<%BANNER%>

UO2-SIC Composite Reactor Fuels with Enhanced Thermal and Mechanical Poperties Prepared by Spark Plasma Sintering

MISSING IMAGE

Material Information

Title:
UO2-SIC Composite Reactor Fuels with Enhanced Thermal and Mechanical Poperties Prepared by Spark Plasma Sintering
Physical Description:
1 online resource (143 p.)
Language:
english
Creator:
Yeo, Sunghwan
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Materials Science and Engineering
Committee Chair:
Baney, Ronald Howard
Committee Members:
Yang, Yong
Subhash, Ghatu
Tulenko, James S
Bowers, Clifford Russell

Subjects

Subjects / Keywords:
compositenuclearfuel -- sparkplasmasintering -- thermalconductivity
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:
The primary ceramic fuel used in nuclear reactors, Uranium Dioxide (UO2), has a low thermal conductivity which results in a decrease in both the energy output and the safety of a reactor. The introduction of high thermal conductivity fuel pellets enables a nuclear reactor to produce more thermal energy while maintaining plant safety due to lower pellet centerline temperature and thermal gradient, resulting in a lower level of fission gas release and thermal cracking.  The main objective of this research is to increase the thermal conductivity of UO2 by the incorporation of Silicon Carbide (SiC) particles or whiskers. Low-temperature oxidative sintering and Spark Plasma Sintering (SPS)techniques have been used to produce UO2-SiC composite pellets.While oxidative sintering failed to achieve enhanced thermal conductivity, the SPS sintered pellet obtained promising features such as higher density, better interfacial contact, and reduced chemical reaction.  The influence of SiC particle size (0.6-55µm) and volume fraction (5-20%) on the thermal conductivity of spark plasma sintered UO2-SiC composites was then investigated. The composites containing larger volume of SiC particles which size is less than 16.9µm showed higher thermal conductivity. The measured thermal conductivity was in an excellent agreement with calculated effective thermal conductivity based on a theoretical model.  Mechanical properties such as hardness and Young's modulus, and the internal stress of SiC particles of UO2-SiC composites containing different sizes and volume fractions of SiC particles were examined.Hardness and Young's modulus were increases up to 53% and 18% with 20vol% of 1µm SiC addition, respectively. The compression stress of SiC particles up to 1.65GPa was successfully measured using Raman spectroscopy.  Lastly, the thermal aging was performed at 1000C and 1500C for 12hours on fabricated UO2-SiC composites to investigate the feasibility of the composites in an operational condition of nuclear reactor. Microstructural defects such as microcracking and interfacial debonding, and SiC particles and whiskers removing from the polished surface were observed after 1500C thermal aging probably due to the chemical reaction between UO2 and SiC. Therefore, the thermal conductivity of composites was decreased by 5-10% after the aging.
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 Sunghwan Yeo.
Thesis:
Thesis (Ph.D.)--University of Florida, 2013.
Local:
Adviser: Baney, Ronald Howard.

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2013
System ID:
UFE0045786:00001

MISSING IMAGE

Material Information

Title:
UO2-SIC Composite Reactor Fuels with Enhanced Thermal and Mechanical Poperties Prepared by Spark Plasma Sintering
Physical Description:
1 online resource (143 p.)
Language:
english
Creator:
Yeo, Sunghwan
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Materials Science and Engineering
Committee Chair:
Baney, Ronald Howard
Committee Members:
Yang, Yong
Subhash, Ghatu
Tulenko, James S
Bowers, Clifford Russell

Subjects

Subjects / Keywords:
compositenuclearfuel -- sparkplasmasintering -- thermalconductivity
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:
The primary ceramic fuel used in nuclear reactors, Uranium Dioxide (UO2), has a low thermal conductivity which results in a decrease in both the energy output and the safety of a reactor. The introduction of high thermal conductivity fuel pellets enables a nuclear reactor to produce more thermal energy while maintaining plant safety due to lower pellet centerline temperature and thermal gradient, resulting in a lower level of fission gas release and thermal cracking.  The main objective of this research is to increase the thermal conductivity of UO2 by the incorporation of Silicon Carbide (SiC) particles or whiskers. Low-temperature oxidative sintering and Spark Plasma Sintering (SPS)techniques have been used to produce UO2-SiC composite pellets.While oxidative sintering failed to achieve enhanced thermal conductivity, the SPS sintered pellet obtained promising features such as higher density, better interfacial contact, and reduced chemical reaction.  The influence of SiC particle size (0.6-55µm) and volume fraction (5-20%) on the thermal conductivity of spark plasma sintered UO2-SiC composites was then investigated. The composites containing larger volume of SiC particles which size is less than 16.9µm showed higher thermal conductivity. The measured thermal conductivity was in an excellent agreement with calculated effective thermal conductivity based on a theoretical model.  Mechanical properties such as hardness and Young's modulus, and the internal stress of SiC particles of UO2-SiC composites containing different sizes and volume fractions of SiC particles were examined.Hardness and Young's modulus were increases up to 53% and 18% with 20vol% of 1µm SiC addition, respectively. The compression stress of SiC particles up to 1.65GPa was successfully measured using Raman spectroscopy.  Lastly, the thermal aging was performed at 1000C and 1500C for 12hours on fabricated UO2-SiC composites to investigate the feasibility of the composites in an operational condition of nuclear reactor. Microstructural defects such as microcracking and interfacial debonding, and SiC particles and whiskers removing from the polished surface were observed after 1500C thermal aging probably due to the chemical reaction between UO2 and SiC. Therefore, the thermal conductivity of composites was decreased by 5-10% after the aging.
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 Sunghwan Yeo.
Thesis:
Thesis (Ph.D.)--University of Florida, 2013.
Local:
Adviser: Baney, Ronald Howard.

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2013
System ID:
UFE0045786:00001


This item has the following downloads:


Full Text

PAGE 1

1 UO 2 SIC COMPOSITE REACTOR FUEL S WITH ENHANCED THERMAL AND MECHANICAL POPERTIES PREPARED BY SPARK PLASMA SINTERING By SUNGHWAN YEO 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 20 13

PAGE 2

2 20 13 Sunghwan Yeo

PAGE 3

3 This work is dedicated to my parents, Inbae Yeo and Y ounsook Jeong and my sister S eu n gmi Yeo for their love, and support.

PAGE 4

4 ACKNOWLEDGMENTS I would like to thank Dr. Ronald Baney my advisor and the supervisory committee chair, for his valuable guidance and support. I would also like to thank Professor James Tulenko and Dr. Ghatu Su bhash my supervisory committee members and research mentors for their patience and instruction. Without their help, this work would not have been possible. I would like to thank Dr. Clifford Bowers for being on my supervisory committee, an d his kindness for providing his equipments to support my research I would also like to thank my supervisory c ommittee member Dr. Yong Yang for his help for me to build knowl edge of nuclear materials I would like to thank Kenneth McClellan at Los Alamos National Laboratory for his advice on my research I would al so like to thank Department of E nergy and AREVA for the funding of this study. Finally, I would like to thank my family for their love and trust through the many years of my education.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ ............... 4 LIST OF TABLES ................................ ................................ ................................ ........................... 8 LIST OF FIGURES ................................ ................................ ................................ ......................... 9 LIST OF ABBREVIATIONS ................................ ................................ ................................ ........ 12 A BSTRACT ................................ ................................ ................................ ................................ ... 13 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .................. 15 2 LITERATU RE AND METHODOLOGY REVIEWS ................................ ........................... 22 Literature Review ................................ ................................ ................................ ................... 22 Thermal Conduc tivity of UO 2 ................................ ................................ ......................... 22 Oxidative (Hyper stoichiometric) Sintering ................................ ................................ .... 24 Methodology Review ................................ ................................ ................................ .............. 25 Spark Plasma Sintering (SPS) ................................ ................................ ......................... 25 Laser Flash Instrument ................................ ................................ ................................ .... 26 3 PREDICTIVE THERMAL CONDUCTIVITY MODELS FOR CERAMIC COMPOSITES ................................ ................................ ................................ ....................... 33 Maxwell Model ................................ ................................ ................................ ....................... 34 Hasselman and Johnson Model ................................ ................................ .............................. 37 Nan's Effective Thermal Conductivity Model ................................ ................................ ........ 38 4 ENHANCED THERMAL CONDUCTIVITY OF UO 2 SIC COMPOSITE FUEL PREPARED BY BOTH OXIDATIVE AND SPARK PLASMA SINTERING (SPS) ......... 42 Background ................................ ................................ ................................ ............................. 42 Experiment ................................ ................................ ................................ .............................. 43 Powder Preparation ................................ ................................ ................................ ......... 43 Characterization Methods ................................ ................................ ................................ 45 Result ................................ ................................ ................................ ................................ ...... 46 Density ................................ ................................ ................................ ............................. 46 UO 2 SiC Microstructure a nd Interface Characterization ................................ ................. 47 Microstructure of UO 2 SiC composite pellets ................................ ......................... 47 Diffusion at the UO 2 SiC i nterface ................................ ................................ .......... 48 Chemical r eaction ................................ ................................ ................................ .... 48 Grain s ize ................................ ................................ ................................ ................. 49

PAGE 6

6 Thermal Conductivity ................................ ................................ ................................ ...... 50 Discussion ................................ ................................ ................................ ............................... 50 Conclusion ................................ ................................ ................................ .............................. 52 5 THE INFLUENCE OF SIC PARTICLE SIZE AND VOLUME FRACTION ON THE THERMAL CONDUCTIVITY OF UO 2 SIC COMPOSITES ................................ .............. 60 Background ................................ ................................ ................................ ............................. 60 Experiment s ................................ ................................ ................................ ............................ 61 Fabrication of UO 2 SiC Composite Pellets ................................ ................................ ..... 61 Characterization Methods ................................ ................................ ................................ 63 Results and discussion ................................ ................................ ................................ ............ 64 Size Effect of SiC Particles on UO 2 5vol% SiC Composite Properties .......................... 64 The Effect of Volume Fraction of SiC Particles ................................ ............................. 67 Conclusion ................................ ................................ ................................ .............................. 7 1 6 MECHANICAL PROPERTIES AND INTERNAL STRESS MEASUREMENT OF UO 2 SIC COMPOSITES ................................ ................................ ................................ ........ 84 Background ................................ ................................ ................................ ............................. 84 Experiments ................................ ................................ ................................ ............................ 86 Sample Preparation ................................ ................................ ................................ .......... 86 Mechanical Testing ................................ ................................ ................................ ......... 88 Raman Spectros copy ................................ ................................ ................................ ....... 88 Results and Discussion ................................ ................................ ................................ ........... 89 Mechanical P roperties ................................ ................................ ................................ ..... 89 Hardne ss of UO 2 SiC composites ................................ ................................ ........... 89 Bulk, shear, and Young's moduli of UO 2 SiC composites ................................ ...... 90 Residue Stress Measurement ................................ ................................ ........................... 93 Conclusion ................................ ................................ ................................ .............................. 96 7 EFFECTS OF THERMAL AGING ON THE MICROSTRUCTURE AND THERMAL CONDUCTIVITY OF UO 2 SIC COMPOSITES ................................ ................................ 107 Background ................................ ................................ ................................ ........................... 107 Experiments and Results ................................ ................................ ................................ ....... 107 Predicted Centerline Temperature of UO 2 10vol%SiC Composite Fuel ...................... 107 Thermal Aging of UO 2 10vol%SiC Composites ................................ .......................... 109 Conclusion ................................ ................................ ................................ ............................ 112 8 CONCLUSIONS AND FUTURE WORK ................................ ................................ ........... 123 Conclusions ................................ ................................ ................................ ........................... 123 Future Work ................................ ................................ ................................ .......................... 124 Challenges in UO 2 Diamond Composites ................................ ................................ ..... 124 P otential R amifications of UO 2 SiC Composites ................................ .......................... 128 LIST OF REFERENCES ................................ ................................ ................................ ............. 137

PAGE 7

7 BIOGRAPHICAL SKETCH ................................ ................................ ................................ ....... 143

PAGE 8

8 LIST OF TABLES Table page 5 1 Details of SiC particle size, volume fraction, and s intering conditions in the SPS ........... 73 5 2 Thermal properties of UO 2 and SiC. ................................ ................................ .................. 73 6 1 SiC powder used to produce UO 2 SiC composites ................................ ............................ 97 6 2 Details of SiC particle size, volume fraction, and s intering conditions in the SPS ........... 97 6 3 Calculated upper and lower bounds of bulk(K), shear(G), and Young's(E) moduli of UO 2 SiC composites using H ashin and Shtrikman model [105] ................................ ....... 98 7 1 Details of composition thermal aging temperature and duration, and relative densities before and after thermal aging of UO 2 10vol%SiC composites. ...................... 114

PAGE 9

9 LIST OF FIGURES Figure page 1 1 The densities of sintered pellets with various vol% of S iC at 1650 o C for 4hours under the conventional sintering process. ................................ ................................ .......... 21 2 1 Thermal conductivity of 95% relative density UO 2 pellet [36]. ................................ ........ 27 2 2 Comparison of thermal conductivities between UO 2 and beta SiC [36, 55]. .................... 28 2 3 Thermal c onductivity of hyper and hypo stoichiometric UO 2 [40, 41] ............................ 29 2 4 Thermal conductivity of UO 2 with different volume porosity [42]. ................................ .. 30 2 5 Spark Plasma Sintering system (Dr. Sinter SPS 1030) in Particle Engineering Research Center (PERC).. ................................ ................................ ................................ .. 31 2 6 Flashline 3000 made by Anter in New Engineering Building (NEB) and its schematic apparatus. Photo courtesy of Sunghwan Yeo ................................ ................... 32 3 1 Spherical particle is subjected to uniform heat flux in an infinite medium ....................... 40 3 2 Model of ellip soidal particles in a composite ................................ ................................ .... 41 4 1 Morphologies of SiC whiskers and powder from manufacturers ................................ ...... 53 4 2 Dr. Sinter SPS 1030 system and schematic d rawing of the sintering chamber ............... 53 4 3 Fabricated UO 2 SiC composite pellets by oxidative sinter ing and SPS technique ........... 54 4 4 Relative density of UO 2 10vol%SiC composite pellets sintered by SPS and oxidative sintering at various temperatures. ................................ ................................ ...................... 54 4 5 Polished surfaces of high density UO 2 10vol% SiC composites sintered by SPS showing uniform dispersion of SiC ................................ ................................ ................... 55 4 6 Microstructure of UO 2 10vol%SiC composites ................................ ................................ 55 4 7 EDS line scan across the interface of UO 2 SiC grains in a composite pellet fabricated by SPS at 1600 o C. ................................ ................................ ................................ .............. 56 4 8 Comparison of XRD spectra of UO 2 70vol%SiC pellets sintered by SPS and oxidative sintering at 1600 o C. ................................ ................................ ............................ 57 4 9 UO 2 grain size in composite pellets resulting from the addition of SiC powder particles and SiC whiskers. ................................ ................................ ................................ 58 4 10 The measured thermal conductivity values of UO 2 and UO 2 SiC composite pellets sintered by SPS and by oxidative sintering. ................................ ................................ ....... 59

PAGE 10

10 5 1 Microstructure of UO 2 5vol%SiC composite pellets containing various diameters of SiC particle ................................ ................................ ................................ ......................... 74 5 2 SiC particles in UO 2 5vol%SiC composite pellets with various mean diameter s ............. 75 5 3 Interfacial debonding in UO 2 SiC composites with various size s of SiC particle ............. 76 5 4 Thermal diffusivity of UO 2 5vol%SiC composite pellets with various SiC particle sizes as a function of temperature. ................................ ................................ ..................... 77 5 5 Thermal conductivity of UO 2 5vol%SiC pellets with various sizes of SiC particles at the selected temperatures ................................ ................................ ................................ ... 78 5 6 Microstructures of UO 2 volume fractions ................................ ................................ ................................ ................. 79 5 7 Relative density of UO 2 size SiC particles. ................................ ................................ ................................ ............... 80 5 8 Thermal diffusivity of UO 2 SiC composite pellets containing various volume ................................ ..... 81 5 9 Temperature dependence of specific heat capacities of UO 2 SiC composite pellets ................................ ......... 82 5 10 Calculated and experimentally determined thermal conductivities of UO 2 SiC ................................ ........ 83 5 11 2 20vol%SiC composite. Interfacial porosities are indicated by arrows. ................................ ................................ ... 83 6 1 Selected SEM micrographs of UO 2 and UO 2 SiC composites showing the transition in SiC particle volume fractions and sizes ................................ ................................ ......... 99 6 2 Vickers and Knoop hardness of UO 2 SiC composite s containing different mean particle sizes and various volume fractions of SiC addition. ................................ ........... 100 6 3 Optical microscopy pictures of Vickers and Knoop indents from the composite containing 5vol% 55 m sized SiC particles. ................................ ................................ ... 101 6 4 Young's modulus of UO 2 SiC composites containing different mean particle sizes and various volume fractions of SiC addition ................................ ................................ .. 102 6 5 Grain size of UO 2 SiC composites containing different mean particle sizes and various volume fractions of SiC addition ................................ ................................ ........ 103 6 6 Raman spectra collected from within SiC particles with different sizes in each UO 2 5vol%SiC composite. ................................ ................................ ................................ ....... 104

PAGE 11

11 6 7 TO peaks in Raman shift and changes in TO peak position compared to that of stress free SiC particles as a function of SiC p article size. ................................ ........................ 105 6 8 Evolution of internal stress as a function of SiC particle size in UO 2 5vol%SiC composites. ................................ ................................ ................................ ....................... 106 7 1 Expected centerline temperature of UO 2 and UO 2 10vol%SiC composite fuels as a function of reactor burnup rate. ................................ ................................ ....................... 115 7 2 Microstructures of UO 2 particles af ter thermal aging for 12 hour...................................................... .................. .. 116 7 3 Microstructures of UO 2 10vol%SiC composites containing SiC whiskers after thermal aging for 12hours ................................ ................................ ................................ 117 7 4 Microstructures of UO 2 10vol%SiC composites after thermal aging for 12hours at 1500 o C. ................................ ................................ ................................ ............................. 118 7 5 Microstructures of UO 2 10vol%SiC composites containing 1 m SiC particles after thermal aging for 12hours at 1500 o C. ................................ ................................ .............. 119 7 6 Microstructures of UO 2 10vol%SiC composites containing SiC whiskers after thermal aging for 12hours at 1500 o C. ................................ ................................ .............. 120 7 7 Thermal conductivity of UO 2 10vol%SiC composites containing SiC whiskers before and after thermal aging at 1500 o C for 12hours. ................................ .................... 121 7 8 Thermal conductivity of UO 2 10vol%SiC composites containing 1 m SiC particles before and after thermal aging at 1500 o C for 12hours. ................................ .................... 122 8 1 Morphologies of diamond particles with 25 m diameter. ................................ ............... 130 8 2 Thermal conductivity of as sintered UO 2 10 vol% diamond composites sintered at different temperatures at 100 o C, 500 o C, and 900 o C ................................ ........................ 131 8 3 Thermal conductivity of UO 2 10 vol% diamond composites after reduction process performed at 800 o C for 6hours in 4%H 2 N 2 +moisture atmosphere. ................................ 132 8 4 Thermal conductivity of UO 2 10 vol% diamond composites after reduction process. The highest density composite (96.2%TD) was reduced and thermal etched. ................ 133 8 5 SEM micrographs of a high density UO 2 diamond composite. ................................ ....... 134 8 6 SEM and EDS elemental maps of a diamond particle surface in the UO 2 diamond composite afte r reduction and thermal etching ................................ ................................ 135 8 7 SEM and EDS elemental spectra of carbon K K along scanned line across a chemical reaction product on the surface of a diamond part icle ... 136

PAGE 12

12 LIST OF ABBREVIATION S ASTM A merican S ociety for T esting and M aterials BET Brunauer Emmett Teller CNT Carbon nano tube CTE Coefficient of thermal expansion DC Direct current EDS Energy d ispersive X ray s pectroscopy FAST Field as sisted s intering t echnique LOCA Lost of coolant accident LWR Light water reactor MOX Mixed oxide fuel NRC Nuclear Regulatory Commission PNNL Pacific Northwest National Laboratory PyC P yrolytic carbon SEM Scanning electron microscopy SiC Silicon carbide SiCp Silicon carbide spherical particles SiCw Silicon carbide whiskers SPS Spark plasma sintering TD Theoretical d ensity TO T ransverse optical TRISO Tristructural isotropic UHP U ltra high purity UO 2 Uranium dioxide XRD X ray diffraction

PAGE 13

13 Abstract o f Dissertation Presented t o t he Graduate School o f t he University o f Florida i n Partial Fulfillment o f t he Requirements f or t he Degree o f Doctor o f Philosophy UO 2 SIC COMPOSITE REACTOR FUELS WITH ENHANCED THERMAL AND MECHANICAL POPERTIES PREPARED BY SPARK PLASMA SINTERING By Sunghwan Yeo August 20 13 Chair: Ronald Baney Major: Materials Science and Engineering The primary ceramic fuel used in nuclear reactors, Uranium Dioxide (UO 2 ), has a low thermal conductivity which results in a decrease in both the en ergy output and the safety of a nuclear reactor. The introduction of high thermal conductivity fuel pellets enables a nuclear reactor to produce more thermal energy while maintaining plant safety due to lower pellet centerline temperature and thermal gradient resulting in a lower level of fission gas release and thermal cracking. The main objective of this research is to increase the thermal conducti vity of UO 2 by the incorporation of Silicon Carbide ( SiC ) particles or whiskers O xidative sintering and Spark Plasma Sintering (SPS) techniques have been used to produce UO 2 SiC composite pellets. While oxidative sintering failed to achieve enhanced thermal conductivity, the SPS sintered pellet contained promising features such as higher density, better interfacial contact, and reduced chemical reaction, and hence, the enhanced thermal con ductivity was obtained. The influence of SiC particle size (0.6 55 m) and volume fraction (5 20%) on the thermal conductivity of spark plasma sintered UO 2 S i C composites was investigated. The composites containing a larg er volume of SiC particles where size is less than 16.9 m showed higher

PAGE 14

14 thermal conductivity. The measured thermal conductivity was in excellent agreement with calculated effective thermal conductivity based on a theoretical model. The m echanical properties such as hardness and Young's modulus, and the internal stress of SiC particles of UO 2 SiC composites containing different sizes and volume fractions of SiC particles were examined. Hardness and Young's modulus showed increases up to 53% and 18% with 20vol% of 1 m SiC addition respectively. The compression stress of SiC particles up to 1.65GPa was successfully measured using Raman spectroscopy. Lastly, thermal aging was performed at 1000 o C and 1500 o C for 12hours on fabricated UO 2 SiC composite s to investigate the performance of the composites in the operational condition s of nuclear reactor. Microstructural defects such as microcracking and interfacial debonding, and SiC particles and whiskers removing from the polished surface were observed after the 1500 o C t hermal aging prob ably due to a chemical reaction between UO 2 a nd SiC. T he therma l conductivity of composites was found to be decreased by 5 10% after the aging test.

PAGE 15

15 CHAPTER 1 INTRODUCTION Over the last few years, there have been enormous changes in the energy sector. Energy consumption has been increasing dramatically. Oil and gas prices have doubled over the last two years. W e are experienc ing exhausting fossil energy resources, environmen tal pollution and climate change which threatens us with diminishing green house gases. Therefore, many countries are meet ing these issues by seeking sustainable and competitive energy sources. Nuclear fuel has been a well known candidate for an ecological and high energy density fuel in response to these challenges. As a result, nearly 20% of electricity in the USA is produced by nuclear reactors, and this power source is growing year by year. Despite the availability of numerous types of nuclear fuels (me tals, MOX, nitrides, etc) commercial reactors in the world are fueled by Uranium Dioxide (UO 2 ). It is the fuel of choice for several reasons such as high melting point (transient accident resistance [1 2]) and enhanced oxidation resistance. Also, UO 2 behav ior has been studied in much more depth throughout different power cycles than other fuel types. Its major disadvantage is a low level of thermal conductivity, which causes both a steep temperature gradient and a high centerline temperature of uranium diox Owing to the steep temperature gradient and high centerl ine temperature, a variety of expected phenomena are induced. Important components of the fuel, such as pores, oxygen, and fission products, are redistribute d from a uniform condition [ 1 ] Thermal stress caused by a large temperature gradient results in either cracking in a low temperature region or plastic deformation in the high temperature r egion. Moreover, in the loss of coolant accident (LOCA) the Zircaloy cladding temperature is rapidly increased due to the high centerline temperature of the fuel pellet leading to significant Zircaloy and water reactions [ 2 ] (ex: Zr+2H 2 O= ZrO 2 +H 2 ). ZrO 2 is produced on the cladding surface, decreasing heat conduction and causing a cladding rupture.

PAGE 16

16 Hydrogen gases, produced by the reaction, increase the internal pressure of the reactor and may cause an explosion. I ncreasing the thermal conductivity of nuclear fuel not only decrease s these harmful phenomena, but it allow s for the output of a reactor to be increased and enhance the safety of a reactor during normal operation and short term excursions The maximum heat output from the reactor core could be increased by the high thermal conductivity of the fuel pellet enabling the reactor to produce more thermal energy while maintaining the plant safety. Moreover, with a decrease in centerline temperatu re caused by the increased thermal conductivity of the fuel pellet, the temperature gradient in the fuel is decreased causing reduced fission gas release and reduced cracked or broken pellets due to thermal stresses while maintaining the desired fuel and c ladding temperatures. Because t he development of the thermal performance of UO 2 fuel has been regarded as a major priority in the research of nuclear power technology other ideas and efforts has been reported. Tristructural isotropic (TRISO) fuel [ 3 ] is a micro sized high density UO 2 particles coated with four different layers such as porous buffer layer made of carbon, a dense pyrolytic carbon (PyC) layer, a fission product retainable Silicon Carbide layer, and a dense outer layer of PyC Various researches [ 4 7 ] regarding TRISO fuel and its constituents are currently being conducted. A more simple and cost efficient idea is to fabricate composite fuels containing UO 2 matrix and high thermal conducting secondary material. The idea of UO 2 composite fuels containing a secondary material can be approached in few ways forming either a high heat conducting percolation pathway or a high thermal conductivity secondary particles distributed homogeneously in the matrix material (UO 2 ). While most thermal ener gy ca n be conducted through a percola tion pathway, the thermal energy must pass through both higher thermal conducting secondary particles and lower thermal conducting UO 2 matrix in

PAGE 17

17 the composite containing dispersed secondary particles. Therefore, forming a high heat condu cting percolation pathway can lead to more efficiently conducting heat because less thermal energy is conducted through the lower thermal conducting UO 2 matrix. However, the manufacture of composite fuel containing such a percolation path way is quite difficult. Sarma et al. [ 8 ] and Singh at al. [ 9 ] attempted to produce UO 2 SiC composites with a percolation pathways. T heir method involved fabrication of a UO 2 pellets with open porosity followed by polymer infiltra tion and pyrolysis of a silicon carbide pre ceramic polymer to form a percolation pathway. The result however, was a pellet with degraded thermal conductivity compared to a pure UO 2 pellet because the formed SiC in the composite pellet had much lower density and thermal conductivity than pure SiC. In our wor k, we fabricate d UO 2 based composite fuels containing homogeneously dispersed high thermal conducting SiC particles. This concept of incorporating high thermal conductivity material int o a UO 2 pellet has been studied [ 10 ] and silicon carbide has been considered a prime candidate for the high thermal conducting secondary particles. This is due to its low neutron cross section, high thermal conductivity, chemical stability (strong resistance to oxidation in air and air moisture environments), and high melting temperature (~ 2973 o C) [ 11 ] Silicon carbide ( SiC) also has the advantage of being non toxic and isotropic over alternatives such as beryllium oxide [ 12 ] Hunt et al. [ 13 ] stated that incorporation of at least 4mol% SiC in UO 2 fuel has a benefit of maintaining O to M ratio near two during irradiation. Allen e t al. [ 14 ] however, found that chemical reactions occur between UO 2 and SiC around 1370 o C which may severely degrade the thermal conductivity of the composite. Solomon and associates [ 15 ] investigated reaction products and suggested the formation of USi 1.88 U 20 Si 16 C 3 UC, CO, and SiO. Originally, it was believed that a continuous phase of SiC throughout the pellet would lead to significantly higher therma l conductivity. However, simulations performed by Latta et al. [ 16 ] revealed that

PAGE 18

18 discontinuous SiC fibers produced nearly the same increase in thermal conductivity as that of a continuous phase. This result is in part due to the higher density and crysta llinity of the SiC whiskers compared to that formed by the pre ceramic polymer. Jiwei Wang [ 17 ] researched the incorporation of whisker type silicon carbide by powder mixing. SiC whiskers are configurationally preferable to pr event fission gas swelling during the irradiation, and more conductive for heat to disperse from UO 2 pellets. However, he did not measure the thermal conductivity values of the UO 2 SiC pellets, and only obtained relatively low density pellets by pressureless sintering at 1650 o C. Hence, supplementary researches for the development of competent sintering methods, and the thermal conductivity characterization of UO 2 SiC composite pellets, are required. Besides SiC, d iamond is a potentially possible c andidate as the incorporating secondary material because of its superior t hermal conductivity (~500 W/mK) Diamond is already a hot research topic material used as a high heat conducting secondary material in ceramic composites used for a heat sink in semi conductor devices [ 18 19 ] Hanada et. al. [ 18 ] increased the thermal conductivity of Cu d iamond composite up to 18% compared to pure Cu by adding d iamond only 1vol%. Abdel et. al. [ 20 ] increased the thermal conductivity of Cu based composite up to 21% by adding 20vo l% d iamond coated Cu particles. Although several researchers are proceeding in that field, composite fuel using d iamond partic les. Because a research project of UO 2 diamond composite fuel is now under active development, only preliminary results of UO 2 diamond compos ites will be introduced in this dissertation. The main objective of this work is an experimental exploration of the properties of UO 2 SiC composite fuel pellets. The objective are achieved by successful outcomes for the following goals.

PAGE 19

19 First, f abricate high density and high thermal conductivity UO 2 SiC composite pellets using a Spark Plasma Sintering (SPS) technique. We have observed that simpl y mixing and c onventional sintering of both UO 2 and SiC powders leads to poor sintering behavior hampering the d ensification as shown in Figure 1. The relative density of UO 2 SiC composite pellets decreases with increasing SiC volume fraction. Similar trend s were observed in various composites [ 21 23 ] It was found that the increase in volume fraction of secondary particles hinders the consolidation of mat rix grains. Maekawa et al. [ 23 ] suggested that pores remain locked around p articles and eventually form pore channels decreasing the three dimensional connectivity of a matrix material to lower relative density. In this dissertation SPS was perform ed to overcome this hindrance. SPS is an advanced sintering technique in which localized heat and uniaxial pressure are generated for the consolidation of ceramic powders. Nevertheless, in light of advantages such as rapid sintering, uniform sintering, an d low running cost, the effect of SPS on the sintering of nuclear fuel has not been empirically demonstrated. T hermal co nductivity measurements was conducted to reveal th e resulting thermal conductivity of the composite fuel pellets Second, i nvestigate the influence of SiC particle size and volume fraction on thermal properties of UO 2 SiC composite s In gener al, it was found that ceramic composites containing second phase particles with larger size (low surface to volume ratio) reduce the inte rfacial a rea inducing thermal resistance and hence, increase the thermal conductivity of ceramic composites [ 24 26 ] The relationship between the thermal conductivity of a two phase mixture and the volume fraction of particles was introduced by Maxwell J.C. [ 27 ] and Bruggeman and Johnson [ 28 ] According to their theoretical work s the thermal conductivity increases with increasing the volume fraction of SiC in UO 2 SiC composite pellets. Thus, we speculate that, for UO 2 SiC composites, higher thermal conductivity composite fuel would be expected with larger size and

PAGE 20

20 higher volume fraction of SiC particles. In our study, a series of UO 2 SiC composite s with different sizes and volu me fraction s of SiC particles was fabricated and t heir thermal conductivities are examined and compared with theoretically calculated thermal conductivities. Third, i dentify the mechanical property of UO 2 SiC composites. Sustainable mechanical properties of a nuclear fuel is required to prevent fuel failures [ 29 ] during lo ng operating cycles of a nuclear reactor. Because of its excellent mechanical properties, SiC have often added in a composite to increase mechanical properties [ 30 31 ] T herefore, UO 2 SiC c omposites are expected to have better mec hanical properties then those of UO 2 Harness and Young's modulus of various composites containing diffe rent sizes and volume fractions of SiC were examined in this study. Fourth t est UO 2 SiC co mposite fuels under the temperature similar to a reactor core to investigate changes in their properties and microstructures. Even if UO 2 SiC composites exhibited outstanding properties, it would be critical to examine the sustainability of their propertie s and sound microstructures under the operating condition of a nuclear reactor. It is difficult to imitate the exact same condition as a nuclear reactor core, and hence, fabricated UO 2 SiC composites were ther mal aged at an expected maximum fuel centerline temperature to reveal the influence of extreme temperature on the composites. Microstructures and thermal conductivity of UO2 SiC composit es were examined and compared before and after the thermal aging test.

PAGE 21

21 Figure 1 1. The densitie s of sintered pellets with various vol% of S iC at 1650 o C for 4hours under the conventional sintering process.

PAGE 22

22 CHAPTER 2 LITERATURE AND METHO DOLOGY REVIEWS Literature R eview Thermal C onductivity of UO 2 As re ported in many other papers [ 32 36 ] the thermal conductivity of UO 2 is very low when compared with other ceramic materials. Fink [ 36 ] reported that the thermal conductivity of 95% relative density UO 2 can be calculated by Equation 2 1, where t = T (K)/1000 and is the thermal conductivity of 95% dense UO 2 in W/m. Figure 2 1 shows the results and Figure 2 2 compares it with the thermal conductivity values of single crystal beta silicon carbide [ 37 ] ( 2 1) The low thermal conductivity of UO 2 compared with other ionic materials, is mainly due to the anharmonic components of crystal vibrations [ 1 ] Since ionic cov a lent bonding of ceramic quantum of Young's vibration of the lattice called a phonon. It is obvious that phonon phonon scattering rapidly decreases thermal conductivity, considering its value at high temperature, as shown in Figure 2 2 In UO 2 the phonon phonon scattering is maintained to a relatively high level throughout the temperature range by lattice anharmonicity which increa ses with the mass difference between the anions and the cations in the ionic solid. Because UO 2 has nearly the greatest mass difference, thermal conductivity of UO 2 is considerably lower than that of other ionic materials. Al ong with the inevitable factors of lattice anharmonicity and bonding structure, specific conditions of UO 2 also impact thermal conductivity. Two condition s are essential to this concept First, UO 2 stoichiometry influences thermal conductivity. The uranium oxide with an O/U ratio greate r than 2.0 is called hyper stoichiometric UO 2 ; the uranium oxide with an O/U ratio

PAGE 23

23 less than 2.0 is called hypo stoichiometric UO 2 The stoichiometric UO 2 and the intermediate oxidation products, U 4 O 9 and U 3 O 7 are the fluorite structures in which uranium atoms are in a face centered pattern contain ed in a cube of oxygen atoms [ 1 ] Compared with stoichiometric UO 2 U 4 O 9 and U 3 O 7 have clustered interstitial oxygen atoms at unoccupie d cubic sites, accompanied by the displacement of neighboring U atoms [ 38 ] On the other hand, a completely oxidized state, U 3 O 8 has an orthorhombic lattice structure [ 39 ] and its density is 8.38 g/cm3, which is 24% less than UO 2 The thermal conductivity of hyper and hypostoichiometric UO 2 is shown in Figure 2 3 [ 40 41 ] The excessive oxygen atoms and the strain field in the matrix surrounding the vacancy act as phonon scattering centers so that thermal conductivity for both hyper and hypo stoichiometric UO 2 is de creased. Second, the increase in volume porosity significantly decreases the thermal conductivity of UO 2 pellets. Since the thermal conductivity of air corresponds to an extremely small value, 0.026 W/m K, it follows that as a pellet has more porosity its thermal conductivity decreases. Fundamental aspects of nuclear reactor material s [ 1 ] if all pores are of equal size and dispersed randomly, the thermal conductivity of UO 2 pellets with porosity can be obtained by Equation 2 2 In the equation, K is the effective thermal conductivity, K S is the thermal conductivity of the fully dense pellet, P is the volume porosity defined by the volume of pores div ided by the entire volume and is a coefficient with a function of temperature. (2 2 ) Van Craeynest and Stora [ 42 ] found th at Equation 2 3 fits their empirical data from 50 o C to 1000 o C. ( 2 3)

PAGE 24

24 Figure 2 5 reveals the calculated K/K S relative thermal conductivity, at 100 o C with different volume porosity. This figure predicts that the UO 2 pellet presents a linear decrease in thermal conductivity values in proportion to an increase in the degree of porosity. Oxidative (H yper stoichiometric ) S intering Uranium dioxide pellets are generally made by a four hour sintering process at 1700 o C in a hydrogen containing gaseous atmosphere to achieve at least 95% relative density. However, due to this high temperature, the energy and maintenance of furnace operation are costly. For decades, oxidative sintering has been cited by many researchers as a s trategy for the lower temperature sintering of UO 2 In 1959, Williams et al. [ 43 ] studied sintering UO 2 with different O/U ratio and atmospheres. He found that hyper stoichiome tric UO 2+x sintering in nitrogen and argon atmosphere lower the temperature to 1400 o C. In 1960, Langrod [ 44 ] attained 95% density UO 2 pellets by sintering hyper stoichiometric powder with the O/U ratio of 2.3 at 1300 o C in a ni trogen atmosphere for 2 hours followed by reduction in a hydrogen atmosphere. Chevron [ 45 ] reported in 1992 that the most appropriate O/U ratio for si ntering is 2.25 Dehaudt [ 46 ] in 2001, found that the activation energy of hyper stoichiometric UO 2.25 for diffusion at grain boundary is lower than that of stoichiometric UO 2 Numerous authors agree that excessive oxygen atoms are the key for enhanced densification [ 47 49 ] In the high temperature sintering process, where uranium ions are heaviest and s lowest, the diffusion rate of uranium ions controls the entire densification rate. Since uranium ions are displaced in the hyper stoichiometric UO 2+x fluorite structure due to the penetration of oxygen atoms, there is more space for uranium ions to diffuse The higher diffusivity of uranium ions improves the total diffusion rate so that hyper stoichiometric UO 2+x performs higher sinterability.

PAGE 25

25 The experimental methods for the oxidation and reduction of UO 2 have been investigated by researchers. In order to create hyper stoichiometric UO 2+x the initial UO 2 powder needs to be oxidized in air. The inverse process, reduction, is also necessary after sintering since stoichiometric UO 2 has a much higher degree of thermal conductivity, as shown in Figure 2 4. Ohas [ 50 ] showed that the ultimate oxidized state U 3 O 8 is achieved by heating the UO 2 powder in dry air at 350 o C for 27 hours. Blackburn [ 51 ] reported that U 3 O 8 is obtained by heating the UO 2 powder in dry air, at 800 o C, for 3hours. A ccording to ASTM C 1430 07, the reduction can be achieved by heating the hyper stoichiometric UO 2+x pell ets at 800 o C for 4hours, in a 4%H 2 N gas, with water vapour atmosphere maintained at 35 o C, using a water bath system. Methodology R eview Spark Plasma Sintering (SPS) SPS is distinguished from the typical sintering method by the use of spark discharge. Figure 2 5 shows our Dr. Sinter SPS 1030 system which is utilized to perform the SPS process. According to the most frequently accepted micro spark, or plasma theory [ 52 ] while DC current pass through the compact powder in a graphite die, the high energy pulse current induces a spark discharges between the fine particles. The plasma momentarily generates intense heat bonding the particles together. The plasma of SPS, however, has not been identified directly, but electric noise was observed and is considered to correspond to plasma generation [ 53 ] A rapid densification process with high diffusivity of SPS is achieved by three factors; 1) mechanical pressure, 2) rapid heating rates, and 3) electric field an d pulsed direct current [ 54 ] Firstly, it is obvious that the application of mechanical pressure during the sintering improves the pore removing process from compacts and enhanc es diffusion Secondly, the fast heating rate, caused by localized spark discharges, facilitates rapid and uniform densification Lastly, the

PAGE 26

26 electric field and pulsed direct current induce spark discharges at the inter particle contacts and remove surface contaminants and absorbed species such as CO 2 and H 2 O from the par ticle surface, thus improving the grain boundary diffusion processes [ 54 ] Laser Flash Instrument The Laser Flash Method is one of the most common methods to measure the thermal conductivity derived by Equation 2 4, where K is the thermal conduc tivity in W/m K, is the thermal diffusivity in m 2 /s, C p is the specific heat capacity in J/kg K, and is the density in kg/m 3 ( 2 4) The derivations of thermal diffusivity, and specific heat capacity, C p are based on the measurement of the rising temperature on the back surface of a sample caused by a pulsed laser 2 6 induced heat from a pulsed laser beam travels throug h a sample, and the heat increase on the rear surface of the sample is measured by an IR detector. The thermal diffusivity in m 2 /s is calculated by Equation 2 5, where L is the thickness of the specimen in m, and t 1/2 is the time in seconds for the rear surface temperature to reach 50% of its maximum value. On the other hand, the specific heat capacity C p is calculated by Equation 2 6, where Q represents the energy of the pulsed laser beam, which can be determined by a reference sample, m is the mass of the specimen, and is the maximum value of the temperature rise. In the Laser Flash Machine, shown in Figure 2 8, this value is obtained by comparing the maximum value of the temperature rise with that of a reference. ( 2 5) ( 2 6)

PAGE 27

27 Figure 2 1 Thermal conductivity of 95% relative density UO 2 pellet [ 36 ]

PAGE 28

28 Figure 2 2 Comparison of thermal conductivities between UO 2 and beta SiC [ 36 55 ]

PAGE 29

29 Figure 2 3 Thermal conductivity of UO 2 A) hyper stoichiometric UO 2 B) hypo stoichiometric UO 2 [ 40 41 ] A B

PAGE 30

30 Figure 2 4 Thermal conductivity of UO 2 with different volume porosity [ 42 ] 0 0.2 0.4 0.6 0.8 1 1.2 0.01 0.04 0.07 0.1 0.13 0.16 0.19 0.22 0.25 0.28 Porosity volume K/Ks

PAGE 31

31 Figure 2 5 Spark Plasma Sintering system ( Dr. Sinter SPS 1030 ) in Particle Engineering Research Center (PERC) Photo courtesy of Sunghwan Yeo.

PAGE 32

32 Figure 2 6 Flashline 3000 made by Anter in New Engineerin g Building (NEB) and its schematic apparatu s Photo courtesy of Sunghwan Yeo.

PAGE 33

33 CHAPTER 3 PREDICTIVE THERMAL C ONDUCTIVITY MODELS F OR CERAMIC COMPOSITE S Heat can be conducted by both lattice vibration waves (phonon) and free electrons in solid materials [ 56 ] Because non metallic materials such as UO 2 or SiC lack free electrons, phonon s are primary responsible for thermal conduction. T ransport by phonons is not as effective as free electrons because phonon scattering frequently occur s due to microstructure defects and lattice imperfections [ 57 ] Since our research materials are ceramic composites con taining two constituents, it is reasonable to consider factors that could lead to the phonon scattering, and hence, have influence on the effective thermal conductivity of composites These include the thermal conductivity of individual constituent s and the crystallinity of the second phase materials ( monocrystalline SiC has higher thermal conductivity than polycrystalline SiC). The second phase particle shape, size, volume fraction, degree of mixing, and orientation in a matrix material also have eff ect s on the thermal conductivity of a ceramic composite. The interfacial characteristics such as bonding between UO 2 and SiC also greatly affect the thermal conductivit y. Because a void at the interface scatter s thermal energy and hinder s the thermal ener gy transport to the high thermal conducting second phase particles, a physically well bonded interface between UO 2 and SiC is essential for high thermal conductivity. However, even though the interface is atomically perfect, some portion of the phonon is s cattered due to the difference in physical and vibrational properties of the composites' constituents. This is well known as Kapitza resistance. Kapitza [ 58 ] researched the heat transfer of liquid helium to capillaries and suggested that temperature change at the interface is proportional to the normal component of the heat flux. (3 1 ) Where is a change in temperature between the matrix and second phase particle at the interface is the heat flow at the interface and is Kapitza resistance.

PAGE 34

34 depends on how different and similar in thermal and physical properties of materials that composed of composites. Maxwell Model Since 1950s, v arious researches have reported the predictive models of the effective thermal conductivity of composites The Maxwell theoretical model [ 27 ] is a basis for many theoretical models. Maxwell originally derived an equation to obtain the electrical conductivity of composites. However, the physical concepts and governing equations could be easily adopted to the effective thermal conductivity of composites. Maxwell assumed a composite containing non interact ed spherical second phase particles in a matrix material. The spatial distribution of temperature outside and inside a spherical particle can be obtained by Laplace equation when the particle is placed in an infinite medium and subjected to an uniform heat flux as illustrated in Figure 3 1 [ 59 ] ( 3 2 ) ( 3 3 ) Where r and z are the directions of heat flux as shown in Figure 3 1, A B and C are unknown coefficients, and is th e angle between r and z axes. O utside a spherical particle, the spatial temperature can be expressed using both the linear temperature change along the direction of heat flux (first term in 3.2 ) and the temperature fluctuation due to the existence of a second pha se particle (second term in 3.2 ). As r fluctuation is disappeared in this equation. The t emperature inside a spherical particle, is linear along the z axis. The following conditions are necessary to obtain the unknown values A and B [ 27 ] = ( 3 4 ) ( 3 5 )

PAGE 35

35 Where and are the thermal conductivities of matrix and second phase particle s respectively. While Equation 3 4 refers to no existence of int erfacial resistance, Equation 3 5 means there is no thermal energy accumulation at the interface. Sol ving these equat ions using Equation 3 2 and 3 3 leading to [ 27 ] ( 3 6 ) ( 3 7 ) Where R is the radius of a particle as shown in Figure 3 1 Eq uation 3 2 c an be reformed using Equation 3 6 as [ 27 ] (3 8 ) If the boundary conditions of particle are given by the continuity of temperature and components of heat flux, the distribution of heat at any point inside and outside a second phase particle can be calculated [ 27 ] When total N identical second phase particles with R o radius are dispersed homogeneously without any particle particle interaction in an infinite medium, a single par ticle can be treated as a sphere in an infinite medium. Then, the temperature at some distance from a particle can be expresse d as [ 27 ] ( 3 9 ) Where N is the number of spheres and is the radius of a sphere. Let us assume that all particles are in a large circle area of the medium with a radius of R 1 Then, the temperature outside the large circle can be expressed [ 27 ] (3 10 )

PAGE 36

36 Where k e is the effective thermal conductivity of the composite. Because the temperature at some distance from a particle and the temperature outside a large circle of me dium are thermally equivalent, E quation 3 9 and E quation 3 10 are same [ 27 ] (3 11 ) Solving this equation, we can obtain the Maxwell equation for the effective thermal conductivity of a composite [ 27 ] (3 12 ) Where is and same as the volume fraction of particle s in the composite. This equation is de rived under several assumptions including that onl y spherical shaped particles are distributed and the i nteraction between particles does not exist [ 27 ] T he volume fraction of second phase particle s ha s to be small enou gh to be isolated from each other. The thermal conductivity of composites containing non spherical shaped second phase particles such as whisker or plate types can't be applied to Maxwell equation. Above all, the interfacial thermal resistance is not consi dered in the calculation. As a result of assumption s t he size of the particles is also not considered in the equation. In an actual case, because different dispersed particle size s changes the area of interface, the effective thermal conductivity is modif ied. However, there is quite a few experimental reports that have confirmed the validity of Maxwell equation. Wong et. al. [ 60 ] found that experimentally determined thermal conductivity of epoxy based composite filled with silica is in good agreement with the calculated value using Maxwell equation when the composite contains less than 40vol% of silica filler Boey et. al. [ 61 ] fabricated Aluminum Nitride (AlN) Yttrium Oxide (Y 2 O 3 ) composite and found that its thermal conductivity is clo ser to the calculated value using Maxwell equation when the low er volume fraction of Y 2 O 3 is added

PAGE 37

37 According to various exper imental studies composites containing second phase particles less than approximately 10 20 vol % and when there is large differenc e in the thermal conductivities between particle and matrix follow the Maxwell equation. Hasselman and Johnson Model Because of the limitations of the Maxwell model Hasselman and Johnson [ 62 ] mod ified Maxwel l 's calculation to derive a new expression for the effective thermal conductivity of composites containing spherical, cylindrical and flat plate second phase particles. They found that the effective thermal conductivity is not only dependant o n the volume fraction of second phase particles but also on the size o f dispersed particles This result is because the interfacial thermal resistance increases with the decreasing size of the dispersed particles due to the lar ger interfacial area They obtained expressions for composites containing spherical and cylindrical dispersed particles by modifying the original Maxwell equation. For fl at plate dispersed particles, they utilized the series circuit approach. Here, we will briefly describe only the expression for the composite containing spherical dispersed particles which is applicable to our UO 2 SiC composites. The effective thermal conductivity of composites is expressed as [ 62 ] (3 13 ) Where a is the radius of particle and h c is the interfacial thermal conductance Hasselman and Johnson introduced the interfacial thermal conductance and considered the effect of particle size on the effective thermal conductivity. The interfacial thermal conductance is a measure of thermal flow at the interface and the inverse form of the inte rfacial thermal resistance R k which is also known as Kapitza resistance. If h c Hasselma n and Johnson model, Equation 3 1 3 and Maxwell model, Equation 3 1 2 become identical.

PAGE 38

38 Nan 's Effective Thermal Conductivity Model Nan et. al. [ 63 ] also used a n effective medium approximation for calculating the effective thermal conductivity where the interfacial thermal resistance and dispersed particle size are included. They developed this theory further than Hasselman and Johnson model by calculating interfacial thermal resistance based on multiple scattering theory [ 64 ] Nan obtained several formulations for composites containing spherical particles, randomly oriented ellipsoidal particles, aligned continuous fibers, and flat plates. Here, we will describe the calculations for spherical and randomly oriented and aligned ellipsoidal particles. For spherical particles, the effective thermal conductivity is given by [ 65 ] (3 14 ) Where is a dimensionless parameter and defined as, (3 15 ) Where R k and a are the interfacial thermal resistance and the radius of dispersed particles, respectively. Considering the interfacial thermal resistance is the inverse form of the interfacial thermal conductance, R k =1/h c the Nan 's model, Equa tion 3 1 4 and the Hass elma n and Johnson model, Equation 3 1 3 are identical. For the formulation of aligned ellipsoidal particles, the structure model of particle in a composite is illustrated in Figure 3 2. An ellipsoidal particle with the thermal conductivity of k p and semi axes a 1 and a 3 is incorporated in a matrix with the ther mal conductivity of k m The directions of thermal conductivity according to the particle geometry and interfacial thermal conductance are also shown in the figure. The thermal conductivity of composite with specific direction is expressed as [ 65 ] (3 16 )

PAGE 39

39 Where is geometrical factor and is defined as, (3 17 ) Where is expressed as, (3 18 ) Geometrical factors is expressed as, (3 19 ) (3 20 ) Where P is the aspect ratio of ellipsoidal particle and expressed as, (3 21 ) The a bove equations are practical for the composite fuel containing aligned ellipsoidal particles such as SiC whiskers Several experiments using the SPS consolidation technique revealed that most ellipsoidal particles in a composite are aligned along a sample plane due to the pressure applied perpendicular to the plane [ 66 68 ] Nan also predicated a calculation for a composite containing completely random oriented ellipsoidal particles. With random orientation, t he effective thermal conductivity becomes in simpler form than that of aligned orientation (3 22 ) Where and are expressed in Equation 3 13 and Equation 3 14, respectively. Th is formulation is also useful to calculate theoretical thermal conductivity of UO 2 SiC whiskers composites fabricated by conventional furnace method.

PAGE 40

40 Figure 3 1. Spherical particle is subjected to uniform heat flux in an infinite medium

PAGE 41

41 F igure 3 2. Model of ellipsoidal particles in a composite

PAGE 42

42 CHAPTER 4 ENHANCED THERMAL CON DUCTIVITY OF UO 2 SIC COMPOSITE FUEL PREPARED BY BOTH OXIDATIVE AND SPARK PLASMA SINTERING (SP S) Background As mentioned in chapter 1 i ncreasing the thermal conductivity of nuclear fuel would allow for the output of a reactor to be incr eased and enhance the safety of a reactor during normal operation and short term accidents. Moreover, with a decrease in centerline temperature caused by the increased thermal conductivity of the fuel pellet, the temperature gradient in the fuel is decreas ed allowing reduced fission gas release and number of cracked or broken pellets due to thermal stresses while maintaining the desired fuel and cladding temperatures. Building on a previous study performed by Wang [ 17 ] we concluded that incorporat ing SiC into UO 2 may increase the therm al conductivity of nuclear fuel. T his study has focused on the utilization of an advanced technique to fabricate fully dense UO 2 SiC composite pellets with excellent thermal properties. B oth low temperature oxidative sintering [ 69 ] and SPS techniques were employed to fabricate UO 2 SiC composite pellets and compare the resulting microstructures and properties. As described in chapter 2, i n the oxidative low temperature sintering, the UO 2 powder is co sintered with SiC powder or whiskers by increasing the O to M ratio of the starting powder [ 43 46 ] to an optimum value of 2.25 [ 45 ] This enhanced sinterability is due to the increased diffusivity of Uranium atoms through vacancies. Spark Plasma Sintering (SPS) or Field Assisted Sintering Technique (FAST) has recently gained significant interest in numer ous research fields [ 70 72 ] In recent years, spark plasma sintering (SPS) has evolved as a promising sintering technique for rapid fabrication of UO 2 pellets of required shape and size [ 73 ] However, literature exists for the fabrication of UO 2 composite fuels As mentioned in chapter 2 i n SPS high amperage (up to 3000 Amp) pulsed DC current is passed through the powder compact

PAGE 43

43 resulting in joule heating at the inter particle contact areas. There is a controversy in the literature whether or not plasma is created due to this spark discharge bet ween the particles [ 53 ] Regardless, intense heat is generated at the particle cont acts which bonds them together in a very short time. With the application of pressure, high density compacts can be produced. Experiment Powder Preparation The uranium dioxide (UO 2.11 ) powder was obtained from Areva, Hanford, WA. The powder was reported to have a bulk density of 2.3g/cm 3 tap density of 2.65g/cm 3 mean diameter of 2.4m, and a surface area from gas absorption method [ 74 ] (BET surface area) of 3.11m 2 /g. Hyper stoichiometric UO 2.25 was produced by heating the starting powder in air at 350 o C for 27 hours to produce U 3 O 8 and then mixing it back with the starting powder (UO 2.11 ) at 30:70 weight ratio. Both SiC whiskers (SiCw) and SiC particles (SiCp) were used in this investigation to produce UO 2 SiC whiskers (3C SiC) were obtained from Advance d Composite Materials, Greer, SC (SC 9D, deagglomerated SiC whiskers) with an aspect ratio, a diameter, and a SiC powder (3C SiC) with the mean diameter of 1m was obtained from Alfa Aesar Inc, Ward Hill, MA. Figure 4 SiC whisker and powder morphologies. The UO 2 and 10vol% (~3.24wt%) SiC were blended with the aid of 2,3 Dihydroperfluoropentane in a SPEX 8000 shaker for 1 hour. After mixing, the blending aid was allowed to evaporate in a fume hood, leaving no residual contamination. This process resulted in homogeneous dispersion of SiC whiskers and powder particles in UO 2 matrix as will be discussed later.

PAGE 44

44 Sintering Both oxidative sintering and spark plasma sintering (SPS) processes were employed. For oxidative sintering green body pellets were made by compressing the blended UO 2 SiC powder at 200 MPa for 10 minutes in a stainless steel die. The die walls were lubricated with a film of stearic acid to prevent fracture of the green body pellets while being taken out. The die and pellet diameters were 12.7mm. The green body pellets were then sintered in an alumina tube furnace with a ramp rate of 2.6 o C/min until the temperature reached 1600 o C where it was held f or 4 hours. To maintain a hyper stoichiometric state, an ultra high purity (UHP) Ar gas atmosphere, with a continuous flowing rate of 2 liter/min, was created in the furnace during the sintering process. Spark plasma sintering was performed in a Dr. Sinte r SPS 1030 system, see Figure 4 2. For SPS, the blended material was loaded into a 12.7mm diameter graphite die. The inner die surface was covered by a thin graphite foil to prevent reaction of the UO 2 powder with the die wall. Cylindrical graphite plugs were inserted into both ends of the die. The end of each plug that contacts the powder was coated with an aerosol of graphite (ZYP Coatings, Inc., Oak Ridge, TN) to prevent reaction of the plug and the powder. The ramp up/down rate was set at 100 o C/min and the hold time at the maximum temperature was set at 5 minutes. An axial pressure of 40 MPa was applied at the beginning of hold time. The maximum sintering temperature was set at 1400 o C, 1500 o C, and 1600 o C for different pellets. After the pellets were si ntered by both oxidative sintering and SPS techniques, they were reduced to stoichiometric UO 2 following the procedure outlined in ASTM C 1430 07. A thermal treatment for the reduction was conducted in a furnace at 800 o C for 6 hours, in a 4%H 2 N gas, with a water vapor atmosphere using a water bath maintained at 35 o C. The UO 2 SiC composite pellets were fabricated with 10vol% SiC at hold temperatures of 1400 o C, 1500 o C, and 1600 o C.

PAGE 45

45 For a comparison of properties between UO 2 SiC composite pellets with those of UO 2 pellets, sintering of UO 2 pellets was also conducted using both SPS and oxidative sintering at the same conditions as described before. While low temperature sintering at 1200~1400 o C yielded poor densities in both methods, the pellets sintered by the SPS process at 1500 o C and 1600 o C yielded 96% theoretical density and were used for comparison of grain size and thermal conductivity. Characterization Methods The characterization of UO 2 SiC composite pellets consisted of density measurements, microstructural analysis using scanning electron microscopy (SEM), analysis of grain size and SiC distribution, and U and Si penetration curves along an interface using EDS, determination of chemical products using XRD, and thermal conductivity measurement s. The density of the UO 2 SiC composite pellets was measured by coating a thin layer of paraffin wax to take into account the open porosity and then using the Archimedean immersion method The paraffin coated pellet was weighed three times in water and th e average density was calculated. For the microstructural observations using SEM, the pellet surfaces were polished with successively smaller grinding medium down to 0.04 micron of colloidal silica. Grain boundary relief was produced by thermal etching at 1340 o C for 4 hours in an Argon atmosphere. Using the secondary electron mode in SEM (JEOL 6335F), 3 5 micrographs of UO 2 SiC pellets were taken and the average grain size was measured in each micrograph by the line intercept method [ 75 ] To determine elemental diffusion ranges, penetr ation curves of U and Si along a line normal to the interface of UO 2 SiC were obtained by Energy Dispersive X ray Spectroscopy (EDS) coupled with high resolution FE SEM.

PAGE 46

46 To determine the reaction products after sintering X Ray Diffraction ( XRD, Philips AP D 3720) was performed on the pellets. To overcome the XRD detection limit of 2~5wt% of chemical compounds, pellets with higher SiC vol% were fabricated. Processing conditions for these UO 2 70vol% (~41.27wt%) SiC pellets were same as those for UO 2 10vol%Si C pellets. The t hermal conductivity of the pellets was measured using an Anter Flashline 3000 system. In this method, the derivation of thermal diffusivity, and specific heat capacity, C p were based on the measurement of the rising temperature on the back surface of a sample caused times each at 100 o C, 500 o C, and 900 o C and the average conductivity at each temperature was calculated. The thermal diffusivity in m 2 /s is given by, 0.1388 L 2 /t 1/2 where L is the thickness of the specimen in m, and t 1/2 is the time in seconds for the rear surface temperature to reach 50% of its maximum value. The specific heat capacity C p is given by, Q /d T m, where Q represents the energy of the pulsed laser beam, which can be determined by comparing the maximum value of the temperature rise with that of a reference, m is the mass of the specimen, and d T is the maximum val ue of the temperature rise. Pyroceram, a glass ceramic material and certified reference, was used as a reference pellet due to its similar conductivity as UO 2 By multiplying density with and C p thermal conductivity was calculated. Result Typical UO 2 pellets were produced via oxidative sintering and SPS and are shown in Figure 4 3. Each pellet was 12.5 mm in diameter and 2 4mm thick. The p ellets were cut and prepared for various characterization methods discussed in the previous section and the results are presented in the following section. Density Porous structures are expected to have lower thermal conductivity compared to fully dense

PAGE 47

47 pellets, and therefore, obtaining high density UO 2 SiC pellets is critical for enhanced thermal conductivity fuel. The measured relative density of the oxidative and SPS sintered UO 2 SiC pellets at v arious sintering temperatures are shown in Figure 4 4. The density of sintered UO 2 10vol% SiC pellets increased with an increase in sintering temperature. However, the highest density among oxidative sintered pellets was still low at 88.91%. On the other hand, all SPS pellets sintered at higher than 1400 o C had higher density between 91.25 and 97.78%. Interestingly, in each of the sintering meth ods both SiC whiskers and p article additions yielded almost the same overall densities of the composite pellets at 1600 o C UO 2 SiC Microstructure and Interface Characterization Microstructure of UO 2 SiC composite pellets Figure 4 5 reveals the distributions of SiCw and SiCp in the composite pellets. Both the whiskers and particles are seen to be uniformly distributed witho ut any agglomeration. This was accomplished with the use of a 2,3 Dihydroperfluoropentane dispersing agent during green compact preparation. UO 2 10vol%SiC pellets sintered at 1500 o C by both processing methods and their micro morphologies are shown in Figure 4 6. It is noted that, in general, a higher level of porosity and poor interfacial contact were observed in oxidative sintered pellet compared to SPS sintered pe llet. In light of the fact that the conducted heat in a pellet can be blocked by the presence of voids at the interface of two grains or poor interfacial contact between the two phases, it is beneficial to produce good interfacial contact for high thermal conductivity fuel. The higher level of porosity in the oxidative sintered pellet shown in Figure 4 6 (a) and (c) also leads to lower density of these pellets shown earlier in Figure 4 4. Conversely, both density and the interfacial contact have improved in the same composition pellets sintered by SPS, see Figure 4 6 (b) and (d). The improved interfacial contact illustrates the

PAGE 48

48 advantage of SPS for sintering high thermal conductivity UO 2 SiC pellets as will be shown later. Diffusion at the UO 2 SiC i nterface Figure 4 7 reveals a high magnification of the UO 2 SiC interface of a SPS sintered pellet processed at 1600 o C for 5 minutes. The separation between the two phases is normally less than 100nm wide. EDS line scanning was performed to determine the uranium a nd silicon concentration profiles across the interface. The concentration profiles, shown in Figure 4 7, revealed approximately 3 m interpenetration of the two elements. This interaction width measurement could be different depending on the electron beam diameter of an instrument. These profiles also illustrate that uranium penetration depth into SiC is around 1.17 m, where as Si pen etration into UO 2 is around 1.83 m. Thus, uranium penetration depth is 36% less than that of silicon. If we assume both materials have similar number of vacancy defects in their lattice structures, since uranium has a greater atomic density and weight than those of s ilicon, the silicon transport would be expected to be greater into UO 2 than that of uranium into SiC. Chemica l r eaction Controlling chemical reactions between SiC and UO 2 during high temperature sintering process is critical to the fabrication o f dense UO 2 SiC pellets, because the formation of various reaction products at the UO 2 and SiC interface may lead to poor thermal properties. A study by Sarma et al. [ 8 ] found that reactions between the two materials could occur at temperature as low as 1370 o C. In our study XRD analysis was used to determine the reaction products at the interface. Figure 4 8 shows two XRD spectra obtained from UO 2 70vol%SiC pellets sintered at 1600 o C for 4 hours and at the same temperature by SPS for 5 minutes hold time. A USi 1.88 peak was clearly seen in the oxidative sintered pellet, and conversely, no such reaction product was detected on the pellet fabricated by SPS. The longer exposure time in oxidative sintering allows

PAGE 49

49 the formation of intermetallics and gas phases such as CO or CO 2 Both could significantly reduce the thermal conductivity of the composite pellet as will be seen in a later section. Grain s ize The average grain size in various composite pellets sintered at 1500 o C by both oxidative sintering and SPS according to their composition is plotted in Figure 4 9. In each pellet, the average grain size was determined from three micrographs from different regions. It is seen that UO 2 without any additives has the highest grain size. The grain size decreased with silicon carbide additions in both sintering methods. This is because when insoluble second phase particles are dispersed randomly in a polycrystalline solid, the grain boundary movement will be pinned by the inclusions resulting in smaller grain size of the matrix. As a result, UO 2 SiCp pellets processed via oxidative and SPS sintering have 62% and 68.5% smaller grains, res pectively, than those of the UO 2 pellet. In general, SPS pellets revealed smaller UO 2 grain size than the oxidative sintered pellets. This result is because of the rapid sintering in the SPS process which provides shorter time for grain growth. A 53.3% red uction in average grain size is observed in the UO 2 pellet made by SPS than in oxidative sintered UO 2 pellet. While the addition of SiC reduces the grain size of UO 2 the SiC particle addition reduced the grain size more severely than the addition of SiC whiskers. The greater amounts of second phase particles increase the pinning effects resulting in the smaller grain size of a composite. While the surface area to volume ratio for the SiC whisker is 4a the ratio for the SiC powder is 6/a where a is the diameter of single particle. Thus for the same 10vol% SiC, the composite with powder particles will have approximately 4.6 times more interface area than the whisker SiC and hence a small UO 2 grain size results with SiC powder.

PAGE 50

50 Thermal Conductivity Figure 4 10 shows the thermal conductivity measurements on UO 2 SiC pellets sintered at different temperatures by SPS and oxidative sintering. The measurements are conducted three times each at 100 o C, 500 o C and 900 o C and the average values were plotted. The avera ge thermal conductivity values of UO 2 from the literature [ 76 ] at various temperatures and the measured values at the above three temperatures are also shown o n the plot. Many observations can be made from this plot. The SPS sintered UO 2 SiC composite pellets have higher measured thermal conductivity than UO 2 pellets. In general, the higher the SPS sintering temperature, the higher the measured thermal conductiv ity. The oxidative sintered composite pellets at 1600 o C exhibited significantly lower conductivity values than previously measured value of pure UO 2 pellets. A maximum thermal conductivity enhancement was observed in UO 2 SiC composites sintered by SPS at 1 600 o C and the increases are 54.9%, 57.4%, 62.1% at 100 o C, 500 o C and 900 o C, respectively, compared to the literature UO 2 value. The SPS sintered composite pellets show a trend similar to that of UO 2 with respect to temperature, i.e., a gradual decrease in c onductivity with increase in temperature. This trend in thermal conductivity is due to the well known phonon phonon scattering phenomena in many ceramic materials [ 77 ] Finally, there was no sign ificant difference in the thermal conductivity values of both UO 2 SiC whisker composites and UO 2 SiC powder particle composites at all temperatures considered in this study. Discussion The sintered pellets by the SPS technique show higher density than oxidative sintered pellets as shown in Figure 4 4. These higher densities indicate that features in SPS such as mechanical pressure and rapid heating can result in fabrication of high density UO 2 SiC composites even w ith significant shorter hold times compared to oxidative sintering method.

PAGE 51

51 In the case of oxidative sintering, the poor interfacial contact (Figure 4 6) and the presence of voids result in low density composites and lead to reduced thermal conductivity. C onversely, the physical contact between UO 2 and SiC grains along the interface is better in the pellets sintered by SPS. The presence of voids at the interface scatter conducted heat in a pellet and therefore, less heat reaches the interior SiC particles. Therefore, thermal conductivity is dramatically reduced in an oxidative sintered pellet. As revealed in Figure 4 5, USi 1.88 peak was clearly seen in the oxidative sintered pellet In the SPS, the pellet stays above 1370 o C only for 9.6 minutes which is a significantly shorter time wh en compared to the 6.9 hours in oxidative sintering. The longer exposure time in oxidative sintering allows for diffusion of chemical species and formation of intermetallics (USi 1.88 ) and gas phases such as CO or CO 2 While intermetallics increase phonon s cattering sites at the interfaces, gas phases may hinder the interfacial contact of UO 2 SiC by forming voids or causing separation. Both these factors could significantly reduce the thermal conductivity of the oxidative sintered pellet as seen in Figure 4 10. It is clear from the results presented in this study that SPS not only offers significantly a shorter sintering time, but also provides a denser UO 2 SiC composite with reduced formation of chemical products better interfacial properties, and above al l, significantly better thermal conductivity than pellets obtained from oxidative sintering. It has been noted in the literature that a smaller grain size yields a lower thermal conductivity [ 78 79 ] However, for reactor applications a larger grain size of UO 2 is preferred due to the potential high diffusivity of fission products along grain boundaries. While the oxidative sintering method provides a larger grain size than SPS, one can increase the UO 2 grain size easily in SPS by simply increasing the hold time for an additional few minutes. Nevertheless, the current results indicate that when SiC is

PAGE 52

52 added to UO 2 the effect of small grain size is not a signi ficant factor when pellets with good density, good interfacial contact and no extraneous chemical products (such as intermetallics) are produced. In the current work SPS seems to produce UO 2 SiC composite pellets with all the above beneficial features. Co nclusion UO 2 SiC pellets fabricated by SPS revealed higher density, better restraint of chemical reaction to form uranium silicide, superior interfacial contact, and smaller grain size compared to those fabricated by oxidative sintering. The SPS technique not only gives higher sintering rate, approximately 30 minutes operating time per pellet, but also allows up to 10% higher density of fabricated pellets compared to pellets prepared by the other method. XRD analysis on UO 2 SiC pellets revealed that the SPS sintering alleviates the concerns of reactions between the two phases that have been reported above 1370 o C. Better interfacial contact between UO 2 and SiC were observed in SEM micrographs. Higher density, restriction on chemical reactions, and good interf acial contact are promising features for enhanced thermal conductivity UO 2 composite pellets. Consequently, despite the small grains, the UO 2 SiC pellets made by SPS revealed enhanced thermal conductivity up to 62.1% compared to UO 2 pellets. On the other hand, the oxidative sintered pellets had lower thermal conductivity than UO 2 pellets and failed to achieve desired density above 95%. The se experimental results suggest that SPS technique is more suitable for sintering of enhanced thermal conductivity UO 2 SiC nuclear fuel.

PAGE 53

53 Figure 4 1. Morphologies of SiC whiskers and powder from manufacturers. Figure 4 2. Dr. Sinter SPS 1030 system and schematic drawing of the sintering chamber. Photo courtesy of Sunghwan Yeo.

PAGE 54

54 Figure 4 3. Fabricated UO 2 SiC c omposite pellets A) oxidative sinter ed composite. B) fabricated by SPS technique Photo s courtesy of Sunghwan Yeo. Figure 4 4. Relative density of UO 2 10vol%SiC composite pellet s sintered by SPS and oxidative sintering at various temperatures.

PAGE 55

55 Figure 4 5. Polished surfaces of high density UO 2 10vol% SiC composites sintered by SPS showing uniform dispersion of SiC A ) Si Cw and B) SiCp Figure 4 6. Microstructure of UO 2 10vol%SiC composites A) and C) sintered by oxidative sintering method at 1500 o C for 4 hours B) and D) sintered by SPS at 1500 o C and 5minutes hold time.

PAGE 56

56 Figure 4 7. EDS line scan across the interface of UO 2 SiC grains i n a composite pellet fabricated by SPS at 1600 o C.

PAGE 57

57 Figure 4 8. Comparison of XRD spectra of UO 2 70vol% SiC pellets sintered by SPS and oxidative sintering at 1600 o C. The peaks contained in dotted circles refer to USi 1.88 phase.

PAGE 58

58 Figure 4 9. UO 2 grain size in composite pellets resulting from the addition of SiC powder particles and SiC whiskers.

PAGE 59

59 Figure 4 10. The measured thermal conductivity values of UO 2 and UO 2 SiC composite p ellets sintered by SPS and by oxidative sintering.

PAGE 60

60 CHAPTER 5 THE INFLUENCE OF SIC PARTICLE SIZE AND VO LUME FRACTION ON THE THERMAL CONDUCTIVITY OF UO 2 SIC COMPOSITES Background In chapter 4, it has been shown that enhanced thermal conductivity of UO 2 10vol%SiC composite fuel pellets can be fabricated by the Spark Plasma Sintering (SPS) technique. In that study, SPS provided higher density composites, better interfacial contact, and reduced chemical reaction between UO 2 and SiC particles, compared to conventional sintering. SPS pellets also revealed a thermal conductivity increase of up to 62.1% at 900 o C compared to the literature value of UO 2 [ 76 ] Because of its unique and superior properties such as high thermal conductivity, low neutron cross section, high melting point, and great chemical stability, Silicon Carbide (SiC) was chosen as the secondary phase in the UO 2 matrix to form heat conducting paths in the ceramic composites. The thermal conductivity of a composite depends on the volume fraction, size, shape, and distribution of second phase particles, as well as the thermal resistance be tween its constituents [ 80 ] The effective thermal conductivity depends on mechanical interfacial contact and interfacial phonon scattering phenomenon which is well known as Kapitza resistance [ 58 ] which was discussed in c hapter 3. Kapitza resistance is always present in a composite even when the interfacial contact is atomically perfect. The influence of particle size and volume fraction of second phase particles on the effective thermal conductivity and its association with the interfacial thermal r esistance have been well documented in literature. However, these principles and models have not been applied to nuclear fuels. Hanada et al., [ 18 ] determined that diamond particle size and volume fraction had a significant influence on the effective thermal conductivity of copper diamond composites. They found th at the composites containing larger diamond particles up to 7.7 m and smaller volume fraction near 1% show higher thermal conductivity.

PAGE 61

61 Hasselman et al., [ 81 ] investigated SiC reinforced alumina composites and found that composites containing larger SiC particles show higher thermal conductivity than that of al umina without SiC. Every et al., [ 19 ] observed that the thermal conductivity of zinc sulphide could be increased by adding large diamond particles and de creased by adding sub micrometer diamond particles due to increased interfacial thermal resistance. Bai et al., [ 24 ] fabricated MoSi 2 SiC composites containing 10, 20, and 30vol% of 0.5 m and 100nm SiC particles and found that the composite showed lower thermal conductivity with decreasing SiC par ticle size and increasing volume fraction due to the interfacial thermal resistance. Chu et al., [ 82 ] sintered Cu Carbon Nanotube (CNT) composites and found that these composites had much lower thermal conductivity than calculated thermal conductivity value based on the rule of mixture which ignored the interfacial resistance. In general, it is found that ceramic composites containing larger particles (low surface to volume ratio) with high thermal conductivity reduc e the interfacial thermal resistance, and hence, increase the effective thermal conductivity of ceramic composites. In the current study, a series of UO 2 SiC composite fuel pellets with different sizes and volume fractions of SiC particles was fabricated using the SPS technique. The thermal conductivities of these composites were measured and compared to the values determined from theoretical formulations available in the literature. During the fabrication process, the sintering p arameters such as hold time, ramp up/down rate, and pressure were kept constant so as to investigate only the effects of SiC particle size and volume fraction on the resulting thermal conductivity of the composite pellets. Experiment s Fabrication of UO 2 S iC Composite Pellets UO 2 SiC pellet fabrication procedure using SPS was described in detail in the previous chapter Therefore, only a brief description is provided here. The uranium dioxide (UO 2.11 )

PAGE 62

62 powder was obtained from AREVA NP, Richland, WA and the SiC powder was obtained from Superior Graphite, Inc., Chicago, IL. The reported SiC particle mean diameters were 0.6, 1.0, 9.0, 16.9, and 55 m. The UO 2 and SiC powders were mixed in a ceramic vial with stainless steel balls and a blending aid 2,3 Dihydro perfluoropentane, and blended in a SPEX 8000 shaker for 1 hour. For each mixing run the SiC mean particle size and the volume fraction of SiC powder in the mixture with UO 2 were varied as shown in Table 5 1 to investigate their effect on the thermal conduc tivity of the resulting UO 2 SiC composite pellet. The SiC particle sizes used in this study are most widely used in various applications and are also available as high purity (>98%) powders, and hence were obvious choice in our study. SiC particles with 1 m size at 5, 10, 15, and 20vol% were chosen to fabricate UO 2 SiC composite pellets. This volume range was selected because 20vol% is the maximum range where the available models are valid (to be described in the next section) so that we can compare the res ulting thermal conductivity of the composites with model prediction. Also, excessive addition of SiC in nuclear fuel is unrealistic due to the exorbitant cost and stringent regulation of U 235 enrichment process which is necessary to compensate reduced fis sile isotope in UO 2 SiC composite fuel. SiC particles dispersed in UO 2 powders were then sintered using a Dr. Sinter SPS 1030 system at 1350 o C and 1450 o C for 5min in a vacuum (~30mTorr). The ramp up/down rate and mechanical pressure at the maximum sintering temperature were held constant at 100 o C/min and 36MPa, respectively. The heat treatment procedure described in ASTM C 1430 07 was conducted on the sintered composite pellet to reduce UO 2+X to stoichiometric UO 2.0 which is known to have the opti mum thermal properties [ 1 ] The ramp up/down rate and maximum temperature were set at 2.6 o C/min and 800 o C, respectively. The heat treatment was performed in a Lindberg alumina tube furnace using 4%H 2 N 2 gas with a dew point ma intained at 35 o C.

PAGE 63

63 Characterization Methods The weight of each pellet in air and water was measured and the average density was calculated from three weight measurements per pellet using the Archimedes principle. The measured density of the composite was t hen compared with theoretical density obtained from the rule of mixture [ 83 ] (5 1) Where , and V p are the densities of UO 2 and SiC, and the SiC volume fraction, respectively. The microstructure of the fabricated composite pellets were observed using a scanning electro n microscope (SEM, JEOL JSM 6335F). The pellets were metallographically polished with successively smaller grit SiC abrasive paper and finally with 0.06 m colloidal silica. The surface was thermally etched at 1340 o C in Ar atmosphere for 4 hours to reveal the grain boundaries of UO 2 matrix in the composite pellets. The measurement of thermal diffusivity was carried out at 100, 500, and 900 o C using a laser flash instrument ( AnterFlashline 3000 ) with a Xenon discharge pulse for 1 s duration. Three measurements were performed at each temperature on each pellet and the average diffusivity was obtained. The specific heat capacity of UO 2 SiC composite pellet was calculated using the Neumann Kopp rule [ 84 ] i. e., ( 5 2 ) Where , and are theoretical specific heat capacities of UO 2 and SiC, and weight fraction of SiC particles, respectively, at a specific temperature. and at 100 o C, 500 o C, and 900 o C are listed in Table 5 2. The thermal conductivity, K, of composite pellets was then determined from the relation. (5 3)

PAGE 64

64 Where, and are the thermal diffusivity and density of the composite, respectively. Results and discussion Size Effect of SiC Particles on UO 2 5vol% SiC Composite Properties The micro morphologies and thermal properties of UO 2 5vol%SiC composite fuel pellets contai ning SiC particles with five different sizes ( Table 5 1) were examined. Fig ure 5 1 shows the microstructures of these composites where the SiC particles appear black and the brighter area indicates the UO 2 matrix. The SiC particles appear to be homogeneous ly dispersed in the UO 2 matrix in all the compo sites. However, as shown in Fig ure 5 1(e), in the composite containing 55 m SiC particles, distinct radial micro cracks were observed originating at the interface between a SiC particle and UO 2 matrix and prop agating towards another SiC particle. The interfaces between the UO 2 matrix and SiC particles in UO 2 5vol%SiC composite pellets with different sized SiC grains are shown in Fig ure 5 2. The micro cracks emanating from the SiC pa rticles are clearly seen in F ig ure 5 2 (c), (d), and (e) indicating that micro cracks evolve in composites with SiC particles larger than 9m in size. However, the micro cracking is less severe in composites with SiC particles of size 9m and 16.9m compared to the composite containin g 55m diameter SiC particles. No visible cracks in the micro structure was seen in the composite pellets with smaller size SiC particles. It is also seen that with increasing particle size there is a larger separation between the SiC particle and the UO 2 matrix. Fig ure 5 3 shows the interfacial debonding between UO 2 and SiC particles in each composite with the three largest size SiC particles. While the interfacial contact between UO 2 grains and 9 m SiC particle is fairly good, a visible gap is observed a t the interface between UO 2 grains and SiC particle when the particle size is 16.9m or greater. Micro cracking and interfacial debonding occur in various composites during the sintering process due to a mismatch in coefficients of thermal expansion (CTE) between the matrix and

PAGE 65

65 the second phase particles [ 85 91 ] Lu et al., [ 90 ] studied thermal matrix cracking in various intermetallic composites and concluded that a critical particle size exists under which micro cracking was suppressed. Todd et al., [ 85 ] investigated alumina 20% SiC composites and found thermal cracks in only the composites containing SiC particles larger than 9 m. Fu et al., [ 91 ] fabricated polypropylene (PP) calcium carbonate (CaCO 3 ) composites and found interfacial debonding in composites containing larger than 0.1m diameter CaCO 3 particles. The relevant thermal properties for UO 2 and SiC ar e shown in Table 5 2. The thermal expansion coefficient of UO 2 is more than twice that of SiC so that the matrix expands into the particles during the cooling process forcing the SiC particles into compression. The larger the particle size, the more will b e the induced compressive stress due to the lower surface area to volume ratio of the larger particles. When the stress intensity at the interface exceeds the grain boundary toughness of matrix material, spontaneous microcracking is initiated from the inte rface into the matrix in a ceramic composite [ 85 ] The induced internal stress caused by a mismatch in CTE of constituents in a composite also leads to a partial interfacial debonding [ 92 ] The degree of interfacial debonding is dependent on the level of mismatch in CTE, elastic properties of the constituents, the temperature range of cooling process, and the energy required to create a new surface [ 93 ] Because these thermal cracks and the interfacial debonding in composite pellets obstruct the pathway for heat conduction, extensive cracking and poor interfacial contact obviously lead to lower thermal conduct ivity. The measured density of UO 2 SiC composite fuel pellets containing 5vol% SiC particles but of different sizes are shown in Table 5 1. The densities of all composite pellets are near 95%TD and appear to not be dependent on the SiC particle size. Beca use the thermal conductivity is directly proportional to the density as seen in Equation 5 3 and the density of the composite pellet is not dependent on the SiC particle size, the measured pellet thermal

PAGE 66

66 conductivity will mostly depend on the size of SiC particles as will be discussed in the following paragraph. Fig ure 5 4 shows the temperature dependence of the measured thermal diffusivity for UO 2 5vol% SiC composite pellets containing various sizes of SiC particles. The red line refers to the literature value [ 76 ] of 95% dense UO 2 In general, the UO 2 thermal diffusivity decreases with temperature due to increased phonon phonon scattering at higher temperatures [ 94 ] This trend is maintained in the thermal diffusivity of UO 2 5vol%SiC composite pellets as well. In general, the larger the SiC particle size the lower the thermal diffusivity. However, the thermal diffusivity of the composite pellets containing 55 m SiC particle shows a significantly lower thermal diffusivity than the literature UO 2 value due to extensive micro cracks and severe interf acial debonding as shown in Fig ure 5 2 and Fig ure 5 3. Fig ure 5 5 shows the thermal conductivity determined by Eq uation 5 3 using the me asured thermal diffusivity (Fig ure 5 4), calculated specific heat capacity, and the measured density (Table 5 1) for composites with different SiC size particles at three temperatures. The specific heat capacities of UO 2 and SiC are shown in Table 5 2 and were utilized to determine that of the UO 2 5vol% SiC composite. Using Equation 5 2, the specific heat capacitie s at 100, 500, and 900 o C were calculated to be 266.5, 317.5, and 328.5 J/kg K, respectively. While the composite pellets containing 0.6, 1.0, and 9.0m diameter SiC particles showed enhanced thermal conductivity, the composite pellets containing 16.9 and 5 5m diameter SiC particles revealed lower thermal conductivity than a UO 2 pellet. The SiC particle size dependence of thermal conductivity at various temperatures is shown in Fig ure 5 5. While marginal reduction in thermal conductivity in the pellets containing 16.9m SiC particles is noted, the reduction in pellets containing 55m diameter SiC particles is particularly large; a decrease of 21~28.3% depending on the testing temperature. These composite pellets with poor thermal conductivity and

PAGE 67

67 containing large SiC particles are the same pellets which exhibited micro cracks and interfacial debonding next to the large SiC particles as shown in Fig ure 5 2 and Fig ure 5 3. These observations support the hypothesis that micro cracking and interfacial debonding are responsible for the reduction in thermal conductivity of UO 2 SiC composite containing large SiC particles. T he 55m diameter SiC particle composite pellet has the most severe micro c racks and the largest interfacial debonding resulting in a greatly reduced thermal conductivity. Based on these observations, a composite pellet with SiC particles smaller than 16.9m diameter is recommended for enhanced thermal conductivity. The Effect o f Volume Fraction of SiC Particles To understand of the influence of volume fraction of SiC particles on the thermal properties of UO 2 SiC composite pellets, we have chosen one SiC particle size (1 varied the volume fraction ( Table 5 1) at 5, 10, 15 and 20%. All the other variables were kept constant during the sintering process. Micro structures of the four composite pellets revealing homogeneously dispersed 1 m SiC particles are shown in Fig ure 5 6. With increase in volume fraction, particle parti cle interaction is noted as seen in Fig ure 5 6 (c) and (d). Fig ure 5 7 reveals a decrease in the relative density of UO 2 SiC composite pellets with increasing SiC volume fraction. This trend has been reported in literature for many composites [ 21 23 ] It is rationalized that increase in particle volume fraction hinders the consolidation of matrix grains. Maekawa et al., [ 23 ] suggested that with increasing volume fraction of particles pores remain locked around particles and eventually lead to f orm pore channels decreasing the three dimensional connectivity of a matrix material to lower relative density. The measured thermal diffusivity of the composite pellets containing various volume fractions of 1 m size SiC particles is shown in Fig ure 5 8. It is seen that the higher the volume fraction of SiC particles, the higher the thermal diffusivity of the composite. This trend indicates

PAGE 68

68 that inclusion of higher thermal conducting particles into UO 2 matrix increases the diffusion of heat energy in the c omposite provided good interfacial bonding is maintained. Fig ure 5 9 reveals t he calculated ( Equation 5 2) specific heat capacity (C p ) of UO 2 SiC composites containing various volume fractions of SiC particles at 100, 500, and 900 o C. As the SiC volume f raction and temperature increase the specific heat capacity also increases. This is because larger number of molecular energy states are available at higher temperature and the specific heat capacity follows Neumann Kopp rule ( Equation 5 2) [ 84 ] Also note that initially there is a significant increase in specific heat from 100 o C to 500 o C but this increase in C p is lower from 500 o C to 900 o C. Hasselman and Johnson [ 62 ] and recently Nan et al., [ 65 ] provided an expression for calculating the effective thermal conductivity of a composite. Their model is based on the effective medium approximation [ 95 ] which includes the influence of size, v olume fraction, and shape of second phase particles, as well as interfacial thermal resistance. For a composite containing spherical shaped particles dispersed homogeneously in a matrix material, the effective thermal conductivity is given by, (5 4) Where is the effective thermal conductivity, subscripts p and m are particle and matrix, respectively, V p is the volume fraction of particles, a is the radius of particle, and h c is the interfacial thermal conductance. The reported interfacial thermal conductance for Al and SiC covers the range of 2.8 4.9 10 8 W m 2 K 1 [ 26 ] So far, the value of accounting for the UO 2 SiC interface has not been reported to our knowledge. However, it can be estimated us ing the acoustic mismatch model of Swartz and Pohl [ 25 ] where the interfacial thermal conductance is giv en by,

PAGE 69

69 (5 5) Where is density, is the specific heat capacity of matrix, is phonon velocity, and subscripts p and m refer to particle and matrix, respectively. The phonon velocities of UO 2 matrix and SiC particle can be estimated using the following equation [ 96 ] (5 6) Where and are the longitudinal and transverse phonon velocities. The referenced UO 2 longitudinal and transverse phonon velocities are 5552.7 and 2841.8m/s [ 97 ] respectively, and those of SiC are 11800 and 7600m/s [ 98 ] respectively. The estimated phonon velocities of UO 2 and SiC using Equation 5 6 are 3272.8 and 8470.9m/s, respectively. Utilizing these values for and =10960 kg/m 3 =3200 kg/m 3 of UO 2 at three different temperatures as listed in Table 5 2 and using Equation 5 5 we obtain =1.69 x 10 8 at 100 o C, 2 x 10 8 at 500 o C, and 2.06 x 10 8 W/m 2 K at 900 o C, respectively. Given these input parameters with particle volume fraction (0.05 0.2), radius (0.5 m), and and as listed in Table 5 1 into Hassel man and Johnson model ( Equation 5 4 ), the comparison between experimentally obtained and theoretically calculated effective thermal conductivity is shown in Fig ure 5 10. Experimentally measured density (Fig ure 5 7), thermal diffusivity (Fig ure 5 8), and the calculated specific heat (Fig ure 5 9) were utilized to determine the experimental effective thermal conductivity using Equation 5 3 The higher the volume fraction of SiC particles, the higher the thermal conductivity of the composite. The average incr ease in thermal conductivity with the addition of 5, 10, 15, and 20vol% of SiC particles are 14.23, 26.44, 43.22, and 49.84%, respectively. Considering the error bar, great agreement between experimentally determined and theoretically calculated effective thermal conductivities

PAGE 70

70 is seen for the composite pellets containing 5, 10, and 15vol% of SiC particles. The agreement is also much better at higher temperature (900 o C) than at lower temperature (100 o C). The lower experimental thermal conductivities of 20v ol% SiC can reflect the relatively lower densification and the abundance of particle particle interactions of the composite containing higher volume fraction of SiC particles. As shown in Fig ure 5 7, the composite containing 20vol% of SiC has only 94.41% r elative density and seemed to be responsible for decreasin g the thermal conductivity ( Equation 5 3). The interaction between SiC particles is seen in Fig ure 5 6 and it is more abundant with increasing SiC volume fraction. The particle particle interaction is not accounted for in Hasselman and Johnson model due to the complexity of the phenomena [ 65 ] Observation of the i nterfacial con tact between SiC particles ( Fig ure 5 11) indicates that pores are predominantly located at the interface reducing the overall thermal conductivity due to the phonon scattering. Non ideal shape of SiC particles and thermal diffusivity measu rement error also can reflect the difference between experimental and theoretical effective thermal conductivities. The irregularities in SiC particle shape can be clearly seen in Fig ure 5 1 and Fig ure 5 2. Because the H asselman and Johnson model (Eq uation 5 4) only accounts for spherical shaped secondary particles, a discrepancy between theoretical model and experimental measurement is expected. Thermal diffusivity measurement error also may contribute to the difference between experimental and theoretical effective thermal conductivities. The differ ence between actual measurement temperature and the set up temperature and change in the density value of pellet at different temperatures may cause some error to the experimental thermal conductivity. Regardle ss, we observed relatively good agreement between the experimental and theoretical effective thermal conductivities of UO 2 SiC composites. This result supports that UO 2 matrix and 1 m SiC particles are mechanically well contacted in UO 2 SiC composites thus

PAGE 71

71 improving the effective thermal conductivity. Moreover, both experiments and the theoretical model revealed that higher effective thermal conductivity is obtained with increasing SiC volume fraction. However, the utilized powder blendi ng procedure and SPS process conditions, see Table 5 1, are only valid for fabrication of UO 2 SiC composites containing up to 15% SiC particles by volume. More acceptable powder mixing procedure such as a longer time blending and better consolidation proce ss such as SPS sintering at higher temperature, pressure, or longer hold time may be required to produce UO 2 SiC composite containing more than 15vol% of SiC at high densities. If the composite possesses reduced particle particle interactions and higher de nsity by doing so, probably higher effective thermal conductivity close to the predictions of Hasselman and Johnson model [ 62 ] could be obtained even in composites with larger particle sizes and at higher volume fractions. Conclusion The microstructure and thermal properties such as thermal diffusivity, specific heat capacity, and thermal conductivity of UO 2 SiC composite fuel containing various SiC particle sizes and volume fractions were investi gated. The SPS technique was utilized to fabricated high density composite pellets in a relatively short time of 5mins at 1350 1450 o C. While the composite pellets containing 0.6, 1 .0 9 .0 m diameter SiC particles showed higher thermal conductivity, those p ellets containing 16.9 and 55 m diameter particles exhibited lower thermal conductivity than the literature UO 2 values. In the latter two composite pellets, extensive micro cracks and interfacial debonding formed due to the mismatch in CTE between UO 2 and SiC and were attributed to be responsible for the observed low thermal conductivity of the composites. The composite pellets containing higher volume fraction of SiC particles revealed lower density, higher diffusivity and specific heat, and higher thermal conductivity. Good agreement was observed between experimentally determined thermal conductivity of the composites containing

PAGE 72

72 5, 10, and 15vol% of 1 m size SiC and theoretically calculated thermal conductivity based on Hasselman and Johnson model [ 62 ] However, the composite containing 20vol% of 1 m size SiC showed lower thermal conductivity than theoretical value due to the lower density and particle particle interaction recognized in microstructu re observation. Irregular shape of SiC particles and thermal diffusivity measurement error also possibly contributed to the difference between experimental and calculated thermal conductivities. Consequently it is concluded that SiC particles smaller than 16.9 m to suppress micro cracking and higher SiC volume fraction to form enough heat conducting paths are required to increase thermal conductivity of UO 2 SiC composite fuel pellets.

PAGE 73

73 Table 5 1 Details of SiC particle size, volume fraction, and sintering conditions in the SPS.* SiC particle mean diamter ( m) SiC volume fraction (%) Maximum sintering temperature ( o C) % TD of the composite pellet SD 0.6 5 1350 95.25 0.24 1 5 1350 95.27 0.3 1 5 1450 96.81 0.39 1 10 1450 96.63 0.35 1 15 1450 95.14 0.23 1 20 1450 94.41 0.3 9 5 1350 95.15 0.09 16.9 5 1350 94.75 0.17 55 5 1350 95.1 0.13 Hold time=5mins; ramp up/down rate=100 o C/min; pressure=36MPa. Table 5 2 Thermal properties of UO 2 and SiC. Materials Thermal expansion coefficient (K 1 ) [ 76 99 ] at 25 o C Specific heat, C p (J/kg K) [ 55 76 ] Thermal Conductivity, K ( W / m K) [ 55 76 ] 100 o C 5 0 0 o C 9 00 o C 100 o C 5 00 o C 9 00 o C UO 2 9.93 10 6 258.17 304.62 314.17 6.83 4.28 3.01 SiC 4.4 10 6 815 88 11 25 5 12 34 28 273.6 4 136.4 2 85.5 3

PAGE 74

74 Fig ure 5 1 Microstructure of UO 2 5vol%SiC composite pellets c ontaining various diameters of SiC particle A) 0.6 m B) 1 m C) 9 m D) 16.9 m E) 55 m Note microcracks originating from large size SiC particles in E

PAGE 75

75 Fig ure 5 2 SiC particles in UO 2 5vol%SiC composite pellets with various mean diameter s. A) 0.6 m B) 1 m C) 9 m D) 16.9 m E) 55 m. The micro cracks in matrix and between two SiC particles are identified by arrows.

PAGE 76

76 Fig ure 5 3 Interfacial debonding in UO 2 SiC composites with various size s of SiC particle. A) 9 m B) 16.9 m C) 55 m

PAGE 77

77 Fig ure 5 4. Thermal diffusivity of UO 2 5vol%SiC composite pe llets with various SiC particle sizes as a function of temperature.

PAGE 78

78 Fig ure 5 5 Thermal conductivity of UO 2 5vol%SiC pellets with variou s sizes of SiC particles at the selected temperatures. The dotted lines refer to UO 2 literature values at each temperature.

PAGE 79

79 Fig ure 5 6 Microstructure s of UO 2 SiC composites containing 1 m SiC particles with various volume fractions. A) 5 %. B) 10 %. C) 15 %. D) 20vol%. Bright and dark areas refer to UO2 matrix and SiC particles, respectively. The red circles indicate particle particle interactions

PAGE 80

80 Fig ure 5 7 Relative density of UO 2 SiC composite pellets containing various fractions o m size SiC particles

PAGE 81

81 Fig ure 5 8 Thermal diffusivity of UO 2 SiC composite pellets containing various volume fractions of size SiC particles as a function of temperature

PAGE 82

82 Fig ure 5 9 Temperature dependence of specific heat capacities of UO 2 SiC composite pellets containing various volume fraction of size SiC particles.

PAGE 83

83 Fig ure 5 10 Calculated and experimentally determined thermal conductivities of UO 2 SiC composites at the selected temperatures with various volume fraction SiC particles. The UO 2 literature values at each temperature are indicated at Vf =0. Fig ure 5 11 Interfacial contact in a UO 2 20vol%SiC composite. Interfacial porosities are indicated by arrows.

PAGE 84

84 CHAPTER 6 MECHANICAL PROPERTIE S AND INTERNAL STRESS MEASUREMENT O F UO 2 SIC COMPOSITES Background In the previous chapter w hen the SiC volume fraction was maintained at 5vol%, t he influence of five different mean diameters of SiC spherical particles (0.6 1, 9, 16.9, and on the therma l conducti vity of composites w ere investigated. T he extensive UO 2 micro crack ing and UO 2 SiC interfacial debonding w ere clearly observed in composites containing large SiC particles (16.9 and 55 ) due to the difference in CTEs ( Figure 5 2) The microstructure of composite containing 9 SiC particles showed relativel y good interfacial bonding ( Figure 5 3) and a few little microcracks ( Figure 5 2). While the resulting thermal properties of the composite containing 16.9 and 55 SiC particles decreased, the thermal properties of composite containing 9 SiC particles showed similar to those of other composites containing 0.6 and 1 SiC particles ( Figure 5 4 and 5 5). Therefore, it is difficult to address the feasibility of 9 SiC particles to produce UO 2 SiC composite fuel s obtaining sound microstructure M icro structural flaws such as microcracks in a nuclear fuel can be a significant problem r eleasing fission gas es during the operation of a nuclear reactor. The production of inert gases such as Helium ( 2 He), Krypton ( 36 Kr), and Xenon ( 54 Xe) by the fission process of Uranium constituents about 15% of total fission product [ 100 ] These gas atoms may fo rm bubbles in the nuclear fuel or be released from the fuel when they reach any open porosity that is connected to a free volume. The m icrocrack is one of major open porosities and the large amount of fission gas can be released through micro cracks. Once t he gas is released to a free space in fuel pin, the pr essure inside fuel pin increase s leading to higher degree of cladding rupture and hence a reduced safety margin. Therefore, more detailed investigation on UO 2 SiC composites

PAGE 85

85 containing various sized SiC particles to verify the onset of micro cracking and the SiC particle critical size where micro cracks can be suppressed. Because microcracks are not always revealed in a polished surface and the thermal property meas ure ments are not very sensitive to mic rocracking, the onset of microcracking can be monitored by the measurement of mechanical properties and internal stress. Pan et al. [ 101 ] investigated SiC TiB 2 composites and found a decrease in Young s modulus due to the microcracking of a composite containing 4.7 sized TiB 2 particles. Green [ 102 ] fabricated Al 2 O 3 ZrO 2 composites and measured Young s modulus and hardness to find the critical ZrO 2 particle size where microcracking can be suppressed. In another study, Watts et al. [ 31 ] examined microcracking in hot pressing sintered ZrB 2 SiC composites. They showed that the strength, modulus, and hardness of composites decreased abruptly due to microcracking when the incorporated maximum SiC partic le size exceeded 11.5 m Todd and Derby [ 85 ] investigated Al 2 O 3 20% SiC composites and fou nd micro cracks in only the composites containing SiC particles larger than 9 m They especially measured the internal stress of SiC particles in composites and showed that matrix microcracking originating from t he interface between Al 2 O 3 and SiC released the compressive stress of SiC particles. In addition to the observation of microcracking, the investigation of mechanical properties of UO 2 SiC composites is necessary to find additional benefits of SiC inclusio n and clarify the values of UO 2 SiC composite fuel. Even though sound mechanical properties of nuclear fuel are essential for successful operation of nuclear reactor [ 1 ] n o reported literature exists for the investigation of U O 2 SiC composites due to the manufacturing difficulty of such composites. Since we have successfully fabricated high density UO 2 SiC composites using SPS technique, it is rationalized to exam their mechanical properties.

PAGE 86

86 M ost mechanical properties of UO 2 fuel are expected to be increased by SiC addition based on other research publications regarding enhanced mechanical properties of composites by adding SiC particles. SiC has excellent mechanical properties such as har dness (20 30GPa) young's modulus ( 40 0 500 GPa) and shear modulus ( 160 190 GPa) [ 103 ] due to its high bonding strength and unique lattice structure There fore, SiC is often utilized to the fabrication of c omposites to fulfill required mechanical properties. Chamberlain et al., [ 104 ] increased the strength and toughness of ZrB 2 from 565MPa to more than 1000MPa and from 3.5MPa to 5.3MPa, respectively, by adding 20 30vol% SiC particles Chou et al., [ 30 ] found that alumina increased young's modulus by near 2 0% by adding 30vol% SiC platelets. In this study, we investigated mechanical properties of UO 2 SiC composites containing various sizes and volume fractions of SiC particles. The mechanical properties consist of Vickers and Knoop hardness, and Young's mod ulus T he internal stress of different sized SiC particles in UO 2 matrix was measured using a Raman spectroscopy. Experiments Sample P reparation The uranium dioxide (UO 2.11 ) powder was obtained from AREVA NP, Richland, WA and the SiC powder was obtained from Superior Graphite, Inc., Chicago, IL. Reported mean diameters of the obtained SiC powders were 0.6, 1, 9, 16.9, and 55 m. To en sure the particle mean diameters and size distribution Laser Diffraction Particle Size Analyzer ( LS 13 320 Beckmen Coulter, Inc., Brea, CA ) was performed. Th is analysis w as conducted three times per SiC powder. P article mean size and particle size distribution were determined using volume distribution as shown in Table 1. In this study, we utilized previously fabricated seven UO 2 SiC composites described in chapter 5 containing the five different sized SiC particles ( 0.6 1, 9, 16.9, and with 5vol%

PAGE 87

87 and two different volume fractions (5 and 10% ) of 1 SiC The compo sites containing 15 and 20vol% of 1 SiC particles utilized in chapter 5 were found to have interact ions between SiC particles ( Figure 5 11) and lower den sity than other composites ( Table 5 2). Because mechanical properties are sensitive to the density of composite and SiC particle interaction, it was required to produce higher density composites with reduced interaction between SiC particles. Basic SPS fabrication method is same as previous chapters except the blending procedure and the sintering temperat ures. H ence, only these modifications are briefly described here. The UO 2 and SiC powder mi xing was performed for 2hours which is 1hour longer than that of previous chapter to reduce the interaction between SiC particles in the composites. After mixing, h igher SPS sintering temperatures than those of previous chapter were set at 1510 and 1550 o C respectively for fabricating higher density composites. The resulting average densities of these composites containing 15 and 20vol% of 1 SiC particles increased to 96.54 and 96.51%TD, respectively. I n addition to that, UO 2 pellets without any SiC addition were also fabricated using SPS at 1350 and 1450 o C to compare the mechanical properties of UO 2 with UO 2 SiC composites. Table 2 lists the composition of fabricated composites and UO 2 with sintering conditions SiC particle size and volume fracti on, and relative densit y. While SPS processing conditions such as mechanical pressure, hold time, and ramp up/down rates were set at 36MPa 5mins, and 100 o C/min, respectively, during all fabrication processes only sintering temperature was varied at each sample to maintain similar resulting densities. Relative densities of the listed composite and UO 2 cover the range of 95~97%. The fabricat ed composites and UO 2 were mechanically polished with successively smaller grit SiC abrasive paper, Diamond particles, and finally 0.06 m colloidal silica. Thermal etching at 1 35 0 o C for 4 hours in UHP Ar atmosphere was performed using a furnace ( Lindb e rg 1700 o C tube furnace ) to reveal grain boundaries of composites. The microstructure of composites was

PAGE 88

88 observed using Scanning Electron Microscopy ( SEM, JEOL JSM 6335F ). The UO 2 grain size was measured three times at different locations for each composite usi ng the line intercept method [ 75 ] Mechanical Testing The h ardness of the composites was measured using both Vickers and Knoop indentations (Model Tukon2100B Instron, Jacksonville, F L ). The measurements were conducted at 500g applied load with a dwell time of 12 seconds. At least 10 measurements were performed at each indentation with different locations and the average hardness values were obtained and utilized in this study. Ultrason ic measurement system (Model 5072PR Olympus, Waltham MA) was performed on each composite to measure both longitudinal and shear velocities to determine Young s modulus, respectively. Five measurements were performed on each composite and the average va lu e was utilized in this study. Raman Spectroscopy Raman spectroscopy (InVia Raman Microscope System, Renishaw, Hoffman Estates, IL) was performed on the composites containing different size SiC particles to measure the internal stress of the SiC particles as a function of particle size. The Raman spectrometer consisted of a Si laser (532 nm) to excite samples, a single spectrograph, and optical microscope (a Leica microscope with a XYZ mapping stage). The spectrometer was initially calibrated wit h a Si (100) standard u sing a Si band position at 520 cm 1 The desired scattering area within SiC grains or at the interface between UO 2 and SiC w ere selected using a 100 x objective lens A maximum power of the laser was set at 25mW. Raman spectra were corr ected from five different SiC particles in each composite and the average peak p osition was used in this study.

PAGE 89

89 Results and D iscussion Mechanical P roperties Hardness of UO 2 SiC com p osites Both Vickers and Knoop hardness, and Young s modulus were measured for all composites and are summarized in Table 6 2. Fig ure 6 2 shows the hardness as a function of the particle size and volume fraction of SiC. In Fig ure 6 2 (a), in general the hardness decreased with increasi ng SiC particle size a nd exhibit ed a discontinuous change, once the SiC particle size excesses 1 From 1 to 9 SiC, the Vickers hardness decr eased by almost 23% (6.9 to 5.3 GPa). A similar trend can be seen in the Knoop hardness; the Knoop hardness decreased by 13% (5.6 to 4.7 GPa) from 1 to 9 SiC. In Fig ure 6 2 (b), with increasing the volume fraction of SiC particles, both Vickers and Knoop hardness values increases. Whil e the Vickers hardness increase d up to 9.2GPa with 20vol% of 1 SiC particles the Knoop hardness i ncrease d up to 8.1GPa. When compared to pure UO 2 these enhancements are 53 and 58%, respectively. Figure 6 3 (a) and (b) show typical shapes of the Vickers and Knoop indentations, respectively. The Vickers indentation is a square pyramidal diamond with a n included angle of 135 o in both direction The Knoop indentation is a elongated pyramidal diamond along one diagonal with an included angle of 170 o on the long axis and 130 o on the short axis. The sharper Vickers indentation more effectively open s the pre existing microcracks or interfacial debonding in the composite than the Knoop indentation. The refore, Vickers hardness decrease s more abruptly than that of the Knoop hardness when UO 2 SiC composites contain microcracks. A s imilar behavior ha s been reporte d in Al 2 O 3 ZrO 2 and ZrB 2 SiC systems [ 31 102 ] Greater than ~10vol% of ZrO 2 addition in Al 2 O 3 ZrO 2 composite resulted in decreased hardness due to mi crocracks [ 102 ] The Vickers hardness of ZrB 2 30vol%SiC composites containing larger SiC particles than 11.5 decreased more than that of Knoop harndess due to the sharper indentation geometry [ 31 ]

PAGE 90

90 In Figure 6 2, the relatively larger error bars of the composites containing large sized SiC particles (9, 16.9, and 55 ) in the Vick ers hardness values may be caused by the limit of the Vickers indentation area. As shown in Figure 6 3 w hile the average diagonal length Vickers indentation covers the 15~20 that of the Knoop indentation covers 60~80 in the long axis and 5~9 in the short axis. The average covered area of Vickers indentation is 160~ 180 2 and that of Knoop indentation is 270~280 2 The average length between SiC particles in the composite containing 1 SiC particles is less than 5 that of composite containing 9, 16.9, and 55 are around 10, 15, and 27 res pectively. Because the Vickers indentation covers a relatively small area, the indentation area may not include any SiC particles in the composites containing large SiC particles. Then, only UO 2 hardness values can be obtained decreasing the average hardne ss value and increasing the error bars. Bulk, shear, and Young's moduli of UO 2 SiC composites Hashin and Shtrikman [ 105 ] derive d an expression for c alculating the effective bulk and shear moduli of a composite. In fact, t he Young s modulus of composites containing finite concentration of second phase spherical particles is very difficult to calculate because detailed elastic field analysis considering interface continuity condition is required [ 106 ] T herefore, t hey had to assume the composite material was statistically isotropic and obtained homogeneous interfacial boundary conditions. Based on variational principles in terms of the elastic polarization tensor described in [ 107 ] they obtained upper and lower boun ds on the effective bulk and shear moduli of a composite containing two phases [ 105 ] ( 6 1 ) ( 6 2 )

PAGE 91

91 Where K is bulk modulus, and are upper and lower effective bu lk moduli, respectivley, V is volume fraction, G is shear modulus, and subscripts p and m refer to second phase particle and matrix, respectively ( 6 3 ) ( 6 4 ) Where and are upper and lower effective shear moduli respectively. These bounds follow typical relation of Young's modulus of E to K and G ( 6 5 ) ( 6 6 ) Where and are upper and lower effective Young s moduli, respectively. The referenced UO 2 bulk and shear moduli are 184 and 76.3GPa [ 108 ] respectively, and those of SiC are 308.6 and 179GPa [ 109 ] respectively. The estimated upper and low er bounds on the effective bulk, shear, and Young s moduli of UO 2 SiC composite using e quation s above with different volume fractions (5 20%) are listed in Table 3. To further investigate the influence of SiC addition on the Young's modulus and verify whether the experimentally determined values are in agreement with the calculated values using the Hashin and Shtrikman model, the Young's modulus values were examined using measured longitudinal and shear velocities. The following equa tions were utilized to determine Poisson's ratio, and Young's modulus, [ 110 ] ( 6 7 ) ( 6 8 )

PAGE 92

92 Where and are shear and longitudinal velocities, and is the density of composite. With measured density of composite, the Poisson's ratio and Young's modulus of composites were determined. E xperimentally measured and theoretically calculated Young s modul i as a function of SiC particle size and volume frac tion followed similar pattern as the hardness as shown in Figure 6 4 (a) and (b), respectively The upper and lower bounds of calculated Young's moduli using Hashin and Shtrikman model [ 105 ] are presented as blue and red dash lines in Figure 6 4 For small SiC particle size s (0.6 and 1 ) the Young s modulus remained constant at 2 09 GPa which value is in an excellent agreement with upper and lower bounds of calculated Young's moduls as shown in Figure 6 4 (a) In contrast with increasing SiC particle size, the elastic modulus markedly decreases to 130GPa when the mean SiC particle size is 55 Because the volume fraction of SiC par ticles is maintained at 5vol%, if there w ere a change in SiC particle size alone wi thout microstructural defects the resulting Young s modulus wou ld not be changed. However, the decrease in Young s modulus for the composite containing 55 particles is al most 39% w hen compared to composites containing 0.6 and 1 particles The starting point where the Young s modulus decreased is consistent with what has previously been observed in Vickers and Knoop hardnesses. These observations support the hypothesis that microstructural defects such as microcracks and interfacial debonding take place in the composites cont aining large r SiC particles than 1 Figure 6 4 (b) shows the comparison between experimentally obtained and theo retically calculated e ffective Young s moduli with different SiC volume fraction s It is clear ly seen that the experimentally determined Young s modulus increase s linearly with increasing 1 SiC volume fraction in UO 2 SiC composites. With 20vol% SiC parti cles, the Young s modulus is increased to 236.5 GPa which is 18 % higher than that of pure UO 2 The difference in effective

PAGE 93

93 Young s moduli between upper and lower bounds of theoretical calculations increases with increasing SiC volume fraction. A verage values of experimental ly determined Young's moduli are fallen into the range between upper and lower bounds of calculated effective Young's moduli. This supports that 1 m SiC particles are homogeneously dispersed in UO 2 matrix and both phases are mechanically well contacted increasing the effective Young's modulus of UO 2 by SiC addition. Because the grain size effects mechanica l properties, it is reasonable to investigate a change of UO 2 grain size in UO 2 SiC composites with different SiC particle sizes and volume fraction s. T he average grain size of UO 2 matrix in composites containing 5vol% SiC increases slowly from 1. 45 to 1.65 with increasing SiC particle size from 0.6 to 55 as shown in Figure 6 5 (a). This moderate increase in grain size is probably due to the decreased surface area to volume ratio of larger SiC particle. The intensity of the pinning effect of embedded SiC particles in UO 2 matrix decreases with higher surface to volume ratio. In chapter 4, we observed larger UO 2 grain size in a UO 2 10vol%SiC composite containing SiC whiskers due to decrease d surface to volume ration ( Figure 4 9). Meanwhile, the average grain size of UO 2 and UO 2 SiC composites decreases with increasing 1 SiC volume fractions as seen in Figure 6 5 (b). The grain size is dramatically decreased from UO 2 to UO 2 5vol%SiC by almost 74% gradually decreased from UO 2 5vol%SiC to UO 2 15vol%SiC, and the value is preserved until UO 2 20vol%SiC. While the initial pinning effect of SiC particles has significantly influence on the reduction in UO 2 grain size, only moderate change in grain size exists by different SiC particle sizes or volume fraction s Residue Stress Measurement Figure 6 6 shows Raman spectra collected f rom a SiC particle in stress free SiC powder and the SiC particles with different sizes, 0.6, 1, 9, 16.9, and 55 m in UO 2 5vol%SiC compsite

PAGE 94

94 pellets. As shown in Figure 6 6 the Raman spectrum from a stress free SiC particle include a transverse optical (TO) peak at 796cm 1 This peak has been reported [ 111 112 ] as a standard TO peak in 3C SiC thin film and directly correspond to Raman peak in free SiC particle without stress. However, changes in TO peaks are observed in various SiC particles in UO 2 5vol%SiC compos ite pellets as sh own in Figure 6 6 While the largest particle size SiC (55 m) exhibit similar TO peak (795.7cm 1 ), the other composites SiC particles containing 0.6, 1, 9, and 16.9 m SiC particles show higher number of TO peaks (800~803cm 1 ) than that of str ess free 3C SiC particle. Fig ure 6 6 also shows a decrease in TO peak Raman shift with increase in SiC p article size. The result suggests that small particle size SiC in UO 2 5vol%SiC composite attain compression stress due to the mismatch of thermal expansion coeff icient between UO 2 and SiC. This compression stress appears during the cooling process when the UO 2 matrix shrinks into the SiC particles as mentioned in chapter 5 With an increas e of the SiC size, the compression s tress of SiC particles decrease d and rea ch a stress free TO position at 55 m size SiC particle. The first compression stress release is observed in 9 m SiC particle and the big drop of stress occurs from 9 to 16.9 m SiC particles. Th is SiC particle size where the initial stress release occurs is consistent with what has previously been observed in both Vickers and Knoop hardnesses and Young's modulus These observations support that microstructural defects at the interface such as microcracks and interfacial debonding release the concentrated st ress and eventually produce stress free 55 SiC particles. Figure 6 7 ( a ) shows the measured TO peak positions as a function of various SiC size in UO 2 5vol %SiC composite pellets. Figure 6 7 ( b ) shows the shift in TO peaks compared with the same peak in a stress free SiC particle (796cm 1 ). In each composite pellet, the Raman TO peaks of thr ee different size SiC particles were observed and the average value was presented in the

PAGE 95

95 plots. For the composite containing 55 m SiC particles, an additional three TO peaks were measured on different areas in a 55 m SiC particle from near the interface to the center of a particle and the average value was regarded as the measured TO peak of the particle. The averaged TO peak of 1 m SiC particle represent the largest positive TO shift corresponding to the highest compression stress. This compression stress is released as the SiC particle size increases from 1 m to 9, 16.9, and 55 m. Micro cracking ( Figure 3(d) and (e)) represents the compression stress release. Althoug h micro cracks were not detected in the microstructure observation of the UO 2 5vol%SiC composite pellet containing 9 m SiC particle, there must be substantial cracks T his is supported by the compression stress release observed in the pellet containing 9 m SiC particles. One can calculate the value of thermal residual stress using the changes in TO peak positions [ 113 ] ( 6 9 ) Where Wo is position of TO peak in a stress free SiC particle (796cm 1 ), S12 and S11 are stress tensor components, p and q are the phonon deformation potentials, and is the shift in TO peak position. p= 0.623 10 6 cm 2 and q= 2.634 10 6 cm 2 are obtained from the mode Gruneisen parameters [ 114 ] for the hydrostatic ( o ) and uniaxial stresses ( s ). ( 6 10 ) ( 6 11 ) The hydrostatic stress ( ) and uniaxial stress ( ) can have different value. However, they were assumed to be same as 1.56 [ 114 ] to calculate p and q S11 =3.7 10 13 and S12 = 1.05 10 13 were utilized to calculate residual stress [ 115 ] Thus, Equation 6 9 reduces to ( 6 12 )

PAGE 96

96 Figure 6 8 shows thermal residual stress within SiC particles in UO 2 5v ol%SiC pellets calculated from Equation 6 12 as a function of SiC particle sizes. The m aximum average compression stress, 649MPa, is found in the composite containing 1 m SiC particles. As the SiC particle size increases, the compression stress is released and the 16 m SiC be comes almost a stress free particle. Micro cracks were directly observed by SEM in the pellet containing larger SiC particles (16 and 55 m ) that coincides with the particle size in which stress relaxation is obvious. Therefore, micro cr acking is responsible for the stress relaxation T he critical SiC size to suppress the micro cracking exists between 1 and 9 m The compression stress is slightly reduced in the middle size SiC particle (9 m ) from that of 1 m particle. In this plot, however, the reduction is too negligible to say there must be micro cracks in the composite pellet. Conclusion The hardness and Young's modulus of UO 2 SiC composites were investigated with different SiC spherical particle sizes and volume fractions. Microcracking and interfacial debonding were monitored by the decrease in mechanical properties. The compression stress of embedded SiC particles with different sizes was estimated using Raman spectroscopy. T he measured compression st ress was up to 1.65GPa in 1 m SiC particle and decreased rapidly to 0Gpa with increasing particle size to 55 m because microcracking and interfacial debonding released the accumulated stress. The hardness and Young's modulus incre ased linearly with increasing 1 m SiC particle volume fraction. With 20vol% of 1 m SiC particle s, while the Vickers hardness increased up to 53% Young's modulus increased up to 18% when compared to that of UO 2 without any SiC addition. Upper and lower bo unds of theoretical Young's modulus of composites containing 0 20vol% SiC particles were calculated using Hashin and Shtrikman model [ 105 ] The measured Young's modulus were in excellent agreement with these calculated upper and lower bounds.

PAGE 97

97 Table 6 1 SiC powder used to produce UO 2 SiC composites. Maximum particle size ( m) Mean size ( m) d10 d50 d90 0.13 0.48 1.54 1.79 0.69 0.32 0.89 1.98 3.07 1.13704 1.94 7.27 22.75 27.35 9.97 3.07 13.06 31.56 40.03 15.85 30.07 54.75 89.00 104.34 57.03 Grade; SiC=99.5% (SiO2=0.2%, C=0.1, Si=0.03...) Table 6 2 Details of SiC particle size, volume fraction, and sintering conditions in the SPS .* SiC p article maximum diamter (m) SiC volume fraction (%) Maximum sintering temperature ( o C) % TD of the composite pellet SD Young's modulus (GPa) Vickers hardness (GPa) Knoop hardness (GPa) N/A 0 1450 96.5 200.2 6 0. 5 5.1 0. 3 0.6 5 1350 95.25 0.24 212.6 7.24 0.2 6.2 0.2 1 5 1350 95.27 0.3 214.9 6.72 0.2 5.9 0.1 1 5 1450 96.81 0.39 216.1 6.98 0.5 6.1 1 10 1450 96.63 0.35 225.6 7.84 0.6 6.9 1 15 151 0 96.5 4 0.23 236.0 8.52 0.9 7.5 1 20 15 50 96.5 1 0.3 249.1 9.2 1.3 8.1 9 5 1350 95.15 0.09 189.2 5.34 0.9 4.7 0.2 16.9 5 1350 94.75 0.17 180.2 5.73 1 4.8 0.1 55 5 1350 95.1 0.13 131.1 5.49 0.86 4.4 0.2 Hold time=5mins; ramp up/down rate=100 o C/min; pressure=36MPa.

PAGE 98

98 Table 6 3. Calculated upper and lower bounds of bulk (K) shear (G) and Young s (E) moduli of UO 2 SiC composite s using Hashin and Shtrikman model [ 105 ] SiC volume fraction (%) Upper bound K (GPa) Lower bound K (GPa) Upper bound G (GPa) Lower bound G (GPa) Upper bound E (GPa) Lower bound G (GPa) 0 183.98 183.98 76.3 76.3 201.1 201.1 5 188.84 188.38 80.05 79.50 210.42 209.09 10 193.82 192.93 83.91 82.83 219.99 217.39 1 5 198.92 197.61 87.88 86.30 229.81 226.02 20 204.14 202.45 91.97 89.92 239.89 234.98 UO 2 : K=184GPa, G=76.3GPa [ 108 ] ; SiC: K=308.6GPa, G=179GPa [ 109 ]

PAGE 99

99 Fig ure 6 1. Selected SEM micrographs of UO 2 and UO 2 SiC co mposites showing the transition in SiC particle volume fraction s and size s. A) pure UO 2 B) 10% 1 m SiC C) 20% 1 m SiC D) 5% 9 m SiC E) 5% 16.9 m SiC

PAGE 100

100 Fig ure 6 2 Vickers and Knoop hardness of UO 2 SiC composites. A) with different mean S iC particle size s. B) with various volume fraction s of SiC addition A B

PAGE 101

101 Fig ure 6 3. Optical microscopy pictures of Vickers and Knoop indents from the composite containing 5vol% 55 m sized SiC particles.

PAGE 102

102 Fig ure 6 4 Young's modulus of UO 2 SiC composites A) with different mean SiC particle size s. B) with various volume fraction s of SiC addition Note that blue and red dot li nes refer to calc u lated Young's modulus based on Hashin and Shtrikman model [ 105 ] A B

PAGE 103

103 Fig ure 6 5 Grain size of UO 2 SiC composites A) with different mean SiC particle size s. B) with different volume fraction s of SiC addition. A B

PAGE 104

104 Fig ure 6 6 Raman spectra collected from within SiC particles with different sizes in each UO 2 5vol%SiC composite

PAGE 105

105 Fig ure 6 7 TO peaks in Raman shift and changes in TO peak position compared to that of stress free SiC particles as a function of SiC particle size.

PAGE 106

106 Fig ure 6 8 Evolution of internal stress as a function of SiC particle size in UO 2 5vol%SiC composites.

PAGE 107

107 CHAPTER 7 EFFECTS OF THERMAL A GING ON THE MICROSTR UCTURE AND THERMAL CONDUCTIVITY OF UO 2 SIC COMPOSITES Background In previous three chapters, we successfully fabricated UO 2 SiC composite fuels and examined their improved thermal and mechanical properties. Because the final operational environment for the UO 2 SiC composite fuel pellet is a nuclear reactor core, it is appropriate to discuss the potential ramifications of using such fuel pellets. The reaction between UO 2 and SiC could be accelerated during plant operation because the current reactor core temperature is higher than the reaction temperature to form USi 1.88 and the irradiation is known to increase the diffusion of both chemical species. Moreover, a mismatch in CTE between UO 2 and SiC may cause microstructural defects such as interfacial debonding and thermal cracks leading to poor nuclear fuel performance during cooling process in a nuclear reactor. It is difficult to induce similar environments as a nuclear reactor core such as steep temperature gradient and high radiation level due to facility restrictions In this study, therefore, only the influence of reactor core temperature for a particular dur ation on UO 2 SiC composites was investigated. For this thermal aging the expected fuel centerline temperature of UO 2 10vol%SiC composites was obtained using the nuclear fuel performance code, FRAPCON. Then, t hermal aging was performed at the predicted cen terline temperature for fabricated high density UO 2 10vol% SiC composites containing SiC whiskers and 1 m SiC particle s. Finally, changes in the microstructure and thermal conductivity of the composites after thermal aging were investigated. E xperiments and Results Predicted Centerline T emperature of UO 2 10vol%SiC Composite F uel A fuel performance code FRAPCON 3.4 [ 116 117 ] was developed by The Pacific Northwest National Laboratory (PNNL) for The Nuclear Regulatory Commission (NRC) to

PAGE 108

108 calculate thermal and mechanical behavior s of Light Water Reactor ( LWR ) fuel rods during long term burnup [ 118 ] The c enterline temperatures of pure UO 2 and UO 2 10vol%SiC composite fuel s can be calculated based on the following equation. ( 7 1 ) Where T = Temperature (K) Bu = Burnup (GWD / MTU ) f(Bu) = 0.00187 Bu g(Bu) = 0.038 Bu 0.28 h(T) = [1+396e Q/T ] 1 Q = 6380 K A = 0. 045 2 m K/W for UO 2 A = 0.038646 m K/W for UO 2 10vol%SiC B = 2.46E 4 m K/W/K for UO 2 B = 2.10E 4 m K/W/K for UO 2 10vol%SiC E = 3.5E9 W K/m F = 16361 K For this calculation, t he measured thermal conductivity of SPS sintered UO 2 and UO 2 10vol%SiC composite fabricated at 1600 o C was utilized ( Figure 4 10). T he calculated centerline temperature of UO 2 and UO 2 10vol%SiC composite fuels as a function of burnup rate is seen in Figure 7 1. T he centerline temperature s rapidly increase with increasing burn up rate and reach to a maximum value in a range of burn up bet ween 3 and 12 MWD/kgU While the predicted maximum centerline temperature of UO 2 is almost 1650 o C, that of UO 2 10vol%SiC composite is only near 1500 o C. The 60% increased thermal conducti vity by 10vol%SiC addition lead s to near 150 o C decrease in the maximum fuel centerline temperature.

PAGE 109

109 Thermal Aging of UO 2 10vol%SiC C omposite s Both UO 2 10vol% SiC composites containing SiC particle s and SiC whiskers were fabricated using SPS technique. While t he 1m SiC particle s (3C SiC) was obtained from Alfa Aesar Inc, Ward Hill, MA t SiC whiskers (3C SiC) were obtained from Advanced Composite Materials, Greer, SC (SC 9D, deagglomerated SiC whiskers) with an aspect ratio, a diameter, and a length exceeding 10:1, 0.65m, and 10 m, re spectively. Because SPS sintering procedures for UO 2 SiC composites are described in both chapter 4 and chapter 5 it is not described in this chapter Bot h composites were sintered at 16 00 o C for 5mins with a mechanical pressure of 36MPa and ramp up/down rate of 100 o C/min. Each sintered composite was cut in half horizontally yielding two disks for 1000 o C and 1500 o C thermal agings Thermal agings were performed in a Li ndberg blue M tube furnace with flowing UHP Ar gas at 1500 o C and 1000 o C for 12hours. The t hermal aging temperature, 1500 o C, was chosen because it was the maximum predicted centerline temperature for UO 2 10vol%SiC composite fuel For a moderate thermal aging 1000 o C was selected to compare the resulting microstructures and thermal conductivities of 1500 o C and 1000 o C aged composites. Table 7 1 shows the composition of aged composites their relative densities before and after thermal aging and thermal aging temperatures and duration While the relative density of composites increased by 1.5 2% af ter 1500 o C thermal aging for 12hours, the relative density of 1000 o C aged composites maintained at similar value. This is because the additional sintering of UO 2 take s place during the thermal aging at 1500 o C, which is in a typical range of UO 2 sintering temperature (1400 1600 o C) [ 119 ] Figure 7 2 shows SEM micrographs of 1000 o C and 1500 o C aged UO 2 10vol%SiC composites containing 1 m SiC particle s While embedded SiC particle s in the UO 2 matrix are clearly seen in the 1000 o C aged composite only traces for SiC particles embedded before

PAGE 110

11 0 thermal aging are observed in the 1500 o C aged composite as shown in Figure 7 2 (a) and (b), respectively. In addition to that, broader and more distinct grain boundaries are found in the 1500 o C aged composite when compared to thos e of 1000 o C aged composite. These phenomena are also observed in the UO 2 10vol%SiC composites containing SiC whiskers. While embedded SiC whiskers are observed in 1000 o C aged composite, all SiC whiskers are removed and only tr aces for whiskers are observed in 1500 o C aged composite as seen in Figure 7 3 (a) and (b), respectively. B road and dis tinct grain boundaries are seen only in the 1500 o C aged composite. A possible explanation for the removed SiC particles and whiskers in the surfac e of 1500 o C aged composites is that brittle and fragile chemical reaction product s arise d at the interface between UO 2 and SiC during the thermal aging led to interfacial debonding and particle removing 1370 o C is a well known chemical reaction starting temperature [ 120 ] between UO 2 and SiC and hen ce, chemical reaction should occur during the 1500 o C thermal aging. In chapter 4, Figure 4 8 showed a reaction product USi 1.88 in a XRD spectra of UO 2 70vol%SiC composite sintered at 1600 o C for 4hours This chemical reaction product at the interface between UO 2 and SiC may play a significant role removing embedded SiC particle s and whiskers during the thermal aging The broad er and more distinct grain boundary in 1500 o C aged composite s is because the grain boundary thermal etching occurred by small oxygen content (~1ppm) in UHP Ar atmosphere during thermal aging Higher temperature and longer duration than typical thermal etching procedure for UO 2 fuel (1340 o C for 3 4hours in UHP Ar) led to extensive grain boundary etching, and hence, broad er a nd distinct UO 2 grain boundaries Meanwhile, the grain boundary of 1000 o C aged composite is much more vague compared to that of 1500 o C aged composite because this temperature is lower than the typical thermal etching temperature.

PAGE 111

111 Figure 7 4 shows the micrographs of 1500 o C aged composite s containing 10vol%SiC particle s and whiskers. Yellow circles indicate particular areas containing large abnormal UO 2 grains. These grains are 4 5times bigger than typical UO 2 grains in other areas as shown in Figure 7 4. T hese areas containing bigger UO 2 grains are only found in 1500 o C aged composites due to the UO 2 grain growth during such high temperature thermal aging for 12hours. Given the fact that traces for SiC particles and whiskers are not found in the parti cular areas containing large UO 2 grains, the reduced pinning effect of embedded SiC seemed like a motivation for the bigger grains. In other areas containing much smaller UO 2 grains, t he pinning effect of SiC particles and whiskers hinder ed the grain growth of UO 2 matrix maintaining UO 2 grains with small size even after thermal aging. Another distinct feature microcracking, due to the 1500 o C thermal aging is seen in Figure 7 5 and Figure 7 6. A microcrack in the aged UO 2 10vol%SiC composit e containing 1m SiC particles is clearly observed in Figure 7 5. A broader microcrack, which width is almost 1 m is observed in the aged UO 2 10vol%SiC composite containing SiC whiskers as shown in Figure 7 6. These cracks were only found in 1500 o C aged composites. They look very similar as thermal microcracks which were observed in UO 2 5vol% SiC composites containing large SiC particles (16.9 and 55 m ) as shown in Figure 5 2. As mentioned in chapter 5, thermal microcracks occurred during the coolin g process of SPS sintering due to the difference in CTE between UO 2 and SiC However, this is not the same case for the 1500 o C aged UO 2 10vol%SiC composites 1 m SiC particles are small enough to suppress the microcracking based on the results in chapter 5. Moreover, before thermal aging, we had not found any similar microcracks i n same composites sintered at 16 00 o C by SPS. If the difference in CTE between UO 2 and SiC were the only cause for microcracking, we would observe similar c racks before t hermal aging because the

PAGE 112

112 microcracking only happen s during the cooling process a nd same composites cooled from 16 00 o C and 1500 o C during sintering and thermal aging processes, respectively A possible reason for the microcracking during 12hours thermal aging at 1500 o C is chemical reaction between UO 2 and SiC. If reaction product s formed during the t her m a l aging at the interface had relatively low fracture toughness, they might not be able to endure the accumulated thermal stre ss d ue to the difference in CTE between UO 2 and SiC In chapter 6, the accumulated compres sion stress in SiC particles was measu re d using Raman spectroscopy ( Figure 6 9 ) The measured compression stress of 1 m SiC particles in UO 2 SiC composites was up to 1.7GPa If this compression stress exceeded the fracture toughness of chemical reaction products at the interface, microcracks nucleate at the interface and propagate spontaneously into UO 2 matrix. Figure 7 7 and Figure 7 8 show the thermal conductivity o f UO 2 10vol%SiC composites containing SiC whiskers and 1 m SiC particles res pectively, before and after thermal aging at 1500 o C for 12hours. While t he enhanced densities of both composites after the thermal aging as listed in Table 7 1 could increase the thermal conductivity ( Equation 5 3) the observed microcracks shown in Figure 7 5 and Figure 7 6 could decrease the thermal conductivity due to the significant phonon scattering. The measured t hermal conductivities of the composites containing SiC whi skers and particles after thermal aging were reduced by 9.5% and 4.5% in average, respectively, when compared to those of same composites before thermal aging. The higher level of reduction in thermal conductivity in the composite containing SiC whiskers i s probably due to more severe cracks as shown in Figure 7 6. Conclusion Thermal aging at 1000 o C and 1500 o C for 12hours were performed on two UO 2 SiC composites containing 1 m SiC spher ical particles and SiC whiskers with an aspect ratio, a

PAGE 113

113 diameter, and a length exceeding 10:1, 0.65m, and 10 m, respectively. While 1000 o C aging didn't change microstructures and properties, the composite aged at 1500 o C showed various modifications such as increased density, SiC particles and whiskers remov ing, microcrackin g, particular areas containing abnormally large grains, and m oderately reduced thermal conductivity. The density increased because the 1500 o C is within a UO 2 sintering temperature range. The removed SiC particles and whiskers and microcracking after therma l aging was probably because brittle chemical reaction product at the interface formed interfacial debonding and microcracks. The abnormally large grains were formulated due to UO 2 grain growth caused by reduced SiC pinning effect on particular areas. Neverthe less of many microstructural changes, we observed relatively good measured thermal conductivity of the composites after thermal aging when compared to that of same composites before aging. The average decrease in thermal conductivity of the composi tes containing SiC particles and whiskers were only 4.5% and 9.5%, respectively.

PAGE 114

114 Table 7 1 Details of composition thermal aging temperature and duration, and relative densities before and after thermal aging of UO 2 10vol% SiC composites. Composition % TD of the composite before aging SD Aging temperature ( o C) Aging duration (hours) % TD of the composite pellet after aging SD SiC particles 96.81 0.39 1000 12 9 6 95 0. 27 SiC whiskers 96.63 0.35 1 000 12 96. 78 0. 33 SiC particles 96.81 0.39 1 500 12 9 8 .14 0.23 SiC whiskers 96.63 0.35 1 500 12 9 8 .41 0.3

PAGE 115

115 Figure 7 1. Expected centerline temperature of UO 2 and UO 2 10vol%SiC composite fuels as a function of reactor burnup rate. 600 800 1000 1200 1400 1600 1800 0.00 10.00 20.00 30.00 40.00 50.00 Centerline Temperature ( C) Burnup (MWD/ kgU ) UO 2 10vol%SiC composite Pure UO 2

PAGE 116

116 Figure 7 2. Microstructures of UO 2 10vol%SiC composites c SiC spherical particles after thermal aging for 12 hours A) thermal etched at 1000 o C B) thermal etched at 1500 o C. A B

PAGE 117

117 Figure 7 3. Microstructures of UO 2 10vol%SiC composites c ontaining SiC whiskers after thermal aging for 12hours at A) 1000 o C and B) 1500 o C. A B

PAGE 118

118 Figure 7 4. Microstructures of UO 2 10vol%SiC composites after thermal aging for 12hours at 1500 o C A) SiC particles B) SiC whiskers Note that yellow circles refer to particula r areas containing large UO 2 grains. A B

PAGE 119

119 Figure 7 5. Microstructures of UO 2 10vol%SiC composites containing 1 m SiC particles after thermal aging for 12hours at 1500 o C.

PAGE 120

120 Figure 7 6 Microstructures of UO 2 10vol%SiC composites c ontaining SiC whiskers after therm al aging for 12hours at 1500 o C.

PAGE 121

121 Figure 7 7 Thermal conductivity of UO 2 10vol%SiC composites c ontaining SiC whiskers before and after thermal aging at 1500 o C for 12hours.

PAGE 122

122 Figure 7 8. Thermal conductivity of UO 2 10vol%SiC composites containing 1 m SiC particles before and after thermal aging at 1500 o C for 12hours.

PAGE 123

123 CHAPTER 8 CONCLUSIONS AND FUTU RE WORK Conclusions Uranium dioxide(UO 2 ) Silicon carbide(SiC) composite fuel pellets were produced by oxidative sintering and spark plasma sintering (SPS) at a range of temperatures from 1400 to 1600 o C. Both SiC whiskers and SiC powder particles were utilized. Oxidative sintering was employed over 4 hours and the SPS sintering was employed only for 5 minutes at the highest hold temperature. It was noted that composite pellets sintered by SPS process revealed smaller grain size, reduced formation of chemical products, higher density, and enhanced interfacial contact compared to the pellets made b y oxidative sintering. T hermal conductivity measurements at 100 o C, 500 o C, and 900 o C revealed that SPS sintered UO 2 10vol% SiC composites exhibited an increase of up to 62% in thermal conductivity compared to UO 2 pellets, while the oxidative sintered composite pellets revealed significantly inferior thermal conductivity values. The particle size (0.6 55 m diameter) and volume fraction (5 20%) of SiC were systematically varied to investigate their influence on the resulting UO 2 SiC composite pellet microstructure and the thermal properties It was found that SiC particle size less than 16.9 m with a larger volume fraction is more effective for improving the thermal conductivity of the fuel pellets. Scanning El e ctron Microscopy examination revealed micro cracking and interfacial debonding in the composites containing larger size SiC particles (16.9 and 55 m) which resulted in reduced thermal conductivity up to 26.5% compared to th at of a UO 2 pellet. For the UO 2 SiC composite s containing 1 m diameter SiC particles, the thermal conductivity increased almost linearly with volume fraction of particles. The experimental thermal conductivity values of the UO 2 SiC composite pellets contai ning 5, 10, and 15vol% are in good agreement with the theoretical values based on the available model in the literature. In the composite pellet

PAGE 124

124 containing 20vol% of SiC particles, the experimental value is lower than the theoretical value by almost 9.9% d ue to both lower density and particle particle interaction. The influence of SiC particle size (0.6 55 m diameter) and volume fraction (5 20%) on the mechanical properties such as hardness and Young's modulus of UO 2 SiC composites were examined. Microcracking and interfacial debonding were monitored by the decrease in mechanical properties. Internal stress measurement using Raman spectroscopy revealed that 1 m SiC particle exhibited 1.65GPa compression stress. This stress decreased rapidly with in creasing particle size until stress free 55 m SiC particle T he hardness and Young's modulus increased linearly with increasing the volume fraction of 1 m SiC particle s. With 20vol% SiC particles, Vickers hardness and Young's modulus increased up to 53% an d 18%, respectively, compared to those of UO 2 The measured Young's modulus were in excellent agreement with theoretically ca lculated upper and lower bounds of Young's modulus. The influence of 1000 o C and 1500 o C thermal aging for 12hours on the microstructure and thermal conductivity of UO 2 10vol% SiC composites containing SiC pa rticles and SiC whiskers was investigated. While no changes in microstructure or thermal conductivity were observed in 1000 o C aged compo sites, SEM revealed removed SiC particles and whiskers from UO 2 matrix, microcracks and particular areas containing 4 5times larger grains in the 1500 o C aged composites. W e observed the measured thermal conductivity of the 1500 o C aged composites was reduc ed 5 10% compared to that of same composites before thermal aging. Future Work Challenges in UO 2 Diamond C omposites Since we have successfully increased the thermal conductivity of UO 2 with SiC additions, the potential benefit of a diamond addition to the UO 2 drew our research group's attention. The measured thermal conductivity of diamond at room temperature is about 22

PAGE 125

125 W /cm K [ 121 ] which is the highest of any solid material. Unlike any sp 2 structure material such as Carbon Nano Tube (CNT), Diamond is composed of sp 3 carbon bonds to be resistant to the radiation environment [ 122 123 ] Therefore, it was expected for UO 2 diamond composites to provide an even higher ther mal conductivity than that of UO 2 SiC composites. In preliminary experiment UO 2 10vol%diamond composite s were fabricated using SPS at 1400 o C, 1450 o C, and 1500 o C for 5mins. The utilized diamond p owder was obtained from Advanced Abrasives Pennsauken, NJ and had a mean diameter of 25 m Figure 8 1 reveals a SEM image of as received diamond powder morpholog y The ramp up/down rate and mechanical pressure were set at 100 o C/min and 36MPa, respectively in the SPS process. The utilized sample preparation and SPS procedure for UO 2 diamond composites are same as those of UO 2 SiC composites and hence, i t is not described here. T he thermal conductivity of fabricated composites was measured before ASTM standard heating to obtain the stoichiometric UO 2 The thermal conductivity measurement procedure used for UO 2 SiC composites and described in chapter 5 was utilized here. The resulting measured densities and thermal conductivities at 100 o C, 500 o C, and 900 o C of the three fabricated com posites were shown in Figure 8 2 Both relative densities and thermal conductivities were increased with increasing sintering temperature. The average increase in thermal conductivity of the highest density composite at those three temperatures were 30.9%, 52.1%, and 55.8%, respectivel y, when compared to a UO 2 literature value. ASTM C 1430 07 procedure was performed to further increase the thermal conductivity of composites by reducing hyper stoichiometric UO 2+x to stoichiometric UO 2 As described in chapter 4 and 5, sintered composites were reduced in a furnace at 800 o C for 6 hours, in a 4%H 2 N gas, with a water vapor atmosphere using a water bath maintained at 35 o C. The thermal conductivity of three composites was measured again after reduction.

PAGE 126

126 Figure 8 3 shows t he thermal conductivity of same UO 2 diamond composites after the reduction process. Thermal conductivities of all composites were reduced dramatically, almost 20 30%, at three temperatures The thermal conductivity of the diamond composites and UO 2 measu red and literature values acquired a similar level. In other words, the benefit of diamond addition for en hanced thermal conductivity almost disappeared The thermal conductivity of the highest density UO 2 diamond composite (96.2%TD) was measured again after thermal etching procedure at 1340 o C for 4hours in UHP Ar atmosphere. Figure 8 4 reveals that the thermal conductivity of the UO 2 diamond composite was again significantly reduced by almost 30 35%, leading to much a l ower thermal conductivity than the UO 2 literature value. We investigated microstructures of UO 2 diamond composites before and after heat processing such as reduction (800 o C, 6hours) and thermal etching (1340 o C, 4hours). Figure 8 5 shows micrographs of diamon d particles in as sintered UO 2 diamond composite (96.2%TD) ((a) and (c)) and same composite after reduction and thermal etching ((b) and (d)) Microcracking and interfacial debonding are clearly seen in the composite after heat processes as indicated by re d arrows We observed similar microstructural defects in UO 2 SiC composites containing large SiC particles (16.9 and 55 m) or small SiC particles after thermal aging (1500 o C, 12hours ) due to the difference in CTE between UO 2 and SiC ( c hapter 5 and 7) The microcracking and interfacial debonding in UO 2 diamond composite can occur more readily because the difference in CTE between UO 2 and diamond is almost five times that of UO 2 SiC (UO 2 : 9.93 x 10 6 K 1 diamond: 1.1 x 10 6 K 1 [ 55 ] ) Unknown features on the surface of diamond particles are observed only in the UO 2 diamond composite after heat processes as indicated by red circles in Figure 8 5. EDS elemental mapping and line scanning were performed on these features to investigate their elemental concentrations

PAGE 127

127 Figure 8 6 shows the SEM micrograph of the unknown features (a) and their EDS elemental counts of carbon K (b) oxygen K (c) and uranium M (d) While unknown features appear ing brighter areas indicated by blue and red circles the darker background refers to the surface of a diamond particle in the SEM micrograph. The c arbon was found in almost the entire area while the background had higher concentration than that of the unknown features. T he number of oxygen K counts in unknown fe atures is more than that of background. T he red circled unknown feature contains more uranium M counts than that of bl ue c ircled unknown features and background. Given these elemental concentrations the red circled feature is probably UO 2 matrix and blue circled features are reaction products between UO 2 and diamond, possibly uranium oxy carbide (UC x O y ) or oxidized diamond. Evan and Phaal [ 124 ] found that the oxidation of diamond can occur at the temperature highe r than 650 o C even at the partial pressure of oxygen as low as 5 x 10 2 mm. Hg. Zhou et. al. [ 125 ] found that uranium oxy carbide (UC x O y ) was formed on uranium surfac e after heat treatment at 500 C for 1hour in vacuum (7 Pa) atmosphere. Therefore, utilized temperatures, heating durations, and oxygen potentials in both reduction and thermal etching processes are able to form both oxidized diamond and uranium oxy carbide (UC x O y ). Figure 8 7 shows the SEM micrograph of a chemical reaction product and EDS elemental counts of carbon K oxygen K M along scanned line across the reaction product. The reaction product has a spherical shape probably because of melting during reduction or thermal etching processes. T his reaction product seems to have a lower melting temperature than 1340 o C. All selected signals such as carbon K oxygen K m M are counted in the chemical reaction product, and hence, its chemical compound is probably uranium oxy carbide (UC x O y ).

PAGE 128

128 In this preliminary study on UO 2 diamond composites, we have observed chemical reaction products between UO 2 and diamond, and microstructural defects such as microcracking and interfacial debonding. These significantly accelerate phonon scattering and can be major disadvantages for enhance thermal conductivity composite fuels. We revealed that the thermal conductivity of UO 2 diam ond composites decreased dramatically during required heating processes such as reduction and thermal etching. Similar reduction in thermal conductivity is inevitable in the final operational environment a nuclear reactor core, because its temperature, du ration, and oxygen potential are extreme. Therefore, the microstructural defects and chemical reaction are essential challenges to be prevented. As we used smaller sized SiC particles to suppress microcracking and interfacial debonding in UO 2 SiC composite s in chapter 5 and 6, similar microstructural defects in UO 2 diamond composites also can be reduced simply by utilizing small diamond particles. Further research is necessary, however, to prevent the chemical reaction between UO 2 and diamond which take pla ce at a temperature as low as 650 o C [ 124 ] P otential R amifications of UO 2 SiC C omposites As mentioned in chapter 7, the severe environment of nuclear core may have negative influence on nuclear composite fuels. The high centerline temperature of nuclear fuel (up to ~1500 o C), steep temperature gradient (2000 3000 o C/cm), and high dose radiation could accelerate the che mical reaction between UO 2 and SiC. Further research is needed to quantify the reaction products in an irradiation environment at appropriate temperatures. Reaction products, fission gas, and fission products could be precipitated along the interface sever ely affecting thermal thermal conductivity of the pellet. Therefore, it is important to investigate interfacial characteristics and the influence of reaction products after irradiation of the pellet. Radiation defects of SiC also have influence on conducti vity of the pellet. However, studies [ 126 127 ]

PAGE 129

129 have shown that there is a reduction in defect concentrations caused by the irradiation damage and the thermal conductivity value is often recovered in the temperature range between 650 to 1400 o C. Further study is needed to verify the recovery of thermal conductivity of irradiated UO 2 SiC composites in a temperature range of reactor core. A ddition of SiC in to a pellet reduces the amount of fissile isotope and hence a higher U 235 enrichment may be required to maintain neutron population sustaining fission process in a reactor core. The exorbitant cost and stringent regulation of U 235 enrichment process have to be considered to reveal the cost detriment of using UO 2 SiC composites.

PAGE 130

130 Figure 8 1. Morphologies of diamond particles with 25 m diameter.

PAGE 131

131 Figure 8 2. Thermal conductivity of as sintered UO 2 10 vol% diamond composites sintered at different temperatures at 100 o C, 500 o C, and 900 o C. Grey and black lines refer to UO 2 measured and literature values.

PAGE 132

132 Figure 8 3. Thermal conductivity of UO 2 10 vol% diamond composites after reduction process performed at 800 o C for 6hours in 4%H 2 N 2 +moisture atmosph ere.

PAGE 133

133 Figure 8 4 Thermal conductivity of UO 2 10 vol% diamond composites after reduction process. T he highest density composite (96.2%TD) was reduced and thermal etched.

PAGE 134

134 Figure 8 5 SEM micrographs of a high density UO 2 diamond composite A) and C) as sintered composite B) and D) same composite after reduction and thermal etching processes. The micro cracks and interfacial debonds are identified by arrows. Unknown features on the surface of a diamond particle are identified by circles.

PAGE 135

135 Figure 8 6 SEM and EDS elemental maps of a diamond particle surface in the UO 2 diamond composite afte r reduction and thermal etching A) SEM image of a diamond particle surface. B) EDS map of carbon K C) EDS map of oxygen K D) EDS map of uranium M Unknown features are identified by circles.

PAGE 136

136 Figure 8 7 SEM and EDS elemental spectra of carbon K oxygen K and uranium M along scanned line across a chemical reaction product on the surface of a diamond particle in the UO 2 diamond composite.

PAGE 137

137 LIST OF REFERENCES [1] D.R. Olander, Fundamental Aspects of Nulcear Reactor Fuel Elements, NTIS, 1976. [2] J.I. Owens, R.W. Lockhart, D.R. Iltis, K. Hikido, U. S. A. E. C. GEAP 3279 (1959) 36 [3] J.J. Powers, B.D. Wirt h, Journal of Nuclear Materials 405 (2010) 74 82. [4] G.E. Jellison, J.D. Hun n, R.A. Lowden, J. Nucl. Mater. 352 (2006) 6 12. [5] T. Nozawa, L.L. Snead, Y. Kato h, J.H. Miller, J. Nucl. Mater. 371 (2007) 304 313. [6] E. Lopez Honorato, P.J. Meadows, P. Xiao, G. Mar sh, T.J. Abram Nucl. Eng. Des. 238 (2008) 3121 3128. [7] L. Tan, T.R. Allen, J.D. Hun n, J.H. Miller, J ournal of Nucl ear Mater ials 372 (2008) 400 404. [8] K.H. Sarma, J. Fourcade, S.G. Lee, A.A. Solomo n, Journal of Nuclear Materials 352 (2006) 324 333. [9] A.K. Singh, S.C. Zunjarrao, R.P. Singh, Journal of Nuclear Materials, 378 (2008) 238 243. [10] J.A. Khan, T.W. Knight, S.B. Pakala, W. Jiang, R. Fan g, J.S. Tulenko, Nucl. Technol. 169 (2010) 61 72. [11] R.A. Verrall, M.D. Vlajic, V.D. Krstic, J ournal of Nuclear Materi als 274 (1999) 54 60. [12] B.S. Hic kman, Stud. Radiat. Eff. Solids 1 (1966) 72 158. [13] R.D. Hunt, G.W.C.M. Silva, T.B. Lindemer, K.K. Anderson, J.L. Collin s, Journal of Nuclear Materials 427 (2012) 245 248. [14] J.A.C. G. C. Allen, T. Swan, 7th Internati onal sympos ium on the reactivity of solids (1972) S4/1 10. [15] A. Solo mon, S. Revankar, J.K. McCoy, Enhanced Thermal Conductivity Oxide Fuels DOE project Report, Purdue University 2006 [16] R. Latta, S.T. Revankar, A.A. Sol omon, Heat Transfer Engineeri ng 29 (2008) 357 365. [17] Wang, J., Developing a high thermal conductivity nuclear fuel with silicon carbide additives, Ph.D. dissertation, University of Florida 2008. [18] K. Hanada, K. Matsuzaki, T. Sano, Journal of Mate rials Processing Technology 153 1 54 (2004) 514 518. [19] A.G. Every, Y. Tzou, D.P.H. Hasselman, R. Raj, Acta Metallurgica et Materialia 40 (1992) 123 129. [20] H.Z. Abdel, S.F. Moustafa, F.A. Morsy, A.K.N. Abdel, M.F. Abdel, Nat. Sci. (Irvine, CA, U. S.) 3 (2011) 936 947. [21] J. Zhang, R Tu, T. Goto, Ceramics International 38 (2012) 351 356. [22] G. Sasaki, H. Nakase, K. Suganuma, T. Fujita, K. Niihara, Journal of the Ceramic Society of Japan 100 (1992) 536 540.

PAGE 138

138 [23] K. Maekawa, Y. Nakada, T. Kimura, Journal of Materials Science 37 (2002) 397 410. [24] G. Bai, W Jiang, L. Chen, Mater. Trans. 47 (2006) 1247 1249. [25] E.T. Swartz, R.O. Pohl, Reviews of Modern Physics 61 (1989) 605 668. [26] Y. Xu, Y. Tanaka, M. Goto, Y. Zhou, K. Yagi, Journal of Ap plied Physics 95 (2004) 722 726. [27] J.C. Maxwell, A T reatise on E lectricity and M agnetism Dov er Publications, 1954. [28] D.A.G. Brugg eman, Ann. Phys. (Berlin, Ger.) 24 (1935) 665 679. [29] D. Dangouleme, V. Inozemtsev, K. Kamimura, J. Killeen, A. Kucuk, V. Novikov, V. Onufriev, M. Tay al, in, Amer ican Nuclear Society 11 ( 2011 ) 199 209. [30] Y.S. Chou, D.J. Green, Journal of the European Ceramic Society 14 (1994) 303 311. [31] J. Watts, G. Hilmas, W.G. Fahrenholtz, Journal of the American Ceramic Society 94 (2011) 4410 4418. [32] R.L. Gibb y, Journal of Nuclear Materials 38 (1971) 163 177. [33] H.E. Schmid t, Journal of Nuclear Materials 39 (1971) 234 237. [34] L.A. Goldsmith, J.A.M. Dougla s, Journal of Nuclear Materials 47 (1973) 31 42. [35] R. Brandt, G. Neuer, Journal of Non Equilibriu m Thermodynamics 1 (1976) 3 24. [36] J.K. Fin k, Journal of Nuclear Materials 279 (2000) 1 18. [37] J. Li, L. Porter, S. Yi p, Journal of Nuclear Materials 255 (1998) 139 152. [38] R.J. McEachern, P. Taylor, Jour nal of Nuclear Materials 254 (1998) 87 121. [39] S. Siegel, Acta Crystallographica 8 (1955) 617 619. [40] I.C. Hobson, R. Taylor, J.B. Ainscough, Journa l of Physics D: Applied Physics 7 (1974) 1003 1015. [41] D.M. Henri Bailly, Claude Prunier, Lavoisier, 1999. [42] J.C. Van Craeynest, J.P. Stor a, Jo urnal of Nuclear Materials 37 (1970) 153 158. [43] J. Williams, E. Barnes, R. Scott, A. Hal l, Journal of Nuclear Materials 1 (1959) 28 38. [44] K. Langrod, American Ceramic Society Bulletin (U.S.) 39 (1960) 366 369. [45] H. Chevrel, P. Dehaudt, B. Francois J.F. Baumar d, Journal of Nuclear Materials 189 (1992) 175 182. [46] P. Dehaudt, L. Bourgeois, H. Chevre l, Journal of Nuclear Materials 299 (2001) 250 259. [47] N. Fuhrman, L.D. Hower, R.B. Holden, Journal of the American Ceramic Society 46 (1963) 114 121

PAGE 139

139 [48] K.W. Song, K.S. Kim, Y.M. Kim, Y.H. Jun g, Journal of Nuclear Materials 277 (2000) 123 129. [49] T.R.G. Kutty, P.V. Hegde, K.B. Khan, U. Basak, S.N. Pillai, A.K. Sengupta, G.C. Jain, S. Majumdar, H.S. Kamath, D.S. C. Purushotham, J. Nucl. Mater. 305 ( 2002) 159 168. [50] H. Ohashi, E. Noda, T. M orozumi, J. Nucl. Sci. Technol. 11 (1974) 445 451. [51] J. Weissbart, P.E. Blackburn, E.A. Gulbransen, Further S tudies on the M echanism of O xidation of U ranium D ioxide DOE Report, Westinghouse Elec. Corp., 1957 [52] Z.A. Munir, U. Anselmi Tamburi ni, M. Ohyanagi, J. Mater. Sci. 41 (2006) 763 777. [53] M. Omori, Mate rials Science and Engineering A 287 (2000) 183 188. [54] S. Zhijian, M. Johnsson, Z. Zhe, M. Nygren, Journal of the American Ceramic Society 85 (2002) 1921. [55] Y.S. Touloukian, Thermophysical Properties of High Temperature Solid Materials Macmillan, 1966. [56] D.D. Joseph, L. Prez iosi, Reviews of Modern Physics 61 (1989) 41 73. [57] W.D. Callister, Materials Science and Engineering: An Introduction, 4th Edition Wiley, 1996. [58 ] P.L. Kapitza, Physical Review 60 (1941) 354 355. [59] L.D. Landau, E.M. Lifshitz, Electrodynamics of Continuous Media Pergamon Press., 1960. [60] C.P. Wong, R.S. Bollampally, Jou rnal of Applied Polymer Science 74 (1999) 3396 3403. [61] F. Boey, A.I.Y. Tok, Y.C. Lam, S.Y. Chew, Mater ials Science and Engineering: A 335 (2002) 281 289. [62] D.P.H. Hasselman, L.F. Johnson, Journal of Composite Materials 21 (1987) 508 515. [63] C. W. Nan, G. Liu, Y. Lin, M. Li, Appl ied Physics Letters 85 (2004) 3549 3551 [64] J.E. Gubernatis, J.A. Krumha nsl, Journal of Applied Physics 46 (1975) 1875 1883. [65] C. W. Nan, R. Birringer, D.R. Clarke, H. Gleiter, Journal of Applied Physics 81 (1997) 6692 6699. [66] Z. Huang, Z. Shen, L. Ling, S. Tan, N. Mats, D. Jiang, Taoci Xuebao 22 (2001) 115 120. [67] D. Lahiri, V. Singh, A.K. Kesh ri, S. Seal, A. Agarwal, Carbon 48 (2010) 3103 3120. [68] S. Nemat Nasser, Y. Su, W. G. Guo, J. Isaacs, Journal of the Mechanics and Physics of Solids 53 (2005) 2320 2346. [69] H. Assmann, W. Drr, M. Peeh s, Journal of Nuclear Materials 140 (1986) 1 6. [70] J. Shen, F.M. Zhang, J.F. Sun, Y.Q. Zhu, D.G. McCartney, Nanotechnology 17 (2006) 2187 2191.

PAGE 140

140 [71] F. Zhang, C. Mihoc, F. Ahmed, C. Lath e, E. Burkel, Chem. Phys. Lett. 510 (2011) 109 114. [72] B. Cheng, Y. Lin, J. Lan, Y. Liu, C. Nan, J. Mater. Sci. Technol. (Shenyang, China) 27 (2011) 1165 1168. [73] L. Ge, G. Subhash, R.H. Baney, J.S. Tulenko, E. McKenn a, Journal of Nuclear Materials 435 (2013) 1 9. [ 74] N. Tang, Guisuanyan Tongbao 7 (1988) 16 18. [75] D.J. Willis, J. S.H. Lake, Scripta Metallurgica 21 (1987) 1733 1736. [76] J.K. Fink, J. Nucl. Mater. 279 (2000) 1 18. [77] P.G. Klemens, Phonon I nteractions in S olids Univ. Connecticut, 1971 [78] C.V. Ramana, M. Noor A Alam, J.J. Gengler, J.G. Jones, ACS Applied Materials & Interfaces 4 (2011) 200 204. [79] G. Soyez, J.A. Eastman, L.J. Thompson, G.R. Bai, P.M. Baldo, A.W. McCormick, R.J. DiMelfi, A.A. Elmustafa, M.F. Tambwe D.S. Stone, Appl. Phys. Lett 77 (2000) 1155 1157. [80] S Yagi, D. Kunii, AIChE Journal 3 (1957) 373 381. [81] D.P.H. Hasselman, K.Y. Donaldson, A.L. Geiger, Journal o f the American Ceramic Society 75 (1992) 3137 3140. [82] K. Chu, H. Guo, C. Jia, F. Yin, X. Zhang, X. Liang, H. C hen Nanoscale Research Letters 5 (2010) 868 874. [ 83] C.M. Friend, J. Mater. Sci. 22 (1987) 3005 3010. [84] A. Rudajevova, B.F. Von, B.L. Mordike, J. Alloys Compd. 292 (1999) 27 30. [85] R.I. Todd, B. Derby, Acta Materialia 52 (2004) 1621 1629. [86] C. Hillman, Z. Suo, F.F. Lange, Journal of the American Ceramic Society 79 (1996) 2127 2133. [87] M. Taya, S. Hayashi, A.S. Kobayashi, H.S. Yoon, Journal of the American Ceramic Society 73 (1990) 1382 1391. [88] B.R. Powell, G.E. Youngblood, D.P.H. Ha sselman, L.D. Bentsen, Journal of the American Ceramic Society 63 (1980) 581 586. [89] J. Bendixen, A. Mortensen, Scr ipta Metallurgica et Materialia 25 (1991) 1917 1920. [90] T.C. Lu, J. Yang, Z. Suo, A.G. Evans, R. Hecht, R. Mehrabian, Acta Metallurgica e t Materialia 39 (1991) 1883 1890. [91] S. Y. Fu, X. Q. Feng, B. Lauke, Y. W. Mai, Composites Part B: Engineering 39 (2008) 933 961. [92] D.P.H. Hasselman, K.Y. Donaldson, J. Wide Bandgap Mater. 7 (2000) 306 318. [93] J.R. Thomas, Jr., D. P.H. Hasselman, The rm. Conduct. 28 (2006) 52 64.

PAGE 141

141 [ 94] N. Mingo, Physical Review B 74 (2006) 125402. [95] P.M. Hui, X. Zhang, A.J. Markw orth, D. Stroud, J. Mater. Sci. 34 (1999) 5497 5503. [96] L.W. da Silva, M. Kaviany, International Jo urnal of Heat and Mass Transfer 47 (200 4) 2417 2435. [97] B. T. Wang, P. Zhang, O. Eriksson, arXiv.org, e Print Arch., Condens. Matter (2012) 1 9, arXiv:1201.5003v1201 [cond mat.str el]. [98] P. riou, C. Denoual, C.E. Cottenot, Internationa l Journal of Impact Engineering 21 (1998) 225 235. [99] Q. Zhang, G. Wu, L. Jiang, G. Chen, Materials Chemistry and Physics 82 (2003) 780 785. [100] J. Hu, A.C. Hayes, W.B. Wilson, u. Rizwan, Nuclear Engineering and Design 240 (2010) 3751 3757. [101] M. J. Pan, P.A. Hoffman, D.J. Green, J.R. Hellmann, Journal of the American Ceramic Society 80 (1997) 692 698. [102] D.J. Green, Journal of the American Ceramic Society 65 (1982) 610 614. [103] J. Brennan, K. Prewo, Journal of Materials Science 17 (1982) 2371 2383. [104] A.L. Chamberlain, W.G. Fahrenholtz, G.E. Hil mas, D.T. Ellerby, Journal of the American Ceramic Society 87 (2004) 1170 1172. [105] Z. Hashin, S. Shtrikman, Journal of the Mechanics and Physics of Solids 11 (1963) 127 140. [106] Z. Hashin, J. Appl. Mech. 29 (1962) 143 150. [107] Z. Hashin, S. S htrikma n, J. Mech. Phys. Solids 10 (1962) 343 352. [108] A.R. Hal l, Journal of Nuclear Materials 37 (1970) 314 323. [109] J.F. Shackelford, W. Alexander, Materials Science and Engineering Handbook, Third Edition CRC, 1999. [110] K.K. Phani, D. Sanyal, A.K. Sengupt a, Journal of Nuclear Materials 366 (2007) 129 136. [111] R. Rahimi, C.M. Miller, S. Raghavan, C.D. Stinespring, D. Korakakis, Journal of Physics D: Applied P hysics 42 (2009) 055108. [112] N. Lee, R.D. Hartschuh, D. Mehtani, A. Kisliuk, J.F. Maguire M. Green, M.D. Foster, A.P. Sokolov Journal of Raman Spectroscopy 38 (2007) 789 796. [113] D. Ghosh, G. Subhash, N. Orlovskaya, Acta Materialia 56 (2008) 5345 5354. [114] Z.C. Feng, W.J. Choyke, J.A. Pow ell, Journal of Applied Physics 64 (1988) 6827. [115] M. Miura, H. Murata, Y. Shiro, K. Iishi, Journal of Physics and Chemistry of Solids 42 (1981) 931 936. [116] A. Karahan, A. Lerch, M.S. Kazimi, Amer ican Nuclear Society ( 2011 ) 488 496.

PAGE 142

142 [117] D.D. Lanning, C.E. B eyer, G.A. Berna, K. Davis, American Nu clear Society ( 1997 ) 601 612. [118] D.D. Lanning, C .E. Beyer, M.E. Cunningham, American Nuclear Society ( 2000 ) 206 220. [119] J. Gao, X. Yang, R. Li, Y. Wang, F. Zhon g, Journal of Materials Science 42 (2007) 5936 5940. [120] G.C. A llen, J.A. Crofts, T. Swa n, Chapman and Hall ( 1972 ) 630 639. [ 121] G.A. Slack, J. Appl. Phys. 35 (1964) 3460 3466. [122] V.M. Kos enkov, S.A. Vorob'ev, At Energy 84 (1998) 23 27. [123] N. Dilawar, R. Kapil, V.D. Vankar, D.K. Avasthi, D. Kabira j, G.K. Mehta, Thin Solid Films 305 (1997) 88 94. [124] T. Evans, C. Phaal, Proc. Conf. Carbon 5th, Univ. Park, Penna. 1 (1962) 147 153. [125] P. Zhou, X. Wang, J. Yang, X. Fu, L. Luo, Xi you Jinshu Cailiao Yu Gongcheng 37 (2008) 94 97. [126] G.E. Youngblood, D.J. Senor, R.H. Jone s, Jou rnal of Nuclear Materials 329 333 (2004) 507 512. [127] T. Maruyama, M. Harayama, Journal of Nuc lear Materials 329 333 (2004) 1022 1028.

PAGE 143

143 BIOGRAPHICAL SKETCH Sunghwan Yeo was born in Seoul Republic of Korea His parents are Inbae Yeo and Y ounsook Je ong He has an younger sister, S eon gmi Yeo Sunghwan graduated with a Bachelor of Science in materials science and engineering from Illinois Institute of Technology in May 200 9 After that, he enrolled in materials science and University of Florida in August of 200 9 He received a non thesis Master of Science degree in May of 20 12 and is scheduled to graduate with a Doctor of Philosophy degree in August 20 13