dissertation is summarized in Chapter 8 and the proposed future work is presented. 23

PAGE 24

CHAPTER 2 BACKGROUND A brief review is given in this chapter that summarizes some of the background and fundamentals required for understanding the present study. 2.1 Nuclear Energy and Fuel Cycle The United States currently has 104 operating nuclear reactors. In 2007 these reactors safely produced over 806 billion kWh of electricity. Unlike many other major energy sources, nuclear energy does not emit any greenhouse gases into the atmosphere. Currently, about one-third of the electricity produced in the US is from carbon-free sources, and nuclear power makes up about 70% of that energy [40]. Figure 2-1 shows the nuclear fuel cycle with different options to manage the spent fuel. The arrows labeled 1, 2 and 3 denote the basic nuclear fuel cycle from the uranium mine to the disposal of the non-reprocessed spent fuel. This is called the once-through fuel cycle. The arrows labeled 6-9 represent the extended fuel cycles to accommodate mixed oxide fuels (MOX fuels) or inert matrix fuels (IMFs). MOX fuels are currently utilized in some of the commercial nuclear power plants to burn Pu. Its drawback is that MOX contains fertile isotope 238 U that transmute into 239 Pu after absorbing a neutron. So while burning the original Pu, new Pu and other radioactive minor actinides are produced in the reactor. As a result, an efficient reduction of the Pu inventory is not possible. A more rapid reduction of the Pu inventory can be achieved by the application of IMF. It replaces the UO 2 phase in the MOX fuel with a neutron transparent matrix. Hence there will be no breeding of additional Pu or any other minor actinides from the matrix. Accordingly, Pu can be burned more efficiently. 24

PAGE 25

2.2 Inert Matrix Fuel 2.2.1 Inert Matrix Candidate Materials The selection of inert matrix material was done following a systematic approach. The proposed methodology used to qualify the IMF candidates for a once-through irradiation in a light water reactor prior to geological disposal is shown in Figure 2-2 [13]. The neutronic properties constitute the first requirement for the material selection. In order to transmute Pu efficiently while not creating more radioactive nuclei, the matrix material should be transparent to neutrons. Other than the neutronic properties, the physicochemical properties are also important for evaluating the reactor fuel behavior. The desired properties of the materials are a high melting point, good thermal conductivity, good compatibility with the cladding, low solubility in the coolant and good mechanical properties. The properties of IMF candidates are often compared to the standard UO 2 or MOX fuel pellets as a reference. One of the first proposed candidate inert matrix fuels is a two phase mixture of a fluorite-type phase (ZrO 2 ) and an alumina phase (Al 2 O 3 ) [41]. The fluorite phase has a high solid solubility of the actinide elements and fission products, a good irradiation and chemical stability [41]. It was determined that such IMFs in LWRs can transmute about 99% of 239 Pu and 85% of the initial loaded Pu [41]. Paratte and Chawla investigated the use of Al 2 O 3 as an inert matrix to burn Pu (in the form of PuO 2 ) [5]. The calculated K-infinity (effective neutron multiplication factor) as a function of burnup for Al 2 O 3 -Pu IMF fuels with different burnable poisons is shown in Figure 2-3 [5]. The UO 2 and MOX data is also given for comparison. In Figure 2-3 [5], the slopes of the K-infinity -vs.-burnup curves are, on average, less steep than those for 25

PAGE 26

the UO 2 cell. A mixture of burnable poisons could be added to provide a more satisfactory reactivity performance than individual components. Table 2-1 [5] shows the reduction of total Pu together with the changes in Pu isotope compositions after 1500 days of irradiation for Al 2 O 3 -Pu IMFs with different burnable poisons and for MOX. The total mass of Pu is reduced by nearly 70% irrespective to the applied burnable poison. This corresponds to a Pu burning rate of 1.14 kg/GWd. On the other hand, the reduction rate for Pu in MOX is only ~30% (0.38kg/GWd). This reflects the generation of the new Pu from the fertile 238 U present in the MOX. The changes in individual Pu isotope densities after 1500 days of irradiation are shown in Table 2-2 [5]. Accordingly more than 95% of the 239 Pu isotope is destroyed in the IMF. On the contrary, MOX fuel can only destroy nearly 55% of the 239 Pu isotope. The destruction of 239 Pu is of exceptional importance, since it can be easily transformed into nuclear weapons. The controllability of an LWR core filled with IMFs has also been calculated [5]. It was concluded that a PWR core design would be quite feasible considering the reactivity control if a suitable burnable poison is selected for the Al 2 O 3 based IMF. The drawback of using Al 2 O 3 as an inert matrix material is the excessive swelling induced by the intense radiation in the reactor [20]. Degueldre et al. investigated the feasibility of using ZrO 2 as a inert matrix material [42]. Different burnable poisons were compared, including: Er, Eu, Ho and Gd. The composition they suggested were: 90-80 atomic % of ZrO 2 as the matrix, 7-14 atomic % of PuO 2 as the fission phase and 3-6 atomic % of Er 2 O 3 as the burnable poison. ZrO 2 with its fluorite-type structure is very stable regarding nuclear processes. It may allow 26

PAGE 27

substitution of foreign atoms (e.g. fission products) without too many structure changes [42]. The drawback of using ZrO 2 as an inert matrix material is the low thermal conductivity, which is about 2 W/Km at 1000C [1]. The lowered thermal conductivity would cause a high centerline temperature and thus a high thermal gradient across the fuel pellets. In order to solve this problem, the employment of annular pellets has been suggested [42]. The calculated radial power density and temperature profiles for both a solid and a annular pellet are shown in Figure 2-4 [42]. The thermal conductivity is 2 W/Km and a linear power of 400 W/cm was assumed. The use of the annular fuel pellet can reduce the centerline temperature about 600 C. The calculated Pu burning rate for these ZrO 2 based pellets is very similar to that of the Al 2 O 3 based IMF pellet. It was calculated that after 1300 days of irradiation, 64% of the reprocessed Pu (94% 239 Pu) is transmuted. The fuel neutronic, thermal hydraulic and transient behavior of IMFs in light water reactor have been studied by Carmack et al. [43]. The candidate inert matrix materials they used in the simulation work were: SiC, yttria-stabilized zirconia (YSZ), and magnesia-zirconia (MgO-ZrO 2 ). The results were compared with a UO 2 fuel. A partial loading of 25% IMF in a UO 2 reactor was used. The burnup in full power day vs. the K-infinity for these fuel materials are shown in Figure 2-5 [43]. It shows that the impact of different matrix materials on the cycle length is slight. Other than the materials mentioned above, numerous candidate materials have been proposed. Some candidate materials are listed in Table 2-3 [19]. The out-of-pile properties of a lot of these candidate materials can be found in a paper by Kleykamp [44]. 27

PAGE 28

Some of the important properties for the selected materials are summarized in Table 2-4 [1]. 2.2.2 Radiotoxicity Issues with IMF Even through the goal of utilizing IMF is to transmute the Pu and reduce the proliferation concerns; the issue of radiotoxicity related to this new type of fuel should be assessed. An increase in the overall radiotoxicity could be a serious drawback and should be avoided. Whether the option of burning Pu in an IMF would produce any increase in the Minor Actinides (MAs) radiotoxicity has been investigated by Lombardi et al. [45] using the SCALE integrated code system. The radiotoxicity of a mixture of different radionuclides was related to the Annual Limited Intake (ALI) of each radionuclide. The overall radiotoxicity was defined as its Number of ALI: Number of ALI = iiiALItA)( (2.1) where, for each radionuclide, A i (t) is the radioactivity at time t after irradiation and ALI i is the Annual Limit of Intake through ingestion for the public [45]. In their experiments, one metric ton of 3.2% enriched uranium was irradiated in the pressurized water reactor to a burnup of 34,400 MWd/tHM. The MAs contained in the spent fuel were labeled as mixture A. For the Pu burning option, the Pu included in mixture A was assumed to be recovered with 2% loss, fabricated as IMF with MgAl 2 O 4 as the inert matrix and irradiated in the same reactor with a 21% IMF loading. The MAs contained in the spent IMF together with lose left in the one metric ton of spent fuel after the Pu recovery process were indicated as mixture B. The decay of MAs radiotoxicity as a function of cooling time for mixture A, which represents the radiotoxicity in the case of simple 28

PAGE 29

storage of the spent UO 2 fuel, is shown in Figure 2-6 [45]. The same plot for mixture B, which represents the total radiotoxicity from MAs for the Pu re-irradiation option, is shown in Figure 2-7 [45]. The radiotoxicity of mixture B was larger than that of mixture A for the first 25 years, because of the large initial contributions of curium and americium. After 25 year, the radiotoxicity of mixture B became lower than mixture A. This was due to the lowered Pu concentration. In summary, the MAs radiotoxicity between the two is fairly small. At least no significant increase in radiotoxicity was observed with the Pu burning option. 2.3 SiC as a Candidate Inert Matrix Material Silicon carbide is a promising inert matrix material because of its high decomposition temperature (~2700C) and very high thermal conductivity (30 W/m-K at 1000 C [21]). It is also known for the resistance against many corrosive agents such as oxygen at high temperatures. Some research has been done to examine SiC as an inert matrix candidate. This section discusses some advantages and drawbacks of SiC as an inert matrix candidate. 2.3.1 Low Neutron Absorption Cross Section One requirement for inert matrix materials is a low thermal neutron absorption cross section of less than 2.7 barns [46]. The small neutron absorption cross section can prevent the formation of new radioactive isotopes resulting form neutron absorption. Additionally, a matrix material with a small neutron cross section will not affect the effective neutron multiplication factor (K eff ) of the fuel significantly. K eff is important since after its value drops below one, the nuclear chain reaction can not be sustained and the fuel pellets have to be replaced. Table 2-5 [47] shows the thermal neutron absorption cross section for some candidate elements that could can potentially be used as inert 29

PAGE 30

matrix materials. It can be seen that both silicon and carbon have a low thermal neutron absorption cross section. 2.3.2 High Thermal Conductivity of SiC SiC is known for its high thermal conductivity compared to other ceramic materials. For synthetic -SiC single crystals, a thermal conductivity of 360 Wm -1 K -1 has been reported at 300 K [48]. Measurement of the thermal conductivity for synthetic polycrystalline SiC IMF pellets with sintering additives and with CeO 2 as a surrogate for PuO 2 has been reported. The results are shown in Figure 2-8 [21]. Price [49] found that the thermal conductivity of SiC irradiated in a fast neutron flux decreased significantly. In his study, the thermal conductivity of pyrolytic -SiC irradiated at 773 K to 1373 K with neutron doses of 2.7 to 7.7 x 10 25 m -2 (E > 0.18 MeV) was measured. Although the thermal conductivity of neutron irradiated SiC was lowered, it is still much higher compared to UO 2 (as shown in Figure 2-8 [21]). The approximate center line temperature of the SiC IMF pellets can be calculated using the values of thermal conductivity with equation 2.2 (2.2) csTTdTkP4 where P is the liner power of the fuel element, T s and T c are the surface and center line temperature of the fuel, k is the thermal conductivity of the fuel and T is the temperature. Assuming a surface temperature of 573 K (typical for a PWR) and a liner element power of 55 kW/m (approximate peak power in a Canada Deuterium Uranium reactor), the center line temperature is only 673 K based on the thermal conductivity values for the 30

PAGE 31

SiC (SAY-6AL) sample shown in Figure 2-8 [21]. Under the same condition, a UO 2 pellet reaches a center line temperature of 1773K. Even considering the lowered thermal conductivity of the neutron-irradiated SiC sample shown in Figure 2-8 [21], the calculated center line temperature wont exceed1100 K. This is still much lower then the center line temperature of a UO 2 pellet (over 1700 K). A lower center line temperature leads to a smaller thermal gradient across the fuel pellet, so the thermal stress can be decreased. A lower center line temperature also means the fuel is operated at temperatures away from its melting point. This increases the operation safety of the nuclear reactors significantly. A lower temperature means a lowered stored heat in the fuel so that in the loss of coolant accident, the cladding temperature is much lower. This can prevent the potential reactions between the cladding and the fuel which can cause cladding failure. The peak cladding temperature for loss of power and loss of primary flow transients has been simulated for different inert matrix materials by Carmack et al. [43]. The results are shown in Figure 2-9 [43]. The peak cladding temperature is much lower for the loss of primary flow accident because of the higher thermal conductivity of SiC compared to other candidate materials. The core average fuel temperature under the same accident is also much lower than other candidate materials as shown in Figure 2-10 [43] Thermal conductivities of SiC IMF samples that underwent several thermal shock cycles have been reported by Lee et al. [50]. The composition of the SiC pellets in their study is: 80.8 weight % SiC, 6.9 weight % Al 2 O 3 5.1 weight % Y 2 O 3 and 5.0 weight % CeO 2 Al 2 O 3 and Y 2 O 3 behaved as sintering aids and CeO 2 acted as a surrogate for 31

PAGE 32

PuO 2 The samples were gas pressure sintered at 2173 K in N 2 atmosphere for 1 hour and a theoretical density of 92% was achieved. Cyclic thermal shock experiments (1, 10 and 30 times) were simultaneously carried out by repeatedly heating the specimens up to 1675 K and cooling down to about 345 K with Ar within 10 min. As shown in Figure 2-11 [50], the thermal conductivity of SiC specimen decreased slightly with the number of thermal shocks. The cyclic thermal shock did not induce any crack in SiC specimen at 1673 K even by 30 cyclic thermal shock cycles. In the same study, ZrO 2 and MgAl 2 O 4 two other IMF candidate materials, have also been investigated. It was found that both ZrO 2 and MgAl 2 O 4 samples showed crack formation due to thermo-mechanical stress after 10 cyclic thermal shock cycles at 1673 K. 2.3.3 Good Irradiation Behavior of SiC The performance of candidate inert matrix materials when exposed to high energy fission fragments is a substantial parameter since swelling of the fuel can cause cladding failure. Verrall et al. [21] have studied this using an accelerator to bombard candidate materials with a beam of 72 MeV iodine ions, which substituted for fission fragments [51]. The bombarded areas were examined for damages, with a focus on materials swelling. The materials tested were: ZrSiO 4 MgAl 2 O 4 CeO 2 CePO 4 ZrO 2 doped with CaO, CeO 2 Er 2 O 3 or Y 2 O 3 and SiC. Temperatures ranging from ambient to 1200 C and ion doses from 10 14 to 10 17 cm -2 were applied. An ion dose of 10 17 cm -2 corresponds to a Canada Deuterium Uranium reactor fuel burnup of 12 mega-watt-day/kgU. A 3 mm diameter beam was used to bombard the samples and the surface relief of the samples was measured with a laser profilometry. Significant height change of the bombarded area indicated that in-reactor swelling would occur. It was discovered that the surface relief was reasonably independent of temperature. Of all materials only 32

PAGE 33

SiC and ZrO 2 didnt show any swelling. Tests of UO 2 under the same conditions did not show any swelling. Some of the results are presented in Figure 2-12 [21]. Figure 2-12 (a) [21] shows clearly that the SiC sample didnt possess any surface relief. On the contrary, the ZrSiO 4 sample did show surface relief as seen in Figure 2-12 (b) [21]. 2.3.4 Good Compatibility of SiC With Coolant (Water) Hirayama et al. [52] measured the corrosion of sintered SiC samples (mainly phase) using boron and carbon as sintering additives. Water solutions of pH 4.6 and 10 at 563K were examined. The effect of deoxygenated and oxygen-saturated solution was also determined. It was found that the oxygenated solutions corroded the SiC faster than the deoxygenated solution by factors of 10, 15 and 100 for pH values of 4, 6 and 10 respectively. The fastest corrosion rate was 3 m over 72 hours with a solution of oxygen-saturated water at pH 10. The rate decreased to 0.03 m over 72 hours with deoxygenated water. The corrosion behaviors of sintered SiC in distilled water at 633 K have been studied by Kim et al. [53]. Both a chemically vapor deposited SiC (CVD SiC) and a solid state sintered SiC (SSiC) with additives of boron and carbon as sintering aids were used and compared in their study. The weight loss as a function of time is shown in Figure 2-13 [53]. In both cases, the weight loss increases with higher exposure time. Moreover, a parabolic relationship was observed between the weight loss and exposure time although the weight loss increased more rapidly at the end of the experiments. The large deviation from the parabolic relation at the longest exposure time in Figure 2-13 [53] can be explained by the micro-structure observations shown in Figure 2-14 [53]. Before the experiment, the polished solid state sintered SiC (SSiC) specimen 33

PAGE 34

showed a featureless surface micro-structure except for some residual pores. However, individual grains can be clearly distinguished as corrosion proceeds (Figure 2-14 c). This suggests a preferential inter-granular attack. The sintered aids used in sintering SiC have been known to segregate at grain boundaries [54]. This segregation of impurities can weaken the grain boundaries, and thus results in the preferential inter-granular attack. As erosion proceeds, the disintegration of grains into water occurs because of the weakening of grain boundaries (Figure 2-14 d). This could explain the abrupt increase in weight loss at longer exposure times. In any case, corrosion of SiC in hot water is not a cause of concern in case of cladding breach. 2.3.5 Good Compatibility of SiC With Zircaloy Cladding Experiments investigating the interactions between SiC and Zircaloy-4 cladding material have been conducted. They showed good compatibility of SiC with Zircaloy-4 [21]. Polished SiC and Zircaloy-4 disc specimens were pressed under a light pressure at an argon atmosphere in a molybdenum cell at 1273 K for an hour, 1773 K for an hour and 1973 K for 15 minutes. Both optical and scanning electron microscopies were used to examine the extent of the interactions. No interactions were found at 1273 K, but a diffusion-based reaction to form ZrC and free Si [21] occurred at 1773 K. These results showed that interactions between SiC and Zircaloy-4 occur at temperatures below the melting point of the un-oxidized Zircaloy-4 cladding (2033 K). On the other hand, UO 2 interacts with Zircaloy-4 at 1473 K [55]. It is concluded that the SiC and Zircaloy-4 interactions are better the UO 2 and Zircaloy-4 interactions [21]. 2.3.6 Chemical Compatibility of SiC and PuO 2 at High Temperatures There are relatively few studies on the chemical reactions of plutonium. It is not 34

PAGE 35

clear (to the authors best knowledge) at what temperatures PuO 2 reacts with SiC. However, SiC is known to react with UO 2 in open systems at temperatures as low as 1643 K [56], and up to 2073 K in closed isothermal systems [57]. CeO 2 is generally used as a surrogate for PuO 2 because of the similarities in their thermodynamic properties [38, 39]. Chemical reactions between -SiC and CeO 2 have been investigated by Suzuki et al. [58] in air and vacuum. Their results showed that chemical reactions of SiC and CeO 2 start at a temperature of 1373 K in vacuum. On the contrary, no reactions were observed between SiC and CeO 2 in air up to 1773 K. This is possibly due to the forming of a passive SiO 2 oxide layer on the surface of SiC, which hindered further reactions between SiC and CeO 2 Considering that chemical reactions occur between SiC and CeO 2 (PuO 2 surrogate) at 1373 K, it is assumed that SiC would react with PuO 2 at temperatures as low as 1373 K. Chemical reactions between SiC and PuO 2 are undesirable because they would make controlling the micro-structure of the fuel pellets more difficult. These chemical reactions might also cause changes in densities, leading to defects in the pellets. If the chemical reactions are inevitable, the stability of the newly formed phase should be investigated. 2.3.7 Density Requirements for SiC IMF Pellets High density is generally a goal for fabricating SiC IMF pellets [21, 50, 51]. The reason is not clear and can not be revolved in the literature search. Some potential issues associated with low density pellets are: in-reactor densification of the pellets, lowered thermal conductivity, lowered mechanical properties, fission gas release problems and neutronic issues. In-reactor densification of the pellets increases the gap 35

PAGE 36

between the cladding and the pellet, which dramatically hinders hear transfer from the fuel pellets to the coolant and thus increase the center line temperature of the pellets and should be avoided. Since SiC has a high thermal conductivity [21], the fuel operation temperature is anticipated to be much lower than that of a UO 2 fuel. Thus, in-reactor densification is not expected to occur in SiC fuel pellets. Pellet density can affect the thermal conductivity and the mechanical properties of the SiC IMF pellets and should be addressed. As for fission gas release problems, the closed pores in the pellet can potentially be the reservoirs for fission gases and could be beneficial. The effect of open porosity on fission gas release phenomenon is unclear. However, in a typical UO 2 pellet, after the average burn-up exceeded about 40 Giga-Watt-day/ metric-ton-uranium, a particular structure called the rim-structure, characterized by significant open porosity, is formed near the peripheral region [59]. Since open pores eventually form in an UO 2 pellet, the initial amount of open pores in the SiC IMF pellets might not be a critical issue. The K-infinity versus burnup curves for SiC IMF pellets with different densities has been calculated using CASMO-3, a multi-group two dimensional transport theory code. SiC pellets with theoretical densities of 100%, 80% and 40% all have very similar K-infinity versus burnup curve as shown in Figure 2-15. This result suggests that the density of the SiC pellet does not affect the reactivity performance of the fuel pellet. Another potential benefit associated with the low pellet density is the decrease of the swelling effect. Pellet swelling caused by nuclear fission reaction puts pressure on the cladding and can possibly cause cladding failure. The pores associated with the low density can potentially reduce the swelling and reduce the pellet cladding interactions. 36

PAGE 37

Nevertheless, the actual behavior of the low density SiC IMF pellet can only be determined in research test reactors. In summary, a high density SiC IMF pellet is not a requirement as long as the pellets have the desired thermal, mechanical, nuclear and chemical properties. 2.4 Fabrication of SiC IMF Pellets Conventional processing of SiC requires high temperatures of over 2073 K. Typically it also involves the use of pressure sintering [60]. The instrument for high temperature and pressure sintering is bulky and not suitable for making IMF pellets since it is not used in the conventional UO 2 or MOX fuel pellets fabrication facilities. SiC pellets can also be fabricated at lower temperatures with the liquid phase sintering route [22, 23, 25]. In Verralls work [21], SiC pellets were synthesized with 7 weight % Y 2 O 3 and 7 weight % Al 2 O 3 as sintering aids. 95% theoretical density was achieved with a sintering temperature of 1700 C. These sintering aids formed a liquid phase during sintering. XRD studies showed that Y 2 O 3 and Al 2 O 3 reacted with CeO 2 (PuO 2 surrogate) during sintering. Two clearly identifiable compounds were 3 Y 2 O 3 Al 2 O 3 garnet (YAG) and cerium aluminate (CeAlO 3 ). Generally, liquid phase sintering is undesirable because the fissile atoms can move around in the liquid phase during sintering. This makes controlling the microstructure and thus controlling the neutronics difficult. Chemical reactions between the fissile phase and the sintering aids are also undesirable for the same reason. A low temperature process for the synthesis of SiC IMF could solve these problems. 37

PAGE 38

Table 2-1. Changes in Pu mass and composition after 1500 days of irradiation [5] Final Pu isotope composition (weight %) Cell type, burnable poison Final/initial Pu (total) 238 Pu 239 Pu 240 Pu 241 Pu 242 Pu MOX 0.697 2.6 39.8 29.7 17.9 10.0 Pu-B 0.315 5.6 7.5 33.0 19.6 34.3 Pu-Dy 0.312 5.8 8.0 31.6 21.5 33.1 Pu-Er 0.314 5.6 7.9 32.5 19.9 34.1 Pu-Hf 0.311 5.4 6.9 33.6 19.6 34.5 Pu-Sm 0.315 5.5 7.4 33.0 19.6 34.5 Pu-Eu 0.315 5.6 7.6 32.8 19.9 34.1 Pu-Ho 0.315 5.6 7.9 32.3 20.5 33.7 Table 2-2. Changes in individual Pu isotope densities after 1500 days of irradiation [5] Final/initial densities Cell type, burnable poison 238 Pu 239 Pu 240P u 241 Pu 242 Pu MOX 1.01 0.469 0.90 1.02 1.75 Pu-B 0.65 0.044 0.46 0.53 1.30 Pu-Dy 0.67 0.046 0.43 0.57 1.24 Pu-Er 0.65 0.046 0.45 0.54 1.29 Pu-Hf 0.62 0.040 0.46 0.52 1.29 Pu-Sm 0.65 0.043 0.46 0.53 1.31 Pu-Eu 0.65 0.044 0.45 0.54 1.30 Pu-Ho 0.65 0.046 0.45 0.55 1.28 38

PAGE 39

Table 2-3. Examples of inert matrix candidates and their chemical formula [19] Inert matrix components Inert matrix formula Elements C, Mg, Al, Si, Cr,V, Zr, Nb, Mo, W Inter-metalics AlSi, AlZr, ZrSi, Alloys Stainless steel, zirconium alloys Carbides B4C, SiC, TiC, ZrC Nitrides AlN, TiN, ZrN, CeN Binary oxides MgO, CaO, Y2O3, ZrO2, CeO2 Ternary oxides Mg(1-x)Al(2+x)O(4-x), Y3Al5O12 (YAG), ZrSiO4 Oxide solid solutions CaxZr1-xO2-x, YyZr1-yO2-y/2 Table 2-4. Some important properties of selected candidate materials for the inert matrix [1] Al 2 O 3 MgAl 2 O 4 CeO 2 ZrO 2 SiC Melting point (C) 2054 2015 ~2400 2710 2830 Vapor pressure (bar) ~10 -6 (total, at 1950 C) 6x 10 -5 (total, at 2000 C) ~10 -8 (total, at 2000 C) 1.4 x 10 -4 (total, at 2000 C) Thermal conductivity at 1000C (W/Km) 8.2 7.7 1.2 2.2 47(), 40() Thermal conductivity at 1500C (W/Km) 5.8 8 0.9 1.5 31(), 30() Compatibility with water Not dissolvable Not dissolvable Compatible 39

PAGE 40

Table 2-5. Thermal neuron absorption cross section for several inert matrix candidate elements and some actinides [47] Element Thermal neutron absorption cross section (barn) C 0.0035 N 1.9 O 0.00019 Mg 0.063 Al 0.231 Si 0.171 Ca 0.43 Zr 0.185 Y 1.28 Ce 0.63 235 U 680.9 238 U 2.68 239 Pu 1017.3 240 Pu 289.6 40

PAGE 41

Figure 2-1. Nuclear fuel cycle showing extension for once through Pu burning [1] Figure 2-2. Methodology used to progress towards IMF qualification for a once-through irradiation in light water reactor prior to geological disposal [13]. 41

PAGE 42

Figure 2-3. K-infinity variation with burnup for the Pu-burnable poison unit cells (a) with B, Dy, and Hf as the burnable poison and (b) with Sm, Eu, and Ho. Results for the reference UO 2 and MOX LWR cells are shown for comparison in each case [5] Figure 2-4. Temperature profiles for solid and annular pellet at 400 W/cm liner power. Conditions: fuel composition 90-80 atomic % ZrO 2 7-14% PuO 2 3-6% Er 2 O 3 [42] 42

PAGE 43

Figure 2-5. K-infinity vs. full power day burnup graph comparison between ZrO 2 SiC, MgO-ZrO 2 and UO 2 [43] 43

PAGE 44

Figure 2-6. MAs radioactivity for mixture A and for nuclides having the larger contributions [45] 44

PAGE 45

Figure 2-7. MAs radioactivity for mixture B and for nuclides having the larger contributions [45] 45

PAGE 46

Figure 2-8. Thermal conductivity of SiC samples with different fabrication routes. Also shown are the thermal conductivity of the neutron-irradiated SiC and the thermal conductivity of a UO 2 pellet [21]. Figure 2-9. Peak cladding temperature for hot pin for loss of power and loss of primary flow transient [43] 46

PAGE 47

Figure 2-10. Core average fuel temperature under loss of flow accident [43] Figure 2-11. The variations of thermal conductivity of SiC IMF specimen as a function of thermal shock temperature and the number of cyclic thermal shock [50] 47

PAGE 48

Figure 2-12. Surface relief of two samples bombarded in a 72 MeV iodine beam with an ion dose of 10 15 cm -2 at 1473K. (a) SiC, (b) ZrSiO 4 The SiC sample shows no evidence of the implantations, in contrast to ZrSiO 4 [21] Figure 2-13. Time dependence of the weight loss for the SSiC and CVD SiC ceramics after erosion testing in water at 663 K [53] 48

PAGE 49

Figure 2-14. SEM images revealing surface micro-structure of the (a) as polished and corroded SSiC specimens after corrosion testing for (b) 1, (c) 5, and (d) 7 days in water at 663 K [53]. 00.20.40.60.811.21.41.61.8010203040506Burn-up (MWD/KgHM)K-infinity 0 100% dense 80% dense 40% dense Figure 2-15. K-infinity versus burnup for SiC IMF (courtesy of Jiwei Wang) 49

PAGE 50

CHAPTER 3 PELLET FABRICATION AND CHARACTERIZATION This chapter describes how the SiC IMF pellets were fabricated through the polymer precursor route and the characterization techniques used to characterize them. 3.1 Materials Used Three types of -SiC powder were used in this study: 1) a coarse SiC powder with a nominal size of 16.9 microns (Superior Graphite, Chicago, IL), 2) a submicron size SiC powder with a nominal size of 0.6 micron (Superior Graphite, Chicago, IL), and 3) a SiC powder with nominal size of 1 micron (Alfa Aesar, Ward Hill, MA). Since Pu materials are highly toxic and radioactive, cerium oxide (CeO 2 ) has been used as a surrogate for PuO 2 [38]. CeO 2 (Alfa Aesar, Ward Hill, MA) with nominal size 70-100 nm was used in this study. A liquid SiC polymer precursor, SMP-10, manufactured by Starfire Systems Inc (Schenectady, NY), was used. SMP-10 is an allyl-hydrido-poly-carbo-silane (AHCPS), which yields a 1:1 Si:C ratio and with H as the only other component. 3.2 Characterization of the as Received Materials 3.2.1 Particle Size Analysis Different sizes of -SiC powder were added to deionized water and placed in an ultrasonic bath for 3 minutes to break the soft agglomerates. The dispersed powder was then characterized with a laser light scattering technique using a Beckman Coulter LS 13320 (Fullerton, CA), which uses a 750 nm diode laser and the Polarization Intensity Differential Scattering (PIDS) assembly. This instrument measures particle sizes from 40 nm to 2000 m. 50

PAGE 51

The particle size distributions for different sizes of SiC particles are shown in Figure 3-1. Both the 0.6 and 1m nominal size powder had a mode at 1.32 m. The 0.6 m nominal size powder had some hard agglomerates in the size range of 3 to 10 m. Furthermore, the 1 m nominal size powder had some small particles in the size range of 50 to 300 nm. The 16.9 m nominal size powder had a mode at 26.1 m and had a much broader size distribution. There were even some big agglomerates with a size of over 100 m. The particles size and shape of all the as-received particles was also characterized by a JEOL 6335 FEG-SEM scanning electron microscope (Peabody, MA) operated at 15 kV with a working distance of ~15 mm. The results are shown in Figure 3-2. All the ceria and SiC particles appear to be irregular shape. The particles size fits well with the size analysis by the laser light scattering technique. 3.2.2 Density of the Particles and the Amorphous SiC from the Decomposition of the Polymer Precursor The densities of all the SiC and ceria powders were measured by a Quantachrome Ultrapyc 1000 Gas Pycnometer (Boynton Beach, FL) with helium as an infiltration gas. The density of the amorphous SiC generated by the decomposition of the polymer precursor was also determined by the same method. In this experiment, 10 gram of the polymer precursor was put in an alumina crucible and sintered in the tube furnace under ultra high purity argon as the protection gas to a temperature of 1050 C. The same sintering temperature profile for sintering SiC IMF pellets was followed and is shown in section 3.3. The amorphous SiC generated by the decomposition of the polymer precursor was ground with a mortar and pestle. The density of the amorphous SiC was 51

PAGE 52

then measured by the Quantachrome Ultrapyc 1000 Gas Pycnometer. The results of all the densities of the powder and the amorphous SiC are summarized in Table 3-1. It was found that all of the SiC and CeO 2 powder were nearly 100% dense. The amorphous SiC was less dense than the crystalline -SiC particles. The density of the SiC generated by the decomposition of the same polymer precursor has also been measured by Lee et al. [61] at different sintering temperatures. The results reported by Lee et al. [61] were found to be similar to the authors experiments and are shown in Figure 3-3 [61]. 3.2.3 Structure and Thermal Analysis of the Polymer Precursor A detailed structure and thermal analysis of the polymer precursor, SMP-10, was not a main thrust of this study. However, information about the structure of SMP-10 in the literature is discussed in this section. Some thermal analysis of SMP-10 has been studied and is reported in this section too. SMP-10 is a partially allyl-substituted hydrido-poly-carbosilane. The structure of SMP-10 has been analyzed by both Kockrick et al. [62] and Sreeja et al. [63]. The 1 H nuclear magnetic resonance (NMR) spectrum together with the predicted structure are shown in Figure 3-4 [62]. The 13 C and 29 Si NMR can be found in the same literature [62]. The FT-IR spectrum of SMP-10 is shown in Figure 3-5 [62]. The thermal gravimetric analysis (TGA) of SMP-10 was tested with a Mettler Toledo TGA/SDTA851 e (Columbus, OH) in a flowing nitrogen atmosphere with a sample weight of ~30 mg. The temperature profile for the TGA test was: 2 C per minutes from room temperature to 250 C, 1 C per minutes from 250 C to 650 C and 3 C per minutes from room 650 C to 1100 C. The result TGA curve is shown Figure 3-6. The major weight loss occurs between 100 52

PAGE 53

C and 1000 C. The weight loss is mainly attributed to the loss of low molecular weight oligomers, silanes and hydrogen. The net ceramic residue after the decomposition of the polymer was found to be around 89% at 1100 C in nitrogen. The TGA curve of SMP-10 has also been reported by Sreeja et al. [63]. Their result showed that the net ceramic residue after the decomposition of the polymer was 73% at 900 C in nitrogen. The difference between the two TGA results is large and the exact reason is not clear. However, some factors that might affect the TGA results are: the size of the sample, the sample heating rate, the type of flowing gas in the TGA instruments and the flowing rate of the gas. Different batches of the polymer precursor, SMP-10, might also have slightly different chemical structures. Moreover, since the polymer precursor decomposes/cross-links slowly at room temperature and the polymer precursor only has a recommended shelf life of six month, the aging of the polymer precursor might also affect the TGA curves. It should be noted that the TGA tests were done with only the polymer precursor. When the polymer precursor was mixed with the SiC particles and pressed into a green body as described in section 3.3, the polymer is geometrically arranged in a different way form the arrangement in the TGA test. Since the weight loss is associated with the loss of gases, the different geometrical arrangements can affect the final weight loss result. The differential scanning calorimetry (DSC) analysis of SMP-10 was tested with a TG/DTA 320 instrument (Seiko Instruments, Torrance, CA) in flowing nitrogen atmosphere. A heating rate of 10 C per minutes was used. The result DSC curve to a temperature of 1050 C is shown in Figure 3-7. A more detailed view of the DSC curve at temperatures below 400 C is shown in Figure 3-8. The exothermal reaction around 53

PAGE 54

250 C is believed to be caused by the cross-linking of the allyl double bonds and the self-cross-linking via the loss of hydrogen. The same exothermal peak was also observed by Sreeja et al. [63] in their DCS analysis of the SMP-10. 3.2.4 Crystallinity Analysis The crystallinity of the amorphous SiC from the decomposition of the polymer precursor at a sintering temperature of 1050 C in argon atmosphere was characterized with an X-ray diffraction technique using a Philips APD 3720 diffractometer (Andover, MA) with Cu K radiation (=1.5418 ). The crystallinity of the 0.6 nominal size -SiC particles was also characterized for comparison. The results are shown in Figure 3-9. No obvious peak is shown for the SiC from the decomposition of the polymer precursor at 1050 C. This confirms that the product from the decomposition of the polymer precursor is amorphous at this sintering temperature. The 0.6 m SiC particles showed strong peaks which verified the crystallinity of the as received crystalline particles. The mean grain size of the 0.6 m SiC particles was calculated using the Scherrer equation (Equation 3.1), Cos B d9.0 (3.1) where d is the mean grain size, is the x-ray wavelength, B is the line broadening at the half maximum intensity and is the Bragg angle. The mean grain size was determined to be ~73 nm using the strongest 111 peak. This indicates that the 0.6 m SiC particles were poly-crystalline. 54

PAGE 55

3.3 Fuel Pellets Synthesis Steps A conventional pellet fabrication route was used, which includes powder preparation, pressing and sintering. Initially, SiC powder with various size compositions and ceria powder were mixed with various amounts of polymer precursor in a mortal and pestle for five minutes. Other mixing methods were also used and will be discussed in more detail in section 5.2. The mixture (slurry) was then transported into a 13 mm die (Figure 3-10) and uniaxially pressed at various pressures with a Carver press (Carver, Wabash, IN) (Figure 3-11) at room temperature for 10 minutes to form the green bodies. The green bodies were then sintered in a Lindberg/Blue Mini-Mite tube furnace (Waltham, MA) (Figure 3-12) at temperature up to 1050 C for an hour under ultra high pure argon as the protection gas. The sintering temperature profile suggested by Starfire Systems Inc. for SMP-10 was followed and is shown in Figure 3-13. Starfire Systems Inc. suggested sintering SMP-10 to at least 850 C to form a stable amorphous SiC matrix. Sintering at a lower temperature might form an amorphous SiC phase that will oxidize more easily at ambient temperature and pressure. Oxidization of the amorphous SiC phase will reduce the thermal conductivity and is unfavorable. Since there are hydrogen atoms in SMP-10, it is expected to have residual hydrogen in the sintered amorphous SiC phase especially at lower sintering temperatures. The concentration of the residual hydrogen atom should be as low as possible since hydrogen can cause embrittlement of the cladding material and results in cladding failure. The TGA curve (Figure 3-6) shows that there is still weight loss from 850 C to 1100 C, which means that heating at these temperatures can potentially reduce the hydrogen content of the formed amorphous phase. The above reasons all indicate that a higher sintering temperature is preferred. On the other hand, a low 55

PAGE 56

temperature route to fabricate SiC IMF pellets is the goal of this study in order to prevent the potential chemical reactions of SiC and plutonium and to avoid the loss of minor actinides due to their higher vapor pressure at higher temperatures. A sintering temperature of 1050 C was chosen because it is below the reaction temperature of SiC and ceria (PuO 2 surrogate), which is around 1100 C [58]. It is higher than the suggested sintering temperature of 850 C. The higher sintering temperature of 1050 C can reduce the residual hydrogen and can form an amorphous SiC matrix with a higher thermal conductivity. 3.4 Fuel Pellet Characterization 3.4.1 Density of the Sintered Pellets The weight of the sintered pellet was determined to an accuracy of 0.1 mg by a lab balance. A caliper was used to determine the bulk volume. The difference of the pellet volume determined by a caliper and the conventional Archimedess method was found to be trivial. The bulk density was calculated from the weight and volume. The percent theoretical density was calculated following the rule of mixtures using the density values reported in section 3.2.2. It was also found that during sintering, the bulk volume shrinkage was nominal. Therefore, the volume of the sintered pellets was assumed to be equal to the bulk volume of the green body 3.4.2 Pore Size Distribution of the Sintered Pellets A Quantachrome Autoscan-60 Mercury Porosimeter (Boynton Beach, FL) was used to measure the open pore size distribution of the sintered pellets. The pellets were put into the sample holder, evacuated with a vacuum pump and filled with mercury using the filling apparatus (Figure 3-14). The sample holder together with the sample was then transferred to an Autoscan-60 Mercury Porosimeter (Figure 3-15). Pore radii, 56

PAGE 57

varying from 7 m to 4 nm, can be measured by forcing mercury to penetrate into evacuated open pores with high pressure. 3.4.3 Microstructure of the Sintered Pellets The pellets were cut with a diamond saw and mounted in epoxy. The cross section was polished to a 1m finish with a serious of diamond polishing pastes. The microstructure was examined in a JEOL SEM 6400 scanning electron microscope (Peabody, MA) operated at 15 kV with a working distance of 15 mm. An optical microscope was also used to examine the microstructure of the polished pellets. Table 3-1. Measured density of different types of powder and the amorphous SiC Material Measured density (g/cc) Standard deviation (%) 16.9 m SiC 3.22 0.007 1 m SiC 3.24 0.006 0.6 m SiC 3.25 0.008 Amorphous SiC 2.40 0.004 CeO 2 7.14 0.001 57

PAGE 58

01234567890.00.11.010.0100.01000.0Particle size (microns)Volume percent (% ) 16.9 micron SiC 1 micron SiC 0.6 micron SiC Figure 3-1. Particle size distribution for SiC particles 58

PAGE 59

Figure 3-2. SEM images of the as-received particles. (a) 70-100 nm ceria, (b) 0.6 m SiC, (c) 1 m SiC, (d) 16.9 m SiC 59

PAGE 60

Figure 3-3. The densities of the SiC from the decomposition of the polymer precursor as a function of the final sintering temperature [61] Figure 3-4. 1 H-NMR spectrum of the commercially available polymer precursor SMP-10 and the predicted structure [62] 60

PAGE 61

Figure 3-5. FT-IR spectrum of the polymer precursor, SMP-10 [62] 889092949698100102020040060080010001200Temperature (C)Weight (%) Figure 3-6. TGA curve of SMP-10 in nitrogen 61

PAGE 62

Figure 3-7. DSC curve of the polymer precursor, SMP-10 Figure 3-8. DSC curve of the polymer precursor, SMP-10, between 150 C and 400 C 62

PAGE 63

0500100015002000250010203040506070802 zetacounts 0.6 micron beta SiC amorphous SiC(111(200 ) (220)(311)(222) Figure 3-9. XRD profiles of the amorphous SiC and the 0.6 m -SiC particles Figure 3-10. Picture of the 13 mm die 63

PAGE 64

Figure 3-11. Picture of the Carver press Figure 3-12. Lindberg/Blue Mini-Mite tube furnace 64

PAGE 65

Figure 3-13. Sintering temperature profile for SiC IMF pellets Figure 3-14. The filling apparatus for the mercury porosimeter 65

PAGE 66

Figure 3-15. The autoscan-60 porosimeter 66

PAGE 67

CHAPTER 4 EFFECT OF POLYMER CONTENT AND COLD PRESSING PRESSURE 4.1 Change of the Polymer Precursor During Sintering The weight change associated with the sintering is determined by comparing the weight of the green body and the weight of the sintered pellet. The SiC particles alone were heated in the same tube furnace to a temperature of 1050 C and there was no detectable weight change. Therefore, it was assumed that all the weight change is from the decomposition of the polymer precursor. The volume measurements before and after sintering showed that there was only trivial bulk volume change during sintering. This means that the bulk volume of the green body is the bulk volume of the sintered pellet. During sintering, the polymer precursor lost about 30% to 35% of its weight, and the density increased from ~1 to 2.40 gram/cc. Hence ~72% of the volume originally occupied by the polymer became pores. From theoretical calculations, a green body with 15 weight percent polymer precursor and without any pores would sinter into a 73% TD pellet. All the pores in the sintered pellet were from the decomposition of the polymer precursor. It should be noted that in the TGA study of the polymer precursor in section 3.2.3, the polymer only lost about 11% of its weight when heated to a temperature of 1100 C. But the polymer in the green body lost about 30% to 35% of its weight during sintering to a temperature of 1050 C (assuming that all the weight loss of sintering was from the decomposition of the polymer precursor). The large weight loss difference might be cause by the different geometrical arrangement of the polymer precursor. When the polymer is dispersed between the SiC particles, it might be easier 67

PAGE 68

for the gases, associate with the decomposition, to diffuse out of the amorphous SiC phase. The crystallinity analysis described in section 3.2.4 shows that SMP-10 itself decomposes into amorphous SiC at a temperature of 1050 C. But, when SMP-10 is mixed with crystalline -SiC particles to form a green body and sintered, the SiC particles can potentially function as nucleation cites for the amorphous phase to nucleate. This effect was studies by XRD analysis of the sintered pellet (with a green body composition of 90 weight % 0.6 m SiC particles and 10 weight % SMP-10). The XRD pattern of the sintered pellet is compared with the patterns of the 0.6 m SiC particles and the amorphous SiC as shown in Figure 4-1. The XRD profile of the sintered pellet is almost identical as that of the SiC particles itself. If the polymer precursor decomposed to nano-crystalline -SiC, a broadening of the XRD peaks should be observed. The same XRD patterns suggest that the polymer precursor in the green body decomposes to amorphous SiC at a sintering temperature of 1050 C. 4.2 Effect of Polymer Content and Pressing Pressure on the Density of the Sintered Pellets The 1 m SiC powder was used to study the effect of polymer content and cold pressing pressure on the final density of the pellets. The powder and polymer precursor were mixed by a mortar and pestle for 5 minutes to form the slurry. The slurry was cold pressed in a die to form a green body. The green body was then sintered to form a pellet. Cold pressing pressures of 70, 200 and 600 mega-Pascal (MPa) and polymer weight contents of 5%, 10%, 12.5%, 15% and 17.5% were used. If the polymer content is lower than 5 weight %, the green body would be too weak to handle because there is not enough polymer to bind the SiC particles together. If the polymer content is too high, 68

PAGE 69

the result slurry will be too wet to press into a green body. For example, the combination of 20 weight % polymer and 70 MPa cold pressing pressure was tried but an intact green body can not be obtained. The sintered pellet density as a function the cold pressing pressure and polymer weight content is shown in Figure 4-2. Data scattering of two separate runs is shown for the 200 mega-Pascal (MPa) data sets. Data scattering was determined to be negligible. Moreover, the final density reached a plateau at approximately 72 percent theoretical density (%TD). This is very close to the theoretical calculations of 73% TD mentioned in section 4.1 When the polymer content was increased, the polymer filled the space between SiC particles and the densities increased until all the space was filled with the polymer precursor (saturation). When the polymer content was further increased, the excess polymer was squeezed out of the die during pressing. This is shown on Figure 4-2 as dash lines. When too much polymer was used, the slurry became too wet to process. The SiC particle packing density in the green body as a function of polymer content and cold pressing pressure is shown in Figure 4-4. The SiC particle packing density is related to both the pressing pressure and the polymer content. It was also observed that increasing the cold pressing pressure would increase the sintered pellet density by increase the packing of the SiC particles. Better packing of the SiC particles also meant that less polymer was needed to fill all the space between particles. Thus, the green body reached saturation with less polymer content. Song et al. [64] demonstrated that random closed packing of hard spheres in three dimensions can not exceed a density limit of similar to 63.4%. In this study, the SiC particles have an 69

PAGE 70

irregular shape and are not mono-sized. Moreover, cold pressing was used. However, the maximum achievable packing of SiC particles is still very close to the theoretical value of 63.4%. 4.3 Polymer Precursor as a Lubricant Figure 4-4 shows the green body volume change while different amounts of polymer precursor were added to 1 gram of 1m SiC particles with different cold pressing pressure. It was observed that the liquid polymer behaved as a lubricant during pressing. As mentioned in section 4.1, higher pressing pressure would lead to a better packing of SiC particles and thus a smaller green body volume. Moreover, when more polymer was added to 1 gram of SiC powder, the volume of the green body first decreased, reached a minimum and then increased. Evidently, the liquid polymer worked as a lubricant in the green body and thus decreased the friction forces between the SiC particles. Therefore, the particles could pack better with more polymer content and the overall volume decreased. However, when there was excess of the polymer, it would just occupy the space between particles during pressing; leaving no space for better packing. Therefore the green body volume would start to increase. The excess polymer would also be squeezed out of the die during pressing. The excess polymer would adhere near the surface of the green body during the die-release process and cause imperfections of the green body. 4.4 Effect of Polymer Content on Pore Size Distribution Figure 4-5 shows the pore size distributions and the cumulative pore volumes for three different pellets produced with 1 m SiC particles, with a pressing pressure of 200 MPa, and varied polymer contents of 5, 10 and 15 weight %. All of the pellets showed a 70

PAGE 71

bimodal pore size distribution. The large pores were in the size range of 110 to 30 nm and the small pores had the size of 30 to 4 nm. The large pores (110 to 30nm) have a size of about 10% of the SiC particles. This indicates that these pores are probably from the porous space between the SiC particles. The small pores are associated with the decomposition of the polymer precursor. As discussed in section 4.3, the polymer behaved as a lubricant to improve the packing of SiC particles; it can also fill the space between particles. Both effects could eliminate large pores in the green body, thus in the sintered pellet. The total volume of the large and the small pores from each pellet is summarized in Table 4-1. The total volume of the small pores being not proportional to the polymer content suggests that the forming of small pores not only relates to the decomposition of the polymer precursor but also relates to the geometric arrangements between the SiC particles and the polymer precursor. A proposed model is shown in Figure 4-6. When the space between SiC particles was only partially filled with the polymer precursor, sintering would result in large pores (Figure 4-6a). This is mostly the case with the 5% polymer precursor pellet. When the space between SiC particles was completely filled with the polymer precursor, sintering would result in small pores (Figure 4-6b). The relative amount of large and small pores can be controlled by the polymer content in the green body. 4.5 Effect of Cold Pressing Pressure on Pore Size Distribution Figure 4-7 shows the pore size distribution of three different pellets with the same polymer content (10 weight %), but different pressing pressure of 70, 200 and 600 MPa. All of the pellets also showed a bimodal pore size distribution with large pores in the 71

PAGE 72

size range of 150 to 30 nm and the small pores in the size range of 30 to 4 nm. The total volume of large and small pores from each pellet is summarized in Table 4-1. As discussed in section 4.4, the large pores are associated with the space between SiC particles. Increasing cold pressing pressure can eliminate large pores. This observation fits the calculated SiC particles packing density very well (Figure 4-3). Higher packing density associated with higher pressing pressure means the elimination of the large pores since the large pores are directly related to particle packing. The largest pore size detected by the porosimeter also decreased from 145 nm for the 70 MPa pellet to 90 nm for the 600 MPa pellet. Moreover, the size distribution of the large pores shifted to a smaller size with increased pressing pressure. The overall volume of the small pores increased with increasing pressing pressure. This result can be explained with the same model (Figure 4-6). With the same polymer content, higher pressure can achieve better SiC particles packing, thus allowing more chance for the space between particles to be completely filled with the polymer precursor. The result would be a sintered pellet with more small pores. Cold pressing pressure can affect the packing of SiC particles and thus affect the amount of large and small pores in the pellets. The largest pores are directly related to cold pressing pressure and can be manipulated. 4.6 Summary and Conclusion The effects of polymer content and cold pressing pressure on the density and pore size distribution of the sintered pellets were investigated using 1 m SiC particles. It was observed that by increasing the cold pressing pressure, the sintered pellet density increased because of the better packing of the SiC particles. It was also found 72

PAGE 73

that the liquid polymer precursor itself behaved as a lubricant to assist the packing of the SiC particles. However, too much polymer content would change the rheological properties of the mixture slurry and resulted in a slurry that is not suitable for pressing into green body pellets. The excess polymer would also just occupy the space between SiC particle thus leave no space for the particles to pack better. A bimodal pore size distribution was observed for all the sintered pellets. A model was proposed to explain how the geometric arrangements between SiC particles and the polymer precursor can affect the pore size distribution of the sintered pellets. Cold pressing pressure can affect the packing of SiC particles and thus affect the amount of large and small pores in the pellets. The largest pores are directly related to cold pressing pressure and can be controlled. The small pores are directly related to the decomposition of the polymer precursor. By controlling both the cold pressing pressure and the polymer content in the green body, the density and the pore size distribution of the sintered pellet could be tailored. 73

PAGE 74

Table 4-1. Total volume of large and small pores for the pellets Polymer content (%) Cold pressing pressure (MPa) Large pore volume (cc/gram) Small pore volume (cc/gram) 5 200 0.145 0.002 10 200 0.093 0.020 15 200 0.025 0.057 10 70 0.119 0.015 10 600 0.062 0.032 Figure 4-1. XRD profiles of the sintered SiC pellet (with a green body composition of 90 weight % 0.6 m -SiC particles and 10 weight % SMP-10), the 0.6 m -SiC particles and the amorphous SiC 74

PAGE 75

Figure 4-2. Percent theoretical density as a function of polymer content and cold pressing pressure Figure 4-3. SiC particle packing density as a function of polymer content and cold pressing pressure 75

PAGE 76

Figure 4-4. The green body volume as a function of the amounts of polymer precursor mixed with 1 gram of 1 micron SiC particles 76

PAGE 77

110100 0.000.010.020.03 5% SMP10, 200 MPa15% SMP10, 200 MPa10% SMP10, 200 MPaPore volume (cc/g) 110100 0.000.010.020.03 110100200 0.000.010.020.03 Pore size (nm)1101000.000.040.080.120.16 1101000.000.040.080.120.16 0.000.040.080.120.16 Cumulative pore volume (cc/g) Figure 4-5. Pore size distributions of pellets with 5, 10 and 15 weight % polymer precursor 77

PAGE 78

Figure 4-6. Schematics showing how geometric arrangements effect the decomposition of the polymer precursor 78

PAGE 79

110100200 0.000.010.020.03 10% SMP10, 70MPaPore volume (cc/g) Pore size (nm)110100200 0.000.010.020.03 10% SMP10, 200 MPa 110100200 0.000.010.020.03 10% SMP10, 600 MPa0.000.030.060.090.120.15 Cumulative pore volume (cc/g)0.000.030.060.090.12 0.000.030.060.090.12 Figure 4-7. Pore size distributions of pellets with 70, 200 and 600 MPa cold pressing pressure 79

PAGE 80

CHAPTER 5 EFFECT OF MIXING COARSE AND FINE SILICON CARBIDE PARTICLES It is well known that mixing of different particles size variations can achieve a better particle packing. A better particle packing in the green body would result in a higher sintered density. The density of the pellets would strongly affect the thermal and mechanical properties. The density of the pellet would also affect its other performance, such as fission gas release issues. Although the required (preferred) percent theoretical density and the mechanical strength for the fuel pellet are not clear, trying to reach a higher density and a lower porosity by mixing coarse and fine SiC particles remains the goal of this chapter. 5.1 Effect of Mixing Coarse and Fine Particles on the Sintered Density, Pore Size Distribution and Microstructure In this section, a coarse -SiC powder with a nominal size of 16.9 m and a fine -SiC powder with a nominal size of 0.6 m were utilized in the fabrication of the pellets (particle size analysis shown in Figure 3-1). The polymer precursor content was fixed at 10% by weight and 5 weight % of CeO 2 was always used as a surrogate for PuO 2 This corresponded to similar PuO 2 volume percent in the MOX fuel. All the powder and the polymer precursor were mixed with a mortar and pestle for five minutes and sintered in the tube furnace to a temperature of 1050 C for one hour (Figure 3-13) in a flowing argon atmosphere. 5.1.1 Effect on Density Different coarse and fine SiC particles volume ratios were tested with 3 different cold pressing pressures of 200, 600 and 800 MPa. The resulted sintered densities are shown in Figure 5-1. 80

PAGE 81

Pellets with 100% coarse SiC particles were tried and abandoned because the slurry did not have a practical rheological property to make an intact green body. Mixing of coarse and fine SiC particles strongly affects particle packing, and thus the sintered density. The maximum theoretical density achieved was 80.0% which occurred at 60% coarse powder and 40% fine powder with a 800 MPa cold pressing pressure. The fine particles can fit inside the pores formed by the coarse particles thus giving a better packing density. The theoretical coarse and fine particles volume ratio to reach the maximum packing density depends on the packing density of the coarse and the packing density of the fine particles. Assuming a random packing of the coarse particles with a packing density of 62%, then the space between the coarse particles that need to be occupied by the fine particles is 38%. Assuming that the fine particles also have a random packing density of 62%, we only need 62% 38% which means only 24% of the fine particles to fill the 38% space. So the overall volume ratio for the coarse particles is 62% / (62% + 24%), which equals to 72% coarse particles. If the packing density of the coarse particles is higher, then less fine particles are needed to fill the space between the coarse ones. Since the packing density of the SiC particles has been shown to be related to the cold pressing pressure, the best coarse and fine particle volume ratio should also be a function of the cold pressing pressure. Figure 5-1 shows that when a pressing pressure of 200 MPa was used, the best volume ratio to get a high density was 40% coarse and 60% fine. But, when a higher pressing pressure of 600 MPa was used, the best volume ratio changed to 60% coarse and 40% fine particles. 81

PAGE 82

The packing of the 1 m SiC particles with different polymer content and pressing pressure is discussed in section 4.2. The packing density is between 56% and 64%. The packing of the 16.9 m SiC particles is not clear, but it is assumed that the packing of the 16.9 m SiC particles is the same as the packing of the 1 m SiC particles at the same polymer content and pressing pressure. With this assumption, the calculated optimum volume ratio to achieve the highest packing is 69% coarse particles and 31% fine particles if both coarse and fine particles pack at a density of 56%. If both the coarse and fine particles have a packing density of 64%, the optimum volume ratio becomes 73% coarse and 27% fine particles. Comparing these values with the values determined in Figure 5-1, which were 40% coarse (for 200 MPa) and 60% coarse (for 600 MPa and 800 MPa), the experiment results show that more fine particles are needed to achieve a higher density than the theoretical calculations. One explanation is the agglomeration of the fine particles. Agglomerates of some of the fine particles means that more fine particles are required to fill the space between the coarse particles. The irregular shape of the SiC particles might affect the packing too. The microstructure analysis is discussed in section 5.1.3. 5.1.2 Effect on Pore Size Distribution Figure 5-2 shows the cumulative pore volumes of nine pellets from Figure 5-1, together with pore size distributions of six pellets. A bimodal distribution was still observed. The pore size distribution was strongly influenced by the composition of the particle size. The large pore size was around 10% of the size of the largest particle size present in the green body. The largest pore size was affected by the coarse and fine particles volume ratios and the cold pressing pressure. 82

PAGE 83

The presence of the 16.9 m SiC particles created large pores in the micron-size range. However, with the right coarse and fine particles volume ratio, most of the micron size pores could be eliminated by a higher cold pressing pressure (Figure 5-2b). The largest pore size also decreased from 1 micron for the 200 MPa pressed pellet to 0.5 micron for the 800 MPa pressed pellet. If there were not enough fine particles to fill the space between the coarse ones, the largest pore size tended to become larger. Moreover, elimination of large pores through pressing did not occur as effectively (Figure 5-2c). However, a significant amount of large pores could still be eliminated by a higher pressing pressure. 5.1.3 Microstructure Characterization SEM images (back scattered) of a polished pellet with 10 weight % polymer precursor and 5 weight % ceria are shown in Figure 5-3. The SiC particles were composed of 40% fine and 60% coarse particles. The pellet was cold pressed at 200 MPa. Figure 5-3 (a) shows that the ceria (white spots) was well distributed. Some agglomerates of the 0.6m SiC particles with size in the 100 micron range can also be seen (black arrow). These areas contain no ceria and would cause higher ceria concentration in other areas. Figure 5-3 (b) shows a more detailed view of the same pellet. Agglomerates of the small particles can be seen on the right side. Some 16.9 m SiC particles are shown on the left side with ceria particles packed between SiC grains. A few micron-sized pores exist near the coarse SiC particle, which agrees with the pore size analysis by the porosimeter. The agglomerates size of ceria is about 1-3 m in diameter. 83

PAGE 84

It proved difficult to break all the agglomerates by a mortar and pestle mixing process. An alternate high energy shaker mill mixing method is discussed in section 5.2. A homogeneous distribution of the coarse and fine particles is preferred to result in a pellet with a higher density, a better ceria distribution and with less large (micron-sized) pores. The observed agglomerates of the fine particles explained the reason that more fine particles were required to achieve a high density. Fine SiC particles in the agglomerates can not fill the space between the coarse particles. And thus more fine particles are required to fill the space. 5.2 Utilization of Different Mixing Methods In section 5.1, it was found that mixing coarse and fine SiC particles can get a better particle packing, and thus results in a higher sintered pellet density. The maximum density occurred at 60% coarse and 40% fine SiC powder. It was also observed that the mortar and pestle mixing method was not efficient enough to break the agglomerates of the fine SiC particles. In this section, two other mixing methods were used and the result sintered pellets are compared with the mortar and pestle mixed pellets. If the agglomerates of the fine particles can be eliminated by the more aggressive mixing method, theoretically the volume ratio to achieve the highest density should change to more coarse and less fine particles. However, this study emphasizes on demonstrating the feasibility of different process methods, but not on fine tuning the process parameters. Thus a particles size composition of 60% coarse powder and 40% fine powder was still used throughout this section. 84

PAGE 85

5.2.1 Different Mixing Procedures A conventional pellet fabrication route was followed which includes: mixing of different particles and polymer precursor to form the slurry, pressing the slurry into a green body and sintering the green body to form a pellet. The amount of polymer precursor was fixe at 10 weight % in the slurry for all the pellets. CeO 2 (PuO 2 surrogate) was fixed at 5 weight %. The amount of SiC powder was fixed at 85 weight %. The composition of the SiC was 60 % coarse (16.9m) and 40% fine (0.6m) since this was determined to be the optimal ratio to achieve a high density in section 5.1.1. Three mixing methods were used to form the slurry: (1) Mixing with a mortar and pestle for 5 minutes as mentioned in section 3.3. This method is hereafter denoted as mortar and pestle. (2) The mixture was put into a stainless steel vial with one 7 mm stainless steel spherical blending media. The vial was inserted into a SPEX high energy shaker mill (SPEX 8000 M Mixer/Mill, SPEX SamplePrep, Metuchen, NJ) and blended for 40 minutes. This mixing method is hereafter denoted as SPEX without the hexane. (3) The mixture was put into the same stainless steel vial with the blending media, but hexane was also added into the vial as a grinding aid. The hexane to SiC powder weight ratio was 1/4. The vial was put into the SPEX high energy shaker mill and blended for 40 minutes. After blending, the slurry was put in a laboratory fume hood for 24 hours for the hexane to evaporate. This method is hereafter denoted as SPEX with the hexane. After mixing, the slurry was pressed in a 13 mm die with two various pressing pressures of 200 and 600 MPa to form a green body. The green body was then sintered in a tube furnace to a temperature of 1050 C as described in section 3.3. The pellets from different mixing methods were then characterized and compared according to section 3.4. Three replicates were done for each condition to determine the average density. 85

PAGE 86

5.2.2 Effect of the Mixing Methods on SiC Particle Size The particles size ratio on the maximum packing density for bimodal powder mixtures has been studies by McGeary [65]. His results showed that for dense random packing of binary mixtures of spheres, the packing density increased as the size ratio increased to ~15 but was relatively unchanged at higher size ratios. The use of the high energy mill is to mix the coarse and fine particles and to break the agglomerates form the fine (0.6 m nominal size) SiC particles. Grinding of the coarse SiC particles in the high energy mill is undesirable because this can potentially reduce the coarse to fine particles size ratios and hence reduce the packing density. The grinding of the fine SiC particles by the high energy mill is assumed to be less effective since it is generally believed that its much more difficult to grind small particles compared to large particles. The effect of a high energy mill on the coarse SiC particles were determined by blank runs using only the coarse SiC powder (both with and without hexane) for 40 minutes. After grinding, the hexane was removed in a lab fume hood over night. The particles size distribution of the ground coarse SiC powder was measured by a laser light scattering technique as discussed in section 3.2.1. The influence of using the SPEX high energy shaker mill (both with and without hexane) on the particle size distributions of the coarse SiC particles are shown in Figure 5-4. The particle size distribution of the fine SiC particles is also shown for comparison. It was determined that the use of the high energy shaker mill without hexane did not reduce the size of the coarse SiC particles significantly, although slight size reduction can be seen. A high energy shaker mill with hexane as a mixing aid can reduce the size of the coarse SiC particles more efficiently than without hexane, but most of the 86

PAGE 87

particles still retained their original size. The effect of grinding on the fine SiC particles is assumed to be much smaller than the effect of the coarse particles and thus is assumed to be trivial. 5.2.3 Effect of the Mixing Methods on Density The densities of the sintered pellets as a function of the mixing method and cold pressing pressure are shown in Figure 5-5. A mortar and pestle method is not an efficient way of mixing the SiC particles and the polymer precursor hence the achieved density after sintered was the lowest. The density of the pellet was more dependent on the cold pressing pressure because a higher pressure is required to get a better packing when the coarse and fine SiC particles are not well mixed. A severe caking effect was observed for the SPEX without the hexane mixing method. The mixture caked near both ends of the stainless vial during grinding, but the achieved density was still much higher than the mortar and pestle mixing method. The hexane worked as a grinding aid and solved the caking problem. The hexane also behaved as a solvent to dilute the polymer precursor. The highest density was achieved with this method (81.2 % theoretical density) with a cold pressing pressure of 600 MPa. The densities of the sintered pellets are less dependent on the cold pressing pressure for pellets mixed with a SPEX mill (both with and without hexane). This indicates that the coarse and fine SiC particles were well mixed. Thus, a higher pressure was not required to obtain a better packing. High pressing pressure is not preferred because during pressing, the particles undergo elastic compression. When the pressure is released, the stored elastic energy leads to an expansion of the green pellets. This is generally referred as springback effect [66, 67] and is directly related to the pressing pressure. The springback effect can cause defects in the green body. At 87

PAGE 88

the low sintering temperature in this study, these defects will remain in the sintered pellets. These defects can potentially reduce the mechanical strength of the pellets. 5.2.4 Effect of the Mixing Methods on Pore Size Distribution Figure 5-6 shows the cumulative pore volume and pore size distribution of the sintered pellets in Figure 5-5. For the mortar and pestle mixed pellets, a bimodal pore size distribution was observed. For the 200 MPa pressed pellet, the large pores are about 1 micron in diameter and the small pores are in the range of 4 to 50 nm. The large pores are related to the packing voids in the green body. The small pores are related to the decomposition of the polymer precursor. A more detailed study about how geometrical arrangement of the polymer precursor and SiC particles in the green body effects the pore size distribution of the sintered pellets is given in section 4.4. For pellet mixed with SPEX without the hexane, most of the large pores were eliminated, thus the overall pore volume was decreases compared to the mortar and pestle mixed pellets. Most of the pores detected by the porosimeter are smaller than 100 nm in diameter. The largest pore size detected was 0.7 and 0.5 micron for the 200 and 600 MPa pressed pellets respectively. The large pores (larger than 50nm) can still be eliminated by applying higher pressing pressure. But the pressing pressure has very little effect on the small pores. Since the majority of the pores are small pores, eliminating large pores has little effect on the overall pore volume of the pellets. The overall pores volume is thus less dependent on the pressing pressure. As a result, the density of the pellet is less dependent on the pressing pressure. This observation agrees with the density measurements shown in Figure 5-5. For pellets made with the SPEX with the hexane mixing method, the pore size distributions are very similar to the pore size distributions of pellets mixed with SPEX 88

PAGE 89

without the hexane method. The largest pore size detected was 0.2 and 0.5 micron for the 200 and 600 MPa pressed pellets respectively. For the 200 MPa cold pressed pellet, utilization of the hexane as a grinding aid greatly reduced the largest pore size from 0.7 m (mixed without hexane) to 0.2 m. For the 600 MPa cold pressed pellet, the reduction of the largest pore size did not occur. This observation suggests that other mechanisms associate with the high pressing pressure may prevent the reduction of the pore size. This topic is discussed in more detail in section 5.2.5. 5.2.5 Effect of the Mixing Methods on Microstructure Optical microscopy images of polished pellets pressed at 200 MPa pressure with different mixing methods are shown in Figure 5-7. The black arrow marks the pressing direction of the pellets. The white arrows indicate the agglomerates of the 0.6 m nominal size SiC particles, which were also observed in the SEM images shown in Figure 5-3. Mixing with SPEX without hexane can eliminate most of the agglomerates in the pellet (Figure 5-7 (c) and (d)) despite the severe caking effect observed. However, some agglomerates can still be seen in Figure 5-7 (c). The better mixing of SPEX without hexane corresponds well with the observed higher sintered density (Figure 5-5) and the elimination of the large pores of the sintered pellets (Figure 5-6). The SPEX with the hexane mixing method can eliminate all the agglomerates of the 0.6 m nominal size SiC particles (Figure 5-7 (e) and (f)). The polished surface did not show any heterogeneous sites. Optical microscopy images of polished pellets pressed at 600 MPa pressure with different mixing methods are shown in Figure 5-8. In general, all of the images look similar to the 200 MPa pressed pellets except that cracks perpendicular to the uniaxial 89

PAGE 90

pressing direction can be seen for all of the 600 MPa pressed pellets. The width of the cracks was in the micrometer size range and the length of the cracks was over 200 m. Since the pore size distribution of the 600 MPa pressed pellets all showed overall pore volume reduction compared to the 200 MPa pressed pellets (Figure 5-6) these cracks were probably mostly closed cracks and did not connect to the surface of the pellets. Hence they did not contribute to the pore volume measured by the mercury porosimeter. Although they can not be detected by the mercury porosimeter, these cracks can potentially work as defects and reduce the mechanical strength of the pellets. The mechanical properties of the pellets are discussed in Chapter 7. These cracks are believed to be caused by the springback effect which related to the release of the elastic energy when the pressing pressure was released [66, 67]. The pellet mixed with high energy shaker mill with hexane had more cracks, but all the cracks had smaller width. This is probably because the pressing strain was more evenly released across the homogeneously mixed pellet. For pellets with heterogeneous sites, the stored elastic energy would release at the weak points during the pressure release. This results in fewer but wider cracks as shown in Figure 5-8 (a) (d). Other than the springback effect, since the pressing pressure does not distribute evenly during pressing, there are high pressure zones and low pressure zones during pressing. The pressure zones cause different packing densities in the green body and thus different densities in the sintered pellet. The different densities can potentially cause defect in the pellet. Very high pressing pressure increases this effect and should be avoided. Back scattered SEM images of polished pellets pressed at 200 and 600 MPa pressure with different mixing methods are shown in Figure 5-9 and Figure 5-10 90

PAGE 91

respectively. The contrast of these images is from the atomic weight difference and the white phase is determined to be cerium-containing phase. Individual SiC particles in the agglomerate can be seen in Figure 5-9 (b), (d), and Figure 5-10 (b). The cerium-containing phase was excluded form the agglomerate and thus was not well distributed except for pellets mixed with the SPEX with the hexane method, which had no agglomerates. The cerium distribution throughout the whole pellet is further examined using back scattered SEM images with an even lower magnification (Figure 5-11). The cerium containing phase was well distributed in the pellets mixed with SPEX with the hexane method as shown in Figure 5-11 (e) and (f). The cold pressing pressure does not affect the distribution of the cerium containing phase. The cracks from the springback effect observed in the optical microscopy images (Figure 5-8) cannot be observed easily with SEM images. 5.3 Summary and Conclusion The effect of mixing coarse and fine SiC particles to further increase the density of the SiC IMF was investigated. Three mixing methods were utilized to mix the particles and the polymer precursor. The sintered density, pore size distribution and microstructure of the pellets were characterized. It was found that mixing the coarse (16.9 m) and fine (0.6 m) SiC particles can greatly increase the sintered density of the pellets. The presence of the large SiC particles (16.9 m) created large pores in the micron size range. However, with the right coarse and fine particle volume ratio, most of the large (micron-sized) pores could be eliminated by a higher cold pressing pressure (Figure 5-2b). 91

PAGE 92

Three different mixing methods were investigated. The mixing methods strongly affect the sintered density, the pore size distribution and the microstructure. The pellets mixed with the high energy shaker mill with the hexane as a grinding aid had the highest density and had no heterogeneous sites. The highest theoretical density achieved was 81.2% with a cold pressing pressure of 600 MPa. The microstructure analysis showed that pellets pressed with a pressure of 600 MPa suffered from springback effect and had numerous cracks that were perpendicular to the pressing direction. These cracks were not detectable by the mercury porosimeter. The 200 MPa pressed pellets showed no cracks because of the lower pressing pressure used. 92

PAGE 93

68707274767880820204060800.6 m SiC wt % (balance with 16.9 m SiC powder)Theoretical density (%) 100 800 MPa 600 MPa 200 MPa Figure 5-1. Percent theoretical density as a function of particles size composition and pressing pressure 93

PAGE 94

1101001000 0.000.010.02 (c)(b)Pore volume (cc/g) (a)0.000.040.080.12 200 MPa 600 MPa 800 MPa 200 MPa 600 MPa 800 MPa 200 MPa 600 MPa 800 MPaCumulateive pore volume (cc/g)1101001000 0.000.010.02 0.000.040.080.12 1101001000 0.000.010.02 Pore size (nm)0.000.040.080.12 Figure 5-2. Pore size distributions and cumulative pore volume of pellets with (a) 100% 0.6 m, (b) 40% 0.6 m, 60%16.9 m and (c) 20% 0.6 m and 80% 16.9 m SiC particles Figure 5-3. SEM images (back scattered) of a pellet with 10 weight % polymer precursor and 5 weight % ceria. The SiC particles were composed of 40% fine particles and 60% coarse particles. 94

PAGE 95

Figure 5-4. The influence of using the high energy shaker mill on the particle size distributions of the coarse SiC particles 95

PAGE 96

Figure 5-5. Density of the sintered pellets as a function of mixing method and cold pressing pressure 96

PAGE 97

11010010002000 0.0000.0040.008 600 MPa600 MPa200 MPa200 MPa200 MPa mortar and pestle, 200 MPa mortar and pestle, 600 MPapore volume (cc/g) pore size (nm)600 MPa11010010002000 0.0000.0040.008 (c)(b) SPEX without hexane, 200 MPa SPEX without hexane, 600 MPa 11010010002000 0.0000.0040.008 SPEX with hexane, 200 MPa SPEX with hexane, 600 MPa0.000.020.040.060.08 cumulative pore volume (cc/g)(a)0.000.020.040.060.08 0.000.020.040.060.08 Figure 5-6. Pore size distributions and cumulative pore volume of pellets mixed with (a) mortar and pestle (b) high energy shaker mill without hexane and (c) high energy shaker mill with hexane as a grinding aid 97

PAGE 98

Figure 5-7. Optical microscopy images of polished pellets pressed at 200 MPa and mixed with (a), (b) mortar and pestle (c), (d) SPEX without the hexane and (e), (f) SPEX with hexane. The arrow marks the pressing direction of the pellets. 98

PAGE 99

Figure 5-8. Optical microscopy images of polished pellets pressed at 600 MPa and mixed with (a), (b) mortar and pestle (c), (d) SPEX without the hexane and (e), (f) SPEX with the hexane. The arrow marks the pressing direction of the pellets. 99

PAGE 100

Figure 5-9. Back scattered SEM images of polished pellets pressed at 200 MPa and mixed with (a), (b) mortar and pestle (c), (d) SPEX mill without the hexane and (e), (f) SPEX with the hexane 100

PAGE 101

Figure 5-10. Back scattered SEM images of polished pellets pressed at 600 MPa and mixed with (a), (b) mortar and pestle (c), (d) SPEX without the hexane and (e), (f) SPEX with the hexane 101

PAGE 102

Figure 5-11. Back scattered SEM images of polished pellets with (a): mortar and pestle mixing, 200 MPa (b): mortar and pestle mixing, 600 MPa, (c) SPEX mill without hexane, 200 MPa, (d) SPEX mill without hexane, 600 MPa, (e) SPEX mill with hexane, 200 MPa, (f) SPEX mill with hexane, 600 MPa 102

PAGE 103

CHAPTER 6 EFFECT OF POLYMER INFILTRATION AND PYROLYSIS CYCLES 6.1 Introduction Polymer infiltration and pyrolysis (PIP) includes infiltration of the pre-ceramic polymer precursor into the pores of the ceramic matrix, and then pyrolyse the polymer to yield ceramic material. The PIP method is an attractive and widely studied technique for industrial production of silicon carbide fiber reinforced silicon carbide matrix composites (SiC/SiC composites). This technique has also been used to fabricate particulate reinforced composite [68]. The main advantage of the PIP method is the potential drastic reduction of fabrication cost because the continuous PIP process is applicable for a large scale production [69]. However, several disadvantages of the conventional PIP produced SiC/SiC composite include: lowered thermal conductivity, insufficient elastic modulus and poor radiation tolerance. These undesirable properties are associated with the amorphous structure and excess carbon/oxygen contents in the PIP produced amorphous SiC after a typical pyrolysis at temperatures of 1473-1673 K. Another disadvantage is that if a high densification level is required, several PIP cycles have to be performed (typically 6 to 10, and even more) which is time consuming and costly [70]. Nechanicky et al. [68] synthesized a -SiC/-SiC particulate composite via PIP processing. In their work, -SiC powder functioned as the particulate reinforced phase and -SiC, from the decomposition of a polymer precursor, functioned as the matrix phase. A modified polymethylsilane (mPMS) was used as a -SiC polymer precursor [71], which showed a ceramic yield of 85% at 1000 C. For the polymer infiltration step, the precursor, mPMS, was dissolved in toluene with a concentration of 10 weight %. 103

PAGE 104

After 12 hours, the samples were dried and heat treated to 1000 C under an argon atmosphere. It was found that the densities reached a plateau after about 6 PIP cycles. PIP process becomes less effective with increasing numbers of PIP cycles because the pores inside the samples become much smaller and it becomes harder for the precursor to infiltrate into the pores. SiC fiber reinforced SiC matrix composite was synthesized by Katoh et al. [69] with multiple PIP cycles. Two kinds of fiber and two kinds of SiC pre-ceramic polymer precursors were used, namely polyvinylsilane (PVS) [72] and polycarbosilane (PCS) [73, 74]. The increase in composites mass density and the decrease in porosity with the increasing number of impregnation and pyrolysis cycles are shown in Figure 6-1 [69]. After the first few PIP cycles, the ability for each further PIP process to increase the density was reduced. Ceramic composite nuclear fuel pellets consisting of UO 2 contained with a SiC matrix have been fabricated with a PIP process by Singh et al. [75]. Eight PIP cycles were used in their study. The density and porosity of the fabricated materials were determined after the 8 th re-infiltration using the buoyancy method with ultra high purity water and are shown in Table 6-1 [75]. They concluded that the PIP process generated closed pores in the consolidated pellets; therefore the bulk densities of the pellets obtained were lower than their true densities. -SiC particulate reinforced SiC composite has also been fabricated and characterized by Lee et al. [61]using a polymer precursor route together with PIP cycles. Their objective was to fabricate a SiC part as a structure material for fusion reactors. They developed a polymer infiltration method where the solid samples were degassed 104

PAGE 105

and the degassed samples were impregnated without exposing to air (Figure 6-2 [61]). The chamber B in Figure 6-2 [61] was evacuated with vacuum and the polymer was then introduced into the samples using the pressure difference between the two chambers. The samples, which were still submerged in the polymer, were then transferred to a pressure vessel and pressurized with gas while submerged. The polymer infiltrated samples were then sintered in ultra high purity argon atmosphere at a temperature of 1573 K. The samples became denser with the successive PIP cycles but reached a plateau. The achievable maximum densities varied with the fabrication conditions and the PIP process. The maximum theoretical density achieved was about 87% after 12 PIP cycles. In summary, several research groups have demonstrated the validity of the PIP method to further increase the density of ceramic parts. However, a high density of over 90 %TD is difficult to achieve. Moreover, higher density is generally achieved by utilizing multiple PIP cycles, which can be costly and time consuming. 6.2 Experimental Procedure Since the goal of the PIP cycles is to close the open pores and to further increase the density, SiC IMF pellets with the least open pore volume and with the highest density were selected to undergo PIP cycles. The pellets selected for PIP studies were fabricated with 85 weight % SiC particles (40% fine and 60% coarse), with 10 weight % polymer precursor and 5 weight % CeO 2 The slurry was mixed the SPEX with the hexane method. The mixed slurry was cold pressed with a pressure of 600 MPa for 10 minutes and sintered to a temperature of 1050C for one hour. The obtained density was about 81.2% TD. Two kinds of infiltration pressures were used to determine the effect of pressure on the infiltration process, namely one atmosphere and 10 MPa argon 105

PAGE 106

pressure. After infiltration, the pellets were sintered in the same tube furnace with the same temperature profile as pellets fabrication process discussed in section 3.3. 6.2.1 PIP with 1 Atmosphere The sintered pellets were put into a vacuum chamber. A vacuum pump was used to evacuate the chamber to a pressure of 10 Torr or lower. The pellets were evacuated in the chamber for 4 hours in order for the gas molecules to diffuse out from the small pores in the pellets. The polymer precursor was then introduced into the pellets while maintaining the vacuum using the pressure difference between the chamber and the ambient pressure. The polymer precursor was introduced until all the pellets were submerged. The pellets were then brought to ambient pressure and infiltrated for 24 hours. After infiltration, the excess polymer on the surface was wiped out and the polymer infiltrated pellets were put into a tube furnace and pyrolyzed under ultra high pure argon as the protection gas to a temperature of 1050 C. The same sintering temperature profile for sintering SiC IMF pellets was followed as shown in section 3.3. The PIP cycle was repeated for one more time to further densify the pellets. 6.2.2 PIP with 10 MPa Argon Pressure The fist few steps were the same as PIP with one atmosphere, but after the samples were brought to ambient pressure; they were transferred into a pressure chamber while submerged. An ultra high purity argon gas was used to pressurize the chamber to a pressure of 10 MPa. The pellets were infiltrated in the pressure chamber for 24 hours. After infiltration, the pressure was released and the same procedures as PIP with one atmosphere were used to pyrolyze the infiltrated pellets. This process was repeated for one more cycle. 106

PAGE 107

It should be noted that the solubility of Ar in the polymer precursor, SMP-10, is a function of pressure and has been investigated by Lee et al.[61]. Their results are shown in Figure 6-3 [61]. When using a high infiltration pressure, the Ar molecules would dissolve in the precursor. After the pressure was released, the dissolved Ar gas was released from the polymer and formed bubbles in the polymer. If the bubbles were forming inside the pellets, they could potentially push the already infiltrated polymer out and reduce the effectiveness of the infiltration process. For this reason, very high infiltration pressure should be avoided. 6.3 Results and Discussion 6.3.1 Effect on Density Density improvements through 2 PIP cycles with 2 different infiltration pressures are shown in Figure 6-4. For the first PIP cycle, a density increase of about 4.8% was observed for both infiltration pressures. For the successive second PIP cycle, the 1 atmosphere pressure PIP can only achieve a density increase of 0.5%, but the PIP with10 MPa Ar pressure can achieve a density increase of about 1.0%. The size of the open pores in the pellets before any PIP cycle was large enough for the 1 atmosphere infiltration process, thus utilizing higher infiltration pressure in the first PIP cycles has little effect on the final density compared to infiltration with just atmosphere pressure. However, utilization of higher infiltration pressure can potentially reduce the infiltration time. As for the second PIP cycle, since the pores become smaller after the first PIP cycle (see section 6.3.2), it becomes difficult for the polymer precursor to infiltrate into the small open pores. In this case, utilizing high infiltration pressure becomes necessary for the polymer to infiltrate into small pores. Even with the high infiltration pressure of 10 107

PAGE 108

MPa, the increase in % theoretical density (1%) is still much smaller than that of the first PIP cycle (4.8%). Comparing our density data with other research groups [61, 68, 69, 75] data, our results showed that a high theoretical density of ~87% was achieved with relatively few times of PIP cycles (2 cycles). The reasons are: 1) a very high pellet density of ~81% was achieved even before the PIP cycles; 2) a liquid polymer precursor was used, so no solvent was required for the polymer infiltration process. The use of the solvent will reduce the actual amount of the precursor infiltrated in the pores and thus reduce the rate of the density increase though PIP cycles. 6.3.2 Effect on Pore Size Distribution The pore size distribution and cumulative pore volume of pellets without any PIP cycle and with 1 and 2 PIP cycles at one atmosphere and at 10 MPa argon pressure are shown in Figure 6-5 and Figure 6-6 respectively. It can be seen in Figure 6-5 that after the first 1 atmosphere PIP cycle, most of the open pores were closed except those pores with diameters smaller than 10 nm. The same pore size distribution was observed for the pellet with the 10 MPa argon pressure PIP process (Figure 6-6). For the second PIP cycle, relatively little change in the pore size distribution and cumulative pore volume was observed for the pellet with one atmosphere PIP. The overall open pore volume decreased form 0.024 to 0.020 cc/gram. For the second PIP cycle with 10 MPa argon pressure, the overall open pore volume decreased from 0.023 to 0.015 cc/gram. Apparently a higher infiltration pressure is necessary to force the polymer to infiltrate into pores that are smaller than 10 nm in diameter. But even with the higher infiltration pressure, it is still very hard to close all the pores that have 108

PAGE 109

diameters smaller than 10 nm. All the pore size and volume data corresponds very well with the density measurement. Without using a solvent to dissolve the polymer precursor and reduce the viscosity of the polymer precursor, this pores size (10nm) might be the limitation for the infiltration process. If closing those small pores is highly desirable, other methods should be utilized, for example chemical vapor deposition [76-78] or chemical vapor infiltration [79, 80]. 6.3.3 Effect on Phase Volume Composition The effects of PIP cycles on the phase volume composition of the pellets were calculated using the densities of each phase, the weight change through PIP cycles and the measured volumes. The phase volume compositions of the pellets before PIP cycle and after one and two PIP cycles with one atmosphere infiltration pressure and with 10 MPa argon infiltration pressure are shown in Figure 6-7 and Figure 6-8 respectively. It is assumed that all the open pores can be detected by the mercury porosimeter. And the pores that are not detected by the mercury porosimeter are closed pores. For PIP cycles with one atmosphere infiltration pressure, the first PIP cycle can effectively close the open pore from 12.1 volume % to 6.7 volume %. The amount of amorphous SiC in the pellet also increased from 8.0 volume % to 12.4 volume %. Since all the pores after the first PIP cycle were already very small, with diameters smaller than 10 nm (Figure 6-5), the second PIP cycle can only close the open pores from 6.7 volume % to 5.9 volume %. The PIP cycles do not increase the closed pore volume % significantly. The closed pore volume increased from 6.6 % to 7.7% after 2 PIP cycles. For PIP cycles with 10 MPa argon infiltration pressure, the first PIP cycle can close the open pore from 12.1 volume % to 5.8 volume %. The second PIP cycle can further 109

PAGE 110

close the open pores to 4.0 volume % and increase the amorphous SiC phase to 14.9 volume %. The PIP cycles still do not affect the closed pore volume significantly. 6.4 Summary and Conclusion In this chapter, the effects of polymer infiltration and pyrolysis (PIP) cycles on the density, the pores size distribution and the phase composition of the pellets have been investigated. Pellets with a composition of 85 weight % SiC particles (60% coarse, 40% fine), 10 weight % SMP-10 and 5 weight % CeO 2 were chosen to study the effects of PIP cycles. Two infiltration pressures (one atmosphere and 10 MPa) were examined. It was found that the first PIP cycle could increase the density of the pellets form ~81% TD to ~86% TD for both infiltration pressures. The pore size distribution showed that most of the open pores were closed after the first PIP cycle except those pores with diameters smaller than 10 nm. The successive second PIP cycle was not as effective on either the density or the pore size distribution of the pellets. This could be explained by the smaller pore size distribution of the pellets before the second PIP cycle, which makes the infiltration process difficult. In this case, the high infiltration pressure showed better results on both density and pore size distribution. However, the pores with diameters smaller than 10 nm were still hard to infiltrate and close by the PIP process even with an infiltration pressure of 10 MPa for 24 hours. The maximum density achieved was ~87% TD with 2 PIP cycles. 110

PAGE 111

Table 6-1. Density and porosity of the samples after 8 th re-infiltration [75] Material Bulk density (g/cm 3 ) Open porosity (%) a-SiC 2.21 0.01 3.39 1.14 a-SiC-U 3 O 8 -A 4.58 0.06 2.77 1.14 a-SiCU 3 O 8 -B 4.70 0.06 1.94 0.56 Figure 6-1. Changes in apparent mass density and porosity with the increasing number of impregnation and pyrolysis cycles [69] 111

PAGE 112

Figure 6-2. Schematics of the polymer infiltration system [61] Figure 6-3. Helium and argon gas solubility depending on the applied gas pressure [61] 112

PAGE 113

80818283848586878801Number of PIP cycles% Theoreticla densit 2 y 1 atm PIP 10 MPa Ar PIP Figure 6-4. % Theoretical density improvements with PIP cycles with different infiltration pressure 113

PAGE 114

1101001000 0.0000.0010.0020.0030.004 600 MPa SPEX with hexanePore volume (cc/g) 1101001000 0.0000.0010.0020.0030.004 1101001000 0.0000.0010.0020.0030.004 After 2 PIP cycles with 1 atm pressure Pore size (nm)11010010000.000.010.020.030.040.05 11010010000.000.010.020.030.040.05 After 1 PIP cycle with 1 atm pressure0.000.010.020.030.040.05 Cumulative pore volume (cc/g) Figure 6-5. Pore size distributions and cumulative pore volume of pellets with PIP cycles using 1 atmosphere pressure 114

PAGE 115

1101001000 0.0000.0010.0020.0030.004 600 MPa SPEX with hexanePore volume (cc/g) 1101001000 0.0000.0010.0020.0030.004 1101001000 0.0000.0010.0020.0030.004 Pore size (nm)11010010000.000.010.020.030.040.05 11010010000.000.010.020.030.040.05 After 1 PIP cycle with10 MPa Ar pressure0.000.010.020.030.040.05 After 2 PIP cycles with 10 MPa Ar pressureCumulative pore volume (cc/g) Figure 6-6. Pore size distributions and cumulative pore volume of pellets with PIP cycles using 10 MPa argon pressure 115

PAGE 116

Figure 6-7. Phase volume composition of SiC pellets before and after PIP cycles with one atmosphere infiltration pressure 116

PAGE 117

Figure 6-8. Phase volume composition of SiC pellets before and after PIP cycles with 10 MPa argon infiltration pressure 117

PAGE 118

CHAPTER 7 MECHANICAL PROPERTIES OF THE PELLETS 7.1 Introduction An understanding of the mechanical properties of nuclear fuel is important since the swelling and fracture of the pellets have a strong influence on the temperature of the fuel, the fission gas release and the mechanical interaction between the fuel and the cladding [81]. There is a lack of published data on the mechanical strength requirement for nuclear fuel pellet. However, there is some information about mechanical properties of simulated MOX fuel pellets and pellets made from inert matrix candidate materials. In this section, some literary background about the mechanical properties of these pellets is reported. Lee et al. [82] analyzed the behavior of UO 2 -5 weight % CeO 2 simulated MOX fuels. In their study, the simulated MOX fuel was sintered at a temperature of 1700 C for 4 hours in a reducing atmosphere (93% N 2 + 7% H 2 ). The % theoretical density was controlled by the amount of pore-former added in the powder mixture. The simulated MOX fuels went through thermal shock and the basic mechanical properties were measured, such as Vickers hardness, fracture strength, fracture toughness and thermal shock resistance. The properties were measured before and after the thermal shock by quenching in a water bath at room temperature. The critical thermal shock resistance (T c ) was determined, which corresponded with an abrupt decrease in fracture strength [83]. UO 2 pellet specimens (98.3% TD) were also prepared in a similar manner for the purpose of comparison. The fracture strength of the simulated MOX pellet specimens was relatively independent of the sintered density [82]. All the fracture strength values were between 118

PAGE 119

40 and 50 MPa for specimens with theoretical densities between 93.7% and 97.6% [82]. The fracture strength for a UO 2 pellet with a theoretical density of 98.3% measured for comparison was 76 MPa, which was higher than those of the simulated MOX pellet specimens. The critical thermal shock resistance (T c ) was determined by the abrupt decrease in fracture strength. It was discovered that T c was 120 C for the simulated MOX pellets without any pore-former and with a theoretical density of 97.6%. When the pore-former was used, and the pellet density was decreased, T c increased to 135 C. The UO 2 pellet with a density of 98.3% TD for comparison had a T c of 100 C. From these results, it is suggested that increasing the porosity can prohibit the propagation of the cracks, and thus increase the critical thermal shock resistance (T c ). Vickers hardness values for the simulated MOX pellet specimens with different densities subjected to thermal shock at different T were shown in Figure 7-1 [82]. It showed that after thermal shock by quenching at various temperatures, the Vickers hardness values did not vary significantly. Most of the pellets had Vickers hardness values of 600 to 700 kgf/mm 2 (6 to 7 GPa). Variation of the fracture toughness of the simulated MOX pellets with different densities subjected to thermal shock at different T (in C) has also been measured [82]. The fracture toughness values did not vary significantly either. Most of the pellets had fracture toughness values of 0.7 to 1 MPam -1/2 Neeft et al. [81] studied the mechanical behavior of some simulated IMF pellets before and after neutron irradiation. All fuel pellets contained about 2.5 volume % UO 2 with a 235 U enrichment of 20%. Uranium was used to simulate some of the effects of the 119

PAGE 120

actinides to be transmuted (e.g. plutonium or americium) since uranium can be handled more easily. The IMF candidate materials tested in the study included CeO 2 MgO, MgAl 2 O 4 Y 3 Al 5 O 12 and Y 2 O 3 The ceramography images showed that irradiated Y 2 O 3 and CeO 2 pellets have only a few cracks (Figure 7-2 A) [81]. These cracks can be attributed to thermal stress in the pellets. The irradiated MgAl 2 O 4 MgO and Y 3 Al 5 O 12 all showed a high density of cracks that run between the UO 2 inclusions (Figure 7-2 B) [81]. XRD studies showed that no chemical interaction occurred between UO 2 and the matrices MgO, MgAl 2 O 3 and Y 3 AL 5 O 12 after sintering. Adversely, some chemical interactions occurred between UO 2 and the Y 2 O 3 and CeO 2 matrices forming solid solutions. The Vickers hardness (H V ) of simulated IMF pellets before and after irradiation was measured at room temperature and is shown in Table 7-1[81]. Considering the distribution of the UO 2 microspheres and a fission product penetration depth of 10 m, it was calculated that most of the matrix is not implanted with fission product for MgO, MgAl 2 O 3 and Y 3 Al 5 O 12 The Vickers hardness indentation tests performed on these materials were therefore in matrices free of fission products. Chemical reactions distributed the UO 2 throughout the matrices of Y 2 O 3 and CeO 2 As a result, the indentation in Y 2 O 3 and CeO 2 were possibly polluted with fission product and pores as a result of fission. A minor increase in the porosity can cause a relatively large decrease of the hardness values [59]. The porosity change of these matrices during irradiation is a complex phenomenon. Therefore, no conclusion has been drawn from the hardness values. However, a decrease in hardness values also decreases the mechanical interaction between the pellets and the cladding decreases. This phenomenon has been 120

PAGE 121

observed in a comparison of the behavior of UO 2 and MOX fuel. The lower hardness values of MOX fuel causes the power at which a MOX fuel rod fails to be higher than that of a UO 2 rod [84]. The fracture toughness (K IC ) of the simulated IMF pellets before and after irradiation was measured using the indentation method [85-89] at room temperature. The results are shown in Table 7-2 [81]. The measured fracture toughness is influenced by the porosity, the microstructure and the Youngs modulus. An increase in the fracture toughness decreases the risk of extensive pellet fragmentation during normal operation or accident conditions [81]. The difference in the fracture toughness of the inert matrix materials is relatively small and the most favorable inert matrix material cannot be decided from these data. In another paper by Lee et al. [90], mechanical properties of simulated ZrO 2 based IMF pellets and simulated MOX fuel pellets were measured and compared. The composition of the ZrO 2 based fuel was: ZeO 2 -10%Y 2 O 3 -7%Er 2 O 3 -15%CeO 2 Two different densities of 88 % and 92 % TD were obtained. The simulated MOX fuel had a composition of UO 2 -5%CeO 2 The Vickers hardness values for IMF with 88 % and 92 % TD are 6.6 0.3 and 7.6 0.7 GPa, respectively. The Vickers hardness value for simulated MOX is 6.5 0.2 GPa. The fracture toughness values for the IMF pellets with 88% and 92% TD and for the UO 2 -5%CeO 2 pellet were 1.7 0.2, 1.5 0.2 and 0.8 0.1 MPam 1/2 respectively. The ZrO 2 based IMF showed higher fracture toughness than the simulated MOX fuel. 121

PAGE 122

7.2 Experimental Procedures 7.2.1 Vickers Hardness The sintered pellets were cut in half with a diamond saw, mounted in the epoxy and polished to a 1 m finish with a series of diamond polishing pastes. The Vickers hardness (H v in GPa) values were measured on the polished surface by a conventional Vickers indentation method and calculated with equation 7.1 21852dFHv (7.1) where F is the load in Newton and d is the average length of the diagonals of the Vickers indents in millimeter. A 1.5 kg load was used throughout the hardness measurements. Five indents were used to determine the average hardness for each sample. 7.2.2 Biaxial Fracture Strength and Fracture Toughness Before the strength and toughness tests, a 1 kg Vickers hardness indent was put on the center of each pellet. The specimens were loaded in a pin on three-ball fixture (Figure 7-3) using a test frame (Instron model 1350, Instron Corporation, Norwood, MA). The face with the indent was pointing down toward the three-ball support ring. The pin moved at a speed of 0.3 mm per minute until the specimen broke. The load at which the specimen broke was recorded. The biaxial fracture strength is calculated according to Equation 7.1 [91] )(432YXdPS (7.1) 22)(21)ln()1( C B C BX 122

PAGE 123

22)()1(])ln(1[)1( C A C AY where is the Poissons ratio, B is the radius of the loaded area, A is the radius of the support circle, C is the radius of the specimen, P is the load and d is the thickness of the specimen. The fracture toughness (K IC ) was calculated using the strength indentation method according to Equation 7.2 [92] 4331)(88.0PKC (7.2) where is the fracture strength and P is the indent load. Six pellets were used for each condition to determine the average fracture strength and fracture toughness. 7.3 Results and Discussion 7.3.1 Vickers Hardness The Vickers hardness values of SiC IMF pellets with different processing parameters are shown in Table 7-3. Optical microscopy pictures of some of the Vickers intents are shown in Figure 7-4. The indent size is about 60 to 110 microns. Because the samples have relatively low density, a very smooth mirror surface after polishing can not be achieved. And thus the intents are not very well defined. However, at least five indents per sample were measured to get a statistically reliable hardness value. When the samples contained coarse (16.9 m) SiC particles, the place to put the indent was chosen carefully in order to obtain a representative hardness value since the coarse particles have sizes close to the indent size. 123

PAGE 124

The Vickers hardness as a function of % theoretical density is shown in Figure 7-5. It can be seen that Vickers hardness strongly depends on the density. For samples with only 1 m SiC powder and the polymer precursor, the highest hardness achieved was 3.26 GPa (sample H7). Although mixing the coarse and fine SiC particles could achieve a higher pellet density, the hardness of the pellets did not increase significantly. However, one PIP cycle could effectively increase the hardness from 3.44 to 5.55 GPa. With the low temperature route (1050 C), the amorphous SiC phase bound the SiC particles and thus formed a pellet. With insufficient polymer precursor in the green body (sample H4), the SiC particles in the sintered pellets would be loosely bound and could be broken more easily. This resulted in a pellet with a low Vickers hardness. The hardness could be increased by either using more polymer precursor or by applying more pressing pressure. The former method can increase the binding by having more amorphous phase around the SiC particles, the latter method can achieve a better particle packing and thus requires less amorphous phase to bind the SiC particles. Even though higher pellet density could be achieved by mixing coarse and fine SiC particles, the particles were still bound by the amorphous SiC phase. The Vickers hardness reaching a plateau suggests that the hardness is controlled by the bonding strength between the amorphous phase and the SiC particles. The intergrainal amorphous phase is the weak spot and is the limiting factor for the hardness. The PIP cycle can further increase the density by filling the pores with more amorphous SiC matrix. This can strengthen the intergrainal amorphous phase and thus increase the Vickers hardness. 124

PAGE 125

Hardness of micro-porous SiC ceramic has been studied by Milman et al. [93]. and Slutsker et al. [94]. In their work, SiC ceramic was sintered at temperatures between 1900 C and 2200 C, moreover, sintering additives were used. Their hardness values together with hardness values from this study are shown in Figure 7-6. A strong dependant of hardness on the density is still observed for density range of 62 % to 99 %TD. It should be noted that in the work of Milman et al. and Slutsker et al. [93, 94], high sintering temperatures and sinter aids were used and the SiC particles became joined together during sintering while the volume decreased. This densification mechanism is quite different from the low temperature polymer precursor route of this study, where the SiC particles were bound by the amorphous SiC phase and only negligible bulk volume shrinkage occurred. It was observed that with the same theoretical density, samples from the work of Milman et al. [93] had hardness values much higher than this study. This suggests that the binding strength (between grains) of their samples, through conventional sintering, is stronger than the binding strength between the amorphous SiC phase and the crystalline SiC particles. A lowered hardness value is beneficial because the mechanical interactions between the pellets and the cladding decrease as the hardness of the fuel pellets decreases [84]. In summary, Vickers hardness values between 1.43 and 5.55 GPa were achieved with different processing parameters. The Vickers hardness strongly depended on the density of the pellets. Even though mixing the coarse and fine SiC particles could increase the density by better particle packing, the Vickers hardness didnt increase significantly because the binding nature between the amorphous SiC matrix phase and the crystalline SiC particles didnt change. However, with the PIP cycle, the amorphous 125

PAGE 126

SiC matrix phase was strengthened because the nanometer-sized pores in the matrix were replaced with the amorphous SiC phase from the PIP process. As a result, the Vickers hardness was significantly increased from 3.44 to 5.55 GPa. 7.3.2 Fracture Strength and Fracture Toughness The biaxial fracture strength of SiC IMF pellets with different processing methods and SiC particle composition is shown in Table 7-4. The biaxial fracture strength as a function of % theoretical density is shown in Figure 7-7. In general, the fracture strength increases with the increase of the % theoretical density. But comparing sample F3 (200 MPa) and F5 (600 MPa), it was found that even though the density increased with the higher pressing pressure, the fracture strength dropped from 57.0 to 33.0 MPa. This can be explained by the springback effect [66, 67] associated with higher pressing pressure which caused crack forming inside the pellets. This has been confirmed with the optical microscopy images of the polished cross section areas of the pellets (Figure 7-8). A large crack, perpendicular to the pressing direction, in the size of ten micron range is observed with sample F5 which was pressed at 600 MPa (Figure 7-8(c)). No cracks were observed with sample F1 (Figure 7-8(a)) or F3 (Figure 7-8(b)), which were pressed at 70 and 200 MPa, respectively. The decrease in fracture strength due to the springback effect observed in sample F5, which contained only 1 m SiC particles, did not happen in sample F9, which was also pressed at 600 MPa but contained both coarse and fine SiC particles. The springback effect in sample F9 resulted in smaller and more evenly distributed cracks as shown in Figure 5-8 (e) and (f). These smaller cracks did not decrease the fracture strength of the pellets. 126

PAGE 127

It should be noted that Equation 7.1 assumed the maximum tensile stress occurred at the center of the bottom surface of the specimen, therefore the specimen should break at the center of the bottom surface. However, in Sample F5, the sample broke from the crack during the test, instead of breaking form the center. Hence the fracture strength value calculated form Equation 7.1 for sample F5 was not the real fracture strength value. Nevertheless, it is still shown here for comparison. Since all the specimens had very similar dimensions, the calculated fracture strength was proportional to the fracture load, which expressed the strength of the specimens. By mixing the coarse and fine SiC particles, the density of the pellets increased. But by using the coarse particles, the largest pore detected by the mercury porosimeter also increased from ~100 nm for pellets without coarse SiC particles to 1 m to 500 nm for pellets with coarse particles. For sample F6, the largest pore was determined to be ~1 m (Figure 5-6 (a)). On the other hand, the largest pore for sample F4 was determined to be ~100 nm (Figure 4-5). Even though sample F6 had larger pores; the fracture strength didnt decrease, but still increased slightly with density. Apparently the largest pores, detected by the mercury porosimeter, did not function as a critical flaw in the fracture strength tests. The large pores did not assist the growth of the crack either. The fracture strength is more related to the density (the total pore volume) of the pellets than the pore size distribution. It is more evident with sample F7 and F9, of which the largest pore is ~400 nm (Figure 5-6 (c)). The fracture strength increased strongly because of the higher density despite the presence of the larger pores. 127

PAGE 128

The PIP process not only increased the density, but also effectively increased the fracture strength. After only one PIP cycle, a ~50% increase in fracture strength was observed with a ~4.8% TD increase (sample F8 and F10). As mentioned in Chapter 7.1, the simulated MOX fuel has fracture strength values of around 40 to 50 MPa [82], and a UO 2 pellet with a density of 98.3% TD has a fracture strength of 76 MPa [82]. The ZrO 2 based IMF has fracture strength values between 40 and 100 MPa [90]. The SiC IMF pellets in this study have fracture strength values as high as 128.0 MPa without any PIP cycles. With one PIP cycle, the fracture strength increased to as high as 201.0 MPa. Higher fracture strength is preferred because the chances of breaking the pellets during handling are reduced. It should be noted that the 1 kg Vickers indent was purposely put on the pellets in order to calculate the fracture toughness from fracture strength using Equation 7.2. The presence of the Vickers indent can potentially reduce the measured fracture strength since the indent is not taken into account in the strength calculation in Equation 7.1. However, the actual fracture strength without the Vickers indent would not be smaller than the measured strength with the indent. The fracture toughness values of SiC IMF pellets with different processing methods and SiC particle composition are shown in Table 7-5. The fracture toughness as a function of the % theoretical density is shown in Figure 7-9. Since the fracture toughness values were calculated from the fracture strength and the indent load (which was fixed at 1 kg throughout this study) according to Equation 7.2, the shape of fracture strength (Figure 7-7) and toughness (Figure 7-9) as a function of 128

PAGE 129

the % TD is very similar. The fracture toughness of the pellets also increased with the % TD. One exception is sample F5 and this can be explained with the springback effect. The PIP process can increase the fracture toughness effectively. After only one PIP cycle, a ~40% increase in fracture toughness was observed with a ~4.8% TD increase (sample F8 and F10). The highest fracture toughness values achieved before and after 1 PIP cycle were 1.88 and 2.63 MPam 1/2 respectively. It should be noted that the validity of using Equation 7.2 to determine fracture toughness lies in the assumption that the critical flow grows from the Vickers indent that was put on intentionally. In the case of sample F5 in Table 7-5, the sample broke from the big crack created by the springback effect, instead of from the indent; thus the fracture toughness determined from Equation 7.2 is invalid. However, the calculated value is still given for comparison. The fracture toughness values of the simulated MOX pellet specimens measured by Lee et al. [82] were 0.7 to 1 MPam 1/2 The fracture toughness values of the simulated ZrO 2 based IMF pellets, measured by Lee et al. [90], were 1.5 to 1.7 MPam 1/2 The fracture toughness values for some IMF candidate materials can be found in Table 7-2 [81]. The values were between 1.2 to 2.5 M Pam 1/2 Comparing these literature values with the fracture toughness values of the SiC IMF pellets fabricated in this study, it is clear that the maximum achievable fracture toughness in this study was higher than the simulated MOX pellets. The pellets with the highest fracture toughness in this study also had higher fracture toughness values than most of the IMF candidate materials. A higher fracture toughness value is preferred because it reduces the chance of pellet fragmentation during operation and accident conditions. 129

PAGE 130

7.4 Summary and Conclusion In this chapter, the Vickers hardness, fracture strength and fracture toughness of the fabricated SiC IMF pellets were investigated. The Vickers hardness strongly depended on the bonding strength between the amorphous SiC matrix and the crystalline SiC particles. Vickers hardness values of 1.30 to 5.55 GPa were achieved with different processing methods. The fracture strength and fracture toughness were found to be more related to the % theoretical density rather than the pore size distributions. It was also found that the PIP process was an effective way to increase the fracture strength and toughness of the pellets. The highest fracture strength and fracture toughness achieved after 1 PIP cycle were 201.0 MPa and 2.63 MPam -1/2 respectively. These values are much higher than the simulated MOX fuels [82] and are also higher than most of the IMF candidate materials [81, 90]. 130

PAGE 131

Table 7-1. Results of the Vickers hardness measurements and the literature data [81] H V (GPa) Measured Matrix Unirradiated Irradiated Literature unirradiated CeO 2 5.32 1.55 # 8.83 1.71 # MgO 10.86 1.02 # 9.98 0.37 7.60 [95] MgAl 2 O 4 16.15 1.13 16.63 1.75 15.00 [96] Y 3 Al 5 O 12 14.46 0.67 10.14 0.60 18.35 0.51 [97] Y 2 O 3 8.43 0.26 10.66 0.94 # 9.75 0.67 739 [95] 612[97] Indentation mass is: # 100 g, 200g. Table 7-2. Results of the fracture toughness measurements and literature data [81] (K IC (MPam 1/2 ) Measured Matrix Unirradiated Irradiated Literature unirradiated CeO 2 1.2 0.2 # 1.5 0.4 # MgO 1.8 0.3 # 2.0 0.5 1.6 1.8 [98] MgAl 2 O 4 2.5 0.4 2.0 0.2 3.0 [98] 1.94 0.10 [99] Y 3 Al 5 O 12 1.2 0.2 1.9 0.5 1.7 0.1 [97] Y 2 O 3 1.7 0.2 2.1 0.4 # 2.1 0.3 Indentation mass is: # 100 g, 200g. 131

PAGE 132

Table 7-3. Vickers hardness values of SiC IMF pellets Sample ID SiC particle composition SMP10 weight % Pressing pressure (MPa) % TD Hardness (GPa) Standard deviation H1 1 m 10.0 70 62.4 1.43 0.40 H2 1 m 15.0 70 67.9 2.37 0.60 H3 1 m 17.5 70 71.1 3.17 0.18 H4 1 m 5.0 200 62.1 1.30 0.10 H5 1 m 10.0 200 66.4 2.22 0.25 H6 1 m 12.5 200 68.3 2.82 0.73 H7 1 m 15.0 200 72.0 3.26 0.24 H8 1 m 10.0 600 68.8 3.21 0.42 H9 0.6/16.9 m* 10.0 200 80.1 3.44 0.27 H10 0.6/16.9 m* 10.0 600 81.2 3.47 0.38 H11 H9 + 1 PIP 10.0 200 86.0 5.55 0.93 All samples contain 5 weight % CeO 2 *: SiC particle composition 40% 0.6 m, 60% 16.9 m. H1H8 samples were mixed with mortar and pestle, H9 and H10 samples were mixed with SPEX high energy mill with hexane as a grinding aid. H11 sample had same processing parameters as F9 but went through 1 PIP cycle with 1 atmosphere pressure as described in section 6.2.1 132

PAGE 133

Table 7-4. Fracture strength of SiC IMF pellets Sample ID SiC composition SMP10 weight % Pressing pressure (MPa) Mixing method** % TD Fracture strength (MPa) Standard deviation F1 1 m 10 70 mortar 62.4 55.1 20.4 F2 1 m 5 200 mortar 62.1 22.7 7.6 F3 1 m 10 200 mortar 66.4 57.0 13.3 F4 1 m 15 200 mortar 72.0 72.5 25.6 F5 1 m 10 600 mortar 68.6 33.0 19.9 F6 0.6/16.9 m* 10 200 mortar 74.3 75.2 25.2 F7 0.6/16.9 m* 10 200 SPEX 80.2 109.9 13.8 F8 F7 + PIP 10 200 SPEX 85.1 167.3 18.4 F9 0.6/16.9 m* 10 600 SPEX 81.2 128.0 7.2 F10 F9 + PIP 10 600 SPEX 86.1 201.0 8.1 A ll samples contain 5 weight % CeO 2 SiC particle composition 40% 0.6 m, 60% 16.9 m ** morta r : mortar and pestle, SPEX: SPEX high energy mill with hexane as a grindingaid F8 samples had same processing parameters as F7 but went through 1 PIP cyclewith 1atmosphere pressure as described in section 6.2.1 F10 samples had same processing parameters as F9 but went through 1 PIP cyclewith1 atmosphere pressure as described in section 6.2.1 133

PAGE 134

Table 7-5. Fracture toughness of SiC IMF pellets ID SiC composition SMP10 weight % Pressing pressure (MPa) Mixing method** % TD Fracture toughness (MPam 1/2) Standard deviation F1 1 m 10 70 mortar 62.4 0.98 0.28 F2 1 m 5 200 mortar 62.1 0.51 0.15 F3 1 m 10 200 mortar 66.4 0.96 0.18 F4 1 m 15 200 mortar 72.0 1.21 0.33 F5 1 m 10 600 mortar 68.6 0.66 0.31 F6 0.6/16.9 m* 10 200 mortar 74.3 1.24 0.35 F7 0.6/16.9 m* 10 200 SPEX 80.2 1.67 0.16 F8 F7 + PIP 10 200 SPEX 85.1 2.29 0.19 F9 0.6/16.9 m* 10 600 SPEX 81.2 1.88 0.08 F10 F9 + PIP 10 600 SPEX 86.1 2.63 0.08 All samples contain 5 weight % CeO 2 SiC particle composition 40% 0.6 m, 60% 16.9 m ** mortar: mortar and pestle, SPEX: SPEX high energy mill with hexane as a grinding aid F8 samples had same processing parameters as F7 but went through 1 PIP cycle with 1 atmosphere pressure as described in section 6.2.1 F10 samples had same processing parameters as F9 but went through 1 PIP cycle with 1 atmosphere pressure as described in section 6.2.1 134

PAGE 135

Figure 7-1. Vickers hardness values for the simulated MOX pellet specimens with different densities subjected to thermal shock at different T (in C) [82] Figure 7-2. Irradiated pellets of (A) Y 2 O 3 + UO 2 and (B) MgAl 2 O 4 + UO 2 (arrow indicates the axial direction of pellet) [81] 135

PAGE 136

Figure 7-3. The schematic of the pin on three ball set up for measuring biaxial fracture strength and fracture toughness (Courtesy of Nadia Rohbeck) 136

PAGE 137

Figure 7-4. Optical microscopy pictures of the Vickers intents from (a): sample H8, (b): sample H9 and (c): sample H10 in Table 7-3 Figure 7-5. Vickers hardness as a function of % theoretical density 137

PAGE 138

Figure 7-6. Vickers hardness values of porous SiC specimens as a function of % theoretical density. Figure 7-7. Biaxial fracture strength as a function of % theoretical density 138

PAGE 139

Figure 7-8. Optical microscopy images of polished pellets with 1 m SiC particles, with 10 weight% polymer precursor, and were pressed at (a) 70 MPa (sample F1), (b) 200 MPa (sample F3), and (c) 600 MPa (sample F5) 139

PAGE 140

Figure 7-9. Fracture toughness as a function of % theoretical density 140

PAGE 141

CHAPTER 8 SUMMARY AND FUTURE WORK 8.1 Summary Silicon Carbide (SiC) based inert matrix fuel (IMF) was successfully fabricated at a low temperature of 1050 C through the polymer precursor route. After one polymer infiltration and pyrolysis cycle, a pellet with a theoretical density of 86%, a hardness of 5.55 GPa, a fracture strength of 201.0 MPa and a fracture toughness of 1.88 MPam 1/2 was achieved. The processing steps of the polymer precursor route included mixing, pressing and sintering. Initially, the -SiC crystalline powder was mixed with the liquid polymer precursor to form the slurry. The slurry was then uniaxially cold pressed to form the green body, which was sintered in a tube furnace to a temperature of 1050 C to form the pellet. The polymer precursor worked as a binder to bind the SiC particles together in the green body and decomposed to amorphous SiC. The amorphous SiC phase worked as a matrix to bind the SiC particles together to form a pellet. The effect of polymer content and cold pressing pressure on the sintered density of the pellets was firstly investigated. It was found that the sintered density increased with the increase of the polymer content in the green body, but reached a plateau. If an excess amount of polymer was used, the slurry would become too wet and the excess polymer would be pressed out of the die during pressing. Increasing the pressing pressure can achieve a better packing of the SiC particles and thus a higher density. When only the 1 m nominal size SiC powder and the polymer precursor were used, the maximum achievable theoretical density was about 72%. 141

PAGE 142

The pore size analysis showed a bimodal distribution for most of the pellets. The large pores were related to the empty space between SiC particles and the small pores were related to the decomposition of the polymer precursor. It was also found that the largest pore size was related to the pressing pressure. The utilization of mixing coarse and fine SiC particles to further increase the density was investigated. Different coarse and fine SiC powder volume ratios were tested and the optimal volume ratio was determined to be 60% coarse and 40% fine SiC particles. Different mixing methods were then investigated to study the effect of mixing. It was determined that the SPEX high energy shaker mill with the hexane as a grinding aid was the best mixing method to achieve a homogeneous and high density pellet. A sintered theoretical density of 81.2% was achieved with a pressing pressure of 600 MPa. The pore size investigation also showed that most of the large pores could be eliminated because of the better packing of the well mixed coarse and fine SiC particles. The microstructure examination showed cracks perpendicular to the pressing direction with all the 600 MPa pressed pellets. This can be explained by the springback effect. The polymer infiltration and pyrolysis process was then applied on the pellets to further increase the density and close the open pores. It was found that the first PIP cycle could increase the theoretical density of the pellets from ~81% to ~86%. Most of the open pores were closed after the first PIP cycle except those with a diameter smaller than 10 nm. The successive second PIP cycle was not as effective on either the density or the pore size distribution of the pellets because of the small pore size distribution before the second PIP cycle. 142

PAGE 143

The mechanical properties, which include the Vickers hardness, the fracture strength and the facture toughness, of the fabricated pellets were determined and correlated with the processing parameters. The Vickers hardness is strongly dependant on the density of the pellets and the binding strength between the amorphous SiC matrix and the SiC particles. Mixing the coarse and fine SiC particles can get a better packing thus a higher density. But the binding nature didnt change. Thus the Vickers hardness didnt change significantly with the increase in density. The PIP process can further increase the binding strength of the amorphous SiC matrix phase by filling the pores in the matrix with more amorphous SiC phase. Thus the Vickers hardness increased significantly. After the first PIP cycle, a Vickers hardness value of 5.55 GPa was achieved. The fracture strength and toughness were also strongly dependant on the density of the pellets. One exception was the springback effect which caused the decrease in strength and toughness. The PIP process can also significantly increase the fracture strength and toughness. Pellets with a fracture strength of 201.0 MPa, and a fracture toughness of 2.63 MPam 1/2 was achieved after one PIP cycle. The fabricated SiC pellets had better mechanical properties than the reference MOX and other IMF candidate materials. 8.2 Future Work There are several aspects of this research that need to be further investigated in the future in order to verify the validity of using SiC as an inert matrix material. The thermal conductivity of the fabricated SiC IMF as a function of processing parameters needs to be determined. The chemical reactions between plutonium and 143

PAGE 144

SiC need to be studied and the in-reactor behavior of the fabricated SiC IMF pellets needs to be studied. The fission gas release issues need to be further investigated since the fabricated pellets in this study are not 100 % dense and have open pores. The closed pores associated with this process can potentially work as reservoirs for the fission gases and need to be studied. The pores can potentially increase the thermal shock resistance and should be investigated. 144

PAGE 145

LIST OF REFERENCES [1] Y. W. Lee, C. Y. Joung, S. H. Kim and S. C. Lee, Met. Mater.-Korea 7 (2001) 159. [2] I. E. Agency, Nuclear Power Plants Information. Number of Reactors Operation Worldwide (2007). [3] L. R. Willett, Transactions of the American Nuclera Society 69 (1993) 88. [4] H. F. Frank von Hippel, Transactions of the American Nuclera Society 69 (1993) 89. [5] J. M. Paratte and R. Chawla, Annals of Nuclear Energy 22 (1995) 471. [6] B. Pellaud, Comptes Rendus Physique 3 (2002) 1067. [7] W. J. Weber, R. C. Ewing, C. R. A. Catlow, T. D. de la Rubia, L. W. Hobbs, C. Kinoshita, H. Matzke, A. T. Motta, M. Nastasi, E. K. H. Salje, E. R. Vance and S. J. Zinkle, Journal of Materials Research 13 (1998) 1434. [8] L. J. Ott, T. Tverberg and E. Sartori, Annals of Nuclear Energy 36 (2009) 375. [9] K. Kurosaki, K. Yamada, M. Uno, S. Yamanaka, K. Yamamoto and T. Namekawa, Journal of Nuclear Materials 294 (2001) 160. [10] Y. Philipponneau, Journal of Nuclear Materials 188 (1992) 194. [11] M. Bober, C. Sari and Schumach.G, Journal of Nuclear Materials 39 (1971) 265. [12] N. Chauvin, R. J. M. Konings and H. Matzke, Journal of Nuclear Materials 274 (1999) 105. [13] C. Degueldre and J. M. Paratte, Journal of Nuclear Materials 274 (1999) 1. [14] R. J. M. Konings, K. Bakker, J. G. Boshoven, H. Hein, M. E. Huntelaar and R. R. van der Laan, Journal of Nuclear Materials 274 (1999) 84. [15] Y. W. Lee, H. S. Kim, S. H. Kim, C. Y. Joung, S. H. Na, G. Ledergerber, P. Heimgartner, M. Pouchon and M. Burghartz, Journal of Nuclear Materials 274 (1999) 7. [16] M. A. Pouchon, M. Dobeli, C. Degueldre and M. Burghartz, Journal of Nuclear Materials 274 (1999) 61. [17] J. J. Carbajo, G. L. Yoder, S. G. Popov and V. K. Ivanov, Journal of Nuclear Materials 299 (2001) 181. 145

PAGE 146

[18] G. Ledergerber, C. Degueldre, P. Heimgartner, M. A. Pouchon and U. Kasemeyer, Progress in Nuclear Energy 38 (2001) 301. [19] I. A. E. Agency, Viability of Inert Matrix Fuel in Reducing Plutonium Amounts in Reactors (2007). [20] H. Matzke, V. V. Rondinella and T. Wiss, Journal of Nuclear Materials 274 (1999) 47. [21] R. A. Verrall, M. D. Vlajic and V. D. Krstic, Journal of Nuclear Materials 274 (1999) 54. [22] A. K. Misra, Journal of the American Ceramic Society 74 (1991) 345. [23] M. A. Mulla and V. D. Krstic, American Ceramic Society Bulletin 70 (1991) 439. [24] A. Nakahira and K. Niihara, Nippon Seramikkusu Kyokai Gakujutsu Ronbunshi-Journal of the Ceramic Society of Japan 100 (1992) 448. [25] M. A. Mulla and V. D. Krstic, Journal of Materials Science 29 (1994) 934. [26] K. H. Sarma, J. Fourcade, S. G. Lee and A. A. Solomon, Journal of Nuclear Materials 352 (2006) 324. [27] Z. Pan, H. J. Seifert and O. Fabrichnaya, in 214th ECS Meeting, edited by E. O. E. Wuchina, J. Fergus, T. Maruyama, D. Shifler (ECS Trans., Honolulu, HI, USA, 2008) p. 65. [28] E. Bouillon, F. Langlais, R. Pailler, R. Naslain, F. Cruege, P. V. Huong, J. C. Sarthou, A. Delpuech, C. Laffon, P. Lagarde, M. Monthioux and A. Oberlin, Journal of Materials Science 26 (1991) 1333. [29] N. R. Dando, A. J. Perrotta, C. Strohmann, R. M. Stewart and D. Seyferth, Chemistry of Materials 5 (1993) 1624. [30] H. Q. Ly, R. Taylor, R. J. Day and F. Heatley, Journal of Materials Science 36 (2001) 4037. [31] Q. M. Cheng, L. V. Interrante, M. Lienhard, Q. H. Shen and Z. Z. Wu, Journal of the European Ceramic Society 25 (2005) 233. [32] K. Moraes, J. Vosburg, D. Wark, L. V. Interrante, A. R. Puerta, L. G. Sneddon and M. Narisawa, Chemistry of Materials 16 (2004) 125. [33] H. J. Wu and L. V. Interrante, Abstracts of Papers of the American Chemical Society 204 (1992) 61. 146

PAGE 147

[34] D. M. Narsavage, L. V. Interrante, P. S. Marchetti and G. E. Maciel, Chemistry of Materials 3 (1991) 721. [35] H. J. Wu and L. V. Interrante, Abstracts of Papers of the American Chemical Society 198 (1989) 149. [36] C. Whitmarsh and L. V. Interrante, Abstracts of Papers of the American Chemical Society 198 (1989) 292. [37] L. V. Interrante, L. E. Carpenter, C. Whitmarsh and M. Hackney, Journal of the Electrochemical Society 133 (1986) C119. [38] H. S. Kim, C. Y. Joung, B. H. Lee, J. Y. Oh, Y. H. Koo and P. Heimgartner, Journal of Nuclear Materials 378 (2008) 98. [39] M. Stan, T. J. Armstrong, D. P. Butt, T. C. Wallace, Y. S. Park, C. L. Haertling, T. Hartmann and R. J. Hanrahan, Journal of the American Ceramic Society 85 (2002) 2811. [40] NEI, Resources and Stats. Retrieved from U.S. Nuclear Power Plants (2008). [41] H. Akie, T. Muromura, H. Takano and S. Matsuura, Nuclear Technology 107 (1994) 182. [42] U. K. C. Degueldre, F. Botta, G. Ledergerber, in 1995 Materials Research Society Fall meeting, edited by D. A. K. William M. Murphy (Pittsburgh, PA, 1995). [43] W. J. Carmack, M. Todosow, M. K. Meyer and K. O. Pasamehmetoglu, Journal of Nuclear Materials 352 (2006) 276. [44] H. Kleykamp, Journal of Nuclear Materials 275 (1999) 1. [45] C. Lombardi and A. Mazzola, Annals of Nuclear Energy 23 (1996) 1117. [46] M. Burghartz, H. Matzke, C. Leger, G. Vambenepe and M. Rome, Journal of Alloys and Compounds 271 (1998) 544. [47] V. F. Sears, Neutron News 3 (1992) 29. [48] L. M. E. Goldberg Yu., Rumyantsev S.L, Properties of Advanced SemiconductorMaterials GaN, AlN, SiC, BN, SiC, SiGe (John Wiley & Sons, Inc, 2001). [49] R. J. Price, Journal of Nuclear Materials 46 (1973) 268. 147

PAGE 148

[50] Y. W. Lee, S. C. Lee, H. S. Kim, C. Y. Joung and C. Degueldre, Journal of Nuclear Materials 319 (2003) 15. [51] M. D. V. R.A. Verrall, V.D. Krstic, in 5th International Conference on CANDU Fuel (Toronto, Canada, 1997). [52] H. Hirayama, T. Kawakubo, A. Goto and T. Kaneko, Journal of the American Ceramic Society 72 (1989) 2049. [53] W. J. Kim, H. S. Hwang, J. Y. Park and W. S. Ryu, Journal of Materials Science Letters 22 (2003) 581. [54] H. Gu, Intergranular and Interphase Boundaries in Materials, Iib98 294-2 (1999) 301. [55] W. Dienst, P. Hofmann and D. K. Kerwinpeck, Nuclear Technology 65 (1984) 109. [56] H. S. Kuchibhotla, Enhanced thermal conductivity oxide fuels: compatibility and novel fabrication techniques using BeO, M.S. Thesis (Purdue University, 2005). [57] W. Lippmann, J. Knorr, R. Noring and M. Umbreit, Nuclear Engineering and Design 205 (2001) 13. [58] H. Suzuki, T. Iseki and T. Imanaka, Journal of Nuclear Science and Technology 14 (1977) 438. [59] H. Matzke and J. Spino, Journal of Nuclear Materials 248 (1997) 170. [60] S. K. Bhaumik, C. Divakar, S. U. Devi and A. K. Singh, Journal of Materials Research 14 (1999) 906. [61] S. G. Lee, J. Fourcade, R. Latta and A. A. Solomon, Fusion Engineering and Design 83 (2008) 713. [62] E. Kockrick, R. Frind, M. Rose, U. Petasch, W. Bohlmann, D. Geiger, M. Herrmann and S. Kaskel, Journal of Materials Chemistry 19 (2009) 1543. [63] R. Sreeja, B. Swaminathan, A. Painuly, T. V. Sebastian and S. Packirisamy, Materials Science and Engineering B-Advanced Functional Solid-State Materials 168 204. [64] C. Song, P. Wang and H. A. Makse, Nature 453 (2008) 629. [65] R. K. McGeary, Journal of the American Ceramic Society 44 (1961) 513. [66] T. Rabe, J. Goebbels and A. Kunzmann, Industrial Ceramics 20 (2000) 193. 148

PAGE 149

[67] J. J. Lannutti, T. A. Deis, C. M. Kong and D. H. Phillips, American Ceramic Society Bulletin 76 (1997) 53. [68] M. A. Nechanicky, K. W. Chew, A. Sellinger and R. M. Laine, Journal of the European Ceramic Society 20 (2000) 441. [69] Y. Katoh, M. Kotani, H. Kishimoto, W. Yang and A. Kohyama, Journal of Nuclear Materials 289 (2001) 42. [70] R. Naslain, Composites Science and Technology 64 (2004) 155. [71] K. W. Chew, A. Sellinger and R. M. Laine, Journal of the American Ceramic Society 82 (1999) 857. [72] M. Itoh, K. Iwata, M. Kobayashi, R. Takeuchi and T. Kabeya, Macromolecules 31 (1998) 5609. [73] A. R. Maddocks, D. J. Cassidy, A. S. Jones and A. T. Harris, Materials Chemistry and Physics 113 (2009) 861. [74 ]R. M. Laine and F. Babonneau, Chemistry of Materials 5 (1993) 260. [75] A. K. Singh, S. C. Zunjarrao and R. P. Singh, Journal of Nuclear Materials 378 (2008) 238. [76] O. Kordina, C. Hallin, A. Henry, J. P. Bergman, I. Ivanov, A. Ellison, N. T. Son and E. Janzen, Physica Status Solidi B-Basic Research 202 (1997) 321. [77] T. Kimoto, A. Itoh and H. Matsunami, Physica Status Solidi B-Basic Solid State Physics 202 (1997) 247. [78] J. A. Powell, J. B. Petit, J. H. Edgar, I. G. Jenkins, L. G. Matus, J. W. Yang, P. Pirouz, W. J. Choyke, L. Clemen and M. Yoganathan, Applied Physics Letters 59 (1991) 333. [79] C. Droillard and J. Lamon, Journal of the American Ceramic Society 79 (1996) 849. [80] K. Sugiyama and E. Yamamoto, Journal of Materials Science 24 (1989) 3756. [81] E. A. C. Neeft, K. Bakker, R. L. Belvroy, W. J. Tams, R. P. C. Schram, R. Conrad and A. van Veen, Journal of Nuclear Materials 317 (2003) 217. [82] Y. W. Lee, S. C. Lee, C. Y. Joung, H. S. Kim and H. L. Lee, Advanced Engineering Materials 4 (2002) 584. [83] M. Oguma, Journal of Nuclear Materials 127 (1985) 67. 149

PAGE 150

[84] W. Goll, H. P. Fuchs, R. Manzel and F. U. Schlemmer, Nuclear Technology 102 (1993) 29. [85] A. G. Evans and E. A. Charles, Journal of the American Ceramic Society 59 (1976) 371. [86] G. R. Anstis, P. Chantikul, B. R. Lawn and D. B. Marshall, Journal of the American Ceramic Society 64 (1981) 533. [87] C. B. Ponton and R. D. Rawlings, Materials Science and Technology 5 (1989) 865. [88] C. B. Ponton and R. D. Rawlings, Materials Science and Technology 5 (1989) 961. [89] K. M. Liang, G. Orange and G. Fantozzi, Journal of Materials Science 25 (1990) 207. [90] Y. W. Lee, H. S. Kim, S. H. Kim, C. Y. Joung, S. C. Lee, S. H. Na, P. Heimgartner and G. Ledergerber, Progress in Nuclear Energy 38 (2001) 231. [91] J. B. Wachtman, W. Capps and J. Mandel, Journal of Materials 7 (1972) 188. [92] P. Chantikul, G. R. Anstis, B. R. Lawn and D. B. Marshall, Journal of the American Ceramic Society 64 (1981) 539. [93] Y. V. Milman, S. I. Chugunova, I. V. Goncharova, T. Chudoba, W. Lojkowski and W. Gooch, International Journal of Refractory Metals & Hard Materials 17 (1999) 361. [94] A. I. Slutsker, A. B. Sinani, V. I. Betekhtin, A. G. Kadomtsev and S. S. Ordan'yan, Technical Physics 53 (2008) 1591. [95] G. V. Samsonov, The Oxide Handbook (Biomedical Information Guides) (Springer, 1982). [96] R. Morrell, Handbook of Properies of Techinical and Engineering Ceramics 1985). [97] G. Dewith and J. E. D. Parren, Solid State Ionics 16 (1985) 87. [98] S. W. F. R.G. Munro, and T.L. Baker, Fracture Toughness Data for Brittle Materials, (National Institute of Standards and Technology, 1998). [99] R. L. Stewart and R. C. Bradt, Journal of the American Ceramic Society 63 (1980) 619. 150

PAGE 151

BIOGRAPHICAL SKETCH Chunghao Shih was born in 1981 in Taipei, Taiwan. After studying really hard for three years in junior high school, he got into the famous Chien-Kuo Senior High School in Taipei. He graduated from Chien-Kuo Senior High School in 1999. After senior high school, he entered National Taiwan University in September, 1999, majoring in chemical engineering. During the 4 years he spent in National Taiwan University, he was an active member of the Mountain Climbing Club. He learned how to survive in the wild mountains in Taiwan. He also became an accomplished rock climber. After graduation from National Taiwan University with a B.S. in chemical engineering, he joined Shei-Pa Nation Park in Taiwan as an alternative military serviceman between 2004 and 2005. He spent a lot of time patrolling the remote mountain trails and hanging out in the aboriginal tribes. After finishing the alternative military service, he joined the Lee Chang Yung Chemical Industry Corp. in Taipei as a research and development engineer. He spent a month working in the chemical engineering plant in order to learn the product line of the electronic grad isopropyl alcohol and acetone. The rest of his time in the company he was in the office in Taipei working on the research projects for optoelectronic display and wafer reclaim industry. He eventually went to University of Florida as a graduate student in August of 2006. He met Dr. Ronald, H. Baney in January of 2007 and joined his research group. He firstly ran the nanofluids project and then switched to the silicon carbide inert matrix fuel project in early 2009. Other than school work, he spent lots of time exploring the mother nature of the southeast part of the United States. He went on a lot of climbing, 151

PAGE 152

caving, canoeing and camping trips through the years. He received his Doctor of Philosophy degree in December 2010. 152