<%BANNER%>

Heterogeneous Sodium Fast Reactor Designed for Transmuting Minor Actinide Waste Isotopes into Plutonium Fuel

Permanent Link: http://ufdc.ufl.edu/UFE0022110/00001

Material Information

Title: Heterogeneous Sodium Fast Reactor Designed for Transmuting Minor Actinide Waste Isotopes into Plutonium Fuel
Physical Description: 1 online resource (309 p.)
Language: english
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: axial, heterogeneous, minor, moderated, sodium, transmutation, transuranic
Nuclear and Radiological Engineering -- Dissertations, Academic -- UF
Genre: Nuclear Engineering Sciences thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: In the past several years there has been a renewed interest in sodium fast reactor (SFR) technology for the purpose of destroying transuranic waste (TRU) produced by light water reactors (LWR). The utility of SFRs as waste burners is due to the fact that higher neutron energies allow all of the actinides, including the minor actinides (MA), to contribute to fission. It is well understood that many of the design issues of LWR spent nuclear fuel (SNF) disposal in a geologic repository are linked to MAs. Because the probability of fission for essentially all the 'non-fissile' MAs is nearly zero at low neutron energies, these isotopes act as a neutron capture sink in most thermal reactor systems. Furthermore, because most of the isotopes produced by these capture reactions are also non-fissile, they too are neutron sinks in most thermal reactor systems. Conversely, with high neutron energies, the MAs can produce neutrons by fast fission. Additionally, capture reactions transmute the MAs into mostly plutonium isotopes, which can fission more readily at any energy. The transmutation of non-fissile into fissile atoms is the premise of the plutonium breeder reactor. In a breeder reactor, not only does the non-fissile 'fertile' U-238 atom contribute fast fission neutrons, but also transmutes into fissile Pu-239. The fissile value of the plutonium produced by MA transmutation can only be realized in fast neutron spectra. This is due to the fact that the predominate isotope produced by MA transmutation, Pu-238, is itself not fissile. However, the Pu-238 fission cross section is significantly larger than the original transmutation parent, predominately: Np-237 and Am-241, in the fast energy range. Also, Pu-238?s fission cross section and fission-to-capture ratio is almost as high as that of fissile Pu-239 in the fast neutron spectrum. It is also important to note that a neutron absorption in Pu-238, that does not cause fission, will instead produce fissile Pu-239. Given this fast fissile quality and also the fact that Pu-238 is transmuted from Np-237 and Am-241, these MAs are regarded as fertile material in the SFR design proposed by this dissertation. This dissertation demonstrates a SFR design which is dedicated to plutonium breeding by targeting Am-241 transmutation. This SFR design uses a moderated axial transmutation target that functions primarily as a pseudo-blanket fuel, which is reprocessed with the active driver fuel in an integrated recycling strategy. This work demonstrates the cost and feasibility advantages of plutonium breeding via MA transmutation by adopting reactor, reprocessing and fuel technologies previously demonstrated for traditional breeder reactors. The fuel cycle proposed seeks to find a harmony between the waste management advantages of transuranic burning SFRs and the resource sustainability of traditional plutonium breeder SFRs. As a result, the enhanced plutonium conversion from MAs decreases the burner SFR?s fuel costs, by extracting more fissile value from the initial TRU purchased through SNF reprocessing.
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.
Thesis: Thesis (Ph.D.)--University of Florida, 2008.
Local: Adviser: Tulenko, James S.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0022110:00001

Permanent Link: http://ufdc.ufl.edu/UFE0022110/00001

Material Information

Title: Heterogeneous Sodium Fast Reactor Designed for Transmuting Minor Actinide Waste Isotopes into Plutonium Fuel
Physical Description: 1 online resource (309 p.)
Language: english
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: axial, heterogeneous, minor, moderated, sodium, transmutation, transuranic
Nuclear and Radiological Engineering -- Dissertations, Academic -- UF
Genre: Nuclear Engineering Sciences thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: In the past several years there has been a renewed interest in sodium fast reactor (SFR) technology for the purpose of destroying transuranic waste (TRU) produced by light water reactors (LWR). The utility of SFRs as waste burners is due to the fact that higher neutron energies allow all of the actinides, including the minor actinides (MA), to contribute to fission. It is well understood that many of the design issues of LWR spent nuclear fuel (SNF) disposal in a geologic repository are linked to MAs. Because the probability of fission for essentially all the 'non-fissile' MAs is nearly zero at low neutron energies, these isotopes act as a neutron capture sink in most thermal reactor systems. Furthermore, because most of the isotopes produced by these capture reactions are also non-fissile, they too are neutron sinks in most thermal reactor systems. Conversely, with high neutron energies, the MAs can produce neutrons by fast fission. Additionally, capture reactions transmute the MAs into mostly plutonium isotopes, which can fission more readily at any energy. The transmutation of non-fissile into fissile atoms is the premise of the plutonium breeder reactor. In a breeder reactor, not only does the non-fissile 'fertile' U-238 atom contribute fast fission neutrons, but also transmutes into fissile Pu-239. The fissile value of the plutonium produced by MA transmutation can only be realized in fast neutron spectra. This is due to the fact that the predominate isotope produced by MA transmutation, Pu-238, is itself not fissile. However, the Pu-238 fission cross section is significantly larger than the original transmutation parent, predominately: Np-237 and Am-241, in the fast energy range. Also, Pu-238?s fission cross section and fission-to-capture ratio is almost as high as that of fissile Pu-239 in the fast neutron spectrum. It is also important to note that a neutron absorption in Pu-238, that does not cause fission, will instead produce fissile Pu-239. Given this fast fissile quality and also the fact that Pu-238 is transmuted from Np-237 and Am-241, these MAs are regarded as fertile material in the SFR design proposed by this dissertation. This dissertation demonstrates a SFR design which is dedicated to plutonium breeding by targeting Am-241 transmutation. This SFR design uses a moderated axial transmutation target that functions primarily as a pseudo-blanket fuel, which is reprocessed with the active driver fuel in an integrated recycling strategy. This work demonstrates the cost and feasibility advantages of plutonium breeding via MA transmutation by adopting reactor, reprocessing and fuel technologies previously demonstrated for traditional breeder reactors. The fuel cycle proposed seeks to find a harmony between the waste management advantages of transuranic burning SFRs and the resource sustainability of traditional plutonium breeder SFRs. As a result, the enhanced plutonium conversion from MAs decreases the burner SFR?s fuel costs, by extracting more fissile value from the initial TRU purchased through SNF reprocessing.
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.
Thesis: Thesis (Ph.D.)--University of Florida, 2008.
Local: Adviser: Tulenko, James S.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0022110:00001


This item has the following downloads:


Full Text





HETEROGENEOUS SODIUM FAST REACTOR DESIGNED FOR TRANSMUTING
MINOR ACTINIDE WASTE ISOTOPES INTO PLUTONIUM FUEL




















By

SAMUEL EUGENE BAYS


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

2008


































2008 Samuel Eugene Bays




























To Nikki.









ACKNOWLEDGMENTS

I would like to thank my faculty advisor, James Tulenko, for his constructive comments

and guidance of this work. I would also like to thank the other members of my advisory panel,

Edward Dugan, Samim Anghaie, Ronald Baney and Steve Herring for their sensible

recommendations. I would like to especially thank and recognize Steve Herring for his technical

guidance and support of this work. I greatly thank and appreciate Idaho National Laboratory for

its financial support of this dissertation project. Thanks go to Douglas Crawford and David Nigg

for their management of projects and activities in which this work supported. I thank Doug

Porter, Mitch Meyer, Steve Hayes, Jon Carmack and Rory Kennedy for their technical feedback

and discussions on transmutation and fast reactor fuels. Special thanks are given to Pavel

Medvedev for his input on establishing the fuel performance criteria of the transmutation targets.

Special thanks are given to Bevin Brush and Thomas Johnson for their technical insight of fuel

handling and separations operations at the EBR-II fuel cycle facility. Special thanks are given to

Steve Piet, Gretchen Matthern and David Shropshire for their technical insight into the system

dynamics and economics of closing the domestic and international fuel cycles. Special thanks

are given to Roald Wigeland, Giuseppe Palmiotti and Massimo Salvatores for their technical

insight into reactor physics and design methodologies of sodium fast reactors. Many thanks are

given to my friends and colleagues of the INL fuel cycle analysis team: Mehdi Asgari, Rodolfo

Ferrer, Benoit Forget and Michael Pope for the excellent studies and results that we have

produced over the last two years.









TABLE OF CONTENTS


page

A CK N O W LED G M EN T S ...................................................... ..............................................

L IS T O F T A B L E S ................................................................................. 9

LIST OF FIGURES .................................. .. .. .... .... .......... ....... 12

A B S T R A C T ............ ................... .................. ........................................ 2 2

CHAPTER

1 INTRODUCTION ............... ...........................................................24

M motivation and O bjectives....................................................................... ... ......................26
Transmutation Physics ................ ........ ....... ................... ....... .. ... ...... .... 31
Neutron Spectrum Influence on Transmutation Behavior.............................................32
Transm utation and N nuclear Stability..................................... .......................... .......... 36
Isotopic A aspects of R epository Im pacts....................................................................... 39
Background on Previously Proposed TRU Burning SFRs..................................................41
The Advanced Burner Reactor Design Concept .................................. ............... .... 43
Transmutation Target Designs: Radial Blankets and Moderated Targets...................45
Transmutation Target Designs: Axial Blankets and Axial Targets.............................48
Transmutation Based Reactivity Control Concept.......................................................49
Design Rationale of an Axial Heterogeneous Fast Transmutation Reactor .........................52
Compensation for Inherent Positive Void Reactivity Feedback ...................................53
Benefits of Axial Targets for a Dedicated MA Burner Core Design ............................54
Technology Com patibilities and Synergies ............................................ .......................... 58
Actinide Partitioning: PUREX and UREX.................................... ...... ............... 59
Pyroprocessing and the Integral Fuel Cycle ................................. ......... .................61
Assumptions for Using Transmutation Targets in SFRs...........................................63

2 COMPUTATIONAL METHODS AND FAST REACTOR PHYSICS.............................66

C calculations and Fuel Cycle M odeling......................... ..................................................66
Fast Reactor Equilibrium Fuel Cycle Calculations Using the REBUS Code .................68
Light Water Reactor Spent Fuel Calculations Using the TRITON Code .......................72
Scoping Calculations and Benchmarking Using the MCNP Code ..............................73
Physics of the Reference Metal Fueled Advanced Burner Reactor.................... .......... 74
Conversion Ratio and H igh Leakage Cores ........................................ .....................75
Conversion Ratio and High TRU Enriched Fuels ................................ ..................... 80
Physics of the Axially Heterogeneous Fast Transmutation Reactor ................................82
Axial Targets and Axial Leakage Recovery ..........................................................82
Axial Targets and Minor Actinide Conversion .................................... ............... 85
Com bining Leakage and Capture Effects.................................................................... ... ..90









3 AXIAL TARGET DESIGN ANALY SIS................................................................... ......93

Waste Management Philosophies and Conversion Ratio Definition........................ ..93
Transm utation B ased R eactor D esign .................................................. ............... .... 96
Transmutation Targets and Accompanying Fuel Cycle ...............................................99
Transm utation Target Physics ......................................................... .............. 103
P aram etric S tu d y ............... .. .............. .......................................................... .............. 10 6
Effects of Pin Diameter and Core Height ............................................. ...............107
Effects of M operating Pins .................................................... .................. ............... 111
Tall and Flattened Axial Heterogeneous Core Designs ............. ................................... 115
R adial and A xial Pow er Profiles ...............................................................................116
R activity Feedbacks .......................................... .............. .. ........ .... 123
Fuel Perform ance Indicators................................................. ............................. 125
Final Down-Selection: The AHFTR Design ............................................ ............... 128
Fuel Cycle Performance of the Final AHFTR Design ...............................................129
Reactor Performance Characteristics of the Final AHFTR Design ............................132
Transmutation Analysis of the Final AHFTR Design ............................................135

4 REACTOR REACTIVITY CONTROL STRATEGY............... ....................................139

Tc-99 versus B-10 as a Control Rod Neutron Poison...............................................139
Control A ssem bly D esign......................................................................... ............... 143
Traditional SFR Control Assembly Design: B4C Ultimate Shutdown Assembly........143
Technetium Based Primary Control Assembly Design............................144
G as E expansion M odule....................................................................... .......... ........... 148
C control R od W north ............................ ... ................... ......... .. ............ ........... .. 150
Reactivity Worth of Boron versus Technetium................... ............155
Top versus Bottom Inserted Shim Rods.................................... ...................156
Axial Power Tilt and Shim Rod Insertion ......................... .......... ...................157
O their R activity Feedbacks ....................................................... ......... ............ 157
A xial Fuel Expansion .......................................... .. .. .................. 160
R adial Fuel B ow ing ................... .... ...... .... .. .... .......... .. ........... .............. 16 1
Technetium Transm station R ate........... ............................. ................ ............... 164

5 DIFFUSION VERSUS TRANSPORT BENCHMARKS ........... ......... ............... 167

D iffusion and Transport M methods ................... ......... ....................................................... 168
F lux Spectrum A naly sis ....................................................................... ...................170
Spatial F lux A nalysis....................................................... .. ................ .............. ..172
Differences between BOEC and EOEC Fluxes.................................. ............... 175
D e p letio n T e st................................................................................................................. 1 7 6
Spatial Self-Shielding Test ............................................... ... ......... ................ 180
Spatial Shadowing Test .................................. .. .. ........ .. ............182
Calculation Validation Rem arks........................... .................. .................................. 186

6 THE AH FTR FUEL DESIGN ......... ................... ....... ...................... ............... 188




6









F u el P in D design .................................. ... ........... .................................... 189
Target Alloy Selection and Design Considerations......................................... 190
Target and Driver Fuel Bumup Criteria............................ ........................ 194
Cladding D am age Criteria.................................................. ............................... 195
Fuel Pin Therm al Perform ance Criterion ........................................ ........................ 197
Fuel Tem perature Criterion .............. ............ .......................... ............... 198
Cladding Tem perature Criterion......................................................... ............... 199
Thermal Analysis......................................... 200
Fuel Assembly Power Peaking ........... .. ......... ....................... 200
Peak Fuel Pin H ot Channel Analysis....................................... .......................... 206
M etallurgical D iffu sion E effects ........................................ .............................................2 15
Inter-D iffusion D ata .......................................... ............. .... ....... 216
Penetration Distance ...... ................................................... .. ... ...... .. 217
Transmutation Gas Generation and Plenum Sizing......................................................218
Helium and Fission Product Gas Calculation..................................... ............... 220
Transmutation and Fission Gas Analysis ........................................... ............... 223
F uel D esign B asis Sum m ary ......................................................................... .................. 226
Fuel Processing Considerations..................... ....... ............................. 227
Higher M ass A ctinide Considerations............................................... ........ ....... 230
R epository C onsiderations.................................................. ............................... 234

7 ECONOMICS OF THE TWO-TIER AHFTR FUEL CYCLE........................................238

E conom ic Issues of R processing ............................ ......................................................238
C capitol C costs .................................................... 240
F u el C o sts ................................................................................................. 2 4 3
First-C ore R processing Cost............................................... ............................. 243
Subsequent Core Reprocessing Cost.......................... ............. ............... 244
H igh Level W aste D isposal Cost........................................................ ............... 247
Fuel Fabrication C ost .......................... .............. ................. .... ....... 248
F ront-E nd U uranium C osts...................................................................... ..................249
Back-End Uranium Costs .............. .......................... ................. 251
D discounting and Financing ............................ ........... ................................. ............... 252
O operations and M maintenance .............................................................. .....................253
C on stru action C capital ................................................................ ................................ 2 54
Fuel Investm ent ................................................................ 256
Fuel Cycle Base Cases ................................ .. ........................... .... ..... 257
The M O X Fuel Cycle .......................... .............. ................. .... ....... 258
The A B R Fuel C ycle ........................... ............................................ ..... .. ............. 262
The Combined "Mixed-Fleet" AHFTR and ABR Fuel Cycle .....................................265
Sensitivity A naly sis ...................... .... ............ ..................... 270
Breakeven Unit Costs for ABRs and AHFTRs.................................................270
Cost Sensitivities for a Combined ABR and AHFTR Fleet................................274
Converting N eutrons into D dollars .......................................... ......... ...................... 278

8 SUM M ARY AND CONCLUSION S................................ ............................................. 282



7









M ethod of R actor D design ....................................................................................... 284
Reactor Control and Safety Features .............................................................................286
C o d e V alid atio n .............................................................................................................. 2 8 8
Transmutation Target Fuel D esign ......................................................... .............. 289
Fuel Processing and Repository Considerations ......................................................... 290
Fuel Cycle Econom ic A nalysis.................................................. ............................... 291
Concluding Rem arks ..................................... ............................292

APPENDIX

A R E B U S IN PU T D E C K ......... .. ............ ....................................................................294

B PARAMETRIC DESIGN ANALYSIS ........................................ .......................... 296

LIST OF REFERENCES ..................... ..........................................300

B IO G R A PH IC A L SK E T C H ............................................................................. ....................309





































8









LIST OF TABLES


Table page

1-1 ABR fission and capture cross section data.................................. ..................... 35

1-2 IM F fission and capture cross section data ......... ..... .......... .................. ....36

1-3 Fuel assembly and pin dimensions for the S-PRISM and ABR designs ...........................46

1-4 Summary of relative studies on transmutation targets................ ...............48

1-5 Examples of real fast reactor plants that have incorporated axial blankets.......................49

1-6 Fuel assembly design of the AHFTR compared to ABR designs................................58

1-7 Core design for the AHFTR compared to similar ABR designs ....................................59

1-8 Waste stream partitioning afforded by various reprocessing technologies .....................60

1-9 Technology compatibility assumptions used for fuel cycle anlaysis.............................65

2-1 Isotopic com position for U OX SNF ............................................................................ 73

2-2 Atom densities and one-group microscopic cross sections for Pu-239 and U-238 ...........76

2-3 AHFTR fission and capture cross sections data ..................................... .................89

3-1 Reactor thermal power and core heights evaluated in parametric analysis .....................97

3-2 Reference fuel assembly design for varying height-to-diameter ratio ............................97

3-3 Transm station utilization factor........... .................. ......... ............... ............... 105

3-4 Single group cross section ratio ......... ................. ......... ..................... ............... 105

3-5 Transmutation half-lives for a preliminary AHFTR design ........................................108

3-6 Transmutation half-life of Am-241 for varying number of moderating rods ..................111

3-7 Core design summary for the reference ABR with tall and flat AHFTR ......................116

3-8 Fuel cycle comparison for the reference ABR with tall and flat AHFTR .....................122

3-9 Physics comparison for the reference ABR with tall and flat AHFTR .........................125

3-10 Fuel performance comparison for the reference ABR with tall and flat AHFTR ...........127

3-11 Mass flow analysis of the downsellected AHFTR ......... .......................................130









3-12 Target and active core region fuel inventory of the downsellected AHFTR ...................131

3-13 Initial reactor physics and fuel performance of the downsellected AHFTR ...................134

3-14 Mass production and destruction rates per installed megawatt per year .......................135

3-15 Isotopic contribution of external supply versus contributing to fission .........................137

4-1 Atomic concentrations of absorber atoms in B4C versus technetium............................142

4-2 Control assembly types and dimensions............. .............. ............. ............... 143

4-3 Core reactivity worth for independently separate reactivity feedback effects...............60

4-4 Summary of Tc-99 consumption by the AHFTR .................................. .....................165

5-1 Select reactor parameters for the AHFTR given by DIF3D and VARIANT. .................175

5-2 Comparison of mean-free-paths of different reaction types ................. .....................185

6-1 Peak burnups for the driver fuel and targets................... ................................. .............. 194

6-2 Average burnups for the driver fuel and targets ................................... ............... 195

6-3 Average and peak fast fluence (E>0.1 MeV) and peak dpa ........................................195

6-4 ABR fuel assembly peak-to-core average LHGR ratio taken at BOEC........................204

6-5 Beginning of Equilibrium Cycle AHFTR fuel assembly peak-to-core average LHGR..204

6-6 End of Equilibrium Cycle AHFTR fuel assembly peak-to-core average LHGR ..........204

6-7 Ternary inter-diffusion coefficients .............................. .............................................216

6-8 AH FTR fuel pin dim ensions......... ................. ................... .................. ............... 224

7-1 Unit costs and fees for the open single-tier MOX fuel cycle........................................260

7-2 Uranium Oxide once through fuel cycle fuel costs........................................................261

7-3 M ixed Oxide fuel cycle fuel costs .............................................................................261

7-4 Unit costs and fees for the closed single-tier ABR fuel cycle ......................................264

7-5 Advanced Burner Reactor fuel cycle fuel costs............... ............................................ 264

7-6 Unit costs and fees for the closed double-tier AHFTR fuel cycle................................268

7-7 Advanced Burner Reactor fuel costs in a mixed ABR and AHFTR fleet .....................268









7-8 Fuel costs for the AHFTR tier of the mixed ABR and AHFTR fleet............................269

7-9 Advanced Burner Reactor two-tier fuel cycle breakeven unit costs............................272

7-10 Axial Heterogeneous Fast Transmutation Reactor fuel cycle breakeven costs ...............272

7-11 M ixed fleet fuel cycle breakeven costs................................................. ....... ........ 274

7-12 Inter-comparison of aqueous reprocessing costs for various scenarios.........................280









LIST OF FIGURES


Figure page

1-1 Synergy between GNEP and IFR fuel recycling strategies ............................................30

1-2 Important transmutation reactions of americium and neptunium into plutonium ............33

1-3 Periodic relationship between fissile and fertile transuranic isotopes. ...........................38

1-4 Decay heat plot for pressurized water reactor SNF ................................. ............... 40

1-5 Core layouts for the S-PRISM and ABR designs ................. .............44

1-6 Cross section plots ofU-238 and Tc-99 total absorption cross sections ........................52

1-7 Core design of the Axial Heterogeneous Fast Transmutation Reactor...........................56

1-8 Modification to the ABR fuel cycle by the axial targets ............................................. 57

1-9 Partitioning and transmutation scenario of an ABR and AHFTR mixed-fleet..................63

2-1 Coupled DIF3D core physics and REBUS fuel cycle algorithm.....................................69

2-2 Flow diagram of data transfer between the MC2-2 and REBUS codes ...........................71

2-3 Material buckling of a simple bare homogeneous SFR.................. ...............77

2-4 Critical radius required to equate geometric buckling with material buckling..................77

2-5 Generalized conversion ratio of simple bare homogeneous SFR............... ............... 79

2-6 Mean-free-path of neutron travel between interactions.............. .... ...............79

2-7 Leakage fraction of a simple bare homogeneous SFR............... .... .................80

2-8 Conversion ratio and the corresponding TRU enrichment for a fixed core size ...............81

2-9 Axial current distribution of the AHFTR and ABR ......................................................84

2-10 Axial flux distribution of the AHFTR and ABR ............................................................85

2-11 Moderated target region binned "microscopic" reaction rate spectra.............................86

2-12 Active core region binned "microscopic" reaction rate spectra......................................87

2-13 Comparison of capture and fission cross sections between Am-241 and Pu-238 .............88

2-14 Comparison of flux spectrums between LWR IMF target region and active core ............91









3-1 Preliminary AHFTR core design used in parametric analyses .......................................98

3-2 Zirconium hydride phase diagram for varying hydrogen content ...................................99

3-3 Representation of the axial target pin-lattice showing the orientation of targets ..........100

3-4 AHFTR fuel cycle scenario for the REBUS calculation .............................................101

3-5 Americium and plutonium cross section plots versus a SFR neutron flux................104

3-6 Spectrum comparison between inner, middle, outer core and targets ...........................106

3-7 Change in target isotope masses as a function irradiation time .........................107

3-8 Active core radial power density profile for a preliminary AHFTR ...........................110

3-9 Inner core axial power density profile for a preliminary AHFTR...............................110

3-10 Lin.-Log Scale: Neutron flux in the target region for varying moderating rods.............112

3-11 Log.-Log Scale: Neutron flux in the target region for varying moderating rods............112

3-12 Percent of the initial Am-241 mass remaining in the target rod.............................. 114

3-13 Percent of the initial Pu-238 mass created in the target rod ............... ...............15

3-14 "Flat" preliminary AHFTR design with eight rows of fuel instead of seven ................17

3-15 Radial power density profile for six axial slices through the core...............................118

3-16 Axial power density profile for each row of fuel.................. ............................... 119

3-17 Target region buildup and depletion curve................................................. ............... 131

3-18 Active core region buildup and depletion curve................................... ..................132

4-1 Total neutron absorption cross section plots for select absorber materials ...................142

4-2 Ultimate shutdown control assembly configuration. ............................... ...............145

4-3 Primary control assembly configuration...................... .... ........................... 147

4-4 Gas expansion module assembly configuration..................... ...................... 149

4-5 Shim control rod bank reactivity worth at BOEC .................. ................................150

4-6 Shim control rod bank reactivity worth at EOEC ................. ................................151

4-7 Safety control rod bank reactivity worth at BOEC............. ....................................152









4-8 Neutron spectrum in the targets as a function of safety rod insertion ...........................152

4-9 Safety control rod bank reactivity worth at EOEC ........................................................153

4-10 Ultimate shutdown system worth at BOEC............................... .................154

4-11 Ultimate shutdown system worth at EOEC ........................................ ............... 154

4-12 Shim control rod bank reactivity worth for metallic Tc-99 compared to B4C ...............56

4-13 Radial flux distributions for shim rods inserted from the top or the bottom .................157

4-14 Axial power distribution for increasing shim rod insertion into the active core ............158

4-15 Neutron spectrum for steady state operation versus a complete loss of sodium ............159

5-1 Core midplane energy spectrum for code validation comparison .................................171

5-2 Target region energy spectrum for code validation comparison ....................................171

5-3 Cod validation comparison of the flux axial profile ........................................................ 173

5-4 Cod validation comparison of the flux radial profile at the core mid-plane.................173

5-5 Cod validation comparison of the flux radial profile at the axial target region............. 174

5-6 Comparison between the BOEC and EOEC neutron spectrums ....................................175

5-7 Comparison between the BOEC and EOEC radial flux profiles................................... 176

5-8 Comparison between BOEC and EOEC axial flux profiles .........................................177

5-9 Reactivity curve comparison between REBUS and MONTEBURNS ..........................178

5-10 Advanced Burner Test Reactor benchmark using "fresh core" composition ..................180

5-11 M C N P sub-lattice m odel......................................................................... ...................18 1

5-12 Neutron spectrum as a function of the zirconium hydride slug radius ............................183

5-13 Neutron spectrum as a function of the target slug radius .............................................183

5-14 MCNP unit-cell model with homogenized fuel annulus......................................184

5-15 Neutron spectrum within the homogenized annulus............................................185

6-1 Conceptual AHFTR Fuel Pin Design .................................. 190

6-2 Optical microscopy for three of the AFC-1B samples................... ...............192









6-3 Fuel swelling performance for the AFC-1B samples ............................................... 193

6-4 AHFTR power density profile as a function of fuel row and axial region ................202

6-5 ABR reference power density profile as a function of fuel row and axial region ..........202

6-6 BOEC LHGR for the driver fuel mid-plane and target regions .........................205

6-7 EOEC LHGR for the driver fuel mid-plane and target regions .........................206

6-8 Peak pin axial BOL LHGR distribution...................... .... ........................... 208

6-9 Axial node procession for Nusselt analysis ........................................ ............... 209

6-10 Assumed thermal properties of TRU-U-Zr metal alloy fuel................. ................213

6-11 Axial temperature profiles for a "typical" fuel rod in the hottest fuel assembly ............214

6-12 Percent contribution to helium production by alpha decay ..........................................219

6-13 Gas plenum pressures resulting from transmutation and fission gas production.............225

6-14 Envisioned power block reactor plant model.......................... ....... ............... 229

6-15 Average neutron emission rate for processed initial TRU in fresh fuel ..........................231

6-16 Average gamma decay energy rate for processed initial TRU in fresh fuel....................233

6-17 Average alpha decay heat rate for processed initial TRU in fresh fuel .........................234

6-18 Decay heat plot for AHFTR fuel cycle losses in a repository ......................................236

7-1 The open single-tier M OX fuel cycle ........................................ ......................... 259

7-2 The closed single-tier ABR fuel cycle............................ ........................ 263

7-3 The closed double-tier AHFTR fuel cycle................................... ........................ 266

7-4 Fuel cycle costs for the ABR and AHFTR for varying aqueous unit costs...................276

7-5 Fuel cycle costs for the ABR and AHFTR for varying pyroprocessing unit costs..........276

7-6 Transuranic-free HLW disposal fee adjustment for breakeven ...............................278

8-1 Cross section data for Am-241, Pu-238 and Pu-239...............................283

8-2 Radial power distribution for the ABR versus the AHFTR............................ .........285

A -1 R E B U S input cards ............................................................................ .. ................ .. 294









B-1 Excess Reactivity of the core design in "tall" for varying core height and p/d ...............296

B-2 Excess Reactivity of the core design in "flat" for varying core height and p/d.............. 297

B-3 TRU enrichment of the core given in "tall" for varying core height and p/d ..................297

B-4 TRU enrichment of the core design in "flat" for varying core height and p/d ...............298

B-5 Cycle length of the core given in "tall" for varying core height and p/d .......................298

B-6 Cycle length of the core design in "flat" for varying core height and p/d .....................299












AAA

ABR

ABTR

AFCI

AHFTR

ALMR

ANL

ARR

ATR

ATW

B4C

BOC

BOEC

BOL

CAPRA

CCCC

CDA

CDF

COMPX

Critical Radius


CRGT

D-9


LIST OF ABBREVIATIONS AND TERMINOLOGY

Advanced Accelerator Applications

Advanced Burner Reactor

Advanced Burner Test Reactor

Advanced Fuel Cycle Initiative

Axial Heterogeneous Fast Transmutation Reactor

Advanced Liquid Metal Reactor

Argonne National Laboratory

Advanced Recycling Reactor

Advanced Test Reactor

Accelerator Transmutation of Waste

Boron Carbide

Beginning of Cycle

Beginning of Equilibrium Cycle

Beginning of Life

Consommation Amelioree du Plutonium dans les Reacteurs Avances

Committee on Computer Code Coordination

Core Disruptive Accident

Cumulative Damage Fraction

CCCC format binary macroscopic cross section data file

Effective core radius required to meet the reactivity criticality requirement
by equating geometric buckling with material buckling

Control Rod Guide Tube

Austenitic steel that can be used for SFR fuel cladding and structural
components









DIF3D


DOE

dpa

EBR-I

EBR-II

EFPD

EFR

ENDF

Enrichment Zone

EOC

EOEC

EOL

FCCI

FCF

FCMI

FFTF

FOAK

GEM

GNEP

HLW

HM

HT-9


IBA

IFBA


Multigroup neutron diffusion code developed for modeling steady-state
reactor physics and other critical systems developed by ANL

Department of Energy

Displacements per Atom

Experimental Breeder Reactor I

Experimental Breeder Reactor II

Effective Full Power Day

European Fast Reactor

Evaluated Nuclear Data File

Same TRU enrichment/composition regions within a SFR

End of Cycle

End of Equilibrium Cycle

End of Life

Fuel-to-Cladding Chemical Interaction

Fuel Cycle Facillity

Fuel-to-Cladding Mechanical Interaction

Fast Flux Test Facillity

First of a Kind

Gas Expansion Module

Global Nuclear Energy Partnership

High Level Waste

Heavy Metal

Ferritic/Martensitic high chromium steel used for SFR fuel cladding and
structural materials

Integral Burnable Absorber

Integral Fuel Burnable Absorber









IFC

IFR

iHM

IHX

ILW

IMF

INL

ISOTXS

JSFR

KALIMER

LANL

LEU

LHGR

LLW

LOCA

LOHS

LWR

MA

Matino Plane

MC2-2


MCNP


mil

MIT

Mixed waste


Integral Fuel Cycle

Integral Fast Reactor

Initial Heavy Metal or fresh fuel going into the reactor

Intermediate Heat Exchanger

Intermediate Level Waste

Inert Matrix Fuel

Idaho National Laboratory

CCCC format binary microscopic cross section data file

Japanese Atomic Energy Agency SFR

Korean Advanced Liquid Metal Cooled Reactor

Los Alamos National Laboratory

Light Enriched Uranium

Linear Heat Generation Rate

Low Level Waste

Loss of Coolant Accident

Unprotected Loss of Heat Sink

Light Water Reactor

Minor Actinide

Contact surface between two inter-diffusing compositions

Multigroup cross section generation code by solution of the neutron
slowing down equations developed by ANL

Monte Carlo N-Particle general geometry, continuous energy and angle
Monte Carlo transport code developed by LANL

one-tenth of one cent

Massachusetts Institute of Technology

Chemically and radiologically hazardous waste









MOEC

MONTEBURNS

MOX

MWD

MWY

NOAK

NRC

NUS

NWF

NWPA

OCRWM

ORIGEN


ORNL

p/d

PHWR

PIE

PUREX

PWR

Pyroprocess


REBUS

SCNES

SFF

SFR

SHORT


Middle of Equilibrium Cycle

Coupling code for MCNP and ORIGEN developed by LANL

Mixed Oxide

Megawatt-day

Megawatt-year

Nth of a Kind

Nuclear Regulatory Commission

Nuclear Utility Service code of account

Nuclear Waste Fund

Nuclear Waste Policy Act

Office of Civilian Radioactive Waste Management

General use isotope buildup, depletion and decay code for solving the
Bateman equations by the exponential matrix method developed by ORNL

Oak Ridge National Laboratory

Pitch-to-diameter ratio

Pressurized Heavy Water Reactor

Post Irradiation Examination

Plutonium Uranium Redox Extraction

Pressurized Water Reactor

Metal fuel reprocessing whereby uranium and transuranics are separated
by electro-deposition on cathodes immersed in a eutectic LiCl-KCl bath

REactor BUrnup System Fuel cycle analysis code developed by ANL

Self-Consistent Nuclear Energy System

Spent Fast reactor Fuel

Sodium Fast Reactor

Shutdown and Heat Removal Test









SNF

S-PRISM

Support Ratio


SWU

THORP

TOP

TREAT

TRU

TRU Enrichment

ULOF

UOX

UP1 & UP2

UREX

UREX+

USEC

UxC

VARIANT

YM-EIS

ZrH1.6


Spent "LWR" Nuclear Fuel

Super Power Reactor Innovative Small Module

Mass balance ratio of installed thermal capacity for the mass consuming
reactor per installed thermal capacity for the mass producing reactor

Separative Work Unit

Thermal Oxide Reprocessing Plant, Britain

Transient Overpower

Transient Reactor Test facility

Transuranic

Concentration by volume of TRU over HM in SFR or MOX-LWR fuels

Unprotected Loss of forced circulation Flow

Uranium Oxide

Reprocessing plants at LaHague, France

Uranium Extraction aqueous reprocessing technology

Uranium Extraction Plus Including waste stream partitioning

United States Enrichment Corporation

Uranium Exchange Consulting Company

Variational Anisotropic Nodal Transport Code developed by ANL

Yucca Mountain Environmental Impact Statement

Zirconium Hydride









Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

HETEROGENEOUS SODIUM FAST REACTOR DESIGNED FOR TRANSMUTING
MINOR ACTINIDE WASTE ISOTOPES INTO PLUTONIUM FUEL

By

Samuel Eugene Bays

May 2008

Chair: James Tulenko
Major: Nuclear Engineering Sciences

In the past several years there has been a renewed interest in sodium fast reactor (SFR)

technology for the purpose of destroying transuranic waste (TRU) produced by light water

reactors (LWR). The utility of SFRs as waste burners is due to the fact that higher neutron

energies allow all of the actinides, including the minor actinides (MA), to contribute to fission.

It is well understood that many of the design issues of LWR spent nuclear fuel (SNF) disposal in

a geologic repository are linked to MAs. Because the probability of fission for essentially all the

"non-fissile" MAs is nearly zero at low neutron energies, these isotopes act as a neutron capture

sink in most thermal reactor systems. Furthermore, because most of the isotopes produced by

these capture reactions are also non-fissile, they too are neutron sinks in most thermal reactor

systems. Conversely, with high neutron energies, the MAs can produce neutrons by fast fission.

Additionally, capture reactions transmute the MAs into mostly plutonium isotopes, which can

fission more readily at any energy. The transmutation of non-fissile into fissile atoms is the

premise of the plutonium breeder reactor. In a breeder reactor, not only does the non-fissile

"fertile" U-238 atom contribute fast fission neutrons, but also transmutes into fissile Pu-239.

The fissile value of the plutonium produced by MA transmutation can only be realized in

fast neutron spectra. This is due to the fact that the predominate isotope produced by MA









transmutation, Pu-238, is itself not fissile. However, the Pu-238 fission cross section is

significantly larger than the original transmutation parent, predominately: Np-237 and Am-241,

in the fast energy range. Also, Pu-238's fission cross section and fission-to-capture ratio is

almost as high as that of fissile Pu-239 in the fast neutron spectrum. It is also important to note

that a neutron absorption in Pu-238, that does not cause fission, will instead produce fissile Pu-

239.

Given this fast fissile quality and also the fact that Pu-238 is transmuted from Np-237 and

Am-241, these MAs are regarded as fertile material in the SFR design proposed by this

dissertation. This dissertation demonstrates a SFR design which is dedicated to plutonium

breeding by targeting Am-241 transmutation. This SFR design uses a moderated axial

transmutation target that functions primarily as a pseudo-blanket fuel, which is reprocessed with

the active driver fuel in an integrated recycling strategy. This work demonstrates the cost and

feasibility advantages of plutonium breeding via MA transmutation by adopting reactor,

reprocessing and fuel technologies previously demonstrated for traditional breeder reactors. The

fuel cycle proposed seeks to find a harmony between the waste management advantages of

transuranic burning SFRs and the resource sustainability of traditional plutonium breeder SFRs.

As a result, the enhanced plutonium conversion from MAs decreases the burner SFR's fuel costs,

by extracting more fissile value from the initial TRU purchased through SNF reprocessing.









CHAPTER 1
INTRODUCTION

Sodium Fast Reactors (SFRs) are currently being evaluated as a means of eliminating the

long-lived transuranic (TRU) waste produced by Light Water Reactors (LWR). Historically, the

inherent neutron surplus, necessary for overcoming neutron losses by leakage, has been used to

breed fissile Pu-239 by transmuting U-238 in a blanket. In the context of a transuranic-burning

SFR, these excess neutrons are applied to destroying, by fission, the transuranic atoms of Spent

Nuclear Fuel (SNF) produced by LWRs.

However, not all transuranic isotopes can be destroyed by fission with the same efficiency.

This is because many of the isotopes in SNF TRU can only fission above a given threshold

neutron energy. Unfortunately, these isotopes, such as Np-237 and Am-241, present in SNF,

pose the largest decay heat and radiotoxicity issues for permanent disposal of this waste in a

geologic repository. The repository space benefit stems principally from the removal of Am-241

from the fuel cycle. This is because the repository's waste emplacement drift spacing is limited

by the maximum rock temperature between drifts. The mid-drift temperature is principally a

function of the decay heat produced by Am-241. Additionally, alpha decay of Am-241 is chiefly

responsible for the buildup of Np-237 in the repository many years after it is closed.

The Am-241 cross section for fission is orders of magnitude less than its capture cross

section at neutron energies less than one MeV. Because the average neutron flux for most

nuclear reactors, including fast reactors, exists at energies below one MeV, fission can not be the

primary mechanism for removing americium from the fuel cycle. In fact, this fission threshold

property for both Np-237 and Am-241 is very pronounced when compared to most other heavy

metal (HM) actinides. Therefore, it is useful to convert these isotopes, especially Am-241, into

more fissionable plutonium isotopes, using specialized transmutation targets. Irradiating Am-









241 in an epithermal or thermal spectrum (below one MeV) maximizes its destruction by neutron

capture. This neutron capture and its subsequent decay chain leads to the breeding of the even

mass number plutonium isotope: Pu-238. This even-plutonium isotope, though not truly fissile,

has a larger fission cross section below one MeV compared with the initial Am-241 atom. The

fast reactor core design proposed in this dissertation uses the plutonium breeding "fertile"

property of Am-241 to maximize the available fissile worth that can be derived from this

otherwise non-fissile transuranic waste.

The relative increase in fissile worth contributed by the transmuted Pu-238 can only be

realized in fast neutron spectra. This is because Pu-238 has a fission threshold of its own.

However, the Pu-238 fission threshold is simply less sharply defined at one MeV than Am-241

(or Np-237). The fissile worth of Pu-238 is only comparable to that of odd mass numberfissile

Pu-239 and Pu-241 in the unresolved resonance range of its fission cross section. At thermal

spectrum energies, the resolved capture resonances of Pu-238 render it effectively non-fissile

with no reactivity benefit. Even though it is possible that a thermal spectrum reactor could be

used to transmute Am-241 into Pu-238, a fast reactor would still be necessary to fully benefit

from this conversion. This scenario would require that both thermal and fast reactors be

collocated with all the necessary fuel recycling facilities at the same location. This concentration

of large infrastructure facilities is considered unfeasible for this work. Instead, this dissertation

will demonstrate a fast reactor design which is dedicated to plutonium breeding via americium

transmutation. This fast reactor design uses a moderated axial transmutation target that functions

as a pseudo-blanket fuel for breeding Pu-238. The transmuted Pu-238 is then co-reprocessed

along with the SFR's active driver fuel in an integrated recycling strategy.









Motivation and Objectives

The United States commercial nuclear power industry currently produces approximately

20% of the nation's electricity. LWRs have become the industry work horse for producing this

power since the first fully commercial LWR was brought online in the United States at

Shippingport Pennsylvania in 1957. Orders for new nuclear power plants grew steadily during

the 1960's and continued through the 1970's. In 1979, a partial reactor meltdown accident at

Three Mile Island, Pennsylvania caused investor confidence in nuclear reactor safety to

essentially disappear. The accident prompted the Nuclear Regulatory Commission (NRC) to

impose stricter safety standards for all existing and future commercial reactors. Though no new

nuclear power plants were ordered after the Three Mile Island Accident, completion of existing

construction projects since then continued to expand the industry's electrical capacity. By 1991,

22% of the nation's electricity was produced by 111 nuclear power plants [1]. Since then,

several nuclear power plants have been decommissioned before the end of their licensed

lifetimes. Despite the loss of these reactors, the current share of nuclear in the nation's electric

resources has stabilized at 20%.

The sustainability of electric generation, even with increasing electricity demands, is due in

part to an overall technological maturity afforded by years of safe operational experience since

the Three Mile Island accident. Gradual improvements in the irradiation integrity of the uranium

fuel have allowed more fission energy to be extracted per initial mass (i.e., burnup) than ever

before. Enhancements in computer simulation have enabled an increased understanding of the

nuclear, mechanical, hydraulic and materials physics within LWRs which allows them to be

operated safely without sacrificing performance. Also, increased attention to human factors has

enhanced operational safety, which in turn, has generally reciprocated public trust in nuclear

energy.









These improvements have allowed nuclear power plant operators to minimize the time the

reactor is shutdown for refueling and maintenance activities. Therefore, the capacity for LWRs

to operate at full power has steadily increased from approximately 55% in the 1960's to over

90% in the 21st century [2]. This "capacity" is equivalent to saying that nuclear power plants

produce 90% of the energy that could be generated if they were operated at full power all of the

time. The remaining 10% roughly accounts for the time required to discharge old fuel and reload

fresh fuel. Therefore, it is arguable that LWRs have achieved their practical limit in nuclear

electricity generating capacity.

Further improvements in LWR fuel technology may allow slightly higher operating

capacities without adding additional fuel costs. However, this is unlikely because the burnup of

enriched fuel is generally linearly proportional to the level of uranium enrichment of the fissile

isotope, U-235, in the fuel [3]. Currently, the NRC enforces a cap on enrichment at five percent

to protect fuel fabrication workers from possible criticality accidents. Due to the criticality

safety implications of exceeding this enrichment limit, it is assumed that additional safety

measures will be required by enrichment and fabricators for achieving higher burnups in LWRs.

These additional safety measures could increase LWR fuel costs Unless technology becomes

available that makes the process of enrichment significantly cheaper, the cost of additional

criticality safety practices will cause the cost of LWR fuel to increase in order to achieve higher

burnups.

These practical limitations extend essentially to the "front-end" of the nuclear fuel cycle.

This part of the commercial industry is responsible for all processes from mining the uranium

from the ground through its irradiation in the reactor. This is also the sector of the commercial

industry in which the private sector has the most control over electricity costs.









The "back-end" of the fuel cycle, and its associated cost, is currently the responsibility of

the federal government in accordance with the Nuclear Waste Policy Act (NWPA). As required

by the NWPA, the nuclear industry pays 0.1 0 per every kW-h of electricity produced by this

fuel to the federal government as payment towards its eventual permanent disposal in a geologic

repository [4]. The LWR fleet currently produces approximately 2,000 metric tones of SNF each

year [5]. Assuming that no additional reactors are built, the adequacy of this levee to pay for

geologic SNF disposal has been assessed as adequate by the Office of Civilian Radioactive

Waste Management (OCRWM) [6]. However if new LWRs are constructed to meet rising

electricity demands, then it becomes possible that several repositories will be required. This

could substantially change the economic compatibility of the NWPA levee with future nuclear

forecasts.

However, if the number of repositories can be limited to one, the NWPA revenue collected

from the existing LWR fleet for the remainder of the lifetime of those plants will be adequate to

pay the full cost of the repository [7]. The utilization of the storage space within the repository

design is critical in determining its effective capacity to store radioactive wastes. The NWPA

stipulates that the repository be designed to not only dispose of SNF, but also for disposing of

separated High Level Wastes (HLW). HLW is comprised of fission product and transuranic

isotopes present in SNF. It is the HLW component of SNF that represents almost the entire

radioactivity produced in the nuclear fuel cycle. In actuality, roughly 90% of the SNF mass is

the original uranium that was mined from the ground. Therefore, substantial improvements in

repository utilization can be achieved by storing only the volume of HLW within SNF.

The cost competitiveness of this option hinges upon the economic advantage gained by

separating HLW versus storing unaltered SNF. Separating transuranic and fission product









isotopes from SNF imply that technologies such as that used for nuclear fuel reprocessing are

required. Reprocessing is an integral step in nuclear fuel recycling which is the process of

separating the TRU isotopes from SNF to create new reactor fuel.

TRU comprises only about 1% of SNF. Additionally, TRU can be further categorized into

major actinides (i.e., plutonium) and minor actinides (MA) (i.e. neptunium, americium, curium,

berkelium and californium) which represent only about 10% of TRU. Therefore, it is the

approximate 0.1% of the total mass of the SNF that strongly controls repository utilization.

Because most of TRU is "fissile" plutonium, it is arguable that it could be recycled in LWRs.

Nevertheless, it is the MAs that create the difficulties for the repository. These isotopes are

much more difficult to destroy in LWRs due to the fact that they have no fissile value in the

LWR's thermal neutron spectrum.

It is because of the improved fissile value of MAs in a fast neutron spectrum that the

United States has renewed its interest in SFR technology. The current Global Nuclear Energy

Partnership (GNEP) program is currently investigating a symbiosis between SFRs and the SNF

generated by TRU (Figure 1-1). Unlike TRU recycling in LWRs, which only recycles TRU once

(limited by fissile concentration in Pu), the GNEP scenario would continuously recycle TRU in

SFRs in a closed fuel cycle after the initial separation from SNF. Because the GNEP scope of

the SFR is tailored for TRU destruction, it is dubbed the Advanced Burner Reactor (ABR) by

this program.

An important note should be made here that the exact GNEP nomenclature refers to the

ABR as a demonstration prototype for proving the technology of building a fleet of TRU burner

SFRs called Advanced Recycling Reactors (ARR) [7]. However, the term "advanced recycling

reactor" is to general descriptor for this dissertation. The term ABR is used regularly in recent









SFR literature to mean specifically a transuranic burning SFR design with a homogenous core

configuration that excludes the use of transmutation targets or uranium blankets [7]. This

definition is adopted by this dissertation.


SNF


Fertile Blanket Material
(MAs replace Uranium)


'E..M'1111111


TRU
(Mostly Pu)


SLWR Reprocessing



SFR


GNEP Recycling


SFR Reprocessing


Spent Targets Join Driver Fuel
for IFR Type Reprocessing


Figure 1-1. Synergy between GNEP and IFR fuel recycling strategies: .Purple represents the
basic IFR closed fuel cycle scenario. Blue represents the LWR input into the IFR
closed fuel cycle for the GNEP recycling scenario. The striped orange represents the
replacement of fertile uranium blankets with MA transmutation targets.1

Reprocessing, fuel fabrication and reactor irradiation in the GNEP closed fuel cycle,

creates a role for the private industry to minimize the "back-end" costs by using LWR and SFR

spent fuel in the "front-end". The further extraction of fission energy from the recycled fuel


1 The original IFR scenario would not require a constant supply of externally reprocessed SNF
TRU because the SFR would create its own plutonium from the external supply of uranium. In
the current GNEP scenario, the uranium blankets are removed and TRU is continuously recycled
in the active core driver fuel without targets.


LWR


lom*









enhances the energy sustainability of nuclear power by making maximum use of the initial

uranium mined. Currently, the GNEP fuel cycle (Figure 1-1) does not have a targeted solution

for MAs. Research and development by the GNEP program has not currently revealed a

conclusive decision to include MA within the closed fuel cycle because the americium in MAs is

easily transmuted (even in the fast reactor) into the higher mass actinides of curium, berkelium

and californium. These higher mass actinides, with relatively shorter half-lives, are highly

radiotoxic and sometimes thermally hot. Additionally, these isotopes do not significantly

contribute fissile worth to the SFR compared to plutonium.

If the MAs are not burned in SFRs then they must be discharged from the fuel cycle as

transuranic HLW. The HLW disposal at the repository is one of the cost contributors of the

closed fuel cycle that hinders economic competitiveness with the option to directly dispose of the

SNF without reprocessing. However, as discussed above, it is possible to transmute plutonium

isotopes from the neptunium and americium in the MAs. Approximately two-thirds of SNF

MAs are Np-237 and Am-241. Therefore, it is necessary to design the SFR in such a way that is

favorable to producing plutonium isotopes as opposed to the higher mass isotopes.

Closed SFR fuel cycles have been technologically demonstrated during the United States

Integral Fast Reactor (IFR) program (-1984 1994). In the IFR scenario, the recycling of the

SFR's uranium blanket and driver fuel was integrated into the same reprocessing step (Figure 1-

1). It is proposed in this dissertation that the MA transmutation can be achieved by irradiating

MA targets in lieu of blankets in an IFR amendment to the GNEP scenario.

Transmutation Physics

Transmutation means literally the conversion of one thing into another and is rooted in the

Latin word "trans-mutare" which is "to change". The modern definition was coined by 17th

century alchemists to define the process of converting baser metals into gold. If the meaning of









transmutation is to transform one isotope into another, then transmutation is accomplished

naturally by radioactive decay. It is the relatively long decay half-life of certain actinide and

fission product isotopes that are the essence of the repository storage problem. This is because

their decay heat, radiological and chemical toxicity are present for many hundreds to thousands

of years. Transmutation can be accomplished artificially by adding neutrons to the long lived

atomic nucleus until a less stable nucleus is created with a disproportionate neutron-to-proton

ratio, which is less stable, thus having a shortened half-life.

There are two primary processes for converting long lived isotopes into shorter lived ones.

These are successive neutron capture or fission. Neutron energy determines the extent that

fission can play in the transmutation of certain actinide isotopes. This is because the fission-to-

absorption ratio for MAs is dominated by the neutron flux available for absorption above the one

MeV fission threshold. Generally, the reactor type and corresponding transmutation behavior can

be categorized by the energy range dominated by the neutron energy spectrum. These are fast

and thermal reactors.

Neutron Spectrum Influence on Transmutation Behavior

Because of the large MA neutron capture cross sections at thermal spectrum energies,

thermal reactors have a high efficiency at transmuting MA isotopes by neutron capture.

However, at the same time thermal reactors accumulate other actinides within the fuel cycle that

are equally hazardous and long lived. A primary example of this is the transmutation of Am-

241, which is the principle americium isotope in SNF (Figure 1-2). Am-241 is generated from

the decay of the plutonium isotope, Pu-241, with a half-life of 14.35 years. Am-241 itself has a

half-life of 432.2 years and decays by alpha particle emission. It is the kinetic energy deposition

of this alpha particle in SNF, which dominates the heat generation in the repository for the first

1000 years. Am-241's well resolved thermal neutron capture cross section resonances, in the










thermal to epithermal energy range, expedites transmuting it into the shorter lived isotope Am-

242. Am-242 in turn beta decays into Cm-242 with a yield fraction of 83%. Cm-242 alpha

decays with a half-life of 163 days into Pu-238. Also, the other 17% of the transmuted Am-242

decays by electron capture into Pu-242 which through successive capture reactions followed by

decay results in the production of Am-243 and eventually Cm-244.

Cm-242 Cm-244 Cm-245
a, 163d P- 82.7% a, 18.1a a, 8500a

85% Am-242
Am-241 16.2hr E.C., 17.3%

p-, 14.34 I5a. 1 412.59% p 10.1hr
-24 I.C9 17.m3 ; 6 ;

Pu-238 Pu-239 Pu-240 Pu-241 Pu-242 Pu-243
3.44% 42.98% 21.41% 10.63% 8.51% P-, 4.956hr
n,y 30%

Np-237 Np-238
6.36% p-, 2.1d

Figure 1-2. Important transmutation and subsequent decay of americium and neptunium into Pu-
238, Pu-242 and curium and the higher mass actinides. The percentages shown in red
are the starting concentrations of these isotopes in LWR SNF TRU.

Proposed thermal spectrum transmutation fuels such as LWR Inert Matrix Fuel (IMF) are

very efficient at destroying americium isotopes by transmuting them into Pu-238 and Pu-242 but

also generate Cm-244 to a large extent. Cm-244 has an 18 year alpha decay half-life that

dominates the near term or first century heat generation for spent IMF. Successive neutron

captures from Cm-244 causes the accumulation of higher mass curium, berkelium and

californium isotopes. These isotopes have very little repository impact because their half-lives

are short compared to the expected repository lifetime given in the hundreds-of-thousands of

years. However, the gamma and neutron radiation fields associated with these isotopes dominate

the near term radiological hazard that IMF fuel handlers would need protection from if spent

IMF were recycled and re-irradiated in multiple reactor passes [8].









Alternatively, a SFR has a higher efficiency at destroying TRU isotopes by fission because

a greater share of the neutron flux exists above the threshold fission energy. Essentially all of the

transuranics are more fissionable in a fast rather than a thermal reactor. This means that all TRU

atoms are supporting the chain reaction by contributing to fast fission. Conversely, in the

thermal spectrum, the transmutation process in MAs acts as a capture sink for neutrons generated

by the fissile isotopes.

The single group fission and capture cross sections for a metallic fueled SFR as well as for

a representative IMF assembly are tabulated in Table 1-1 and Table 1-2 respectively. The

neutron physics simulation code Monte Carlo N-Particle was used to generate the single group

cross sections by tallying over all energies within the fuel region of the ABR. This code is used

periodically throughout this dissertation for analysis and verification of results produced by other

codes and is described in more detail in the next chapter. The SFR driver fuel is based on the

ABR with a metallic alloy consisting of 30TRU/60DU/10Zr metallic alloy by weight percent.

The IMF fuel is an 8 v/o weapons grade PuO2 mixed with 2 v/o AmO2 in a Nd2Zr207 matrix [9].

Notice that the fission and capture cross sections are orders of magnitude less for the SFR

than the IMF. It is also interesting to note that the fission-to-capture ratios for the SFR odd-Pu

isotopes are only slightly greater than that for IMF. However, the fission-to-capture ratios for

the SFR MAs are an order of magnitude greater than the IMF MAs. The fission-to-capture ratio

is a strong indicator of whether fission or capture is the primary mode for removing an isotope

from the fuel cycle.

This is especially true when the capture cross section for any given MA is comparable in

magnitude to that of Pu-239. For example, the ratio of Am-241 capture to Pu-239 fission in the

IMF fuel is 1.64 versus 0.83 for the SFR. This indicates that Am-241 capture is a stronger










competitor against Pu-239 fission for neutrons in the IMF case as opposed to the SFR case.

Furthermore, the ratio of total absorption in Am-241 over total absorption in Pu-239 is 1.1 for

IMF versus 0.81 for the SFR.

Table 1-1. SFR fission and capture cross section data based on the Argonne National Laboratory
Advanced Burner Reactor design with a conversion ratio of 0.5.


Fission
(barns)
0.34
1.73
0.10
0.04
0.32
3.80
3.47
1.10
1.71
0.38
2.28
0.26
0.27
3.55
0.20
0.16
2.37
0.42
2.14
0.26
1.93
0.30
0.16
2.43
1.15
2.21
0.63
0.05


Capture
(barns)
0.50
0.47
0.37
0.24
1.31
0.12
0.53
0.62
0.38
0.41
0.37
0.36
1.42
0.31
1.27
0.25
0.20
0.72
0.27
0.19
0.27
0.20
1.09
0.58
0.33
0.27
0.25
0.06


Capture per
(Pu-239)
Fission
0.29
0.27
0.22
0.14
0.77
0.07
0.31
0.36
0.22
0.24
0.22
0.21
0.83
0.18
0.74
0.15
0.11
0.42
0.16
0.11
0.16
0.12
0.64
0.34
0.19
0.16
0.15
0.03


Considering these two facts, americium is a stronger competitor against Pu-239 for

neutrons in the IMF case as opposed to the SFR. Therefore, the thermal spectrum allows more

neutrons to be absorbed into americium rather than plutonium because the thermal americium

capture cross section is larger than the plutonium fission cross section. A higher absorption rate


U-234
U-235
U-236
U-238
Np-237
Np-238
Pu-236
Pu-238
Pu-239
Pu-240
Pu-241
Pu-242
Am-241
Am-242m
Am-243
Cm-242
Cm-243
Cm-244
Cm-245
Cm-246
Cm-247
Cm-248
Bk-249
Cf-249
Cf-250
Cf-251
Cf-252
Combined


Fission per
Capture
0.67
3.71
0.26
0.16
0.25
32.76
6.50
1.78
4.45
0.91
6.19
0.73
0.19
11.51
0.15
0.63
12.14
0.58
7.87
1.36
7.12
1.52
0.15
4.16
3.52
8.22
2.49
0.85


Absorption per
(Pu-239)
Absorption
0.40
1.05
0.22
0.13
0.78
1.87
1.91
0.82
1.00
0.38
1.27
0.30
0.81
1.84
0.70
0.19
1.23
0.55
1.15
0.22
1.05
0.24
0.60
1.44
0.71
1.19
0.42
0.05


Fission per
Absorption
0.40
0.79
0.21
0.14
0.20
0.97
0.87
0.64
0.82
0.48
0.86
0.42
0.16
0.92
0.13
0.39
0.92
0.37
0.89
0.58
0.88
0.60
0.13
0.81
0.78
0.89
0.71
0.46










facilitates greater removal from the fuel cycle and in the case of IMF is caused by nuetron

capture reactions rather than fission.

Table 1-2. IMF fission and capture cross section data based on a typical 17x17 ressurized Water
Reactor fuel assembly design fueled with a TRU-02/Nd202ZrO7 Matrix.


Fission
(barns)
0.62
12.55
0.36
0.13
0.64
54.01
30.31
1.87
19.23
0.70
28.05
0.54
0.82
127.14
0.54
0.48
55.59
1.04
36.23
0.71
16.50
0.90
0.85
60.43
1.12
140.51
4.59
0.44


Capture
(barns)
19.09
4.84
7.68
8.16
19.99
2.29
8.33
8.39
10.27
49.27
8.74
33.07
31.54
23.68
40.02
3.57
6.85
13.16
4.89
2.17
10.29
7.36
71.05
15.13
208.03
54.98
1.66
0.58


Fission per
Capture
0.03
2.59
0.05
0.02
0.03
23.63
3.64
0.22
1.87
0.01
3.21
0.02
0.03
5.37
0.01
0.13
8.12
0.08
7.41
0.33
1.60
0.12
0.01
3.99
0.01
2.56
2.77
0.76


Transmutation and Nuclear Stability

The actinide cross sections in the SFR neutron energy spectrum are almost completely in

the unresolved resonance range. Therefore, the resolved resonance cross section of one isotope

self-shielding the reaction rate of another does not play a significant role. Hence, the ratio of

plutonium versus americium fission is more related to the number density of each isotope and the

percent of neutrons above the threshold energy. In fact, the magnitude of the fission-to-


U-234
U-235
U-236
U-238
Np-237
Np-238
Pu-236
Pu-238
Pu-239
Pu-240
Pu-241
Pu-242
Am-241
Am-242m
Am-243
Cm-242
Cm-243
Cm-244
Cm-245
Cm-246
Cm-247
Cm-248
Bk-249
Cf-249
Cf-250
Cf-251
Cf-252
Combined


Fission per
Absorption
0.03
0.72
0.05
0.02
0.03
0.96
0.78
0.18
0.65
0.01
0.76
0.02
0.03
0.84
0.01
0.12
0.89
0.07
0.88
0.25
0.62
0.11
0.01
0.80
0.01
0.72
0.73
0.43


Capture per
(Pu-239)
Fission
0.99
0.25
0.40
0.42
1.04
0.12
0.43
0.44
0.53
2.56
0.45
1.72
1.64
1.23
2.08
0.19
0.36
0.68
0.25
0.11
0.54
0.38
3.69
0.79
10.82
2.86
0.09
0.03


Absorption per
(Pu-239)
Absorption
0.67
0.59
0.27
0.28
0.70
1.91
1.31
0.35
1.00
1.69
1.25
1.14
1.10
5.11
1.37
0.14
2.12
0.48
1.39
0.10
0.91
0.28
2.44
2.56
7.09
6.63
0.21
0.03









absorption ratio in the epithermal to fast energy range is strongly tied to the relative magnitudes

of the unresolved capture cross section below and the fission cross section above the one MeV

threshold. For example, the capture cross section for Am-241, Am-243 and Cm-244 are at least

two orders of magnitudes greater than fission at energies between one keV and one MeV (Figure

1-3). In addition, the magnitude of the capture cross sections is roughly the same as that for the

Pu-239 fission cross section. If the neutron spectrum for the target is made epithermal, a greater

americium destruction rate can be achieved by the neutron capture mechanism below the one

MeV threshold. A softer spectrum would enhance the neutron capture importance relative to

fission in the competition for neutrons in this energy range. The general increase in the total

absorption probability shifts the competition in favor of the fertile MA isotopes.

As discussed in the previous section, removing MAs by neutron capture transmutes them

into higher mass actinides. However, similar to IMF, the transmutation of Am-241 into Am-242

is followed by a decay path into curium and eventually plutonium. A similar transmutation

behavior is exhibited by Am-243. Am-243 is transmuted by neutron capture into Am-244 which

in turn beta decays into Cm-244 with a half-life of 10.1 hours. Cm-244 decays by alpha particle

emission with an 18.1 year half-life into the even plutonium isotope Pu-240.

As identified earlier, Am-241 has a long half-life and has a fission threshold of one MeV.

Transmuting Am-241 produces the fissile Am-242,242m atom. By fissile, it is meant that

virtually zero kinetic energy is required to overcome the energy barrier for fission to occur. This

fissile characteristic is exhibited by essentially all actinides with an odd neutron number.

Eighty-five percent of this transmuted Am-242 is the ground state which is not long lived

enough to be used as fuel in the reactor [10,11]. Am-242's primary daughter by beta emission is

Cm-242 whose short lived alpha decay results in Pu-238 (Figure 1-3). This Pu-238 has a larger













fission-to-capture ratio in the fast spectrum than the starting Am-241. However, Pu-238 is not


truly fissile because some kinetic energy is required to cause it to fission. But, the amount of


added kinetic energy needed for Pu-238 to fission is much less than that needed for Am-241.


This explains why the Pu-238 fission cross section is so much higher than Am-241 at energies


below one MeV. In addition to being more fissionable than the starting Am-241, Pu-238 is only


one neutron capture away from becoming the long lived Pu-239 atom which is fissile.


Cm-245
8.5E3
7.87
Isotope 7.41
fic fast Am-244 Cm-244
f therm. 10.1hr 18.1yr
flc fast 0.58
f/c therm 0.08
Pu-243 Am-243 Cm-243
4.956hr 3.73E3yr 28.1yr 10
flc fast 0.15 12.14
fie therm 0.01 8.12 E
Pu-242 Am-242 Cm-242l l I
3.75E5yr 17% 16.02hr 83% 162.8d
0.73 11.51 0.63
0.02 5.37 0.13 i .,
Pu-241 Am-241 W )
14.4yr 432.7yr 1 1
6.19 0.19 U
3.21 0.032
Pu-240
6.56E3yr
0.91
0.01
Capture
2.410E4yr 10 10
10- 10- 1
4.45
1.87 Incident Energy (MeV)

Green = Am-241 Capture
2.117d 87.7yr
32.76 1.78 1 Red=Am-243 Capture
23.63 0.22 Gold=Cm-244 capture
p-27 Blue=Am-241 Fission
2.14E6yr Purple=Cm-244 Fission
0.25 Teal=Am-243 Fission
0.03


Figure 1-3. Periodic relationship between fissile and fertile transuranic isotopes


This circular behavior is also exhibited by Am-243 (Figure 1-3). The long-lived Am-243


nucleus is transmuted into Am-244 (fissile) which decays by beta emission into Cm-244. This


Cm-244 has a short half-life compared to the starting Am-243. Its 18.1 year half-life alpha decay


produces Pu-240 which is more fissionable (but not truly fissile) than the starting Am-243 and is


only one neutron capture away from the fissile Pu-241 which is fissile.









Isotopic Aspects of Repository Impacts

For this dissertation work the motivation behind transmutation of Am-241 (and Np-237) is

rooted in the site selection and design constraints of the proposed repository site at Yucca

Mountain in Nevada. Yucca Mountain is a ridge line located in the desert region of Amargosa

Valley approximately 90 miles from the city of Las Vegas. The mountain ridge is formed by

layers of volcanic rock called tufff'. This rock was deposited by falling ash from successive

eruptions of nearby super-volcanoes many millions of years ago. These volcanoes are now

extinct.

The heat generation rate produced by the alpha decay by Am-241 is the main governing

parameter that determines the separation distance between drifts. The drift spacing is determined

by the thermal heat rate produced by the SNF waste packages. The spacing between drifts is the

minimum separation needed to keep the mid-drift rock temperature beneath the boiling point of

water (96C) at the elevation of the Yucca Mountain repository site [12]. The reason for the

prevention of complete "dry-out" between drifts is to ensure that rain water, which is transported

via fractures in the tuff layers, is allowed to flow freely through the repository to the water table

below it.

The second heat generation limit is the drift tunnel wall surface temperature. The drift wall

surface temperature limit is established to be below 2000C to prevent crystalline alteration of the

rock. Wigeland et al indicated that this limit is not strongly influenced by the presence of Am-

241 but rather from the barium and yttrium decay products of the cesium and strontium fission

products [13]. Cs-137 (T1/2=30.07) and Sr-90 (T1/2=28.78) both have half-lives less than 100

years. Therefore, similar to Wigeland, it is assumed for this work that these isotopes could be

separated and diverted from repository storage during the fuel cycle [13]. Figure 1-4 shows the

decay heat contribution for the principal heat generating SNF isotopes.










1E+04

1E+03

1 E+02

1E+01

1E+00

I 1E-01

1 E-02

1E-03

1E-04
1E


+00


1E+01 1E+02 1E+03 1E+04
Time After Irradiation (Years)


1E+05


1E+06


-A-Np-237 Pu-239 Pu-241 --Am-241 -Cm-245 -*-total
Figure 1-4. Decay heat plot for pressurized water reactor SNF (5 w/o starting enrichment and 45
MWD/kg discharge burnup)

There is a third repository design consideration. Because of its long half-life,

radiotoxicity, high solubility and low sorption in Yucca Mountain tuffs, Np-237 is the principle

environmental concern to the biosphere if water does come into contact with the SNF [12]. The

natural and engineered waste package barriers are designed to minimize the likelihood that water

may contact Np-237 for at least 10,000 years. The other soluble SNF isotopes of interest are:

Tc-99, 1-129 and U-234. However, the radiation doses from these isotopes are not expected to

be as significant as from Np-237. This is principally due to the fact that Np-237 is the alpha

decay product of Am-241. Therefore, Am-241 destruction in the SFR is not only important to

the emplacement drift spacing but also important for controlling the Np-237 accumulation in the

repository.

Three repository design benefits are sought for SNF isotope transmutation within the

heterogeneous SFR design evaluated in this dissertation. The first is the removal of Am-241 by









neutron capture in a heterogeneous target. The second is the removal ofNp-237 by neutron

capture and fission within the driver fuel. Also, Tc-99 is evaluated as a control rod neutron

poison for reactor performance purposes with the added benefit that this isotope is not sent to the

repository.

Background on Previously Proposed TRU Burning SFRs

The net TRU destruction of the SFR increases for decreasing conversion ratio. The

conversion ratio is commonly defined in fuel cycle terms as the ratio of the TRU production rate

divided by the TRU destruction rate averaged over the reactor and irradiation cycle. The most

straight forward way to decrease the conversion is by decreasing the parasitic capture of neutrons

in U-238. This prevents the breeding of Pu-239 by transmutation from U-238. Decreasing

parasitic capture can be done by simply removing uranium from the fuel. It can also be achieved

by shifting the neutron balance between parasitic capture and neutron losses by leakage.

Generally, for low conversion ratio SFRs with high MA loaded fuels, there is a tradeoff

between the optimal Doppler and void coefficients and the attainable TRU destruction

efficiency. The two most basic SFR reactivity feedback mechanisms are the Doppler feedback

provided by U-238 in the fuel and blankets; and the increase in axial and radial neutron

streaming that occurs during coolant voiding. The tradeoff stems from the fact that the

mechanisms commonly used to remove neutrons from the reactor during transients also remove

them in steady state operation. For example, enhancing axial streaming with a pancake geometry

or a reduced fuel pin diameter (large sodium fraction) makes the void coefficient more negative.

But, the increased overall leakage reduces the available excess reactivity. Alternatively,

increasing U-238 increases resonance capture and makes the Doppler coefficient more negative.

However, this produces further TRU from captures in the same resonances that provide the

beneficial feedback. In both cases, more fissile plutonium is required to compensate for









reactivity lost by the modifying strategy. The added plutonium and lack of fertile material

increases the rate that the reactor losses reactivity as a function of the burnup of the TRU.

During coolant voiding, the lack of neutron down scattering in sodium causes the neutron

spectrum to harden. This hardening causes an increase in the fission-to-capture ratios and hence

an increase in multiplication [14]. SFR designs typically do not have a passive sodium void

negative reactivity feedback. In case of a sodium density reduction, passive negative reactivity

feedback is normally achieved by thermal expansion of the fuel. Engineered and inherent

negative leakage reactivity feedback mechanisms will be explained in a later discussion on SFR

reactor control.

Most SFR designs are fairly insensitive to changes in the sodium density. However, the

spectrum hardening effect, caused by a complete void of sodium coolant, is more pronounced for

higher MA loadings. The increase in void worth is because more neutrons are absorbed at

energies greater than the threshold fission energy which causes an increased multiplication

feedback. Past parametric studies in the literature showed that the void coefficient increased too

much when the MA content exceeded roughly 10% (MA per mass of HM) [15]. This is because

the multiplication increase given by Np-237 and Am-241 creates more neutrons than the Doppler

broadening feedback from U-238 can absorb. In some instances the constraint on MA density is

as low as 5% or 2.5% such as the Consommation Amelioree du Plutonium dans les Reacteurs

Avances (CAPRA) type core [16,17].

Because the MA loading (in the fuel) is limited by a void coefficient constraint, the

attainable MA destruction rate for that design suffers. Several pathways to optimize total TRU

destruction, and hence MA destruction, while maintaining acceptable reactivity coefficients and

reactivity swing have been considered. First and most importantly, the rate of total TRU









destruction by fission is fixed by the fission reaction rate of the core. This occurs because the

rate at which the overall mass is destroyed by fission is fixed by the power rating of the reactor.

On average, the amount of energy produced by one megawatt of power produced in one day

"Megawatt-day" (MWD) requires approximately one gram of HM to undergo fission.

Therefore, reducing the TRU generation is the only possibility for maximizing the net

TRU destruction, which can only be reduced by eliminating Pu-239 creation from U-238.

Enhanced leakage is accomplished by altering the geometric buckling. Axial buckling can be

increased by reducing the core height with respect to its diameter giving a configuration known

as a "pancake" core. Parasitic neutron capture is increased by the addition of an alternate

resonance absorber. This raises the unique possibility of using fission products such as Tc-99 as

an alternate epithermal absorber for replacing U-238, thus reducing the conversion ratio.

In any case, the reduced "poorer" neutron population in the core caused by these

modifying strategies ultimately results, in most cases, a higher fissile loading to compensate for

neutron losses. Therefore, the burnup reactivity swing increases as a result or the refueling cycle

length being reduced. The first option has the drawback of requiring an unrealistic control rod

worth to manage the excess reactivity. The second option has the drawback of limiting the

irradiation time allotted for burning MAs. This second option is important because the

probability for absorption in the MA is much less than that for plutonium (Table 1-1), even in the

fast spectrum, meaning that more neutrons are being consumed to fission plutonium in the fuel

instead of transmuting MA in the fuel.

The Advanced Burner Reactor Design Concept

The ABR fuel, core internals, and power plant are essentially identical to the Advanced

Liquid Metal Reactor (ALMR) and Super Power Reactor Innovative Small Module (S-PRISM)

designs that were produced during the IFR program [7,18,19]. In order to achieve a low









conversion ratio, the ABR draws upon several of the modifications discussed above. First, to

reduce parasitic capture in U-238, the internal rows of uranium blankets were removed from the

S-PRISM design (Figure 1-5).

These blankets were an integral component to the plutonium self-sustainability of the IFR

fuel cycle but are now contradictory to the TRU burning philosophy of the ABR. Second, the

ABR core height-to-diameter ratio is only 0.5 to enhance axial buckling. Typically, in the

absence of internal rows of blankets, the active core (driver fuel only) of a SFR breeder would

need a height-to-diameter ratio closer to unity [14].

S-PRISM ABR






O Blanket Assembly
SDriver Fuel Assembly
D Middle Enrichment Driver
Outer Enrichment Driver
Reflector Assembly
Shield Assembly
0 Ultimate Shutdown Rod Assembly
0 Primary Control Rod Assembly


Figure 1-5. Core layouts for the S-PRISM and ABR designs2


2 The ABR uses the same driver fuel assembly design throughout the core but with different
TRU enrichments for Inner, Middle and Outer core regions. The three different "enrichment
zones" are necessary for radial power profile flattening. The reflector and shield are both
dummy assemblies filled with steel rods to reflect some of the neutrons back into the core and
also to protect the rector vessel from neutron damage by fast neutrons. Some variations of the
shield include neutron absorbing material. The Gas Expansion Module is a special reflector
assembly designed to provide negative reactivity feedback by increasing leakage in the event of
loss of coolant flow









Parametric studies performed by Morris et al showed that further reductions in height

(from one meter) do not yield significant reductions in conversion ratio [20]. Therefore, the

ABR core height is the same as the one meter height of the S-PRISM. However, the removal of

the blankets allowed the radius to be reduced from 2.7 to 2.2 meters. Finally, the ABR fuel pin

pitch-to-diameter ratio was reduced from that of the S-PRISM. This had the attribute of adding

more sodium volume to the core which also allows for more neutrons to stream out of the core.

The volume of uranium in the pin was reduced in step with the reduction in fuel pin volume.

Hence, the loading of TRU is roughly kept constant across the S-PRISM and ABR designs but

the volume of uranium is decreased. This had the effect of increasing the fuel concentration of

TRU per total HM in the core, also known as the TRUenrichment. The increase in TRU

enrichment dictates a significantly different fuel assembly design in terms of fuel pin diameter,

pin spacers, etc. than the S-PRISM. The fuel assembly and pin dimensions of the S-PRISM and

the ABR reference core design used for comparisons in this dissertation are given in Table 1-3.

As stated by Hoffman et al, the reduction in pin diameter and the corresponding increase in

TRU enrichment cause the fuel excess reactivity to increase and the irradiation cycle length to

decrease [7]. Because, of cycle length reduction, large excess reactivity and large enrichment,

the ABR has a correspondingly different refueling interval, control rod worth requirement and

fuel composition from the current experience database based on past SFR experience.

Transmutation Target Designs: Radial Blankets and Moderated Targets

The principal drawback of having an enhanced leakage SFR core design, such as the ABR,

is that the reactor purposely wastes neutrons through leakage instead of re-investing them in the

fuel. This is contrary to previous plutonium breeders, which used uranium blankets to recover

the neutrons leaked from the active driver core.










Table 1-3. Fuel assembly and pin dimensions for the S-PRISM and ABR designs
Ref. ABR
Fuel Type S-PRISM Driver S-PRISM Blanket CR 5
(CR=0.5)*
Approximate core conversion ratio 1.00 0.50
Fuel type Metal Metal Metal
Fuel alloy composition Pu/U/10-Zr U/10-Zr TRU/U/20-Zr
TRU Enrichment (%) 20 0 35
Assembly pitch, cm 16.142 16.142 16.142
Inter-assembly gap, cm 0.432 0.432 0.432
Duct outside flat-to-flat, cm 15.710 15.710 15.710
Duct material HT-9 Steel HT-9 Steel HT-9 Steel
Duct thickness, cm 0.394 0.394 0.394
Pins per assembly 271 127 324 (7 support)
Fuel pin spacer mechanism Helical wire wrap Helical wire wrap Grid
Spacer wire wrap diameter, cm 0.142 0.094 n/a
Bond material in cladding gap Na Na Na
Overall fuel pin length, cm 407.040 407.040 407.040
Top fission gas plenum height, cm 191.140 191.140 191.140
Middle Active fuel "core" height, cm 101.600 101.600 101.600
Bottom Axial reflector height, cm 114.300 114.300 114.300
Fuel smeared/ fabrication density, % TD 75/100 85/100 75/100
Pin outer diameter, cm 0.744 1.201 0.6230
Cladding thickness, cm 0.0559 0.0559 0.0559
Pin pitch-to-diameter ratio 1.191 1.078 1.293
*The ABR reference design is currently not yet a finalized point design in the SFR community.
The pin diameter, spacer type, fuel composition, etc. vary depending on the desired conversion
ratio of the reactor design [7].

In general, the long mean-free-path of fast neutrons causes SFRs to have a high neutron

escape probability. In fact, the active driver fueled region of most plutonium breeders purposely

had a high leakage. However, the high leakage was created only to maximize the amount of

neutrons invested in the uranium blankets. Therefore, the overall leakage, including the

blankets, would be less than the active core by itself. The use of the blankets increases the

overall utilization of neutrons to produce new fuel in the SFR fuel cycle.

Remembering back to Table 1-1 and Table 1-2, though the fission-to-absorption ratio of

the even neutron number MAs (i.e., Am-241, Am-243, Cm-244) are greater in the fast spectrum

than in the thermal spectrum, they are still much less than that of the plutonium isotopes. In fact,

the fast spectrum fission-to-absorption ratio of most of these MAs is actually closer to that of U-









238. Considering that the transmutation path of most of the MAs leads directly to much more

fissionable material (i.e., Pu-238, Pu-242 and Cm-245) within one to two-neutron captures, it is

possible that they could be a suitable fertile material for replacing the uranium in the blankets.

This heterogeneous approach has been proposed by previous authors studying fuel cycles

similar to that of the ABR. Buiron et al studied wrapping a radial external blanket of un-

moderated target assemblies around the active core region of the European Fast Reactor (EFR)

[21]. Similarly, Fujimura et al and Sanda et al proposed a SFR core where moderated target

assemblies containing neptunium and americium were scattered throughout the active core and

also in a radial blanket surrounding the active core of the Self-Consistent Nuclear Energy System

(SCNES) [22,23,24].

The moderation was provided by a U/MA/Zr/H dispersion matrix. These target assemblies

were first heterogeneously scattered throughout the center core region and eventually shuffled to

the first row of the radial reflector. The shuffling was done to ensure that the plutonium atoms

generated from neptunium and americium transmutation did not create an unacceptable power

peak at the end of the one year irradiation. This study demonstrated that the overall destruction

efficiency is enhanced by including moderation to the target design. It also demonstrated a fuel

cycle scenario where the transmuted plutonium from the targets was recycled into the driver

assemblies. Similar results were observed separately by Eliseev et al and Rome et al. Eliseev et

al proposed a target assembly having the outer two rows of the assembly comprised of

neptunium and americium target pins with all inside rows being moderator rods filled with

zirconium hydride [25]. Rome et al also proposed a heterogeneous target assembly where the

moderating rods and transmutation target rods were equally distributed throughout the assembly

[26,27].










A table of most relevant transmutation target design studies is offered in Table 1-4. The

transmutation target design proposed in this work is a hybrid compilation of:

Transmutation targets located on the top axial top core periphery to have a close
proximity of the target and active core regions for flux sharing

A heterogeneous lattice of MA pins and zirconium hydride moderating pins to take
advantage of existing metal fuel technology for the target transmutation matrix

A slight uranium content in the MA axial blanket/target region of the core (i.e., in
the MA loaded pins) for minimizing power swing in the targets during irradiation

In all of the radial heterogeneous core designs with target assemblies, the moderating effect

of neutrons leaving the target and entering the driver fuel assembly caused localized power

peaking in the neighboring fuel pins. To compensate, all the moderated target assembly designs,

found in the literature, used a thermal neutron filter that encompassed the target assembly to

ensure that thermal neutrons did not leave the target. The filter used by Fujimura et al replaced

the peripheral ring of rods within the target assembly with Tc-99 bearing rods. Eliseev et al used

a cadmium laced fuel assembly shroud as the thermal filter.

Table 1-4. Summary of relative studies on transmutation target assembly: matrix compositions
and moderating strategies
Target Moderator Reference
Author Reactor Type Target Location Target Moderator Reference
Composition
Buiron et al EFR Outer radial blanket MA-O2/UOX n/a 21
Both active core and
Fujimura et al SCNES Both active core and U/MA/Zr/H Hydride Matrix 22,23,24
radial blanket
Both active core and
Sanda et al SCNES ative e and U/MA/Zr/H Hydride Matrix 28
radial blanket
MA-02 "rock-
Eliseev et al BN-800 Outer core 2 "r- ZrH2 Pins 25
like"
Internal and outer
Rome et al EFR Ieral an oter MA-02 CaH2 or ZrH2 Pins 26,27
radial blankets

Transmutation Target Designs: Axial Blankets and Axial Targets

The issue of localized power peaking may be avoided altogether by taking advantage of the

large spatial gradient typical of most SFR radial and axial power profiles. A large flux and

power profile gradient is typical of most SFR core designs because of the large geometric









buckling caused by high leakage. Given that the power on the periphery is significantly less than

the peak power at the core's axial and radial center, localized power peaking from moderation

can actually be favorable as a means to flatten the power profile across the core. In fact, Rome et

al found that the issue of power peaking was minimized when the moderated targets were placed

exclusively on the core periphery and not in the core center. In this work a moderated axial

target/blanket is proposed for capturing the neutrons leaked from the active core. As it will be

shown in Chapter 3, the affect of this moderation actually decreases the axial power gradient in

the target region.

Axial blankets are actually not a new concept in SFR design. Table 1-5 gives a list of

examples of past reactor cores that incorporated axial blankets into their designs [29]. Axial

target designs are also not a completely foreign concept for transmutation purposes. Kuraishi et

al and Arie et al proposed replacing radial and axial blankets with fission product targets [30,31].

Table 1-5. Examples of real fast reactor plants where axial blankets have been incorporated into
the SFR core design
Reactor Location
EBR-I (United States) Above and Below
EBR-II (United States) Above
FERMI (United States) Above and Below
Clinch River (United States) Above and Below
JOYO (Japan) Above and Below
MONJU (Japan) Above and Below
Dounrey (Great Britain) Above and Below
Rapsodie (France) Below
Phenix (France) Above and Below
Super-Phenix (France) Above and Below
BN-350/600/800 (Russia) Above and Below

Transmutation Based Reactivity Control Concept

Due to the fast neutron's long mean-free-path, spatial heterogeneities on the dimensional

level of a fuel pin virtually have zero impact on the local neutron flux. Therefore, the major

geometry features of SFR are on the dimensional level of the fuel assembly. The fuel









enrichment zoning in the ABR is possible due to this "smearing" effect. This smearing effect

over the fuel assembly is also why the SFR has whole "primary control" and "ultimate shutdown

control" assemblies as shown in Figure 1-5. This smearing principle of neutron absorbers in a

SFR makes introduction of burnable poison materials problematic unless they can be

implemented in standalone fuel assembly structures.

Kim et al explored the homogeneous pin distribution and heterogeneous assembly

distribution of burnable poisons in the Korean Advanced Liquid Metal Cooled Reactor

(KALIMER) [32]. The burnable poison used was boron carbide (B4C) which was enriched to 90

w/o in the highly absorbing B-10 isotope. The purpose of the KALIMER study was not only to

determine the B4C's potential excess reactivity suppression benefit, but also to determine the

affect that the neutron poison would have on the void and Doppler feedback.

The first case studied by Kim was a homogeneous loading where 30 of the 271 fuel pins

were replaced with burnable poison rods. The second case was a parametric study to determine

the optimal location for burnable poison assemblies instead of homogeneously placed rods.

Homogeneous and heterogeneous burnable poison options were compared to a non-poisoned

base case. It was found that the homogeneous case actually made the sodium void worth more

positive than the base case. This is because the U-238 fast fission contribution increases in the

harder spectrum attained during voiding. The sodium void worth increases because the thermal

(resonance) absorption cross sections of B-10 (and also the actinides) are significantly greater

than U-238. Therefore, the neutron spectrum becomes harder during voiding. Thus, undesirable

fast fission neutron multiplication surpasses the desirable neutron consumption by parasitic

capture. The heterogeneous loadings had the opposite effect because the burnable poison

assembly is closer in physical dimensions to the fast neutron mean-free-path. Since the burnable









poison becomes more visible to fast neutrons, the achievement of lower sodium void worth is

primarily owing to the parasitic capture in the burnable poison assembly. However, just as in

most "typical" SFR designs the void coefficient was still positive.

This concept of burnable control assemblies is applied to the primary control assembly

design discussed in Chapter 4. However, the fission product technetium (Tc-99) is used instead

of B4C. Metallic Tc-99 is chosen as for reactivity control shim because:

* The atom density of Tc-99 is higher than that of B-10 in enriched-B4C.

* Tc-99 has an unresolved cross section structure in the fast neutron energy range similar to
U-238 making it suitable for a gray absorber rod in a SFR.

* The combination of long fast neutron mean-free-path with the gray absorbing Tc-99 in
discrete primary control assemblies is ideal for reactivity shim control.

* The magnitude of the Tc-99 unresolved cross section resonances falls off sharply at
energies above one MeV which could promote the above-threshold fission of MAs.

* Transmuting Tc-99 reduces the amount of this radiotoxic and long lived SNF (or HLW)
isotope that would otherwise be destined for the Yucca Mountain repository.

Messaoudi et al investigated burning Tc-99 as a potential replacement for all of the U-238

in the CAPRA core design [33]. The CAPRA design achieved a low conversion ratio by

replacing fuel pins from the driver fuel assemblies with dummy "dilution" pins of stainless steel

(or Tc-99 as proposed by Messaoudi et al). Without changing the loading of TRU in the CAPRA

core, the dilution rods displaced uranium from the core. This strategy was similar to what was

done for the ABR but without changing the fuel pin diameters.

However, in the Messaoudi et al CAPRA design all U-238 was displaced from the core

and was replaced by Tc-99 as a potential resolved capture resonance surrogate. It is important to

note that Tc-99 has some resolved and unresolved resonances similar to U-238 (Figure 1-6).

Messaoudi et al first loaded Tc-99 homogenously in all fuel pins. It was found that this

homogeneous loading increased the void coefficient compared to the base case with U-238











present. Similar to the results achieved by Kim et al, the increase was explained by the fast

fission multiplication feedback becoming greater than the compensation of the Doppler

broadening feedback. Unlike Kim et al, Messaoudi et al chose not to concentrate the Tc-99 in

specialized standalone assemblies. Instead, a calcium hydride moderator was placed in dilution

pins of discrete fuel assembly locations throughout the core. For the CAPRA design these

dilution assemblies comprised every third row of fuel assemblies in the core. The moderation

allowed more neutrons to be down-scattered to lower energies where the Tc-99 resolved

resonances could have more effect which reduced the magnitude of the positive void coefficient.


1E+04
1E+03 -
1E+02
1E+01
S1E+00 -
r 1E-01 -
O 1E-02
O
( 1E-03
1E-04
-U-238
1E-05 -Tc-99
1E-06
1E-07
1E-07 1E-06 1E-05 1E-04 1E-03 1E-02 1E-01 1E+00 1E+01 1E+02
Energy (MeV)

Figure 1-6. ENDF-VI plot ofU-238 (blue) and Tc-99 (red) total absorption cross sections

Design Rationale of an Axial Heterogeneous Fast Transmutation Reactor

Because the SFR's resonance feedback is highly dependent on the presence of sodium, it is

important to have a design that can compensate with alternative neutron sinks during the event of

a loss-of-coolant-accident (LOCA). The large mean-free-path of fast neutrons allows feedback

by neutron streaming as a common solution to this problem in many SFR designs.

However, designing core geometry for high streaming neutron losses during LOCA's also

requires a high neutron leakage during steady-state operation. Therefore, implementing axial









targets will have a dual role: (1) MA conversion into more fissionable plutonium isotopes, and

(2) recovering the axial leakage lost by the active core during steady-state operation.

Compensation for Inherent Positive Void Reactivity Feedback

It was discovered in the operational experience of the Experimental Breeder Reactor I

(EBR-I) that the SFR core's reactivity was highly sensitive to core geometry. During overpower

tests, the EBR-I exhibited oscillatory power characteristics and periods of prompt positive

(increasing with increasing power) reactivity feedback. Operating the reactor in order to

quantify these prompt reactivity feedbacks caused a partial meltdown of the EBR-I Core-II fuel.

It was later found that the positive power coefficient was due to thermal-mechanical bowing of

the fuel pins towards the core center. The bowing was caused by axial and lateral temperature

differentials across the fuel tubes. The inward bowing made the reactor more compact thus

causing the reactivity increase [34]. The positive power feedback was resolved in the

Experimental Breeder Reactor II (EBR-II) by removal of the upper grid plate [35]. The lack of

a lateral constraint at the top of the core allowed the fuel to bow outwards, instead of inwards,

under thermal gradients. The increase in outward bowing with increasing temperature thus

enhanced neutron streaming. Hence, the bowing and subsequent enhancement in neutron losses

during a LOCA provided an inherent negative reactivity feedback which was demonstrated in

tests performed at EBR-II.

The inclusion or concentration of MAs in the central region of the core can actually

increase the positive void feedback. Compared to U-238, most MAs do not possess resolved

resonances at energies high enough in the SFR fast flux spectrum to provide useful amounts of

negative reactivity feedback. Also, because the fast flux energy spectrum becomes harder during

a LOCA, the above-threshold multiplication contribution from MAs further complicates and

increases the positive reactivity void feedback by the fuel.









The positive reactivity feedback by the MAs creates a need to further enhance the negative

streaming feedback of the core design. Finding a core geometry that possesses an enhanced

leakage property during a LOCA without removing too much reactivity from the core during

steady-state operation is essentially the Holy Grail of SFR design. Traditionally, flattening the

core height-to-diameter ratio into a "pancake" geometry has been the most straightforward route

for achieving this higher leakage. Remember back to the previous discussion on core geometry

modifications for attaining a TRU burner, a pancake geometry has also been proposed for

achieving a "very low" conversion ratio. However, as discussed earlier, the high leakage

necessitates a high TRU enrichment and a shortened cycle length [36]. Therefore, there is a

practical limit on how flat the pancake core can be made.

In this work, a hybrid pancake core design is sought that slightly reduces the height of the

ABR design while preserving the volume of its active core. The flatter core design:

* Enhances neutron losses during a LOCA compared to a taller core

* Reduces the conversion ratio of the active core compared to the ABR

* Invests the neutron leakage from the active core during steady-state operation into
transmutation of new plutonium in the axial targets

Benefits of Axial Targets for a Dedicated MA Burner Core Design

Earlier studies indicate that a low conversion ratio, attained by a high leakage core design,

is necessary to destroy the undesired transuranics waste isotopes found in SNF. From a physics

standpoint, a low conversion ratio is ideal for reducing the production of plutonium and MAs by

reducing parasitic capture. This work makes a fundamental change in philosophy regarding MA

waste management in the SFR. Previous repository studies establish the limitation that the Am-

241 isotope puts on the quantity of SNF that can be stored in Yucca Mountain [13]. Because this

isotope is not fissile, its removal from the fuel cycle can best be achieved by increasing its









transmutation rate by neutron capture. Because this neutron capture in americium leads to a

transmutation path ending in Pu-238 (and the other plutonium isotopes to a lesser extent), it can

be used as a fertile blanket material.

The reactor design proposed and analyzed in this dissertation uses both a moderated

epithermal transmutation target fuel and an un-moderated fast spectrum driver fuel to achieve the

maximum MA transmutation rate. Therefore, the design has a heterogeneous fuel configuration

where excess neutrons from the fast driver zone are used to transmute fertile MA in the

epithermal target zone. For the remainder of this dissertation, this hybrid MA-to-plutonium

converter reactor is given the name: Axially Heterogeneous Fast Transmutation Reactor

(AHFTR)

This breeding of transmuted isotopes can be accomplished with a MA fast flux trap at the

axial top periphery of the active core. The goal of such a flux trap is to recover the fast flux

leakage from the active core and moderate it to a softer spectrum which enhances neutron

capture in the MA rich target. The relationship of this flux trap (target) to the active core (driver)

is shown in Figure 1-7. This symbiotic arrangement allows the driver fuel to have a low MA

content while the MA concentration in the targets is significantly higher. The moderation in the

targets increases the capture cross section magnitude (especially in Am-241) relative to fission

(especially in Pu-239). Thus, the active core neutron flux entering the axial targets is invested

into capture reactions leading to transmutation as opposed to being lost to leakage.










0
0

0

O0
0
0
0
0
C


- Handling Socket


c P



CD
o &t
tl


Coolant Inlet Nozzle


Figure 1-7. Core design of the Axial Heterogeneous Fast Transmutation Reactor


Gas Expansion Module
Ultimate Shutdown Rod Assembly
Primary Control Rod Assembly
Inner Enrichment Zone Driver
Middle Enrichment Zone Driver
Outer Enrichment Zone Driver
Radial/Axial Reflectors
Shield Assembly
Axial Target Portion of Driver Assembly
Fuel Rod Gas Plenum


Z
---









When the transmuted product from the targets is recycled and used later as driver fuel, the

Pu-238 (and other transmuted fissile material such as Pu-238 and Cm-245) will be exposed to the

fully fast flux of the active core. There, the transmuted Pu-238 will have higher "fast spectrum"

fissile worth than the initially loaded MAs (Table 1-1). To take advantage of this fast fissile

improvement, the spent axial targets are co-reprocessed with the spent driver fuel and included in

the fresh driver fuel fabrication for the next reactor pass (Figure 1-1 and Figure 1-8).

TRU
LWR Reprocessing

I.. .. 1111



Am+Cm+Bk+Cf AHFTR
-----------------------
Target Fuel Fabrication --------
- - - .


Driver Fuel Fabrication Fuel Assembly
Np+Pu Manufacturing r r


TRU

SFR Reprocessing I--i ""


Figure 1-8. Modification to the ABR fuel cycle by the axial targets: Solid lines represent the
reference ABR fuel cycle. Dashed line represents modifications to include axial
targets.

The key technical attribute of converting MAs into plutonium isotopes is the relaxation of

the requirement to have a low conversion ratio. As will be discussed in the next chapter,

homogenous core designs, such as pancake designs or the ABR, require a high TRU enrichment

in order to achieve a low conversion ratio. The AHFTR axial blanket converts MAs into

plutonium. Because, one type of TRU is being traded for another, the targets have a negligible

impact on the overall core conversion ratio. However because the MAs are preconditioned into










plutonium before being applied to the active core, the fissile worth of TRU used by the driver

fuel is increased. This effect allows the AHFTR fuel composition to be more comparable to IFR

fuel testing experience than ABR designs of equal conversion ratio. The fuel assembly and pin

dimensions of the AHFTR and two ABR core designs are given in Table 1-6.

Table 1-6. Fuel assembly design of the AHFTR compared to similar ABR designs
Ref ABR
Fuel Type AHFTR ABR (CR=0.75) (CR 0.
(CR=0.5)
Approximate core conversion ratio 0.70 0.75 0.50
Fuel type Metal Metal Metal
Driver fuel alloy composition TRU/U/10-Zr TRU/U/10-Zr TRU/U/20-Zr
Axial target alloy composition MA/Pu/U/40-Zr n/a n/a
TRU Enrichment (%) 20 25 35
Assembly pitch, cm 16.142 16.142 16.142
Inter-assembly gap, cm 0.432 0.432 0.432
Duct outside flat-to-flat, cm 15.710 15.710 15.710
Duct material HT-9 Steel HT-9 Steel HT-9 Steel
Duct thickness, cm 0.394 0.394 0.394
Pins per assembly 271 271 324 (7 support)
Helical wire Helical wire .
Fuel pin spacer mechanism Grid
wrap wrap
Spacer wire wrap diameter, cm 0.1329 0.1329 n/a
Bond material in cladding gap Na Na Na
Overall fuel pin length, cm 407.040 407.040 407.040
Top fission gas plenum height, cm 191.140 191.140 191.140
Middle Active fuel "core" height, cm 91.600 101.600 101.600
Bottom Axial reflector height, cm 114.300 114.300 114.300
Fuel smeared/ fabrication density, % TD 75/100 75/100 75/100
Pin outer diameter, cm* 0.755 0.755 0.623
Cladding thickness, cm 0.0559 0.0559 0.0559
Pin pitch-to-diameter ratio 1.176 1.176 1.293
*The pin diameter varies depending on the desired conversion ratio

A more practical driver fuel composition allows the fuel assembly design to be more

similar to what was proposed for the S-PRISM. Table 1-7 compares the main differences

between the AHFTR core design and the ABR designs.

Technology Compatibilities and Synergies

One of the technology goals of the GNEP program is to establish a higher level of

accountability for HLW mass streams than previous fuel cycles. Aqueous reprocessing (using









liquid organic solvents) of SNF has been performed in Europe, Russia and Japan on a

commercial scale using the Plutonium-Uranium Redox Extraction (PUREX) process.

Table 1-7. Core design for the AHFTR compared to similar ABR designs
Fuel Type AHFTR ABR (CR=0.75) ABR (CR=0.5)
Driver Assemblies 192 144 144
Inner (Lowest Enrichment) 42 30 42
Middle (Medium Enrichment) 66 42 66
Outer (Highest Enrichment) 84 72 36
Primary Control Assemblies 16 16 16
Ultimate Shutdown Assemblies 3 3 3
Gas Expansion Modules 6 0 0
Reflector Assemblies 96 90 90
Shield Assemblies 66 60 60
Core Diameter (m) 2.6 2.3 2.3
Reactor Diameter (m) 3.2 3.0 3.0
Active Core Height (m) 71.6 101.6 101.6
Total Core Height (m) 91.6 101.6 101.6

However, this technology extracts relatively pure uranium and plutonium products and

leaves behind a highly radioactive "mixed waste" of fission products and MAs which are also

considered HLW. From the discussion on the isotopic aspects of the repository design, it is

evident that not all of these isotopes carry the same importance for needing disposal in the

repository. Therefore, the GNEP program and its predecessors have developed a suite of

addition separations steps to aqueous reprocessing, called Uranium Extraction Plus (UREX+)

[37]. The UREX+ reprocessing technology, and its associated sub-steps, divides the HLW

stream into individualized mass streams. Uranium Extraction (UREX) by itself is an aqueous

partitioning technology designed only for creating a highly pure uranium stream from SNF.

UREX is expanded into UREX+ by adding additional extraction steps for partitioning:

technetium, iodine and Cesium/Strontium.

Actinide Partitioning: PUREX and UREX

An additional advantage of the UREX or UREX+ suite is the ability to control the level of

elemental actinide partitioning. The only partitioning possible by the PUREX process is the









creation of a plutonium and uranium product. The production of pure plutonium raises some

debate over the nuclear weapons proliferation resistance of this technology. The UREX+ suite

was designed to provide elemental partitioning such that plutonium is always diluted in another

actinide. For example, the UREX+la process separates all the TRU together as one product so

that plutonium is diluted over all the transuranics. As seen in Table 1-8, additional separations

steps to the UREX+la process creates an increasing number of product streams.

Table 1-8. Waste stream partitioning afforded by various reprocessing technologies*
Process Prod. 1 Prod. 2 Prod. 3 Prod. 4 Prod. 5 Prod. 6 Prod. 7
PUREX U Pu MA+all FP
UREX+1 U Tc Cs/Sr TRU+Ln FP
UREX+la U Tc Cs/Sr TRU FP+Ln
UREX+2 U Tc Cs/Sr Pu+Np Am+Cm +Ln FP
UREX+3 U Tc Cs/Sr Pu+Np Am+Cm FP+Lanth
UREX+4 U Tc Cs/Sr Pu+Np Am Cm Cm+Ln
Pyroproc. U TRU All FP
*Prod. is product. Tc is technetium. Cs is cesium. Sr is strontium. Ln is lanthanides. FP is
fission products

Of these options, the UREX+la process has the most resemblance to PUREX. The

primary difference is that plutonium is diluted by all of the MAs found with it in SNF. As

mentioned in the motivations and objections section, inclusion of these MAs in the SFR driver

fuel can lead to an accumulation of curium, berkelium and californium [38]. The associated

gamma and neutron radioactivity, as well as thermal heat, associated with decay of these

actinides may significantly complicate fuel handling and fabrication of recycled fast reactor fuel.

These high radiation fields raise the possibility that expensive hot-cell facilities would be

necessary for all fuel handling operations. Hot-cells are large monolithic shielded facilities

where radioactive material is handled remotely using mechanical master-slave manipulator arms.

Because of their large size and relative complexity, it is expected that hot-cell facilities are

significantly more expensive than glove-box facilities. The PUREX fuel cycle allows fabrication

in glove-box environments because the MAs and higher mass actinides are removed during









reprocessing. Also, as a general rule, the commercial PUREX fuel cycles to date have not dealt

with the multi-recycling scenario. Hence, the accumulation of highly radioactive curium,

berkelium and californium has not been dealt with on a commercial level.

To avoid the fuel handling penalty, it has been proposed by Pillon et al that the

complexities of MA management in hot-cells could be limited to a small fraction of the fuel

cycle infrastructure [39]. Because the MAs constitute only a small fraction of TRU, the

associated hot-cell infrastructure (MA) can be made significantly smaller than the glove-box (Pu)

infrastructure. By constantly partitioning the MAs from the plutonium driver fuel, after each

reloading cycle, and continuously recycling them in targets the driver fuel fabrication process

could be performed with less difficulty.

To achieve the MA partitioning, the UREX+3 process from Table 1-8 is selected. This

option provides that plutonium is not separated by itself but diluted by neptunium. Also the

MAs, of highest importance to repository and hot-cell criteria (Am+Cm+Bk+Cf), are partitioned

and sent to target fabrication in Figure 1-8. The fuel cycle in Figure 1-8 is slightly different than

that proposed by Pillon et al. Instead of continuously separating MAs from the driver fuel and

sending them to targets, it is assumed that the transmutation conversion efficiency of americium

into plutonium is sufficient enough to not require that the MAs be multi-recycled in the targets.

Therefore, the MAs separated from SNF are irradiated only once in the targets before this mass

is co-reprocessed with the driver fuel into the next batch of driver fuel. Special attention is given

to the sizing of the target region in the AHFTR to ensure that the concentration of MAs in the

driver fuel is small.

Pyroprocessing and the Integral Fuel Cycle

The small concentration of MAs in the driver fuel and the single-pass target irradiation

makes feasible the choice of metallic fuel and metallic fuel reprocessing. Metal alloys of









uranium, plutonium and zirconium were successfully used as reactor fuel at both EBR-I and

EBR-II for over 40 years. Metal fuel reprocessing was demonstrated at EBRR-II during the IFR

program using the pyrometallurgical process involving metal alkaline salts (Table 1-8) [40]. The

benefits of applying pyroprocessing to any SFR fuel cycle are:

* Pyroprocessing does not require water, avoiding many criticality safety issues.

* Pyroprocessing is performed at high temperatures and does not use organic solvents. From
the PUREX experience, organic solvents decompose when exposed to the radiation fields
of SNF/HLW.

* The reprocessing machinery of pyroprocessing is considerably more compact than aqueous
reprocessing which makes it more suitable for deployment in hot-cells.

* Similar to UREX+la, pyroprocessing does not allow for actinide partitioning which
increases its general proliferation resistance.

Because of these attributes, pyroprocessing sufficiently meets the requirements for

reducing the infrastructure dealing with MAs. The primary difference from the approach taken

by Pillon et al, is that the entire AHFTR and its associated IFR style co-reprocessing (targets plus

driver) perform the task of dedicated MA burning. Hence, the AHFTR and its associated

reprocessing machinery constitute a small fraction of the overall fuel cycle and ABR/AHFTR

mixed-fleet. In this fuel cycle, the AHFTRs and ABRs would work in parallel to accomplish the

overall directive to burn SNF TRU. The relationship between ABRs and AHFTRs in this hybrid

partitioning and transmutation strategy is given in Figure 1-9.

In Figure 1-9, the AHFTRs are the only reactors committed to destroying the MAs

separated from SNF. For the purpose of this dissertation, the ABRs in the mixed-fleet are also

assumed to use pyroprocessing technology which may or may not require hot-cell facilities due

to an eventual buildup of higher mass actinides (Cm+Bk+Cf). The rate of this buildup for

various partitioning strategies is currently the topic of ongoing studies in the transmutation

analysis field. Given the suitability of metal fuel with pyroprocessing, it is assumed that if these










technologies are adopted for the AHFTR they will reach economical maturity and would

ultimately be adopted for the ABRs as well.

Pyroprocessing
Reference
ABR ....
Recycling TRU ..R 11111'
Strategy


Np+u Driver Fuel Fabrication Fuel Assembly
Manufacturing



ABR
TRU

-,, LWR SNF

UREX+3 Aqueous Processing

Am+Cm+Bk+CfFTR
AHF TR
---------------- -------
Target Fuel Fabrication [-


Np+Pu Driver Fuel Fabrication Fuel Assembly
S Manufacturing

Proposed I
Addition to TRU
the ABR
Strategy SEE N
Pyroprocessing I--i
.... l11111

Figure 1-9. Partitioning and transmutation scenario of a combined ABR and AHFTR mixed-
fleet

Assumptions for Using Transmutation Targets in SFRs

It is important to mention that the choice to use pyroprocessing in hot-cells which are

collocated with the ABR or AHFTR is based on the fact that the need to transport spent SFR fuel

to a centralized reprocessing plant is eliminated. Avoiding the need for a large centralized









reprocessing plant may have a significant economic advantage because the cost of transportation

is eliminated. Also, it can be argued that the fast reactor fuel cycle services will require

collocation with the reactor in order to simplify the handling issues associated with "hot" fuel.

Though centralized reprocessing and fuel fabrication facilities benefit from economy of scale,

they have not been demonstrated on a large scale with the high fissile concentrations and

possible radiation fields inherent with fast reactor fuels. LWR SNF is stored for a period no less

than approximately five years in order to minimize the radiation fields from fission products (and

Cm+Bk+Cf). When spent fuel (LWR or SFR) comes out of the reactor it is thermally hot due to

the decay energy of these fission products. To cool the SNF during the decay time, LWR

operators store the spent fuel assemblies in large pools of water before transportation them

offsite. Fast reactor operators will not have the luxury of using water to cool the fuel because of

the high fissile content in the driver assemblies. Therefore, the cost to cool spent fast reactor fuel

could be a foreseeable additional non-trivial cost if an interim decay period is required before it

can be reprocessed. These assumptions are stated in Table 1-9.

Using Table 1-9 as a guide, fast reactors with heterogeneous targets are chosen because

similar core configurations have been used in the past explicitly for the purpose of transmutation.

Historically, the main transmutation process in SFRs was the conversion of U-238 into fissile

Pu-239 atoms. However, regardless of whether or not the transmutation product is fissile, the

inherent neutron economy of fast reactors enables bombardment of heterogeneous targets using

excess neutrons. Furthermore, fast-fission by all actinides, including MAs and their

transmutation products, contributes to reactivity which gives them a neutronic advantage over

LWRs where transmutation is concerned. As discussed earlier, moderating pins in the target

region is adopted to provide a spectrum benefit that can enhance transmutation. Finally,











pyroprocessing of metallic fuel at a hot-cell facility which is co-located with the SFR is chosen.


These technologies have been demonstrated during the IFR program to have an industrial


synergy when implemented together.


Table 1-9. Technology compatibility assumptions with Pros and Cons (Technology options
indicated by t represent technologies adopted for this dissertation)*


MA targets in
LWRs


MA targets in
SFRst

Moderated
targets
Un-moderated
targets
Once-Through
Deep Bum of
Targets
Multi-
Recycling of
Targets
Aqueous
reprocessing
with SFRs

Pyroprocessing
with SFRst

Co-location of
reprocessing
with SFRs
Centralized
Reprocessing
Oxide fuel and
targets
Metal fuel and
targets


Using existing reactors utilizes existing
technology. Thermal capture cross
sections are high for most SNF MAs
Fast reactors derive all of its reactivity
from fast-fission enabling all actinides to
be a neutron source (as opposed to a
sink)
Utilization of neutron capture enhances
conversion into even-plutonium isotopes
Simplifies materials science and reactor
physics issues
Eliminates recycling/handling of higher
mass actinides

Ultra-long irradiations are not necessary.
Closed fuel cycle allows a net-zero MA
mass balance

Economic advantages due to the
economy of scale
Pyroprocessing has been demonstrated
with SFRs and is more adaptable to
handling thermally "hot" spent fast
reactor fuels

Transportation of spent fast reactor fuel
is eliminated

Economic advantage due to the economy
of scale
Oxide fuels are well recognized in LWR
commercial experience
Metal fuels have been demonstrated as
compatible with SFRs but not
demonstrated on a commercial scale


MAs and most of their transmutation daughters are neutron
sinks. Conventional LWRs do not possess the surplus of
neutrons required to continuously irradiate MAs (and transmuted
products) once the initial fissile material has been exhausted.
Implementation of SFRs requires reprocessing technology in
order to fully utilize the reactivity investment made by
transmuting fertile isotopes including the MAs.
Moderated targets require placement on the core periphery to
minimize adverse power peaking and kinetics feedbacks. Less
flux on periphery.
No spectrum benefits that could potentially enhance
transmutation.
Achieving a net-zero MA mass balance with "deep-bum" is
difficult in any real reactor system. Material science does not
currently exist to withstand ultra-long irradiations.

Shielding requirements of higher mass actinides essentially
requires a hot-cell infrastructure for MA recycling.

Spent fast reactor cooling will be considerably more expensive
than in LWRs because cooling methods other than spent fuel
pools "water" will be required for criticality safety.
Pyroprocessing does not allow for elemental transuranic
partitioning which requires that the targets and driver fuel be
reprocessed together thus preventing complete segregation of
MA and plutonium mass streams within the fuel cycle

Fuel reprocessing and fabrication facilities for every reactor site
adds extra capital cost

Transportation of fuel is required which also necessitates a spent
fast reactor fuel cooling facility
The higher bumup and temperatures of SFRs mutes many
advantages such as fission gas retention and pellet integrity
Metal fuels are not widely recognized by the commercial nuclear
industry


*Further advances in any of these technologies may change the validity of the assumptions made.









CHAPTER 2
COMPUTATIONAL METHODS AND FAST REACTOR PHYSICS

Many of the methods for cross section preparation and flux calculations used to study

LWR core physics today, have their origins in the analysis of fast reactors. This is partly

because fast reactors were essentially one of the first nuclear energy systems. It is also due to the

fact that the spatial smearing approximations made to simplify lattice and pin-cell calculations

for these early methods were usually sufficient to achieve acceptable accuracy in fast reactor

applications. Due to the lack of SFR commercialization, the demand to update fast reactor

simulation and computational methods has been low. Furthermore, due to the general adequacy

of smearing fuel pin and assembly heterogeneities over large core regions, fast reactor analysis

methods have not evolved to the extent of those used for LWR analysis over the past two

decades. Several calculation codes are used throughout this dissertation. Figure s and Table s

will be labeled according to the calculation method used. Sometimes the results produced by one

code are used as the input data for a secondary code. If this is the case, special mention of the

coupling process is given in notes below the graphic or within the nearby text.

Calculations and Fuel Cycle Modeling

The Argonne National Laboratory (ANL) fast reactor codes MC2-2, DIF3D and REBUS

are used for the reactor physics and fuel cycle calculations [41,42,43]. These codes have been

developed together by ANL for SFR design since the mid-1970's and are well benchmarked

against SFR operational data [44]. The MC2-2 code was used to generate region dependent, 33

energy-group cross sections at hot fuel, cladding and coolant temperatures based on an ultra-fine

group cross section library. The MC2-2 ultra-fine libraries were pre-processed at ANL from

evaluated nuclear data files (ENDF) and using a standard fast spectrum distribution for

appropriately weighting the fine group generation. Starting with the ultra-fine group libraries,









MC2-2 creates a collapsed coarse group cross section set by performing a zero dimensional

infinite dilution critical buckling search using the extended P1 method [41]. The MC2-2 code

also performs a resolved resonance broadening treatment, to account for energy shielding only,

at user defined material and fuel temperatures. A continuous slowing down calculation is

performed at lower energies to handle elastic scattering [45]. A direct multigroup spectrum

calculation accounting for inelastic and anisotropic scattering and upscattering is performed at

higher energies.

It should be noted that the calculations performed by the publically available release of this

code are limited to using an older ENDF/B-V.2 release of cross section data because the publicly

available release ofMC2-2 does not come standard with the preprocessed ultra-fine group

libraries generated from newer ENDF/B-VII data. Also, it should be noted that there are two

additional code packages (SDX and DB2), which are typically used by ANL in concert with

MC2-2 and REBUS, but are not publicly available, and thus were not used in the calculations

performed by this work, which was performed at the Idaho National Laboratory (INL). The first

of these codes, SDX, accounts for the spatial effect of heterogeneity introduced by pins in an

assembly, which is modeled initially in MC2-2 as a homogenized region. The second, DB2,

takes the fission energy spectrum from areas on the periphery of the core and collapses the cross

sections for the neighboring control rod, reflector, and shield regions based on the leakage

spectrum.

The approach taken by this dissertation (and parallel ABR core physics studies performed

by INL) was to disregard the spatial heterogeneity introduced by pins in the assembly. This

zero-dimensional approach in the cross section collapsing does not account for spatial resonance

shielding effects between the various core regions. However, for fast reactor calculations this is









generally sufficient due to the long neutron mean-free-path. The fast flux is almost entirely in

the unresolved resonance range, thus making unresolved energy shielding the dominating effects

in the group collapsing. This assumption is supported by the good agreement between the INL

benchmarking effort (INL-EXT-12466) of the ABR and the ANL scoping calculations for the

ABR (ANL-AFCI-177) [7,38]. Also for this work and the INL benchmark, the reflector and

shield region cross sections were collapsed, without DB2, using a generic Pu-239 fission energy

spectrum in MC2-2.

These 33-group cross section sets produced by MC2-2 are then used by the DIF3D code to

perform the actual core physics and criticality calculation. The DIF3D diffusion code was used

to solve the multigroup steady state neutron diffusion equation using a hexagonal-z nodal

coordinate system [42]. In the nodal discretization, each hexagonal node in the lateral direction

represents a fuel assembly. Because the mean-free-path is on the dimensional level of the fuel

assembly, the individual fuel rods are homogenized across this hexagon.

Fast Reactor Equilibrium Fuel Cycle Calculations Using the REBUS Code

The actual fuel depletion and fuel cycle modeling, including the mass balance between

reactors and reprocessing plants, is performed by the REactor BUrnup System (REBUS) code

[43,46]. REBUS uses DIF3D to generate reaction rate and flux information at each time step

"burn step" in its fuel depletion algorithm. Once the reactor physics calculation is completed,

the fluxes from DIF3D are fed to a depletion solver routine. This solver uses the exponential

matrix method to solve the Bateman equations for isotopic buildup and decay within the

discretized burn step. REBUS also performs the in-core fuel management and out-of-core

cooling, reprocessing and re-fabricating for each reactor cycle. The REBUS depletion and fuel

management algorithm is given in Figure 2-1.










Reactor and
Fuel Cycle Homogenize batches across
Inputs user defined regions


Microscopic
cross section Create macroscopic Modify the TRU
libraries / cross sections enrichment of the
libraryes --fresh fuel batch


Has the End of Cycle been
reached according to the
pre-defined burnup limit?
(taken to be 18 at. %, or
nominally 215 EFPD)


No


Does k-eff = 1 at
End of Cycle?


Yes


Has the cycle length
converged within the
convergence criteria
from the previous cycle?


Figure 2-1. Coupled DIF3D core physics and REBUS fuel cycle algorithm

In the REBUS fuel cycle model, individual fuel assemblies are homogenized into "like

neutron spectrum" representative regions. Therefore, independent batches of fuel are tracked


Create
Output
File


Perform in-core fuel
management and
out-core fuel cycle
activities









within the external fuel cycle but not explicitly spatially represented in the physics calculation.

These regions may represent all of the batches of fuel having common transuranic enrichment,

defined as an enrichment zone. However, it is not essential to divide the same-spectrum regions

according to the enrichment zoning. The AHFTR fuel management model assumes that each

row in the core is a same-spectrum region. Because fuel is typically not shuffled in SFRs, the

assumption is made that each row in the AHFTR core is filled with fuel assemblies of the exact

same fresh fuel specifications but having different levels of depletion. Therefore, the average

fuel composition within each row (or more generally region) is the volumetric average of the

same fuel assembly at the different stages of its depletion.

The in-core fuel management and out-of-core fuel cycle activities are carried out until the

equilibrium cycle was achieved. The equilibrium cycle is defined as the equilibrium or "steady-

state" condition of the fuel cycle when the reactor performance properties (i.e., k-eff, enrichment,

cycle length, etc.) become invariant from cycle to cycle. This equilibrium mode calculation was

performed for both the ABR and the AHFTR. In a typical REBUS equilibrium fuel cycle

calculation, the fuel management operations are carried out until the excess reactivity and

equilibrium cycle length are found. A maximum fuel burnup of 18 a/o after 6 reactor cycles

(seven cycles for the outer core of the ABR) was used to constrain the search procedure to

determine the equilibrium cycle length. This burnup constraint nominally gives an equilibrium

cycle length of approximately 215 Effective Full Power Days (EFPD) for both the ABR

reference and the AHFTR. This burnup closely correlates to the maximum exposure limitations

of SFR metallic fuel and cladding integrity. These limitations will be discussed in Chapter 6.

To account for the spectral changes during fuel depletion, the 33-group cross section

library was updated using a coupled corrector-predictor loop between MC2-2 and REBUS. This










loop is initialized by a first guess of the region and depletion averaged fuel composition in the

core. This guess is made by using a SNF TRU composition, depleted uranium composition and

approximating the fresh fuel TRU enrichment. This composition data is used by MC2-2 to

generate a guessed cross section set. Then REBUS uses the guessed cross section set to perform

a guessed fuel cycle calculation. Finally, the region and depletion averaged fuel composition is

extracted from the REBUS output and imported into a new MC2-2 calculation. An automated

scripting system is used to continuously re-calculate the cross sections for each enrichment zone

based on that zone's fuel inventory at equilibrium. Figure 2-2 shows the flow of cross section

and isotopic composition data between MC2-2 and REBUS.

Initial guess Region
of fuel dependent
enrichment batch and
and time averaged
composition compositions


MC2-2 Cross MC2-2 Cross
Section Generation Section Generation

Final Rebus
Cross Cross YesNoutput
Section Set Section Set


REBUS Equilibrium REBUS Equilibrium
Fuel Cycle Fuel Cycle Check: k-eff, cycle
Calculation Calculation length, average fuel
enrichment, etc.

Rebus Output Rebus Output




Figure 2-2. Flow diagram of data transfer between the MC2-2 and REBUS codes

Once this equilibrium fuel cycle was attained, the Beginning-of-Equilibrium-Cycle

(BOEC) and End-of -Equilibrium-Cycle (EOEC) total core void and Doppler coefficients were

calculated. The total core void worth was attained by taking the BOEC number densities from









each isotope in each axial and radial region from the REBUS output file into a new DIF3D (no

depletion) input file. The number density for sodium for each region was reduced by the coolant

volume fraction leaving only the bond sodium in the fuel rod gap with the coolant sodium

voided. The sodium fraction was also reduced for the corresponding MC2-2 calculation. The

Doppler coefficient was calculated in a similar procedure without the change in sodium number

density. Instead, a new MC2-2 calculation was performed with all cross sections broadened at a

temperature 100 K greater than that for the REBUS calculation. Then a new DIF3D calculation

was performed with the BOEC number densities unchanged from the REBUS output file.

Light Water Reactor Spent Fuel Calculations Using the TRITON Code

Since REBUS only deals with the closed portion of the fuel cycle involving fast reactors,

the external supply of TRU to this closed fuel cycle must be generated externally. To perform

this task, the Oak Ridge National Laboratory (ORNL) code package, Scale 5.1, was used to

generate the composition of SNF by performing a single fuel assembly depletion and decay

calculation [47]. Using the geometry of a (Pressurized Water Reactor) PWR 17x17 pin bundle

design, a two-dimensional lattice calculation was performed to represent the in-core irradiation.

The SCALES.1 depletion code, TRITON, was used to perform the physics and depletion

calculations for this single assembly model. TRITON uses several other scale codes within

SCALES.1 to perform the cross section generation and depletion of the fuel in this single

assembly calculation.

* BONAMI was used to apply Bonderenko factors to correct for energy shielding between
un-resolved resonances

* NITAWL was used to apply the Nordheim integral treatment for spatial shielding between
resolved resonances

* NEWT is a two-dimensional discrete-ordinance code used to solve the neutron transport
equation and create reaction rate and flux data.









ORIGEN-S is a fuel depletion solver used to carry out the isotope buildup and decay
process by solving the Bateman equations for each bum step and also for the post-
irradiation decay period

Using TRITON simulation, a uranium oxide (UOX) fuel composition, enriched to 4.5%,

was irradiated to a bumup of 50 MWD/kg. After this irradiation, the fuel was allowed to decay

for five years to represent the spent fuel pool cooling time. It is assumed that the fuel will be

partitioned into the Np+Pu and Am+Cm+Bk+Cf streams after this cooling off period. After

partitioning, an additional decay time of two years was assumed for the time after separation

which includes: reprocessing, fuel fabrication and transportation to the AHFTR. The isotopic

composition of the SNF that is used as the external feed to the SFR is given in Table 2-1.

Table 2-1. Isotopic composition in weight percent for UOX SNF
Spent Fuel Pool and Transportation to+3 Repr
UREX+3 Reprocessing
Reprocessing Center
Discharged from LWR After Five Year Decay Np+Pu Am+Cm
Np-237 5.46% 5.54% 5.88%
Pu-238 2.47% 2.55% 2.66%
Pu-239 45.79% 46.16% 49.01%
Pu-240 22.61% 22.58% 24.01%
Pu-241 13.18% 10.28% 9.94%
Pu-242 7.05% 7.01% 7.44%
Am-241 0.50% 3.28% 1.01% 55.75%
Am-242m 0.01% 0.01% 0.17%
Am-243 1.92% 1.91% 32.37%
Cm-242 0.18% 0.00% 0.00%
Cm-243 0.01% 0.01% 0.09%
Cm-244 0.78% 0.64% 10.89%
Cm-245 0.04% 0.04% 0.65%
Cm-246 0.00% 0.00% 0.08%

Scoping Calculations and Benchmarking Using the MCNP Code

Given the dissimilarities between the core physics of LWRs and SFRs, it is prudent to test

the computational methods available for fast reactor analysis using state-of-the-art tools that can

simulate both reactor types. The Los Alamos National Laboratory (LANL) code Monte Carlo N-

Particle (MCNP) is a general purpose physics simulation tool that uses the Monte Carlo method

to recreate the exact particle physics of neutrons (and photons and electrons) in any arbitrary









three-dimensional geometry over a continuous energy range [48]. All scoping calculations and

benchmarking analysis in this dissertation are performed with the MCNP code. MCNP allows

the user to tally the neutron flux, and energy spectrum, as well as fission and capture reaction

rates, in any region of the core. This neutron tally feature is used to indicate certain core physics

parameters such as leakage and transmutation performance in the axial targets.

The MCNP code is also used to benchmark the accuracy of the deterministic diffusion

method used by DIF3D to model the heterogeneity between the active core (fast spectrum) and

the axial targets (epithermal spectrum). To model the accuracy of the coupled reactor physics

and depletion algorithm of the MC2-2/DIF3D/REBUS scripting system, the LANL depletion

code package MONTEBURNS was used in conjunction with MCNP [49]. MONTEBURNS

uses MCNP tally data to produce single group fission, capture and n,2n cross sections and fluxes.

These cross sections are used for neutron flux determination at a particular time step. These

fluxes and cross sections are then fed into the ORNL fuel depletion code ORIGEN2 for burnup

and decay calculations to the next time step [50, 51]. Then MONTEBURNS feeds the updated

isotopic composition back to MCNP and the process begins anew.

Physics of the Reference Metal Fueled Advanced Burner Reactor

In the introduction, it was established by references to literature that a SFR conversion

ratio is essentially a function of parasitic capture by uranium and neutron escape by leakage.

* A SFR's conversion ratio can be decreased by increasing axial leakage, as in a pancake
design.

* The conversion ratio can also be decreased by decreasing the ratio of TRU to uranium
loaded in the core (i.e., TRU enrichment) as was done by the ABR.

In order to establish the motivation for using a heterogeneous design, such as axial targets,

it is important to first qualify these statements. The computational tools described in this chapter

are used to explore these heterogeneous qualities and their affect on axial leakage. However, the









ABR, unlike the S-PRISM or the axial target design proposed here, exploits very little

heterogeneity (no blankets or transmutation targets). Because the homogeneous ABR design

does not take advantage of blankets or transmutation targets, decreases in conversion ratio are

made by changes in the neutron economy (balance between neutron losses and neutrons

contributing to fission or fissile production).

Conversion Ratio and High Leakage Cores

Consider the criticality condition that the geometric and material buckling must be equated

in order for a reactor core to be critical [45].

B2 B2 (2-1)

Where: Bg2 is the flux curvature's geometric buckling and Bm2 is the material buckling.

For a bare right circular cylinder, the geometric buckling (without reflection) is defined by

Equation 2-2:


B2 =B2 + B2 2.405 + (2-2)
g r R +z H+

Where: Br2 and Bz2 are the radial and axial component of the geometric buckling,

respectively. R and H are the critical radius and height of the core respectively taking into

account the extrapolation distance:

1 1 1
zo = 0.71tr = 0.71x = 0.71x = 0.71 x (2-3)
Y, Yt + goZY- Ntott + yoNtotso

Where: Zo is the extrapolation distance. 4tr is the transport mean-free-path. Xtr, Et and Ys

are the macroscopic cross sections for transport, total interaction and scattering, respectively.

Ntot is the total number of atoms in the homogenized composition. ,to is the average cosine of the

neutron scattering angle assuming elastic isotropic scatter in the lab system. The material

buckling of a homogeneous mixture is defined as:










B V2 Y + Z vYZ + v + (2-4)
S D 1 1
3(,r) 3(Y, T-o j

Where: v is the average number of neutrons produced per fission. D is the diffusion

coefficient. Yf and Ea are the macroscopic fission and absorption cross sections, respectively.

For the purpose of this discussion, consider only the Pu-239 and U-238 isotopes.

Therefore, the macroscopic cross sections in Equation 2-4 can be represented in terms of the

enrichment and the total HM atom content by:

B2 x(rRUNtot f ,239 + ( rRU )NtotC2 )+ (TRU NtotC,239 + ( rRU )Ntot a,238 (2-5)
B 1 (2-5)
m 1/3((rTRUNtot t,239 TRU tot t,238 ) o TRU tot s,239 TRU tot 238

Where: rTRU is the TRU enrichment which for this sample calculation is equal to Pu-239

divided by the sum of Pu-239 and U-238 atoms. oTf,239, Gf,238, Ga,239, Ga,238, Gt,239, Gt238, Gs,239 and

Os,238 are the fission, absorption, total interaction and scattering microscopic cross sections for

Pu-239 and Pu-238, respectively.

The microscopic cross sections for Equation 2-5 were generated by tallying the ABR

neutron flux and capture and fission reaction rates over all energies and collapsing using MCNP.

These tallies were used to create representative one-group cross sections for U-238 and Pu-239.

These cross sections and the combined number density Ntot of U-238 and Pu-239 are given in

Table 2-2. For a constant total HM atom density, increasing the TRU enrichment also increases

the material buckling, as shown in Figure 2-3.

Table 2-2. Reference ABR atom densities and one-group microscopic cross sections for: Pu-
239 and U-238
N (atom/b*cm) .
N (atom/*cm) One-group Microscopic Cross Sections (barns)
of ref ABR
Total 0.0062 Fission Scatter Absorption Fission
Pu-239 0.0007 11.27 8.28 2.09 1.71
U-238 0.0054 10.88 9.53 0.28 0.04












4.0E-03

3.5E-03

3.0E-03

S2.5E-03

S2.0E-03

0 1.5E-03

1.0E-03

S5.0E-04

O.OE+00
0% 20% 40% 60% 80% 100%
TRU Enrichment

Figure 2-3. Material buckling of simple bare homogeneous SFR as a function of TRU
enrichment (Hand Calculation)

Therefore, for increasing TRU enrichment, the geometric buckling required for criticality


must also increase. Thus, for a given height "H" the axial buckling can be increased by


decreasing the core's critical radius "R". The critical radius required to equate geometric and


material buckling will vary depending on core height. The critical radius for different core


heights and varying enrichment is given in Figure 2-4.


120


100


E 80


60 -


r 40


20
-- H=200 cm -- H=100 cm H=75 cm

0
20% 30% 40% 50% 60% 70% 80% 90% 100%
TRU Enrichment

Figure 2-4. Critical radius required to equate geometric buckling with material buckling for
increasing TRU enrichment (Hand Calculation)









The converse of the above statement is also true. If the geometric buckling is increased

(either by reducing height or radius) the TRU enrichment that is necessary for criticality must

increase. Note that this is the reason for flattening the SFR's height-to-diameter ratio to make a

pancaked design. The flattened design increases buckling and decreases neutron economy.

Hence, in order to achieve criticality, the TRU enrichment must be high.

Increasing TRU concentration and decreasing uranium concentration equates into a

reduction in conversion ratio. For the purpose of this example, consider that the only source of

TRU breeding results from neutron capture in U-238. Also assume that any neutron absorption

in Pu-239 for this simplified model results in TRU destruction. This is not entirely the case in

reality because some neutron absorptions in Pu-239 result in Pu-240 and the rest of TRU.

Therefore, the conversion ratio as a function of enrichment is given by:

R g TRU production ,238 ~f,238 (1- 3TRU a0,238 f,23s8)
Cgenerazed TRU destruction Ya,239 rTRU a,239

Where: CRgeneralized is the generalized conversion ratio definition for the two isotopes

considered in this sample calculation. This equation is plotted using the values from Table 2-2

for varying TRU enrichment in Figure 2-5.

Therefore, an enrichment of 30% yields approximately a CR of 0.25 in this bare (un-

reflected) sample problem. From Figure 2-4, this corresponds to a critical radius of over 120 cm

for a core height of one meter. This generalized bare reactor conversion ratio is significantly less

than the actual reference ABR (CR=0.5) due to the absence of reflectors. This draws attention to

the importance of reflection in a SFR design. The mean-free-path between interactions of HM

atoms is significantly high in a SFR (Figure 2-6). Therefore it is probable that a neutron can

travel great distances in the reactor core without interacting with another fuel atom which











increases the likelihood of escape by leakage. The axial and radial reflectors in SFR designs are

necessary to maintain some base level of neutron economy.


0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
TRU Enrichment

Figure 2-5. Generalized conversion ratio of a simple bare homogeneous SFR as a function of
TRU enrichment (Hand Calculation)


14.85

14.80

14.75

E 14.70

14.65

e 14.60

L 14.55

0 14.50

14.45

14.40

14.35
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
TRU Enrichment

Figure 2-6. Mean-free-path of neutron travel between interactions (of any reaction type)
between HM atoms (Hand Calculation)

The correlation between leakages, TRU enrichment and CR can be observed by plotting

the leakage fraction which is given by:











1 1
LF = 1-0 leake = 1 = 1 (2-7)
lekg +L2Bg2 +(D )Bg2


Where: Pnon-leakage is the probability of non-leakage and L is the diffusion length. The

leakage fraction is plotted in Figure 2-7.


3.5E-03

3.0E-03

S2.5E-03
U
2.0E-03 -
LL.
v 1.5E-03

1.0E-03

5.0E-04

0.OE+00
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
TRU Enrichment

Figure 2-7. Leakage fraction of a simple bare homogeneous SFR as a function of TRU
enrichment (Hand Calculation)

Conversion Ratio and High TRU Enriched Fuels

Assume that a very low CR is desired. But the prospect of core flattening and a pancake

design may be unattractive for other reasons such as the expense of a large core barrel.

Therefore, it would be necessary to fix the geometric buckling with a fixed set of core

dimensions: H=100 cm, R=113 cm. Because the geometric buckling has been fixed, the

material buckling is also a fixed quantity. As previously stated, in order to achieve a low CR, a

very high TRU enrichment is necessary (e.g., rTRU>90%). Because the geometric/material

buckling and TRU enrichment is fixed by Equation 2-5, the density of HM atoms (Ntot) in the

core must decrease to meet the criticality condition. The decrease in HM atom density is caused

by the fact that as the TRU enrichment increases, the fuel becomes mostly fissile. Therefore,










less fuel (i.e., heavy metal) is needed to be critical. Figure 2-8 shows this relationship between

the CR, TRU enrichment and the HM core loading.

This is the approach taken to achieve decreasing low conversion ratios by a homogeneous

ABR design. The ABR height is 101.6 cm and its active core radius is 109 cm. A final decision

on the design conversion ratio of the ABR has not been made in the GNEP program. Currently,

there are several variations of the ABR design that range from having a CR of 0.25 to 0.75.

1 -

-0.9 h
o -- Conversion Ratio -E-TRU Enrichment
S0.8

0.7
cE


n 0.5
S- 0.
W 0.4

U- 0.3
o.
0.2
0 .
0.1


0.003 0.0035 0.004 0.0045 0.005 0.0055 0.006 0.0065 0.007
Heaavy Metal Charge Loading (atom/b*cm)

Figure 2-8. Conversion ratio and the corresponding TRU enrichment as a function of HM atom
charge density for a bare homogeneous SFR of fixed size (Hand Calculation)

The decision on conversion ratio is primarily dictated by the feasibility of fabricating and

irradiating high TRU with low uranium concentration in the SFR driver fuel. In the

homogeneous burner reactors proposed for the ABR, reductions in CR are achieved by removing

from the fuel pin, the volume corresponding to uranium [7,52,53,54]. This decrease in volume

for constant fissile inventory increases the TRU enrichment which reduces the CR. Since the

volume reduction is done such that the fuel pin radius is decreased, the fabrication feasibility of

very small diameter fuel pins becomes a design factor. There is also a separate feasibility issues









related to reactor kinetics and safety of a low conversion ratio SFR with virtually zero uranium

content.

Physics of the Axially Heterogeneous Fast Transmutation Reactor

The AHFTR design concept seeks to increase the neutron economy of the overall SFR

design by using axial targets to recover the axial leakage produced by the active core. This

enables the active core to maintain some level of high leakage which is important for

establishing a baseline conversion ratio. Establishing a "reasonably" low conversion ratio

without the targets is important because, as will be shown, the axial target region operates in a

converter mode (i.e., MA converted into Pu with little decrease in total TRU). This is an

important point to make because the design rationale of the targets is to convert MAs into

plutonium which is a different philosophy from a purely MA burning target or a plutonium

breeding blanket.

The enhanced neutron economy without significantly sacrificing conversion ratio allows

the AHFTR to have a TRU enrichment that is more akin to past SFR fuel experience than the

lower conversion ratio forms of the homogeneous ABR. The more feasible TRU enrichment

also solves the problem of fabricating small pin diameters. This is because the higher uranium

loading in the fuel also adds volume and hence diameter to the fuel pin. Also, increasing the

uranium content of the driver fuel generally enhances the resonance feedback attributes of the

reactor design.

Axial Targets and Axial Leakage Recovery

To highlight the potential for neutron economy improvements, the ability of the axial

targets to "trap" axial leakage is tested. A series of parametric studies on core and pin geometry

are performed in the next chapter using the REBUS code. These analyses ultimately culminated

in the AHFTR design described in the Introduction. For the purpose of showing the axial target









effectiveness in this chapter, the region homogenized atom density data was imported to an

MCNP calculation to recreate the BOEC DIF3D calculation from the REBUS fuel cycle

simulation. No attempt was made to unfold the actual batch or exact fuel assembly compositions

from this region-by-region data. The MCNP calculation was performed for both the AHFTR and

ABR reference design.

Within the MCNP model, neutron current tallies were made for increasing heights in the

core geometry. In MCNP, the current is tallied by simply totaling the number of particles,

crossing a given surface within a specified angular range. This range corresponds to the limits of

integration of the summation over dQ about Q in the equation:

J = fdE dtfd dA Q ni (iE, E, ,Q t) (2-8)
E t C0 A

Where: J is the net leakage. E is neutron energy, t is time, A is the area of the tally

surface, Q is the solid angle over which the tally is taken. n is the unit normal vector to the tally

surface and 4 is the region, energy and time dependent angular flux.

Integration of Equation 2-8 gives the net current with a positive sense in the direction of

the normal vector. For the following calculation, this normal direction was chosen to be positive

upward in the direction towards the top of the core. For neutrons leaving through the bottom of

the core, the absolute value of the current is taken so that the current appears positive at all

locations. The current and flux distribution for the AHFTR is given in Figure 2-9 and Figure 2-

10.

The change in current per change in height is, for all practical purposes, the average axial

leakage. It can be observed from Figure 2-9 that the net number of neutrons moving out of the

core is less at the top of the targets (90 cm) than it is at the top of the active core (70 cm). This

represents the combined effect of reflection and absorption by the targets. To quantify reflection










at the top of the core, the targets were replaced by: (1) a region of sodium, (2) axial reflector or

(3) driver fuel. When compared with a sodium or steel reflector, the targets have less net current

at the top of the target region than either of the two reflector compositions. This indicates the

added effect of absorption in the targets. When the axial target region is replaced by fuel, the

shape of the axial current profile flattens out as the neutrons enter the gas plenum above the core.

Therefore, it is apparent that even though the target region contains some TRU, its neutron

production is less than that produced by the active core region.


2.0E+14 AHFTR
-ABR Reference
1.8E+14 Target Material Replaced by Sodium
Target Material Replaced by Steel Reflector Target
a 1.6E+14 Region
S--Target Material Replaced by Fuel IeglOn
1.4E+14 -___

S1.2E+14 Active Core

0 1.0E+14 -

= I 8.OE+13 -

6.0E+13 -

0 4.0E+13 *

2.0E+13

0.0E+00
0 10 20 30 40 50 60 70 80 90 100
Axial Distance from Bottom of Active Core (cm)
Figure 2-9. Axial current distribution of the AHFTR and ABR (MCNP)

Thus, it can be inferred from this comparison that the decrease in current in the axial

targets is caused by a combination of axial reflection as well as capture reactions by the MAs.

The reduction in the net current from the top of the active core to the top of the target region

equates into a reduction in leakage by the target region. This reduction in leakage can also be

seen by comparing the curvature of the scalar flux in the target region compared to that of the

active core (Figure 2-10).











5.0E+15

4.5E+15

4.0E+15

' 3.5E+15
In
1 3.0E+15
E
. 2.5E+15

2.0E+15

S 1.5E+15
I-
1.0E+15

5.0E+14
n nE.nn


0 20 40 60 80 100 120
Axial Distance from Core Bottom (cm)

Figure 2-10. Axial flux distribution of the AHFTR and ABR (DIF3D)

Axial Targets and Minor Actinide Conversion

The capture effect of the MAs in the moderated target region can best be qualified by

examining the energy dependence of the capture and fission reaction rate of Am-241. To do this,

the MCNP model used for the above neutron current calculation was modified to tally the

neutron flux in the target region and in the active core region. This flux tally was discretized into

33 equal lethargy sized bins (same as that used by MC2-2) in order to create a point-wise neutron

energy spectrum for both regions. This binned flux tally was weighted with a cross section

multiplier corresponding to the isotopes and reaction rates of interest. Using the Monte Carlo

method, this multiplier effectively integrates the reaction rate in each bin from an energy

continuous flux and cross section. To calculate a volume averaged flux, MCNP sums the length

of all neutron track-lengths crossing the tally region and then divides this summation over the

region's volume. The flux multiplier is used in a similar way except the track-length of each


-AHFTR
- ABR Reference










tally is multiplied by the continuous energy microscopic cross section of the desired reaction and

isotope type.


H+AE Lt0rack-length (E) x 0, (E)
R(E 1,2)= J(E)o(E)dE= -E (2-9)
E.

Where: Ltrack-length is the total length of travel of a neutron as it passes through the tally

region. V is the volume of the tally region. oi(E) is the microscopic cross section for the

reaction and isotope of interest for the tally multiplier.

Using the 33 energy bins, the binned reaction rates give an approximately smooth

distribution of the reaction rates as a function of energy. The binned capture and fission reaction

rate spectra for Am-241, Pu-238 and Pu-239 are plotted in Figure 2-11 and Figure 2-12. It is

important to mention that these plots give the microscopic reaction rate which is normalized per

atom and not the macroscopic reaction rate which would be weighted by the atom density of

each isotope.

1E+16




0 1E+14 -
1E+153 -



S1E+143 pt I





.- 1E+11 ,
O -A Capture: Pu-238 -A -Fission: Pu-238
'M 1E+10 ---Capture Pu-239 3 Fission: Pu-239
0 0

iL --Capture: Am-241 C Fission: Am-241

1E+09 ...
1E-07 1E-06 1E-05 1E-04 1E-03 1E-02 1E-01 1E+00 1E+01 1E+02
Energy (MeV)
Figure 2-11. Binned "microscopic" reaction rate spectra of select isotopes as a function of
incident neutron energy on an atom in the moderated target region (MCNP)










As can be seen in Figure 2-11, the fission reaction rate of Am-241 is several orders of

magnitude less below one MeV than for any of the plutonium atoms. However, the Am-241

capture reaction rate is of the same order of magnitude as any of the plutonium fission reactions

in plutonium. Therefore it is evident that transmutation of an Am-241 atom into a plutonium

(i.e., Pu-238) atom will increase its fissile worth. It is useful to note that the Pu-238 fission rate

falls off sharply at energies below 100 eV (Figure 2-13). This is due to the fact that the Pu-238

fission cross section falls off sharply below 100 eV.

1E+16 --Capture: Pu-238 -A Fission: Pu-238
Capture Pu-239 -H *Fission: Pu-239
Capture: Am-241 *Fission: Am-241
S1E+15

0 1E+143





S< 1E+11 -

110 0
U-



1E+09
1E-07 1E-06 1E-05 1E-04 1E-03 1E-02 1E-01 1E+00 1E+01 1E+02
Energy (MeV)
Figure 2-12. Binned "microscopic" reaction rate spectra of select isotopes as a function of
incident neutron energy on an atom in the active core (MCNP)

It is also noteworthy to point out that neutron absorptions in Pu-238, that do not result in

fast fission transmutes into fissile Pu-239, which has a high fission cross section at all energies.

Unlike the purely fast spectrum (Figure 2-12), isotopes in the axial target are equally exposed to

neutron fluxes in both energy ranges. Because the practical energy range for fission reactions

increases with each successive neutron capture (Figure 2-11), the target is essentially breeding

fissile worth, even though fissile atoms are not a direct result of the first neutron capture.










1E+05
-Am-241 Capture
1E+04 --Am-241 Fission
-Pu-238 Fission
1E+03

E 1E+02

1E+01
Co
a 1E+00

2 1E-01

1E-02

1E-03

1E-04
1E-08 1E-06 1E-04 1E-02 1E+00 1E+02
Energy (MeV)
Figure 2-13. Comparison of capture and fission cross sections between Am-241 and Pu-238
(ENDF-VI)

If the MA atom concentration in the targets is sufficiently high, the Am-241 capture

reactions will energy shield the fission reaction of both Pu-238 and Pu-239. Because neutrons

are being absorbed in Am-241 instead of plutonium, this energy shielding effect is roughly

analogous to the spatial resonance shielding provided by a burnable poison in an LWR fuel pin.

Because of this energy shielding, the fissile worth created via transmutation in the target region

can not be truly realized until the transmuted plutonium atoms are reprocessed and placed in the

fast spectrum as driver fuel.

The flux in the active core is entirely fast (Figure 2-12). Therefore, the corresponding

reaction rate spectra are also in the fast energy range. Hence, the energy shielding ability of Am-

241 over Pu-238 or Pu-239 is not available. The capture cross section of Am-241 falls off

sharply after one MeV, whereas the fission cross sections of Pu-238 and Pu-239 are much higher

in this energy range. Also, the concentration of Am-241 is significantly less in the driver fuel

than it was in the targets because of the much smaller Am-241 concentration in the driver fuel










than in the targets. Hence, without the energy shielding by Am-241, the fissile worth of Pu-238

and Pu-239 is higher in the active core than in the target region. The change in fissile worth can

be seen by observing the differences in fission-to-absorption ratio between the target region and

the active core region (Table 2-3).

Table 2-3. Fission and capture one-group cross sections and fission-to-absorption ratios for the
AHFTR axial target and active core regions (MCNP)
Axial Targets Driver Fuel/Active Core
Fission Capture Fission per Fission Capture Fission per
(barns) (barns) Absorption (barns) (barns) Absorption
U-234 0.32 3.00 0.10 0.35 0.46 0.44
U-235 3.93 1.69 0.70 1.64 0.42 0.80
U-236 0.12 1.73 0.06 0.10 0.33 0.23
U-238 0.04 0.63 0.06 0.04 0.22 0.15
Np-237 0.31 4.83 0.06 0.34 1.19 0.22
Np-238 10.16 0.36 0.97 3.52 0.11 0.97
Pu-236 8.54 1.66 0.84 3.30 0.50 0.87
Pu-238 1.36 2.02 0.40 1.10 0.56 0.66
Pu-239 3.58 1.91 0.65 1.67 0.33 0.83
Pu-240 0.37 2.19 0.15 0.39 0.38 0.51
Pu-241 5.74 1.45 0.80 2.18 0.33 0.87
Pu-242 0.26 1.78 0.13 0.27 0.33 0.46
Am-241 0.28 4.75 0.06 0.28 1.29 0.18
Am-242m 11.54 1.71 0.87 3.31 0.26 0.93
Am-243 0.20 4.68 0.04 0.20 1.15 0.15
Cm-242 0.15 1.57 0.09 0.16 0.21 0.44
Cm-243 8.58 1.20 0.88 2.24 0.17 0.93
Cm-244 0.45 2.24 0.17 0.44 0.66 0.40
Cm-245 6.31 0.83 0.88 2.00 0.25 0.89
Cm-246 0.28 0.93 0.23 0.27 0.17 0.61
Cm-247 2.59 0.98 0.73 1.94 0.25 0.88
Cm-248 0.35 1.96 0.15 0.31 0.18 0.64
Bk-249 0.18 5.14 0.03 0.17 0.97 0.15
Cf-249 6.03 1.45 0.81 2.33 0.56 0.81
Cf-250 0.95 5.78 0.14 1.19 0.29 0.81
Cf-251 6.09 1.66 0.79 2.16 0.25 0.90
Cf-252 1.94 0.80 0.71 0.62 0.23 0.73
Combined 0.05 0.12 0.27 0.06 0.06 0.47

Based on fission-to-capture ratio data from Table 2-3, one Pu-238 atom is produced for

every 1.38 neutron captures in Am-241 in the target region. This newly formed Pu-238 atom

than fissions for every 2.49 neutron absorptions. This is a smaller neutron investment compared

to LWR IMF; which is 5.47 absorptions per fission (Table 1-2). An additional capture in Pu-238









produces Pu-239 which has a fission-to-absorption ratio in the targets of 0.65. In the fast active

core region, 1.586 neutrons are required to convert an Am-241 atom into a Pu-238 atom. This

newly formed Pu-238 atom than fissions for every 1.51 neutron absorptions. The much high

fission-to-absorption rate of Pu-238 in the active core (i.e., 1/1.51=0.66) than the axial targets

(i.e., 1/2.49=0.40) shows the utility of the spectrum shift from the thermalized to the fast

spectrum. An additional capture in Pu-238 produces Pu-239 which has a fission-to-absorption

ratio in the active core of 0.83.

Combining Leakage and Capture Effects

As can be seen from Figure 2-14, even though the target region contains moderating pins,

the moderating effect places most of the flux in the epithermal to fast energy range. In this

energy range, most of the resolved resonances are very close to each other and not sufficiently

wide to create local flux depressions in the core or pin geometry. This is different from thermal

spectrums where neutron slowing down places most of the neutrons in the thermal energy range

where the resonances are much wider and well separated. Therefore, spatial self-shielding

effects between well resolved and well separated resonances have less importance in the axial

target region than for completely thermal spectrums such as for LWRs. Hence, the "same-

spectrum" region-wide homogenization assumptions used in cross section and flux calculations

by the MC2-2 and DIF3D/REBUS codes are adequate for approximating the lattice physics in the

active core and target regions.

Also, it was demonstrated in this chapter that SFRs inherently have a high TRU

enrichment for criticality reasons because of their inherent high rate of neutron leakage.

Therefore, the relative change in concentration of fissile atoms (i.e., TRU enrichment) as a

function of burnup can be small, even for high burnups. Hence, it should be expected that the

neutron spectrum in the core is fairly insensitive to the effect of fissile atom depletion.










Therefore, the need for spectrum updated cross section sets (i.e., micro-depletion) in analyzing

the ABR and AHFTR is not as strong as it is for LWR spectrums. In LWR fuels, the enrichment

is small compared to SFRs. Therefore as the LWR fissile atoms are depleted, the percent change

in the fissile concentration is significantly more than it is for SFRs. The burnout of isotopes with

resonances that are well resolved and well separated changes the magnitudes of flux depressions

that are caused by those resonances. Therefore, for LWR fuels it is important to create several

cross section libraries corresponding to different stages in depletion. Because, SFRs such as the

ABR and AHFTR do not have significant changes in the neutron spectrum with burnup, this

spectrum updating or micro-depletion is not necessary.

1E+16
-Target Region
-Active Core Region
1E+15 --LWRIMF

1E+14

UIE+13

E
1E+12


1E+11

1E+10
1E-07 1E-06 1E-05 1E-04 1E-03 1E-02 1E-01 1E+00 1E+01
Energy (MeV)
Figure 2-14. Comparison of flux spectrums between LWR IMF (thermal), AHFTR target region
(epithermal-fast) and AHFTR active core region (fast) (MCNP)

In fact, the ability of the AHFTR to breed plutonium through Am-241 transmutation is

made possible because the target region's spectrum does not change significantly with burnup.

The moderation effect causes the energy shielding of plutonium unresolved and poorly resolved

resonances by Am-241. This energy shielding minimizes the in situ burnup of plutonium

isotopes in the target region. Therefore, transmuted plutonium isotopes are allowed to









accumulate with irradiation. The fissile usefulness of this transmuted plutonium is only fully

realized when these isotopes are placed in the active core where the spectrum is much harder.

The negligible spectrum change in the target region is mostly due to the fact that the target

region's epithermal spectrum is not strongly affected by the well resolved and well separated

resonances occurring at thermal energies. The epithermal flux can be explained by the fact that

the target volume (despite the presence of moderating rods) is still mostly filled with sodium,

steel and HM atoms. In addition, the curvature of the flux (i.e., buckling) in the targets is not

completely flat and is similar to the active core. Therefore, the axial leakage through the targets

is still fairly significant. This fact suggests that the target region is not a black absorber and that

a significant number of neutrons are lost through leakage before they can be moderated and

absorbed. Hence, the axial target region behaves more as an integral component of the core than

a neutron sponge outside of the core.

The target composition contains a starting amount of plutonium and uranium that makes

the total plutonium content and corresponding spectrum roughly constant throughout the

irradiation. This constant level of plutonium (regardless of isotopic composition) helps minimize

spectrum shifts in the target region. A detailed explanation of the target composition of MAs,

some plutonium and some uranium is given in the next chapter.









CHAPTER 3
AXIAL TARGET DESIGN ANALYSIS

As discussed in the previous chapter, in order to give a high net TRU destruction rate it is

necessary to greatly reduce the SFR's conversion ratio. This has the advantage from a waste

management point of view of decreasing the amount of net TRU production from neutron

capture in U-238. The first necessary requirement for reducing the amount ofU-238 capture

reactions is to decrease the amount of U-238 in the SFR. This decision leads to an ABR design

with no U-238 radial blankets. Furthermore, the removal of blanket assemblies reduces the

overall reactor size from an S-PRISM design which has nine rows of driver and blankets to the

current 1000 MWth ABR design with seven driver rows [7,19,55,56].

Waste Management Philosophies and Conversion Ratio Definition

Traditionally, the fissile CR (fCR) has been defined as the fissile atom production rate

divided by the fissile atom destruction rate [14]. The fissile production rate is defined as the sum

of neutron capture reactions (including reactions of fissile atoms) that lead to production of a

fissile atom. The fissile destruction rate is the sum of neutron capture and fission reactions

which remove fissile atoms. Therefore, the fCR is simply the ratio "or balance" between the

mass production rate (source) divided by the rate of mass destruction (sink). This more

traditional CR definition is the one calculated and reported by the REBUS code [57].

ZNP PO
S Fissile Atom Production (3
SFissile Atom Destruction N" Nd +C f
d

Where: (d) is the daughter product resulting from the neutron capture and subsequent

decay (transmutation) form the (p) parent isotope. N is atom density and oc and of are the single

group neutron capture and fission cross sections respectively.









This definition has been altered slightly by authors in recent years to mean the net TRU

produced from uranium divided by the net TRU destroyed by fission giving a transuranic

burning CR (tCR) [58]. The difference between the fCR and tCR is that the tCR does not give

credit to capture reactions in the denominator because these reactions only transmute existing

TRU into other TRU. From a waste management standpoint only fission removes transuranic

waste from the fuel cycle. Additionally, only capture reaction of uranium (specifically U-236

and U-238) that lead to TRU atoms are allowed in the denominator of the tCR because these

isotopes directly transmute into Np-237 and Pu-239.

TRUProduction U-236 U-236 U-238U -238
tCR = (3 -2)
S TRUFission NTR (mr Ru
d

The change in the definition of CR is made to reflect the ABR's principle design objective,

which is to destroy TRU. The only real discrepancy between these two definitions is the

accounting of MA transmutation. The tCR does not consider fissile plutonium generation from

MAs as a source term in the numerator. The fCR gives an indication of the balance between

transmutations over total absorption reactions in the fuel but does not give an indication of the

actual TRU burning performance of the core. In both cases, MA transmutation is simply treated

as existing TRU isotopes being converted into other TRU. For example, if only plutonium and

no uranium are supplied to the SFR, than the tCR is zero. This is because the net TRU

production in the numerator of the tCR is zero. The fCR would also be near zero because the

transmutation parent of Pu-239 is U-238. The tCR and fCR would be in good agreement because

the contribution of U-238, which is the dominating parent isotope in the numerator of the fCR,

would be near zero.









However, if neptunium and americium is the only fertile species and the only fissile

species is their transmutation daughter Pu-238 (and ultimately other TRU through successive

neutron capture), than the tCR is still zero. Yet, the fCR would be near unity. In this

hypothetical situation, there is no net production of TRU. However, this hypothetical core would

generate its own fissile worth, in the form of transmuted plutonium, in order to meet the excess

reactivity and cycle length requirements. Therefore, the core would be self sufficient from a

reactivity standpoint and thus require no external supply of fissile isotopes. Similar to a breeder

reactor, only an external supply of fertile MA isotopes would be required.

It is important to remember that transmuted plutonium (Pu-238) is not fissile in the

classical sense because in order for it to fission, additional kinetic energy must be added to

overcome the critical energy for fission. This is why the fission cross section of Pu-238 is so

much higher in the fast spectrum than in the thermal spectrum (Table 1-1). Pu-239 is fissile by

definition because fission can occur simply by contact with a neutron of virtually zero kinetic

energy.

The AHFTR design concept uses the fertile nature of SNF MAs to reduce the amount of

externally supplied SNF Pu-239 needed to supplement the plutonium bred internally. This has

the overall effect of reducing the tCR while increasing the fCR. Meeting the reactivity demands

of the fuel cycle using MAs reduces the demand for externally supplied fissile material from

SNF. The resulting surplus of SNF plutonium could be better used to fuel other SFRs such as

the ABR or perhaps be used to make plutonium based Mixed Oxide (MOX) fuel for LWRs.

Odd mass number plutonium isotopes (Pu-239,241) are fissile in any spectrum which

makes them a viable fuel for thermal reactors. In effect, SNF plutonium in a MOX fuel form can

have a tCR between 0.6 and 0.7 when irradiated in a PWR [14]. Given that several of the newer









classes of LWR's (Generation III and III+) are designed with the capability to be fueled with full

MOX cores, these reactors are a possibility for burning the plutonium component of SNF.

However, for the purpose of bounding the scope of this work, only the ABR and AHFTR are

considered as the final destination for SNF Pu and MAs in the closed fuel cycle analyses. For

the closed fuel cycle analyses to follow, the emphasis of the AHFTR is on MA consumption,

whereas the ABR focus is to destroy the plutonium component of SNF TRU.

This chapter will show that the AHFTR is well suited for the task of burning the MAs from

TRU. Therefore, a combination of ABRs and AHFTRs can be used to burn the TRU produced

by LWRs. In this scenario, most of the TRU mass, which is plutonium, would be burned by

ABRs. A fraction of the mixed-fleet (ABR and AHFTR) would be reserved for AHFTRs which

are dedicated to MA burning (Figure 1-9). An economic evaluation that compares this hybrid

ABR and AHFTR fuel cycle with the single reactor ABR fuel cycle will be evaluated in Chapter

7.

Transmutation Based Reactor Design

In this chapter, a parametric design study is conducted on the effects of core flattening,

moderation and heavy metal fuel content upon the transmutation performance of the axial

targets. The plutonium generation in the axial targets is also a function of reactor and fuel

geometry. Therefore, the primary purpose of the parametric study is to evaluate the tradeoff

between core performance factors and the maximum MA destruction rate. Also evaluated is the:

cycle length, excess reactivity, and "total core" Doppler coefficient and sodium void worth. All

these factors are related by core geometry and fuel composition.

In this parametric study, the target's height and composition are held constant. The targets

comprise a 20 cm tall axial blanket placed between the plenum and active core. In this

parametric study, the lateral core layout is assumed to be identical to the ABR case, reported by









Hoffman, with a tCR of 0.5 [7]. The lateral core layout for the ABR with the axial target

modification is shown in Figure 3-1. Starting with the ABR reference height of 101.6 cm, the

active driver core height was varied to 51.6 cm in 10 cm increments. For each core height, the

total core power was reduced by the fraction of volume removed from the active core by the

height reduction. The reduction in power with height was done in the parametric study to give

an equally comparable power density, depletion rate and cycle length for each height. Table 3-1

gives the reactor power of the core shown in Figure 3-1 for each increment in active core height.

Table 3-1. Reactor thermal power and core heights evaluated in parametric analysis
Total core height (cm) 121.6 111.6 101.6 91.6 81.6 71.6
Active core height (cm) 101.6 91.6 81.6 71.6 61.6 51.6
Reactor Thermal Power (MW) 1000 900 800 700 600 500

For each one of these active core heights, a range of pin pitch-to-diameter ratios, which

correspond to variable CR, is evaluated. For comparison purposes, the ABR reference (Figure 1-

5) was evaluated for each increment in pin pitch-to-diameter ratio. These ABR reference cases

provide a benchmark for comparison of reactor parameters for down-selecting to a practical

AHFTR core height and pin diameter.

The pin pith-to-diameter ratios and the corresponding fuel assembly volume fractions are

shown in Table 3-2. These values represent the fuel pin dimensions, used by Hoffman, to vary

the ABR tCR from 1.0 to 0.25.

Table 3-2. Reference fuel assembly design for varying pitch-to-diameter ratio.
p/d=l.1 p/d=1.176 p/d=1.293 p/d=1.357
tCR of the reference ABR 1.00 0.750 0.500 0.250
Fuel pin diameter (cm) 0.808 0.755 0.623 0.464
Fuel pin pitch (cm) 0.888 0.888 0.806 0.630
Pins per assembly 271 271 324 540
Fuel Assembly Volume Fraction
Fuel 34.27% 29.30% 22.08% 17.40%
Bond 11.42% 9.77% 7.36% 5.81%
Structure 25.73% 25.68% 26.41% 29.15%
Coolant 28.79% 35.25% 44.15% 47.60%









0
0

0


O0
0
0
0
0
0


Ultimate Shutdown Rod Assembly
Primary Control Rod Assembly
Inner Enrichment Zone Driver
Middle Enrichment Zone Driver
Outer Enrichment Zone Driver
Radial/Axial Reflectors
Shield Assembly
Axial Target Portion of Driver Assembly
Fuel Rod Gas Plenum


Handling Socket










(JQ
CD

C



A



Coolant Inlet Noiile


w

-

B
C-


Figure 3-1. Preliminary AHFTR core design used in parametric analyses

Finally, a core height and pin diameter that gives the highest transmutation efficiency

while at the same time ensuring reactivity coefficients within the boundaries of the homogeneous










reference was selected. Using this core height and pin diameter combination, additional driver

assemblies were added to the AHFTR core radius to bring the thermal power rating back to 1000

MWth as it is for the ABR design. This gives the AHFTR a commercial scale by roughly

equating the overall active core volume and power to that of the ABR.

Transmutation Targets and Accompanying Fuel Cycle

Target rods are placed adjacent to zirconium hydride (ZrHi.6) "dilution" rods in an axial

blanket configuration above the driver fuel. The term "dilution" was adopted to describe blank

fuel pins within the CAPRA core design that were filled with steel instead of fuel. A zirconium

metallic fuel alloy is assumed for both driver and target rods. Zirconium hydride was selected as

the moderator for its high thermal conductivity and melting temperature. A hydrogen-to-

zirconium stoichiometric ratio of 1.6 was selected for zirconium hydride's delta phase which

retains composition for temperatures up to 1000C (Figure 3-2) [59].


at. H
5o S 0 6 1 62 6 36 64 6S B6


f/ e
800 / ,
/
tooa / !








10
Iiw


1.3 1. 1. 1 1, 1.9 7,0
H/Zr
Figure 3-2. Zirconium hydride phase diagram for varying hydrogen content [59].

These driver fuel and axial target rodlets, "slugs", are collocated within the same fuel pin

and share the same plenum space. The term "slug" is common in metal alloy fuel literature

which indicates the injection casting process used in fabrication. Therefore, the fuel rod contains









a stack of metal slugs as opposed to the ceramic oxide pellets typical of LWRs. The ratio of

target to ZrH1.6 containing fuel pins is approximately five to one (Figure 3-3). Therefore, for a

typical hexagonal fuel assembly, 226 of the 271 pins contain target and 45 contain ZrH1.6 slugs.


O ZrH1.6 Moderating Rod


* Transmutation Target Rod


Figure 3-3. Representation of the axial target pin-lattice showing the orientation of targets (red)
and zirconium hydride pins (green)1

The accompanying fuel management strategy allows the targets and driver fuel to be

discharged and recycled together in the same pyroprocessor batch. After pyroprocessing, the

mixed target and driver mass streams supplies the fabrication of fresh driver fuel. Fresh targets

are fabricated from the americium and curium component of SNF. This americium and curium

stream is provided by the UREX+3 separations process (Table 1-8) for LWR fuel. The UREX+3

neptunium and plutonium stream provides the external fissile feed to the driver fuel as well as

the targets. The neptunium is kept with the plutonium by the UREX+3 process for the purpose

of enhancing proliferation resistance. Some plutonium is supplied to the targets to minimize the

power swing as a result of Pu-238 and Pu-239 generation during the irradiation. The reasons and

effectiveness of this approach are discussed in Chapter 6.


1 The color pattern indicates the five-to-one relationship between targets and moderating rods.
However, the actual number of pins in the picture is 210 targets and 61 zirconium hydride rods
which is 77% and 23% of the total 271 pins respectively.










As depicted in Figure 3-4, the SNF americium and curium is irradiated only once in targets

before it rejoins the neptunium and plutonium in the driver fuel. The fuel cycle scenario in

Figure 3-4 (and also in Figure 1-8) is used as the model for the out-of-core fuel management

operations performed by the REBUS code (Figure 2-1). The combined fuel cycle represented in

Figure 1-9 is performed by two separate REBUS calculations: one for the ABR and one for the

AHFTR.


AHFTR
spent fuel






Spent Fast Reacto
AHFTR Pyroprocessing, Blei
Driver Fuel Cas





Fresh Fast Reactor Target Casting an
Fuel Assembly Assembly Fabrici


Reprocessing
Losses


Np+Pu & RGU Make-up
Material for Driver Fuel


Np+Pu & Am+Cm & RGU
Target Material


ABR Fu
Light Water Reactor LWR SNF "UREX+"
Separation and Fabrication Plant

Figure 3-4. AHFTR fuel cycle scenario showing connections between partitioning and
transmutation technology2


el


2 Fuel cycle calculations performed for the LWR-through-UREX+ portion of the fuel cycle were
performed by TRITON. Equilibrium fuel cycle calculations performed for the boxed portion of
the flow chart were performed by REBUS. A separate REBUS calculation is performed to
model the equilibrium cycle of the ABR. (See Figure 1-1 and Figure 1-9).


r Fuel
ending &
ting




I
d Fuel
nation


l T lrJu


TT I T -









Therefore, sufficient americium must be destroyed in the targets such that the unburned

americium is less than 5 w/o of the HM in the driver. The 5 w/o limit for MA driver fuel

concentrations is selected as a guideline for ensuring acceptable reactivity kinetics features of the

fast reactor. This limiting MA concentration also ensures that the expected irradiation

performance of the driver fuel does not significantly deviate from the current experience base on

metallic SFR fuels. The decision to use 5 w/o of MA per total HM is based on results obtained

by the CAPRA program conducted in Europe to evaluate the EFR [16,17].

Similar to the ABR, the AHFTR has no radial blankets and uses three enrichment zones to

flatten the radial power profile across the core [7]. For the purpose of simplification of

nomenclature, the term "enrichment zone" is dropped and only the terms: inner core, middle

core and outer core are used instead. The driver fuel composition is a TRU/U/O1Zr (by weight)

alloy. The ratio of middle and outer core enrichments to that of the inner core is: 1.25 and 1.50

respectively. The driver fuel composition definition is the same as the homogeneous ABR

reference with the exception of the external feed being UREX+3 Np+Pu. The external feed for

the reference metallic fueled ABR design groups all of TRU elements together using the

UREX+la process at the SNF aqueous reprocessing facility.

The target fuel fixed composition is 10Np+Pu/10Am+Cm+Bk+Cf/40U/40Zr by weight.

Some of the UREX+3 separated Np+Pu stream (i.e. external feed) is diverted from the active

driver fuel supply to the targets. None of the pyroprocessed plutonium is used to refuel the

targets. This is mainly because the current pyroprocessing technology does not readily allow for

elemental separation. The choice to use pyroprocessing allows the higher mass actinides of

curium and californium that were generated during the target irradiation to be diluted over a

larger fuel inventory. This dilution is intended to reduce the intensity of radiation fields and









hence shielding requirements to fuel fabrication workers dealing with recycled material. The 10

w/o Am+Cm+Bk+Cf concentration in targets was selected to retain the fuel performance

characteristics observed in the AFC-1 irradiation experiments which were performed at INL.

The AFC-1 tests and the reasons for proposing this specific combination of Am+Cm+Bk+Cf

with Np+Pu, U and Zr is discussed in detail in Chapter 6.

Transmutation Target Physics

As mentioned in the introduction, the repository space benefit stems primarily from the

removal of americium from the fuel cycle. This is because the repository's waste emplacement

drift spacing is limited by the maximum rock temperature at the midpoint between them. This

rock temperature is principally a function of the decay heat produced by Am-241 in the SNF

[13]. Because of its long half-life, radiotoxicity, high solubility and low sorption in Yucca

Mountain tuffs, Np-237 is the principal environmental concern to the biosphere, if water does

come into contact with the SNF. Because, Np-237 is the alpha decay product of Am-241,

americium destruction in transmutation targets also minimizes the Np-237 accumulation in the

repository.

Am-241 and Np-237 have an even neutron number and thus the binding energy

contribution of an absorbed neutron is not sufficient to overcome the critical energy required for

fission. In fact, the addition of a neutron to an odd-neutron nucleus (fissile isotope) to form an

even-neutron compound nucleus gives a binding energy change that is about one MeV greater

than for changing an even-neutron nucleus into an odd-neutron compound nucleus. This

explains the fission threshold at one MeV for the long lived MAs, Np-237 and Am-241. A

neutron capture in Np-237 generates the fissile Np-238 nucleus. However, Np-238 quickly beta

decays into Pu-238 with a short 2.117 day half-life. A neutron capture in Am-241 produces the

fissile Am-242,242m isotopes. The yield fraction to the ground state is estimated to be










approximately 85% in the fast spectrum [10]. This Am-242 ground state beta decays into Cm-

242 with a branching ratio of 83%. The other 17% of Am-242 electron captures to become Pu-

242. Cm-242 decays into Pu-238 with a 163 day half-life.

Because of the fission threshold, the Am-241 fission cross section below one MeV and

above the resonance range (where the SFR neutron spectrum is very small) is two orders of

magnitude less than for fission above one MeV (Figure 3-5). However, the capture cross section

is almost as high as the Pu-239 fission cross section in the same energy range.

1 E+06


1E+04
-c

-1E+02



1E+00
,x


E 1E-02 -Am-241 Capture
-Am-241 Fission
-Pu-239 Fission
-Neutron Flux
1E-04
1.E-07 1.E-05 1.E-03 1.E-01 1.E+01
Energy (MeV)
Figure 3-5. ENDF-VI americium and plutonium cross section plots versus a metal fuel SFR
neutron spectrum

Therefore, the AHFTR axial target neutron spectrum is moderated slightly in order to

reduce the neutron energy to just below one MeV. This increases the neutron capture in

americium relative to plutonium fission in the targets. The effect can be seen by evaluating the

ratio of total absorption in americium over total absorption in Pu-239. This increase can be seen

by observing a capture and a fission based fuel utilization factor for the driver and the targets. In

Table 3-3, a transmutation utilization factor is used to quantify the reaction probability of capture









in a given isotope divided by any absorption in the region of space being studied. The

transmutation utilization factor is defined in the equation below.

Nro
transmutation utilization = (3-3)
Y N al +Naj

Where: i represents a given isotope in targets or driver fuel and r represents capture or fission

Similarly, the fission utilization factor is defined as the reaction probability of fission in a

given isotope divided by any absorption in the region of the core being studied. Also supplied in

Table 3-4 is a single group cross section ratio. The cross section ratio shows the spectral effect

on the microscopic cross sections alone without being weighted by number density.

Table 3-3. Transmutation utilization factor (MCNP*)
Driver Target
CAP/ABS FISS/ABS CAP/ABS FISS/ABS
Am-241 1.59% 0.34% 15.64% 0.93%
Pu-239 6.02% 30.36% 8.24% 15.43%
*Single group cross sections for Table 3-3 and Table 3-4 were performed by MCNP benchmarks
of REBUS calculations using the final AHFTR core design.

Table 3-4. Single group cross section ratio (MCNP)
Am-241 CAP over Pu-239 FISS Am-241 ABS over Pu-239 ABS
Driver 0.77 0.78
Target 1.33 0.91

It is important to note that the ratio of the Am-241 capture to the Pu-239 fission cross

section is greater in the target than the driver fuel because of moderation. It is also apparent that

the total absorption cross section ratio of Am-241 over Pu-239 is greater in the targets than the

driver fuel because of this increased neutron capture. This effect explains how the capture

utilization factor for Am-241 is nearly equivalent to the fission utilization in Pu-239 in the

targets.

The slight target moderation results in a relatively epithermal flux compared to that for the

active driver regions as shown in Figure 3-6. It is important to note, that though the total flux in

the targets is less than in the active driver, much of it is being depressed by resonance absorption










in the epithermal range. Much of this resonance absorption is in Pu-239 and U-238. However,

resonance absorption in Pu-239 leading to fission serves to generate more neutrons. These

neutrons have a relatively short mean-free-path because of the softer spectrum. Therefore, they

remain locally within the target's neutron population. Moreover, resonance absorption in U-238

serves to generate more Pu-239. The combination of epithermal spectrum and plutonium

breeding by both Am-241 and U-238 creates enough plutonium to replace the initial plutonium

loaded in the targets at BOEC.


1E+16
x Targets
> Inner Core
1E+15 Middle Core "
Outer Core *
1E+14 X

< 1E+13
E X
S1E+12 *





1E-05 1E-04 1E-03 1E-02 1E-01 1E+00 1E+01
Energy (MeV)
Figure 3-6. Target flux spectrum compared with the inner, middle and outer core for a 101.6 cm
tall AHFTR with p/d of 1.1 (REBUS)

This breeding gives a rCR in the targets close to one but a tCR more near to 0.75. As can be

seen by Figure 3-7, the total amount of Pu-239 stays relatively constant. However, the total

plutonium content increases slightly, mainly from Pu-238,242 production from Am-241

transmutation. However, the total transuranic mass is being reduced as americium is being

transmuted.

Parametric Study

Because the transmuted plutonium remains in the fuel cycle via the pyroprocessor, the

amount of space reserved in the reactor for targets plays a significant role in the demand for the
1E+10O-----------------------
























amount of space reserved in the reactor for targets plays a significant role in the demand for the










UREX+3 supplied Np+Pu. From a geometry standpoint, the ratio of target volume to core

volume is dictated by the active core axial height. From a physics standpoint, the amount of Pu-

239 breeding in the active core is a function of core flattening. Since, breeding excess Pu-239 in

the active core is not a primary objective of the AHFTR, it is desirable to minimize the active

core's CR. Conversely, since breeding even-plutonium is an indicator of americium destruction,

it is not obligatory to minimize the conversion ratio in the target region.


3.5E+01
*CM245
3.0E+01 CM244
O CM242
2.5E+01 0 AM243
U AM242m
2.0E+01 AM241
S PU242
O1.5E+01 e PU241
SPU240
.S 1.0E+01 0 PU239
CO
O PU238
5.0E+00 0 NP237

0.0E+00
0.00 0.92 1.85 2.77 3.69 4.61 5.54
Effective Full Power Year

Figure 3-7. Isotope masses as a function irradiation time within the target region a 101.6 cm tall
AHFTR with p/d of 1.1. Each color bar represents an isotope's mass (REBUS)

Effects of Pin Diameter and Core Height

Because of the vast range of options presented that affect the reactor's overall physics and

transmutation performance, the following assumptions are made for the purpose of confining the

parametric analysis of pin and core dimensions.

* Core design from Figure 3-1.
* Core height varied from 101.6 cm to 51.6 cm (Table 3-1)
* Fuel pin diameter varied from 0.808 cm to 0.464 cm (Table 3-2)
* Number of fuel pins per assembly (both driver and target region): 271 (Table 3-2)
* Peak fuel burnup constrained to 18 at. % (Figure 2-1)
* Ratio of zirconium hydride to target rods is: 1/5









The rate of americium destruction is a function of the target spectrum, active core axial

leakage and target rod volume. Increasing the diameter of the target rod also increases the

volume of americium charged to the AHFTR per cycle. If the reactor cycle is shortened, then the

time rate of americium fed to the core is increased because the target mass loaded per cycle is

constant due to the fact that the target rod composition has been fixed. This does not, however,

indicate the spectrum effect of the zirconium hydride content or flux intensity in the targets.

Because the target irradiation time is not held constant, simply comparing the ratio of charged to

discharged target americium is not an adequate indication of the destruction efficiency. Instead,

a half-life is defined for Am-241 transmutation which quantifies the magnitude of the capture

reaction rate. Because the rate of Am-241 production rate via beta decay from Pu-241 is

relatively slower than its destruction rate via neutron capture, the transmutation half-life can be

approximated using a first order differential equation. Therefore, Am-241 destruction can be

characterized with a purely exponential behavior just as with radioactive decay. This

exponential behavior is given by the Am-241 production and decay rate equation. Table 3-5

shows the irradiation time required to reduce the americium mass by half.

u 2a 41 A tu u1 T 1 transmutation ) (3-4)
b = Zdecay N h ive Nfo a p AF di(captuYas,,)---- (3B-4)S
ldt al Am-2

Table 3-5. Transmutation half-lives for a preliminary AHFTR design (Years) (REBUS)
Pitch/Diam. p/d=1.1 1.176 1.293 1.357
101.6 cm 2.49 2.31 2.04 1.87
70.6 cm 2.73 2.54 2.24 2.06
50.6 cm 3.23 2.99 2.64 2.44

Notice the half-life decreases for decreasing pin diameter. This can be explained by the

higher sodium fraction in the active core for decreasing fuel pin diameter. Increased leakage

invests more neutrons in the targets. The decreasing half-lives with height are also related to

axial leakage. However, the expected result is a decrease in half-life due to an increase in









leakage with the core height reduction. This contradiction in the half-life trend with core height

can be explained by more evenly distributed neutron utilization in the active driver core.

Because of the enhanced axial leakage, the driver fuel radial power profile develops a

depressed region in the inner core with decreasing height. The increase in the americium

transmutation half-life with decreasing core height in Table 3-5 is caused by a reduction in the

intensity of neutrons leaving the inner core region.

It is important to note the conformity of the three power profiles plotted in Figure 3-8 in

the area of the outer core. This behavior is indicative of the cosine shape of the radial flux

profile as neutrons leave the core. The increased axial leakage shown in Figure 3-9 affects the

curvature of the radial power profile in the inner core region more so than in the outer region.

This is because the axial flux gradient is decreased more in the inner core than the outer core.

The flux in the outer core regions are already suppressed everywhere by radial leakage. This

explains the conformity of the three different radial power profiles in Figure 3-8 at the outer

edge. Because the power density becomes more evenly distributed axially and radially, the

power density on the active core top surface becomes reduced.

A flatter radial power profile also reduces the radial power peaking. The reduction in

power peaking in the inner core makes possible raising the reactor power, which was reduced in

the parametric study for decreasing core height. Even though the decrease in core height

increases axial buckling and hence axial leakage, the larger buckling in the radial direction

relative to the axial direction causes flux in the inner core to be reduced which gives the flat

radial power profile. This reduction of the inner core flux is also a result of the reduction in

power for each reduction in core height which was done for the purpose of the parametric study.











Because of the depressed flux in the inner core, the power of the final flattened AHFTR core


geometry is increased to give a more realistic power density.


400
-+101.6 cm Tall
350 -71.6 cm Tall
350
S-=-51.6 cm Tall

E 300

250

S 200

I 150
0
0.
100

50
1 2 3 4 5 6 7
Fuel Assembly Row Number

Figure 3-8. Active core radial power density profile for a preliminary AHFTR with active core
height: 101.6, 71.6 and 51.6 cm (p/d=l.1) (REBUS)


470 101.6 cm Tall
-71.6 cm Tall
420 -*-51.6 cm Tall

370
E
320

270

o 220

S170
0

120

70
20% 30% 40% 50% 60% 70% 80% 90% 100%
Percent Height from Bottom of Core

Figure 3-9. Inner core axial power density profile for a preliminary AHFTR with active core
height: 101.6, 71.6 and 51.6 cm (p/d=l.1) (REBUS)

After increasing reactor power back to the practical limit, the increased power density in


the targets will translate into a higher flux density for transmutation. Therefore, the


transmutation half-life for the targets with a more realistic thermal power rating will be higher









than for the parametric study. A following section describes two possible AHFTR core

geometries with two different active core heights and a 1000 MW power level.

Effects of Moderating Pins

The affect of moderation on the transmutation efficiency of the target region was also

evaluated. Using a pin pitch-do-diameter ratio of 1.176 from Table 3-2 and a core height of

101.6 cm, the number of zirconium hydride rods per target rods was varied. The ratio of

zirconium hydride rods per target rods was varied from: 1/5, 1/6, 1/7, 1/8 and 1/9. All of the

other design variables used in the previous section were kept the same:

* Core design from Figure 3-1.
* Core height of 101.6 cm, Fuel pin diameter of 0.755 cm (Table 3-2)
* Number of fuel pins per assembly (both driver and target region): 271 (Table 3-2)
* Peak burnup constrained to 18 at. % (Figure 2-1)
* Vary the number of zirconium hydride to target rods from 1/5 to 1/10

It was found that the affect of moderator did have an affect on transmutation efficiency.

Table 3-6 gives the transmutation half-life of Am-241 for each moderator-to-target rod ratio.

Table 3-6. Transmutation half-life of Am-241 for varying number of moderating rods per target
rods in the target region for the preliminary AHFTR design (REBUS)
Description* Transmutation Half-life (Years)
One ZrH1 6 per Nine Target Rods 3.903
One ZrH1 6 per Eight Target Rods 3.759
One ZrH1 6 per Seven Target Rods 3.605
One ZrH1 6 per Six Target Rods 3.411
One ZrH1 6 per Five Target Rods 3.193
*Total number of pins is held constant by the fuel assembly design

It is important to note that the transmutation half-life decreases for increasing moderator.

This means that the addition of hydrogen to the target region does contribute a spectrum effect to

the transmutation efficiency. Figure 3-10 and Figure 3-11 shows the average neutron spectrum

in the target region for each moderator-to-target rod ratio.












4.0E+14
-_- One ZrH1.6 Per Five Targets
3.5E+14 -- One ZrH1.6 Per Six Targets
-- One ZrH1.6 Per Seven Targets
>. -E+1 One ZrH1.6 Per Eight Targets
e 3.OE+14
S-e- One ZrH1.6 Per Nine Targets

2.5E+14

2.0E+14

E 1.5E+14

2 1.0E+14
ul.

5.0E+13

1.0E+11
1E-06 1E-05 1E-04 1E-03 1E-02 1E-01 1E+00 1E+01
Energy (MeV)

Figure 3-10. Average neutron flux in the target region of a preliminary AHFTR design for a
varying number of moderating rods per target rods (lin.-log. scale) (REBUS)


1E+15




P 1E+14




1E+13


*- One ZrH1.6 Per Five Targets
2 1E+12 / -- One ZrH1.6 Per Six Targets
-- One ZrH1.6 Per Seven Targets
One ZrH1.6 Per Eight Targets
-- One ZrH1.6 Per Nine Targets
1E+11
1E-06 1E-05 1E-04 1E-03 1E-02 1E-01 1E+00 1E+01
Energy (MeV)


Figure 3-11. Average neutron flux in the target region of a preliminary AHFTR design for a
varying number of moderating rods per target rods (log.-log. scale) (REBUS)


Notice that the fast flux decreases for increasing concentration of moderating rods. This is


a sign that the fast component of the flux is being reduced by the moderation. However, because


the y-axis of Figure 3-10 is linear, it is difficult to see from the plot whether or not there is an









increase in the neutron flux at lower energies. Figure 3-11 shows the same data that was plotted

in Figure 3-10 but on a logarithmic scale.

Notice that the thermal-epithermal component of the spectrum increases for increasing

number of moderator per target rods. The increase in thermal-epithermal flux for increasing

zirconium hydride rods explains the improvement in transmutation half-life. As the spectrum

softens, more of the flux is placed in the resolved-unresolved resonance range below one MeV.

Recalling the cross section plot in Figure 3-5, the Am-241 capture cross section in this range has

the same overall magnitude of that of the Pu-239 fission cross section. Because, the Am-241

loading and Pu-239 loading are approximately equal, the Am-241 capture reaction rate provides

a degree of energy shielding effect over the fission rate of Pu-239 (and the other plutonium

isotopes). The small amount of energy shielding gained by Am-241 allows neutrons invested in

the target region by leakage from the active core to be invested in neutron captures in Am-241 as

opposed to fissions in the Pu-239 which enables high transmutation efficiency.

The transmutation efficiency gained by the moderation can be visualized by tracking the

change in the relative Am-241 content of the fuel normalized to the initial Am-241 loading of the

fresh fuel, which is shown in Figure 3-12. In addition to the destruction of Am-241, the gain in

transmutation efficiency by increasing neutron captures (i.e., transmutation) can be visualized by

tracking the change in the relative Pu-238 content of the fuel normalize to the initial Pu-238

loading in the fresh fuel, which is shown in Figure 3-13.

Because the case with one zirconium hydride rod per five target rods demonstrated the

highest transmutation efficiency over all other cases, this combination was selected for the final

down selection of the AHFTR design. A decision was made to limit the concentration of

moderator rods to one per five targets due to a concern that too much moderation would cause










power peaking effects in the target region and near the interface with the driver fuel. Increasing

the thermalizing effect decreases the neutron mean-free-path. The concern over power peaking

arises from the possibility that discontinuities in the neutron flux between regions with dissimilar

neutron spectrums (due to isotope depletion) could emerge if the neutron mean-free-path

becomes too short. These discontinuities would not only lead to unacceptable power peaking,

but would also invalidate the accuracy of the fast reactor codes, MC2-2 and DIF3D, for the target

analysis. These codes assume minimal flux discontinuity from region-to-region in order to

homogenize the fuel over large regions of the core.


120%
LLI
0
m 100%

80% -
.~ -
< 60%

S40% One ZrH1.6 per Five Targets
S---One ZrH1.6 per Six Targets
-- One ZrH1.6 per Seven Targets
S20% -*- One ZrH1.6 per Eight Targets
S-e-One ZrH1.6 per Nine Targets
0%
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5
Effective Full Power Year

Figure 3-12. Percent of the initial Am-241 mass remaining in the target rod as a function of
irradiation time for varying number of moderating rods in a preliminary AHFTR
(REBUS)

There is an additional feasibility limitation on the practical number of pins that can be

loaded with zirconium hydride as opposed to targets. The addition of moderator rods reduces the

number of target rods available for loading MAs. Therefore, to bur an equal amount of mass in

the targets, the MA concentration in the target slug must increase for decreasing number of target

pins. As will be discussed in Chapter 6, for metal alloy fuels there is a practical limit on the










concentration of MAs that can feasibly be loaded into the fuel which are related to the volatility

of americium in its melted form.


800%
LLI
O 700%

0 600%

S500% -

400%

S300%
S- One ZrH1.6 per Five Targets
--One ZrH1.6 per Six Targets
200%
2- 0 -One ZrH1.6 per Seven Targets
a -*-One ZrH1.6 per Eight Targets
P 100% -e-One ZrH1.6 per Nine Targets

0%
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5
Effective Full Power Year

Figure 3-13. Percent of the initial Pu-238 mass created in the target rod as a function of
irradiation time for varying number of moderating rods in a preliminary AHFTR
(REBUS)

Tall and Flattened Axial Heterogeneous Core Designs

Using the reference ABR, with pin diameters from Table 3-2, as a standard for

comparison, the AHFTR core height and pin diameter were varied. The reactor performance

traits considered were the excess reactivity, total core void worth and Doppler coefficient. The

largest diameter fuel pin design (p/d=l.1) from Table 3-2 is selected (for this section) to give a

Doppler coefficient with increased negativity than the reference case. This is expected because

the volume of fuel in the active core is increased and the percent of that fuel being uranium is

also increased. After down-selection, the final active core geometry is 91.6 cm with a 1.1 pin

pitch-to-diameter ratio. The 10 cm reduction in active core height was found to give a total core

void worth that was slightly less than the homogeneous reference ABR with the 1.1 pin pitch-to-

diameter ratio (Table 3-2). Though the active driver core height is 10 cm less than the reference

ABR, the total core height is 10 cm "taller" due to the addition of the 20 cm height of the targets.









This core design, as well as a much "flatter" core design, is compared to the reference ABR in

Table 3-7. It should be noted that these comparisons are made for a pin pitch-to-diameter ratio

of 1.1. The flat AHFTR core design, discussed next, is given in Figure 3-14.

Table 3-7. Core design summary for the reference ABR with tall and flat AHFTR (REBUS)
ABR (Ref.) Tall AHFTR Flat AHFTR
Total Core Height 101.6 110.6 91.6
Active Driver Height 101.6 91.6 71.6
Total Core Volume (m3) 3.3 3.6 4.0
Active Core Volume (m3) 3.3 3.0 3.1
Rows of Driver Fuel 7 7 8
Pitch-to-diameter ratio 1.1 1.1 1.1
Inner Core Enrichment 14.92% 16.07% 18.65%
Enrichment Split (IC/MC/OC) 1.0/1.25/1.50 1.0/1.25/1.50 1.0/1.12/1.25
Cycles per Enrich. Zone (IC/MC/OC) 6/6/7 6/6/7 6/6/7
Cycle Length (EFPD) 322.86 319.90 350.58

For the flattened version of the AHFTR, a 71.6 cm height is evaluated to observe a much

larger target-to-driver volume ratio. An additional row of outer core drivers is added to this flat

core so that the power density is made roughly comparable to the "tall" AHFTR (Figure 3-14).

Radial and Axial Power Profiles

Because of the inherently flatter power distribution, the gradient of enrichment splitting for

the flat AHFTR can be decreased from that used for the ABR. Therefore, the "flat" core's

middle and outer zones are enriched to 1.12 and 1.25 times that for the inner core respectively.

This tailored enrichment splitting gives the flat core a much more evenly radial power

distribution than the tall core. The axial leakage works in parallel with the enrichment splitting

to create an almost completely flat radial power profile across the inner and middle enrichment

zone. Figure 3-15 shows the radial power distribution for six axial slices through the tall and

flat versions of the AHFTR.

In Figure 3-15 and Figure 3-16, the power density of the flatter core geometry is noticeably

greater than that of the tall core geometry. This greater power density is linked to the axial








leakage coming from the active core. As was seen in the transmutation half-life discussion in the

parametric study, flattening the core reduces the radial curvature, or buckling, of the power (and

flux) profile.


O Ultimate Shutdown Rod Assembly

O Primary Control Rod Assembly
O Inner Enrichment Zone Driver

O Middle Enrichment Zone Driver
O Outer Enrichment Zone Driver
SRadial Reflector Assembly or Axial
Reflector Region of Driver Assembly
SShield Assembly

O Axial Target Portion of Driver Assembly

O Fuel Rod Gas Plenum
Handling Socket


n H

& -t
C-

s cD
'CT


SCoolant Inlet Nozzle

Figure 3-14. Preliminary "Flat" AHFTR design with eight rows of fuel instead of seven












5.0E+02

4.5E+02

4.0E+02

3.5E+02
E
3.0E+02

S2.5E+02
c
o 2.OE+02

o 1.5E+02

1.0E+02

5.0E+01

0.OE+00


4.5E+02


4.5E+02

4.0E+02


E
3.5E+02


2.5E+02

" 2.0E+02

S1.5E+02

. 1.0E+02

5.0E+01

0.0E+00


1 2


-20%
-=-40%
- -60%
80%
-- 100%
--120%

2 3 4 5 6 7
Radial Fuel Assembly Row Number
















-*-20%
--40%
- 60%
80%
--100%
-120%


3 4 5 6
Radial Fuel Assembly Row Number


7 8

B


Figure 3-15. Radial power density profile for six axial slices through the core: Axial height is
represented as a percentage of the full core height. (A = "tall", B= "flat" core designs)
(REBUS)


As can be seen in Figure 3-15, the curvature of the radial power profile is less for the flat


core than it is for the tall core. Remembering back to the discussion in Chapter 2 on the critical


buckling of a hypothetical bare cylindrical SFR, the sum of radial and axial geometric buckling


must be equal to the material buckling. Therefore for approximately equal materials buckling


between tall and flat geometries, a decrease in radial buckling requires an increase in axial












buckling. Hence, the axial buckling of the flat core is greater than that of the tall core. It is


because of this increase in axial buckling that the flat core has more axial leakage than the tall


core. The greater axial leakage gives the flat core a greater power density in the target region


than in the tall core, which is shown in Figure 3-16.


5.0E+02

4.5E+02

4.OE+02

E 3.5E+02 -

g 3.0E+02

2.5E+02

o 2.0E+02

S1.5E+02 -ROW1 \
S1.5+02- ROW 2
0. ROW 3
1.0E+02 ROW 4
-ROW5
--- ROW 5
5.0E+01 ROW6
-- ROW 7
O.OE+00 -
20% 40% 60% 80% 100% 120%
Axial Level (20%=bottom,100%=top,120%=target) A


4.5E+02

4.0E+02

3.5E+02

E 3.OE+02

2.5E+02

2 2.0E+02

1.5E+02 -- ROW 1
ROW 2
0 ROW 3
1.0E+02- ROW 4
-- ROW 5
5.0E+01 -ROW 6
ROW 7
ROW 8
0.OE+00
20% 40% 60% 80% 100% 120%
Axial Level (20%=bottom,100%=top,120%=target) B


Figure 3-16. Axial power density profile for each row of fuel: Axial height is represented as a
percentage of the full core height. (A = "tall", B= "flat" core designs) (REBUS)









The increased power density in the target region also provides for a reduced transmutation

half-life of 2.22 years. This transmutation half-life corresponds to a transmutation efficiency

which is comparable to the small pin diameter (p/d=1.357) from Table 3-5. This small pin

diameter was used by Hoffman et al to attain a tCR equal to 0.25. Therefore, the flat AHFTR

core design achieves the MA destruction efficiency of a core with a thin fuel pin but does this

using a larger pin diameter. The larger pin diameter used for the flat version of the AHFTR is

more representative of the ABR with a tCR= 1.0.

This higher efficiency combined with an increased volume of the target region (physically

more fuel assemblies) gives a higher destruction rate of Am+Cm+Bk+Cf when compared with

the tall core. This higher Am+Cm+Bk+Cf destruction rate equates into a higher plutonium

transmutation breeding rate. This increases the amount of Am+Cm+Bk+Cf and reduces the

amount of Np+Pu drawn from the UREX+3 plant. Hence, the size of the surplus Np+Pu feed

(outside the boxed off portion of Figure 3-4) decreases.

In fact for both the tall and flat designs, the external Np+Pu feed for the active core is

reduced to zero for the equilibrium fuel cycle. Consequently, for both the tall and flat core

designs, the UREX+3 plant only needs enough Np+Pu and Am+Cm+Bk+Cf to produce fresh

targets.

For either case, the radial power profile falls off sharply in the outer two rows of fuel. This

is attributed to the dominating radial leakage effect on the flux gradient in the outer core. It is

important to note, the volume ratio of these outer two rows to the rest of the fuel is 0.5 and 0.44

for the flat and tall cores respectively. So the volume of the target material located in the low

flux of these outer two rows decreases as the core radius is increased. Hence, the flat AHFTR

has the smallest fraction of targets located above the low flux, low power driver fuel. This gives









the flat AHFTR the advantage of having the highest achievable transmutation rate over a greater

share of the targets than the tall AHFTR.

In addition to the geometrical improvement in target exposure in the radial direction, the

axial volume ratio of target to driver fuel increases with core flattening. This is because the

target volume in the numerator of this ratio is fixed for a given radius by its 20 cm height, but the

denominator is decreasing as the core height is decreasing.

Target eRre x (20 cm) (20 cm)
(3-5)
driver co achve or e aactve core

Where: Vtarget is the approximate volume of the targets in the core. Vdriver is the

approximate volume of driver fuel in the core. Rlore is the core radius. active core is the height of

the active core. The height of the targets is held constant at 20 cm.

Therefore, the overall MA charge rate per cycle increases as the active core height is

decreased. The combined effect of the increased axial leakage from the active core, the

increased share of fuel having a higher power and an increase in the MA charge rate per cycle

makes the flat AHFTR design the most attractive transmutation system because of its physical

ability to consume MAs.

Table 3-8 shows the fuel cycle parameters for the homogeneous reference compared to the

tall and flat AHFTR designs. As expected, the flat AHFTR has a greater amount of

Am+Cm+Bk+Cf being destroyed in the targets. This can be observed by examining the

americium transmutation half-life for the flat versus the tall core design. In addition to a

decreased transmutation half-life (increased capture reaction rate), the overall Am+Cm+Bk+Cf

target destruction rate is also increased. The enhanced destruction rate is the combined effect of

transmutation half-life and the target assembly charge rate.









Despite having a larger destruction rate in the targets, the flat core has a smaller overall

Am+Cm+Bk+Cf consumption rate (Table 3-8). This is because the flat core has a longer cycle

length (i.e., refueling interval) than the tall core which is also equivalent to a smaller fresh target

charge rate. Though the tall AHFTR has a more efficient target (shorter transmutation half-life),

these targets discharge a larger Am+Cm+Bk+Cf volume to the pyroprocessor. The tall core has

a less efficient target but the shorter cycle length increases the rate that Am+Cm+Bk+Cf is

charged to the targets and consequently the rate that un-transmuted MA mass is received by the

pyroprocessor. The difference in target destruction rate and total core consumption rate

illuminates the importance of distinguishing between the transmutation rate of a given isotope

and the actual charge rate of that isotope. The isotope destruction or reaction rate is a function of

neutron flux and energy in any given region of the core, whereas the charge or consumption rate

is more influenced by the logistics of the fuel cycle operating at steady state.

Table 3-8. Fuel cycle comparison for the reference ABR with tall and flat AHFTR (REBUS)
ABR (Ref.) Tall AHFTR Flat AHFTR
Active Driver Height 101.6 91.6 71.6
Maximum MA in Driver Fuel
OC Am/HM 0.00% 1.72% 1.79%
OC MA/HM 1.06% 3.20% 3.40%
TRU Externally Supplied Feed Rate
Am+Cm+Bk+CfFeed* Rate (kg/EFPY) 2.79E+00 3.63E+01 3.31E+01
Np+Pu Feed Rate* (kg/EFPY) 4.84E+01 3.60E+01 4.31E+01
HM Feed Rate (kg/EFPY) 3.75E+02 3.80E+02 3.75E+02
Transmutation Half-Life
IC Target Am Half-Life (Yr) -- 2.49 2.22
*See Figure 3-4 for a pictorial representation of the source of the external feed.

In both cases, the amount of MAs diluted in the driver fuel is comparable to that of the

reference ABR and is less than the 5% limit set by the CAPRA recommendation as discussed

earlier. Note that the MA concentration in the outer core is higher than the other two zones

because the transuranic enrichment is the highest in that region due to enrichment splitting









(Table 3-8). The Am+Cm+Bk+Cf in the driver fuel is effectively burned in the active core

region and does not accumulate in the AHFTR fuel cycle.

Reactivity Feedbacks

The Doppler coefficient and total core void worth are compared in Table 3-9. These

performance parameters are calculated for the cores at BOEC, using the calculation methods

discussed in Chapter 2. The Doppler coefficients for these cores are fairly comparable with the

values reported by Hoffman et al [7]. It should be noted that the ABR reference considered in

these scoping calculations is not the exact same core design proposed by Hoffman et al [7]. The

core layout, cycle length, number of batches, etc. were not conserved between this analysis and

that of Hoffman's. The core layout given in Figure 1-5 is the one proposed by Hoffman for a

tCR of 0.5. The core design that Hoffman proposed for the larger fuel pins (p/d=l.1) is slightly

different (Table 3-9). The tCR=0.5 core layout was considered a middle-of-the-road design that

could be used for comparison purposes in the scoping calculations in this chapter.

The reference ABR design and the flat version of the AHFTR both have a slightly negative

Doppler coefficient. However, the tall version of the AHFTR has a slightly positive Doppler

coefficient. The positive void coefficient of the tall core is due to the higher concentration of

MAs in the driver fuel than the reference ABR and an absence of core flattening. The affect of

fuel temperature increase, for the tall core causes resonance broadening of the U-238, TRU and

fission products. This resonance absorption can cause the neutron spectrum to harden which

causes an increase in multiplication. The positive worth of this multiplication can be greater than

the negative feedback of the U-238 resonance absorption if the core has inadequate leakage to

remove this extra reactivity. Hence, a degree of leakage is beneficial for Doppler feedback to

ensure that the inner regions of the core are not overly reflected. An increase in axial leakage









also decreases the positive void worth, because of an improvement in axial streaming compared

to the other two more symmetric cores.

As discussed in Chapter 2, the conversion ratio is also sensitive to core flattening. As

leakage is enhanced, the concentration of fissile material must increase. Hence, the TRU

increases. The excess reactivity may be viewed in the same way. As leakage increases, the TRU

enrichment required to achieve the maximum fuel burnup (taken to be 18 at. % in these scoping

calculations) also increases. This can be seen in Table 3-9. For the homogeneous case, this

would normally reduce the cycle length because a reduction in U-238 capture also reduces the

production of Pu-239 which is needed to achieve the fuel burnup. However, the flat AHFTR

design has the longest cycle length. This is attributed to the reactivity breeding effect of the axial

target region. Since, the overall core size is larger for the flat versus the tall or ABR cases, there

is actually less neutron escape when the targets are factored into the consideration. Therefore,

neutrons are efficiently invested in creating fissile material in the targets as the initial plutonium

charge in the driver bums out.

Also shown in Table 3-9 is the effect of total core size on the AHFTR dominance ratio.

The dominance ratio gives an indication of the closeness of the first and second lambda-mode

eigenvalues and is generated by the DIF3D calculation [42]. A larger dominance ratio (nearer to

1.0) indicates less neutronic coupling between regions of the core. Because the mean-free-path

for fast neutrons is much larger than for thermal neutrons, fuel regions in an SFR are normally

well coupled. This is confirmed by the small dominance ratios in Table 3-9. It is desirable to

have a tightly coupled fast reactor from the standpoint of transient response. Since, the effective

delayed neutron fraction for fast reactors is significantly less than for thermal systems; it

becomes advantageous to reduce the complexity of the types of reactivity feedbacks that are









expected to occur in the fast reactor. The flat AHFTR has the largest total core volume

compared to the other two designs and also has a moderated region (reduced mean-free-path).

Nevertheless, the increase in dominance ratio from the reference ABR is not significant.

Table 3-9. Physics comparison for the reference ABR with tall and flat AHFTR (REBUS)
Hoffman [7] ABR (Ref.)* Tall AHFTR Flat AHFTR
Active Driver Height 101.6 101.6 91.6 71.6
Pin Pitch-to-Diameter Ratio 1.1 1.1 1.1 1.1
Total Core Void Worth ($) 6.29 9.16 8.82 8.27
Doppler Coefficient (c/K) -0.11 -0.13 0.06 -0.14
Excess Reactivity (%) -0.06 1.31 1.12 1.70
BOEC k-eff 1.0 1.01325 1.01312 1.01727
Cycle Length (EFPD) 370 322.86 319.90 350.58
Max. Bumup (at.%) 11% 18% 18% 18%
Dominance Ratio** -- 0.3798 0.3813 0.3958
Conversion Ratio (tCR)
Core Conversion Ratio 0.89 0.87 0.84
Target Conversion Ratio -- 0.91 0.90
Inner Core Conversion Ratio -- 1.01 0.97 0.90
Middle Core Conversion Ratio -- 0.85 0.82 0.81
Outer Core Conversion Ratio -- 0.77 0.75 0.77
*The performance characteristics for the reference ABR listed here are not that of the actual
point design of the GNEP ABR design which is currently evolving and not final. See Chapter 2
for an explanation of kinetics calculations. **The dominance ratio is not reported by Hoffman et
al in the report, ANL-AFCI-177.

Fuel Performance Indicators

One issue identified in previous heterogeneous target studies, by Sanda et al, was that the

transmutation of MAs into plutonium can create an undesirable power peaking effect [28]. This

issue was most pronounced for target irradiations in high flux regions in the inner core or for

long "deep-burn" irradiations in the outer core. This work sought to diminish or eliminate any

changes in the target thermal power throughout its life in the core (i.e., from beginning-of-life

(BOL) to end-of-life (EOL)). For axial targets this is especially important because the targets

share the same coolant channel flow orifice as the driver fuel.

SFR fuel assemblies often are, by-and-large, designed with a metal shroud that

encompasses the fuel pins. This shroud prevents cross-flow of sodium to adjacent fuel









assemblies. Therefore, sodium coolant flow in the fuel assembly is controlled on an assembly

basis using an orifice at the bottom of the fuel assembly near where it meets the bottom grid

plate. The shroud prevents differences in pressure losses in each fuel assembly that would

otherwise cause sodium coolant to move laterally in order to equalize the lateral pressure

gradient. Therefore, despite a large radial power gradient (even for the AHFTR) compared to

LWRs, the flow orifices are used to make the coolant outlet temperature at the top of the core

fairly constant.

If there were to be a wide change in thermal power of the target region, the amount of flow

could be sufficient to cool the target at BOL and inadequate for EOL. Also, undesirable power

peaking in the driver fuel could result if the power sharing between driver and targets shifts

significantly throughout the irradiation. Power peaking and shifting was minimized by choosing

a target composition with a small amount of starting plutonium and uranium to ensure a slow net

destruction of Pu-239 over the course of the irradiation. Also, the high importance of the Am-

241 capture cross section in the epithermal flux ensured that neutrons could be equally absorbed

in americium, as could be absorbed into Pu-239 (Table 3-3 and Figure 2-11). This gives the

americium the purpose of a burnable poison in the target region as well as a fertile source for

breeding even-plutonium isotopes. The power produced by these even-plutonium isotopes are

sufficiently suppressed by the capture reactions in Am-241. Table 3-10 evaluates the fuel

performance indicators for the AHFTR.

The peak Linear Heat Generation Rate (LHGR) in the targets was highest for the flat case.

In all cases, the peak LHGR for the entire core occurred in the driver fuel not the targets. This

peak driver LHGR occurs in the innermost row of fuel at the active core's mid-plane. This is

also true of the volume and exposure integrated fission density for the target and driver fuel.









Table 3-10. Fuel performance comparison for the reference ABR with tall and flat AHFTR
ABR (Ref.) Tall AHFTR Flat AHFTR
Active Driver Height 101.6 91.6 71.6
Linear Heat Generation Rate (kW/m)
Peak Driver LHGR (kW/m) 38.1 36.9 35.5
Peak Target LHGR (kW/m) -- 20.1 25.3
Peak Fast Fluence: E>0.1 MeV (1E23 cm-2)
Inner Core 5.72 5.66 5.53
Middle Core 5.58 5.49 5.61
Outer Core 5.05 4.95 5.88
Max Integral Fission Density: (f/cm3)
Driver Fission Density (1E21 f/cm3) 6.96 7.67 7.09
Target Fission Density (1E21 f/cm3) -- 3.64 4.62
Average Discharge Bumup (MWD/kgiHM) 124.4 127.3 134.1

Historically, the fission density has been used as an indicator of irradiation induced

swelling and gas release in metal fuels. It allows equal comparison of fuel performance when

the amount of HM loading is not constant in the comparison. This is the case for the AHFTR

because the metal driver fuel and targets have different zirconium contents in their alloying. For

all cases, the peak fission density is less in the targets than it is in the fuel. Hence, the expected

swelling and gas release, including transmutation helium, should be similar to that of past

experience with metal driver fuel for equivalent fission density. This assumes that the rate of

void formation and interconnected porosity in the target is roughly equivalent for the lower

zirconium content driver alloy. The irradiation induced interconnected porosity of fission gas

bubbles coming together early in the irradiation was one of the significant developments of the

Mark II driver fuel used at the EBR-II [60,61]. The effect slows the rate of swelling by allowing

fission gas to escape to the plenum and increases the time before cladding interaction. Because

of the high LHGR and burnups associated with both metallic and oxide based SFR fuels,

typically 60-80% of all fission gas is released to the plenum [62]. This explains a plenum height

of about 1.5 to two times the height of the core in Figure 1-7, Figure 3-1, and Figure 3-14.









The smaller radial power gradient for the flat case gives its driver fuels the smallest peak

LHGR. The flatter power profile also explains the higher peak fast fluence in the outer core for

the flat case than the other core designs. A more evenly distributed fast flux also results in a

more level fast flux exposure between the inner, middle and outer core regions. Note that the

peak fluence limit for all the core designs mentioned is greater than 4x1023 cm2. This number is

the fast fluence limit assumed in the Hoffman report in consideration to the maximum

displacements-per-atom (dpa) for fast reactor grade steel. HT-9 is a fast reactor grade

martensitic/ferritic steel that was used for cladding and in-core structures for EBR-II and the Fast

Flux Test Facility (FFTF) [63]. For the final down selection discussed in the next section, a limit

of 200 dpa is assumed. The design basis and performance criteria for the AHFTR driver fuel and

targets are discussed in detail in Chapter 6.

Similar to the LHGR distribution, a flatter fluence distribution allows more of the fuel to

be irradiated to a level closer to the 4x 1023 cm2 limit. As discussed in Chapter 2, the

constraining parameter on the fuel cycle for the parametric study was the peak fuel burnup at the

midpoint of the first row of fuel. The limit on fuel burnup as opposed to maximum cladding

damage during the equilibrium cycle convergence process is a limitation of the REBUS code

(Figure 2-1). The next section discusses a final down selection that reduces the cycle length such

that the technical irradiation damage limits for the HT-9 structural components are met.

Final Down-Selection: The AHFTR Design

A final down selection in the AHFTR core design is made based on the transmutation and

reactor physics attributes of the above flattened design case. In order to give a more favorable

fluence and dpa, the cycle length from this case is reduced. In order to give a shorter cycle

length, the pin diameter of the flat design is reduced to increase the TRU enrichment. The only

parameter changed in the fuel assembly design is an increase in the pin pitch-to-diameter ratio









from 1.1 to 1.176. The resulting pin diameter (0.755 cm) is equal to the ABR version with a tCR

of 0.75 which was proposed by Hoffman et al (Table 1-6 and Table 3-2) [7]. This pin diameter

and number of pins per assembly also closely matches the S-PRISM driver fuel assembly design

(Table 1-3) [8]. The slight increase in enrichment also had the benefit of decreasing the

AHFTR's CR (both fCR and tCR) from the case analyzed above.

The core geometry of the flat case was also modified slightly. Due to the low neutron flux

in that part of the target region residing in the outer core (row 7 and 8), the target region above

these assemblies was removed. Without these outer targets, the core design shown in Figure 3-

14 now becomes the core design originally shown in Chapter 1 in Figure 1-7. Additionally, the

removal of these outer core targets resulted in a reduction in the un-transmuted MAs discharged

from the target region and sent to the pyroprocessor. Because of the higher efficiency, a higher

concentration of transmuted plutonium and a smaller concentration of un-transmuted MA in the

mass flow were sent to pyroprocessing from the target region.

Fuel Cycle Performance of the Final AHFTR Design

The generation of plutonium isotopes by the transmutation of MAs can be seen by

comparing the mass flow rate of Am-241 entering and the Pu-238 exiting the target region. The

resulting mass flow rate of plutonium exiting the target region can then be viewed as a supply

rate (via pyroprocessing) of internally provided plutonium to the driver fuel. This internal

plutonium supply rate can then be compared with the external plutonium supply rate from the

aqueous reprocessor.

From Table 3-11, it can be seen that the discharge rate of Am-241 coming out of the target

region (0.0202 g/MWD) is 40% that of the Am-241 charge rate entering the targets (0.0580

g/MWD). Therefore, 60% of the Am-241 mass entering the target region is destroyed by

transmutation or by fission. Also, it is important to note that the target region's Pu-238 discharge










rate (0.0175 g/MWD) is 6.7 times the charge rate going into the targets (0.0026 g/MWD). Also,

it can bee seen from Table 3-10 that the target region's Pu-238 discharge rate (0.0175 g/MWD)

is three times the externally supplied Pu-238 feed rate from the aqueous reprocessing center

(0.0058 g/MWD). This means that the target region provides more than three times the supply of

Pu-238 to the active core via pyroprocessing than the Pu-238 supplied by the "make-up"

transuranic supply from the aqueous SNF reprocessing.

Table 3-11. Final AHFTR mass flow of each isotope entering and exiting the target region and
active core region per MWD
Charge and Discharge Rate of Mass Flows Normalized to grams/MWD
Entering Target Exiting Target* Entering Core** Exiting Core*
Np-237 0.0053 0.0024 0.0264 0.0207
Pu-238 0.0026 0.0175 0.0969 0.0736
Pu-239 0.0480 0.0471 0.9963 0.9201
Pu-240 0.0217 0.0284 0.6255 0.5802
Pu-241 0.0105 0.0086 0.0947 0.0850
Pu-242 0.0062 0.0108 0.1457 0.1312
Am-241 0.0580 0.0202 0.0829 0.0577
Am-243 0.0302 0.0138 0.0761 0.0623
Cm-244 0.0116 0.0188 0.0815 0.0666
Supply Rate from Aqueous Supply Rate from Aqueous
Reprocessing to Target Region Reprocessing to Active Core Region
Np-237 0.0053 0.0033
Pu-238 0.0026 0.0058
Pu-239 0.0480 0.0291
Pu-240 0.0217 0.0169
Pu-241 0.0105 0.0011
Pu-242 0.0062 0.0037
Am-241 0.0580 0.0050
Am-243 0.0302 0.0000
Cm-244 0.0116 -0.0039
*The sum of these two columns is the total mass stream seen by the pyroprocessor. **This
column is the total mass of externally supplied and the internally supplied "pyroprocessed" TRU.

The balance between Am-241 destruction versus Pu-238 generation can also be seen by

examining the isotopic inventory of the core at BOEC and EOEC which is given in Table 3-12.

Also from Table 3-12, it can be seen that the total concentration of MAs per HM is maintained to

be much less than the imposed 5% limit imposed on the active core composition.










Table 3-12. Target and active core region isotopic fuel inventory compared with the ABR at
BOEC and EOEC (REBUS)
Target Region Active Core Region ABR (CR=0.5)
BOEC EOEC BOEC EOEC BOEC EOEC
Heavy Metal Loading (kg) 714 698 12,583 12,380 9,368 9,127
TRU Loading (kg) 244 240 2,791 2,741 3,021 2,913
MA Loading (kg) 115 106 374 360 343 332
Pu Loading (kg) 129 134 2,417 2,381 2,678 2,581
Pu-238 Loading 13 16 110 105 100 97
Pu-239 Loading 61 61 1,236 1,220 1,140 1,088
Am-241 Loading 50 42 92 86 88 83
MA/HM Ratio 16.11% 15.23% 2.97% 2.91% 3.66% 3.64%

The buildup and depletion curve for both the target region and the active core region is

given in Figure 3-17 and Figure 3-18, respectively. These curves give the change in fresh driver

fuel and target composition as a function of irradiation time. Notice the small relative change in

the overall plutonium content of the fuel with irradiation time. This verifies the statement made

in Chapter 2 regarding the lack of spectrum shifting as a result of depletion in the SFR. Because,

the "relative" fuel composition of the core does not change considerable during the course of the

irradiation, the change in resonance and unresolved resonance shielding between various

isotopes is negligible.


3.5E+02

3.0E+02
| Np-237
I-
p 2.5E+02 -Pu-238
Pu-239
2.0E+02 -Pu-240
C. Pu-241
o
S1.5E+02 Pu-242
S-------- Am-241
4--- -
o 1.0E+02 Am-243
Cm-244
5.1E+01

1.0E+00
0 1 2 3 4
Effective Full Power Year

Figure 3-17. Buildup and depletion curve for fresh fuel in the target region of the final AHFTR
design (REBUS)










The same can be said about the target region to a limited extent. Though the change in

MA concentrations is significant as intended, the Pu-239 concentration varies slowly with

irradiation. This slow Pu-239 depletion was intended by design by adding a small amount of

uranium to the metal alloy of the target slug in order to buffer changes in spectrum with

irradiation. The insensitivity of the target region's neutron spectrum to burnup is verified by

calculations performed in Chapter 5.


1E+03


I I- Np-237
I--
i- Pu-238
| 1E+02 -Pu-239
Pu-240
C. Pu-241
Pu-242
S1E+01 Am-241
o Am-243
-Cm-244


1E+00
0 1 2 3 4
Effective Full Power Year

Figure 3-18. Buildup and depletion curve for fresh fuel in the active core region of the final
AHFTR design (REBUS)

Reactor Performance Characteristics of the Final AHFTR Design

A comparison is made between the final AHFTR down-selection and the homogeneous

metal ABR (CR=0.5) design proposed by Hoffman et al [7]. It was the lateral core layout of this

ABR (CR=0.5) design case that was adopted for the ABR reference cases which was used for

comparative analysis of the earlier parametric study. This is also the lateral core layout shown in

Figure 1-5 and Figure 3-1. To achieve the conversion ratio of 0.5, Hoffman used a pin diameter

of 0.623 (p/d=0.623) which is less than the pin diameter of 0.808 (p/d=l.1) which was used in

the previous section to compare the tall and flat versions of the AHFTR. Also, Hoffman









increased the number of fuel pins from 271 (tCR=1.0 and tCR=0.75 designs) to 324 for a tCR of

0.5.

Because of the significant change in fuel pin size and assembly design, the ABR (CR=0.5)

design is somewhat more exotic than the fuel designs used in past reactor designs such as EBR-

II, FFTF and the proposed S-PRISM. This is why the AHFTR fuel assembly design is based on

the ABR with a tCR=0.75 instead of the tCR=0.5 design. Table 3-13 gives a comparison of

various reactors and fuel cycle attributes of both the final AHFTR and two ABR designs.

The AHFTR's transuranic enrichment is just inside the current experience database with

Pu-U-Zr fuel alloys tested during the IFR program in the 1980's [64]. Also because of the

smaller enrichment, the excess reactivity for the AHFTR is less than the ABR. It is important to

note the selection of the pin diameter of 0.755 cm (p/d=1.176) for the AHFTR was not optimized

to make the fluence and dpa limit as close to the 4x1023 cm-2 and 200 dpa limits as possible.

Instead, it was statically selected on the basis of most probable and feasible fuel composition and

assembly design. Extending the cycle length is possible by decreasing the driver fuel

enrichment. However, as discussed in Chapter 2, an increasing conversion ratio results from

decreasing TRU enrichment.

The AHFTR core design was made to have approximately the same cycle length as the

ABR (CR=0.5) case. However, because of the higher neutron leakages from the AHFTR active

core, due to the flattened geometry, a larger uranium fraction in the core is required to increase

internal breeding, which extends the fuel burnup, in order to meet the cycle length requirement.

However, per the discussion Chapter 2, the addition of uranium (i.e., reducing the transuranic

enrichment) does not necessitate the removal of fissile TRU. It does however indicate that the

fuel fraction is increased and the sodium fraction in the core is decreased (i.e., a larger fuel pin










diameter). Therefore, the transuranic enrichment of the AHFTR driver fuel is less than that of

the ABR (CR=0.5), but the transuranic loading is comparable and actually slightly larger (Table

3-12).


Table 3-13. Initial reactor physics and fuel performance comparison between the AHFTR and
ABR design proposed by Hoffman et al
ABR (CR=0.5) ABR (CR=0.75) AHFTR
Fuel and Reactor Dimensions*
Total Core Height 101.600 101.600 91.6(
Active Driver Height 101.600 101.600 71.6(
Pin Pitch-to-Diameter Ratio 1.293 1.176 1.1'
Pin Diameter (cm) 0.623 0.755 0.7
Pins per Assembly 324 271 2'


00
00
76
55
71


Calculated Conversion Ratio
Fissile Conversion Ratio (fCR) 0.64 0.84 0.84
Transuranic Conversion Ratio (tCR) 0.53 0.77 0.72
Excess Reactivity and Cycle Data
Inner Core Enrichment 26.6 16.1 20.77%
Excess Reactivity 2.85% 1.47% 1.23%
Cycle Length (EFPD) 219 232 214
Fuel Assembly Residence Times (cycles)
Inner Core 6 6 6
Middle Core 6 6 6
Outer Core 7 6.5 6
Enrichment Splitting Factor (multiple of Inner Core)
Inner Core 1.00 1.00 1.00
Middle Core 1.25 1.25 1.12
Outer Core 1.50 1.50 1.25
Reactivity Kinetics and Feedbacks
Delayed Neutron Fraction** 0.0035 0.0035 0.0033
Total Core Void Worth ($) 9.17 6.82 8.34
Doppler Coefficient (c/K) -0.08 -0.10 -0.14
Peak Fast Fluence: E>0.1 MeV (1E23 cm-2)
IC: Row 1 Mid-plane 4.00 3.86 3.55
Displacements per Atom in HT9 (dpa)
IC: Row 1 Mid-plane 182.7 176.3 160.4
*These dimensions are also given and discussed in more detail in Table 1-6 and Table 1-7 in
Chapter 1. **The delayed neutron fraction was calculated with MCNP using fission source
libraries with and without delayed neutron fraction data [65].

Because of the higher concentration of uranium in the AHFTR fuel, it should be noted that

the Doppler coefficient of the AHFTR design is slightly more negative than that of the ABR

(CR=0.5). Also, because the transuranic, particularly MA, concentration in the AHFTR driver









fuel is less than that of the homogeneous core, the positive void worth of the AHFTR is slightly

less than that of the homogeneous reference (Table 3-13).

Transmutation Analysis of the Final AHFTR Design

Table 3-14 shows the rates of MA and plutonium consumption for the AHFTR versus the

ABR designs. From Table 3-14, the AHFTR burns less total TRU per fission energy generated

than the ABR (CR=0.5) but roughly twice the amount of MAs. Also the AHFTR bums almost

as many MA as it does plutonium. For comparison, Table 3-14 also shows the amount of TRU

produced for a typical PWR design fuel assembly with the composition given in Table 2-1.

Table 3-14. Mass production and destruction rates per installed megawatt per year
ABR (CR=0.5) ABR (CR=0.75) AHFTR
External Supply Mass Streams Provided by UREX+3 Aqueous Plant*
External Np+Pu Supply (kg/MWY) 0.1596 0.0695 0.0549
External Am+Cm+Bk+Cf Supply (kg/MWY) 0.0092 0.0051 0.0370
External Mass Supply Broken Down by Pu and MA
Pu Consumption (kg/MWY) 0.1516 0.0654 0.0519
MA Consumption (kg/MWY) 0.0172 0.0092 0.0400
UOX PWR** MOX PWR**
Pu Consumption (kg/MWY) -0.0758 0.2033
MA Consumption (kg/MWY) -0.0074 -0.0482
* Mass streams calculated from the equilibrium cycle search calculation performed by REBUS
on the fuel management scenario in the boxed off region of Figure 3-4. **UOX PWR
calculations were performed using TRITON for an initial enrichment of 4.5% and burned for 50
MWD/kg. MOX PWR calculations were performed using TRITON for an initial Np+Pu
concentration of 10% and burned for 50 MWD/kg.

Also shown in Table 3-14, is the amount of plutonium consumed and MAs produced in a

typical PWR MOX fuel assembly. As a side note, it is worth mentioning that the MOX-PWR

fuel assembly has approximately the same net TRU consumption rate as the ABR. In fact, it can

be easily calculated that most PWR MOX fuels have a conversion ratio between approximately

0.6 and 0.8 depending on the TRU enrichment. However, the MOX has a negative net

consumption (i.e., production) of MAs. This emphasizes the fact that a reactor design can have a

low CR with a high net destruction of TRU but still do very little for reducing the isotopes most









important to HLW management and the repository design (i.e., MAs and long lived fission

products)

By comparing the mass flows per unit of installed thermal reactor capacity, the ABR

(CR=0.5) would require approximately one MWth to destroy the transuranics produced by two

MWth ofUOX PWRs. Coincidently, if the fuel design and core layout of the metal ABR

(CR=0.75) is used, the ratio of ABR installed power to PWR installed power is near unity.

The AHFTR also destroys TRU at about the same rate that it can be produced in a PWR

UOX. However, the TRU destroyed by the AHFTR is more than 40% MAs. Hence, the

AHFTR's MA consumption is roughly 2.3 times greater than that of the ABR (CR=0.5) design.

These comparisons demonstrate that the transuranic consumption of the AHFTR is comparable

to the range of ABR options. More importantly, the fissile (i.e., plutonium) requirements of the

AHFTR are significantly less than the ABR. The AHFTR requires only a third of the externally

supplied plutonium than the ABR (CR=0.5) and 70% of the plutonium burned in the ABR

(CR=0.75). Note that the use of the term "externally supplied" can be used synonymously with

the "consumption rate". The fact that the fission cross section for most MAs is much smaller

than that of fissile plutonium suggests that the 40% MA concentration in the AHFTR TRU

requires transmutation before ultimately being destroyed by fission. In order to show that the

MAs are not producing 40% of the fission reactions in the core, the contribution to the reactor

average fission rate for each actinide isotopes is broken down in Table 3-15.

Because 81% of the reactor's thermal power is produced by fissions of plutonium isotopes

but only 14% of the externally supplied HM is actually plutonium, Table 3-15 shows that the

remainder of the plutonium that is undergoing fission is provided by transmutations from the

fertile isotopes: U-238 and the MAs. A closer look at the amount of plutonium supplied to the









core versus actually undergoing fission shows that 92% of the TRU that is undergoing fission at

any given time is plutonium. In actuality, only 56% of the TRU being supplied to the core is

plutonium. One important fact to observe is that the total rate that HM is actually supplied to the

core is 0.00103 kg/MWD (1.03 g/MWD). This is equivalently the amount of HM mass that is

destroyed by fission in order to produce one megawatt of power in one day. Am-241 is supplied

to the core at a rate of 5.8E-05 kg/MWD. The fraction of Am-241 in the HM make-up feed is

5.65%. This is equivalent to saying that 5.65% of all fission reactions in the core are derived

from the introduction of Am-241. However, only 1.01% of the fission power is being produced

by Am-241.

Table 3-15. Concentration of isotopes supplied to replace the mass destroyed by fission
compared to the contribution to fission by each isotope (REBUS)
External Fraction of Fraction of Each Fraction of Fraction of Each
Heavy Metal Each Isotope Isotope per HM of Each Isotope Isotope per TRU
Supply per HM of All Fission per TRU of of All Fission
(Kg/MWD) External Feed Reactions External Feed Reactions
U-234 7.61E-08 0.0074% 0.16%
U-235 3.96E-06 0.3860% 0.48%
U-236 2.34E-06 0.2278% 0.07%
U-238 7.68E-04 74.8441% 11.57%
Np-237 8.38E-06 0.8170% 0.33% 3.33% 0.37%
Pu-238 4.15E-06 0.4047% 3.58% 1.65% 4.09%
Pu-239 7.65E-05 7.4637% 59.36% 30.42% 67.66%
Pu-240 3.46E-05 3.3718% 8.55% 13.74% 9.75%
Pu-241 1.68E-05 1.6334% 7.64% 6.66% 8.71%
Pu-242 9.96E-06 0.9714% 1.43% 3.96% 1.62%
Am-241 5.80E-05 5.6534% 1.01% 23.04% 1.15%
Am-242m 1.97E-07 0.0192% 1.17% 0.08% 1.33%
Am-243 3.02E-05 2.9460% 0.70% 12.01% 0.79%
Cm-242 1.92E-09 0.0002% 0.03% 0.00% 0.03%
Cm-243 9.33E-08 0.0091% 0.14% 0.04% 0.16%
Cm-244 1.16E-05 1.1342% 1.34% 4.62% 1.52%
Cm-245 1.01E-06 0.0989% 2.21% 0.40% 2.52%
Cm-246 1.19E-07 0.0116% 0.16% 0.05% 0.18%
Total 0.00103 100% 99.9% 100% 100%

As an appropriate comparison, for the same analogy applied to the ABR (CR=0.5), the

Am-241 concentration in the HM make-up feed is only 1.5% which is set by the isotopic content









of SNF TRU and the TRU enrichment in the fuel. Also, the fraction of fission power contributed

by Am-241 in the ABR (CR=0.5) is only 0.8%. Hence, about half of the Am-241 externally

supplied to the core is destroyed by fission (as Am-241) and the remainder is transmuted. In

contrast, the AHFTR destroys only 20% of the externally supplied Am-241 through direct fission

of Am-241 atoms. The remaining 80% of the Am-241 external mass supply is transmuted. This

result demonstrates that the AHFTR use MAs for conversion into fissile isotopes. This strategy

is different from the ABR which does not use transmutation targets to precondition the

americium into plutonium.









CHAPTER 4
REACTOR REACTIVITY CONTROL STRATEGY

The standard control rod poison for most SFR designs has historically been natural boron

or boron enriched in the more absorbing isotope: B-10 [29]. The preference of boron over other

poison materials, such as silver or gadolinium, is primarily due to the order of magnitude larger

unresolved resonance cross section of B-10 over other conventional neutron poisons in the fast

spectrum. The necessity to enrich the boron in B-10 is determined by the excess reactivity of the

core. The ABR, like the EBR-II and many other SFR designs, has been proposed to have

enriched B4C as the control rod material. As a side note, FFTF used a sintered boron powder as

the control rod material.

Because the AHFTR has a smaller excess reactivity than the ABR reference case (Table 3-

13), it requires less control rod worth to suppress this reactivity. Therefore, boron enrichment is

not required. As demonstrated by calculations performed by Hoffman et al and Morris et al, the

excess reactivity of an actinide burner SFR (such as the ABR) increases with decreasing

conversion ratio [7,20]. The increase in excess reactivity is a result of the requirement to

enhance neutron losses by leakage in order to attain the low fCR. As mentioned before, for

homogeneous cores, the fCR is closely coupled to the tCR because transuranic breeding is a

function of the neutron balance between parasitic absorption and neutron escape losses. The

AHFTR has less neutron losses because the axial targets lessen their probability of escape.

Tc-99 versus B-10 as a Control Rod Neutron Poison

Though it will be demonstrated in this chapter that B4C would be an acceptable neutron

poison for the AHFTR, the fundamental objective of this work is to destroy TRU, and if possible

other HLW, separated from SNF such as fission products. As identified in the introduction, Tc-

99 has a non-trivial concentration in SNF (approximately equal to the MA concentration). Also,









because of its radiotoxicity, and long half-life (in a geologic time frame), it requires a permanent

disposal solution, such as in a geologic repository. Because of this need to remove Tc-99 from

the fuel cycle, previous studies have focused on ways of transmuting it in a SFR. It is

noteworthy that the UREX+ aqueous reprocessing technology was developed partly for the

purpose of removal of Tc-99 from the repository destined SNF waste stream. For destroying Tc-

99 by transmutation, Yang et al proposed burning metallic Tc-99 in moderated targets within an

accelerator driven SFR system [66]. However, Yang's studies concluded that the transmutation

half-life of Tc-99, even in a moderated target within the SFR, is in the range of decades.

This result precludes the use of Tc-99 as an Integral Burnable Absorber (IBA). IBA's are

burnable poisons that are an integral part of the fuel assembly that can not be removed once it

has been manufactured [3]. An example of an IBA is the Westinghouse Integral Fuel Burnable

Absorber (IFBA) fuel rod which has a coating of zirconium diboride on the fuel pellets [3].

IBAs are used extensively in many PWR fuel designs for the purpose of leveling power peaking

and power shifting that can occur between beginning-of-cycle (BOC) and end-of-cycle (EOC).

Using IBAs allows all the fissile material in an LWR to be consumed at a more level rate which

enhances fuel utilization and extends the fuel's reactivity limited burnup. In order to control pin

and assembly power peaking between BOC and EOC, an IBA must be able to be almost

completely depleted by the end of the first cycle. Otherwise, the presence of the IBA would

penalize the excess reactivity of the core during the next irradiation cycle. Because a Tc-99

target can not be completely depleted during a single irradiation cycle, it is unlikely that it could

be made a feasible IBA.

Despite the difficulty of burning Tc-99 as an IBA, there is still a cost advantage to using

technetium as a neutron poison in control rods. The isotopic enrichment of boron is an









expensive process that could be avoided if Tc-99 were used instead. Since, the UREX+ process

produces a technetium waste stream that is separate from the other fission products; it is a

material that would be readily on hand in the LWR-to-SFR fuel cycle (Figure 1-9). If not used

for the production of control rods, technetium would require geologic disposal as a HLW stream.

If the AHFTR consumes Tc-99 as a service to the fuel cycle, it would be entitled a credit towards

the cost of its fuel purchase as payment for the destruction of Tc-99, which is considered HLW.

A detailed explanation of the economic incentive to destroy HLW is given in Chapter 6.

Tc-99 has a virtually identical unresolved resonance neutron capture cross section to that

ofU-238 in the fast spectrum (Figure 1-6). By comparison, the unresolved resonance capture

cross section of Am-241 has a magnitude three times as great (Figure 4-1) as Tc-99. This fact

explains why Tc-99 is more difficult to transmute than americium in a SFR. Similarly, when

compared to B-10, the Tc-99 capture cross section is also less by approximately a factor of three

(Table 4-1).

It is important to remember that it is B-10's highly absorbing property that makes enriched

B4C the primary candidate for reactivity control in the ABR. However, the excess reactivity

hold down requirement (i.e., excess reactivity) of the AHFTR is less than that of the ABR

(CR=0.5) by roughly a factor of two (Table 3-13). Therefore, enriching the B4C is not as

significant of a prerequisite for the AHFTR. Also it is important to note that metallic Tc-99 has

an atomic density that is roughly 3.5 times greater than that of the B-10 constituent of 90%

enriched B4C (Table 4-1). Therefore, Tc-99 can be considered a candidate for the neutron

poison in the movable control rods in the AHFTR. Because a control rod can be inserted or

withdrawn per the excess reactivity requirements of the core, it is not necessary to completely

burn out the Tc-99 by EOC as it would be if were implemented as an IBA.










Table 4-1. Atomic concentrations of absorber atoms in B4C versus technetium metal
(atom/barnxcm)
Atom Concentration Natural B4C Enriched (90%) B4C Metallic Technetium
B-10 4.39E-03 1.98E-02 --
B-11 1.76E-02 2.20-03 --
C 5.49E-03 5.49E-03
Tc-99 -- 6.69E-02


1E+04
1E+03
1E+02
1E+01
1E+00
S 1E-01
O 1E-02
(n 1E-03


1E-05 -Am-241


1E-07
1E-07 ----------------------------
1E-07 1E-06 1E-05 1E-04 1E-03 1E-02 1E-01 1E+00 1E+01 1E+02
Energy (MeV)

Figure 4-1. ENDF-VI total neutron absorption cross section plots for select AHFTR absorber
materials

Because Tc-99 is not as strongly absorbing as B-10 or Am-241, it is more suited as a gray

absorber material. As mentioned in the Introduction, Messaoudi et al proposed that Tc-99 could

be used as a neutron absorber in dilution pins within the driver fuel assembly for the purpose of

resonance feedback [32]. To minimize the affect of energy shielding with U-238, which could

happen during a sodium void, Kim et al has shown that the poison rods should be concentrated in

their own standalone poison assemblies as opposed to evenly distributed in the driver fuel. In the

case of the study performed by Kim et al, B4C was proposed as an IBA [31]. However, it should

be expected that the same principle applied by Kim et al also applies to control rods. Therefore,

distributing the control rods in evenly distributed clusters throughout the fuel, as is done in

LWRs, is not practical in SFRs. In fact, in most SFR designs control rods are commonly









bunched together into a movable subassembly that travels inside a fixed control assembly that

takes its own discrete location within the core. Therefore, the technetium control rods will be

located in dedicated control assemblies as shown in Figure 1-7.

Control Assembly Design

The AHFTR reactivity control strategy consists of three different types of control rod

mechanisms: (1) a reactivity shim control rod bank, (2) a safety shutdown rod bank and (3) an

ultimate shutdown rod bank for achieving shut down margin. Technetium is selected to be the

absorber material for the shim and safety rod banks. The shim and safety rod clusters operate

independently of each other, but are together co-located in the primary control assembly

locations (Figure 1-7). The ultimate shutdown rods compose a separate control assembly (Figure

1-7). The dimensions of each control system, which will be discussed in the following sections,

are given in Table 4-2.

Table 4-2. Control assembly types and dimensions
Gas
Ultimate Gas
Control Assembly Type Shutdown Primary Control Expansion
Shutdown Module
Module
Control Assembly Pitch (cm) 16.142 16.142 16.142
Assembly Duct Thickness (cm) 0.394 0.394 0.394
Inter Assembly Gap (cm) 0.432 0.432 0.432
Subassembly Duct Thickness (cm) 0.394 -- --
Intra Assembly Gap (cm) 1.000 -- --
Rod Type Shutdown Shim Safety Reflector
Poison Type B4C Tc-99 Tc-99 Void
CRGT Outer Diameter (cm) -- 0.755 0.755 0.755
CRGT Inner Diameter (cm) -- 0.643 0.643 0.643
Control Rod Cladding Outer Diameter (cm) 0.755 0.455 0.455 --
Poison Slug Outer Diameter (cm) 0.643 0.343 0.343 --
Length of Poison Section (cm) 91.100 56.900 91.100 91.100
Number of Rods 271 133 138 271

Traditional SFR Control Assembly Design: B4C Ultimate Shutdown Assembly

It is important to indicate the current state of typical SFR control assembly design such as

that proposed for the ABR before discussing the details of the technetium primary control. To









illuminate this point, the design of the ultimate shutdown control assembly is selected to be

based on that of the ABR. Therefore, the ultimate shutdown control assembly draws its design

from more conventional SFR control designs. Similar to the ABR, a small number of the control

rod assemblies are designated for attaining the ultimate shutdown margin of the core. This

margin is considered necessary to bring the core reactivity down to the point of the cold

shutdown condition [55].

Most of the length of the ultimate shutdown assembly consists of just the outer hexagonal

HT-9 shroud (Figure 4-2). Sodium coolant is allowed to flow through this empty assembly

structure. The ultimate shutdown rods make up a 271 pin bundle array that is located inside a

secondary hexagonal HT-9 shroud which is concentric with the outer assembly shroud. This

secondary shroud and pin bundle comprise a sub-assembly which is allowed to move freely

inside of the outer assembly shroud. This control subassembly is connected to a drive shaft that

extends to control motors outside of the reactor vessel.

Just as with traditional SFR core designs, the neutron poison material for the ultimate

shutdown system is B4C. For the AHFTR, natural boron is used instead of enriched boron. For

the AHFTR primary control rods, technetium metal is adopted for the poison material for its high

atomic density of Tc-99 atoms. Tc-99 metal was also proposed for the technetium transmutation

targets by Yang et al. However, unlike the Tc-99 targets proposed by Yang et al, the primary

control assemblies proposed in this dissertation are not implemented with a moderating material

in their design.

Technetium Based Primary Control Assembly Design

Originally, an internal subassembly design, exactly the same as that proposed for the ABR,

was considered for the AHFTR primary control system. In the ABR, the downward motion of a

primary control hexagonal subassembly into the top of the active core is responsible for











reactivity shim. These rods used for reactivity shim also have a dual role as the ABR's safety


rods. Thus, the ABR uses the same primary control subassembly to provide enough negative


reactivity to shut the core down to a safe condition in case of an accident situation. Therefore,


the ABR uses this subassembly design to provide reactivity shim as well as part of the shutdown


margin.


SDrive Shaft

Concentric Hexagonal
HT-9 Ducts Handling Socket

Rod Bundle Sub-Assembly Wall





"natural boron" D i
Direction of
Sodium Ultimate
outside the Shutdown
rods ISubassembly
Movement (D



Control Assembly Wall 0 -

Region of
the Active 0
Sub-Assembly Fully Withdrawn Core 0
0,


Figure 4-2. Ultimate shutdown control assembly configuration

When a top loaded primary control subassembly is inserted into the top of the AHFTR, the


resulting absorption of the neutron flux in the upper half of the core also decreases the


transmutation efficiency of the axial targets. Therefore, the mechanics of the primary control









assembly design diverges slightly from the ABR. This decision was made to allocate a portion

of the primary control assembly rods, to be used for shim control, to be inserted through the

bottom of the core so that they would not significantly perturb the flux in the axial target region.

Instead of a moveable subassembly filled with control rods, two separate clusters of

control rods are created. Rods from both a shim rod cluster and a separate safety rod cluster are

evenly distributed throughout the primary control assembly. Rods of the same cluster are

attached to one another through a spider-like tie rod subassembly similar to that used in PWRs.

Therefore, there is a spider-like tie rod subassembly for the shim rods and a separate one for

safety rods within each primary control assembly. Because the shim rods are inserted into the

core from the bottom, the shim rod cluster is attached to long tie-rods that are equal to the height

of the core.

To insert shim rods into the core, its spider subassembly is pulled upwards by the action of

the rod drive. When the shim rods are inserted to their desired location, a mechanical clamping

mechanism can be actuated to prevent them from falling out of the core. Alternatively (or

additionally), the shim rod drive mechanism can be designed with sufficient internal friction in

the gearing system that unintentional rod drop out of the core is highly improbable. It is

important to note that the incorporation of both top and bottom loaded control rods does not

increase the physical size of the reactor and associated core externals because neither shim or

safety rods are withdrawn beyond the extent of the gas plenum (top) or the axial reflector

(bottom). Figure 4-3 shows the mechanical details of the primary control assembly design.

The AHFTR primary control assembly consists of a tube bundle of control rod guide tubes

(CRGT) that are wrapped in a hexagonal assembly shroud. These CRGTs are capped at the

bottom to prevent sodium coolant from entering the tube. Half of these CRGTs are allocated for











the shim control rods whereas the other half are allocated for the safety control rods. The shim


and safety rod clusters move independently of each other in the CRGTs through the locomotion


of their respective connecting spider subassemblies. These spider-like connectors are then each


connected to two concentric drive shafts that extend to the control motors outside of the reactor


vessel.


Spider-Like Tie Rod Subassembly Concentric De
Concentric Drive Shafts

CRGT Bundle within the
Primary Control Assembly H
Handling Socket

Control Assembly Wall |



Tie Rod Connecting to Shim Rod t_

Plenum


Tc-99 Safety Rod d

Direction of
Safety Rod
Movement O
Tie Rod Connecting to Shim Rod


Region of O
the Active -
CRG-
Empty CRGT CD




Empty CRGT Direction of
Shim Rod
Movement
Tc-99 Shim Rod O

Region of
the Axial
Sodium Reflector
outside the /
CRGTs

S Coolant Inlet Nozzle

Figure 4-3. Primary control assembly configuration

It is envisioned that a rack-and-pinion style control rod drive motor would be used for the


inner safety rod shaft. A worm-screw type drive mechanism would be used to move the outer









annular shim rod shaft. The safety system would operate on an electric servo with a circuit

interrupter such that interruption of power to the servo would cause the safety rod clusters to fall

into the core. The shim system would only move the shim rod cluster when power was applied

to the worm gear servo. Hence, if power is interrupted to the system, the shim rods would not

fall out of the core, but rather be held in place by the friction in the worm-gear system.

It is envisioned that the CRGT tube would be filled with helium. The lack of sodium flow

through the tube prevents the action of hydraulic frictional drag on the control rod in the tube

that could hinder its motion. This is an important design feature because of the possibility of a

stuck control rod cluster due to a control assembly being axially bowed or warped. The lateral

and axial temperature gradients, which are typical in a SFR core, frequently create expansion and

bending forces that cause assembly bowing. In fact, assembly bowing is sometimes relied upon

as a negative reactivity feedback mechanism related to leakage. This feedback mechanism will

be discussed later in this chapter.

An additional design consideration is attributed to gamma ray heating of structural

components, which is the result of gamma rays that are produced by fission and neutron

activation, imparting kinetic energy into the atoms of structural materials. The number and

dimensions of the CRGT tubes in the primary control assembly are exactly the same as that of

the fuel cladding used in the fuel assemblies. Because the CRGTs receive the same level of

sodium flow that fuel pins receive, it is expected that gamma ray heating of control assembly

structures will be sufficiently dealt with by the coolant outside the tubes.

Gas Expansion Module

A measure of passive safety control is included by incorporating six gas expansion

modules (GEM) in each six covers of the core periphery (Figure 1-7). GEMs are special

assemblies placed at the perimeter of a SFR which are designed to enhance neutron leakage in










the event of the loss of primary coolant flow. GEMs were originally designed for reactors

concepts of the IFR program [67]. A typical GEM design is shown in Figure 4-4.

Handling Socket

Control Assembly Wall







Helium Bubble




0-

Region of
the Active
Core
Sodium Filled into Tube .




Sodium

coolant flow

Sodium
outside the
rods

Coolant Inlet Nozzle
Figure 4-4. Gas expansion module assembly configuration

The lower end of the GEM assembly consists of a lower rod bundle of HT-9 pins that is an

integral component of the core's lower axial reflector. A separate bundle of tubes fills the space

above the axial reflector. These tubes are open at their bottom end to allow sodium coolant to

enter the tube. A helium gas bubble fills the upper end of the tube, which is capped at the top

near the handling socket to prevent the sodium from passing through the tube. In the event that

the bulk coolant circulation pumps fail, the pressure differential between the non-circulating

sodium and the helium gas bubble forces sodium out of the tube causing the gas bubble to











expand. The displacement of the sodium and expansion of the helium creates a voided space in

the radial reflector that extends down into the active core region. The removal of reflection by

the sodium at the core periphery enhances neutron streaming from the active core.

Control Rod Worth

The shim rods are inserted through the bottom of the active core by the withdrawal of the

shim cluster spider subassembly. The purpose of the shim rods is to absorb the excess reactivity

produced by the addition of fresh fuel at BOC. As can be seen in Figure 4-5, the reactor core's

BOEC excess reactivity can be completely suppressed when the rods are inserted 55 cm. This

distance is 75% of the height of the active core.


4.00

3.50

3.00
0
re 2.50

^ 2.00 ---

I 1.50-

O 1.00-
0
0.50

0.00
0 10 20 30 40 50 60
Rod Tip Position from Bottom of Active Core (cm)

Figure 4-5. Shim control rod bank reactivity worth at BOEC

As the fuel is irradiated, the excess reactivity is consumed by the burnup of the fuel.

Therefore, the shim rods are pushed out of the bottom of the core, by the downward motion of

the spider subassembly, until they are fully withdrawn by the end of the irradiation cycle. The

reactivity worth of the shim rod bank at EOEC is shown in Figure 4-6. Figure 4-5 and Figure 4-

6 show the excess reactivity worth of the core at BOC and EOC as a function of the length of










control rod inserted into the core. Notice, that the relative reactivity worth of the shim rod does

not change considerable between BOC and EOC.


0.00

-0.50

-1.00
0
W -1.50 -

o -2.00
03
-2.50 ,

-3.00

-3.50

-4.00
0 10 20 30 40 50 60
Rod Tip Position from Bottom of Active Core (cm)

Figure 4-6. Shim control rod bank reactivity worth at EOEC (DIF3D)

The safety control rods are inserted through the top of the core above the axial targets.

These rods do not move throughout the irradiation. Instead they are raised to the top of the core

during startup where they remain during normal operation. In the event that rapid shutdown is

required, the reactor monitoring and safety systems can trigger the drop of these rods into the

core by interrupting power to the control rod drive servo. The safety rod bank reactivity worth at

BOEC and EOEC is shown in Figure 4-7 and Figure 4-9 respectively.

There is a slight positive reactivity insertion as the rod tip passes through the axial targets.

This positive reactivity feedback is caused by the spectrum hardening of the flux in the targets

due to the epithermal absorbing worth oftechnetium. Figure 4-8 shows the flux spectrum in the

targets when the safety rod tip is inserted into the axial target region. However, it is important to

note that the preferential absorption of neutrons in Tc-99, as the control rods pass through the

targets, can only be observed in the epithermal energy range. This is because, the capture cross

section of Tc-99 falls off sharply at neutron energies above one MeV (Figure 4-1). It is also











important to note that the spectrums plotted in Figure 4-8 are virtually indistinguishable.


Therefore, the spectrum hardening from Tc-99 entering the target region can be considered


negligible. Hence, the initial positive increase in reactivity in Figure 4-7 has a total value of only


ten cents. It is likely that this ten cents can be compensated by allowing for passive safety


feedbacks such as the leakage control afforded by the GEMs.


0 10 20 30 40 50 60 70 80 90
Rod Tip Position from Top of Axial Target (cm)

Figure 4-7. Safety control rod bank reactivity worth at BOEC (DIF3D)

1.0E+15 I


I 1.0E+14



C 1
E
- 1.OE+13
x
LL


1.0E+12 1 1
1E-05 1E-04 1E-03 1E-02 1E-01 1E+00 1E+01
Energy (MeV)

Figure 4-8. Neutron spectrum in the targets as a function of safety rod insertion (DIF3D)










It is important to note that the positive insertion exists only at the BOEC. As the

americium is depleted in the targets, the local spectrum dependence on neutron capture is also

lessened. This is because the reactivity suppression provided by the combination of moderation

and americium capture is lessened. Therefore, the insertion of the rod and resulting absorption

of neutrons in the epithermal range has less impact on reactivity at EOEC. This is reflected in

the safety rod reactivity worth curve at EOEC in Figure 4-9.

As can be seen from Figure 4-7 and Figure 4-10, the safety rod bank has sufficient

reactivity worth by itself to completely shut the reactor down to at least one dollar below critical.

Therefore, even if for any reason, the entire shim rod bank were to be absent from the core

during an emergency shutdown, the shim rod bank could completely take its place and still have

one dollar of shutdown margin remaining.


0.00

-1.00
..,
0 -2.00
LU

-3.00

S-4.00 --

0
-5.00


-6.00
0 10 20 30 40 50 60 70 80 90
Rod Tip Position from Top of Axial Target (cm)

Figure 4-9. Safety control rod bank reactivity worth at EOEC (DIF3D)

Additional shutdown margin is provided by the ultimate shutdown subassembly bank.

Like the safety rods, the ultimate shutdown subassemblies are inserted through the top of the

core. These subassemblies are withdrawn through the top of the core during startup where they

remain during normal operation. At the end of the irradiation cycle, the safety and ultimate









shutdown subassemblies are inserted to provide for the full shutdown margin of the core. Figure

4-10 and Figure 4-11 show the reactivity worth of the ultimate shutdown system for BOEC and

EOEC respectively.


4.00
3.50
3.00
2.50
2.00
1.50
1.00
0.50
0.00


0 10 20 30 40 50 60 70
Rod Tip Position from Top of Axial Target (cm)
Figure 4-10. Ultimate shutdown system worth at BOEC (DIF3D)


80 90


0.00

Z -0.50
LU
-1.00

-1.50

L -2.00


-2.50
o -2.00 --- --- --- --- --- --- ---__ ~ .

-2 .5 0 ---- ---- --- ---- ---- --- ---- ---- ---
0 10 20 30 40 50 60 70 80 90
Rod Tip Position from Top of Axial Target (cm)
Figure 4-11. Ultimate shutdown system worth at EOEC (DIF3D)

It is important to note that the overall reactivity worth of the ultimate shutdown system is

$1.75 despite being composed of natural boron in B4C. The three ultimate shutdown assemblies

are located in the inner most region of the core where the flux is strongest. Therefore, despite


Ii~-









being composed of natural boron, the ultimate shutdown assemblies have the largest average

reactivity worth per assembly. There are three ultimate shutdown assemblies with a combined

worth of $1.75 which gives an average reactivity worth of 580 per assembly. Conversely, the

safety rod system and shim rod systems both only have an average reactivity worth of 220 and

280 per primary control assembly, respectively.

By inspection of Figure 4-7 and 4-10, it is apparent that the total shutdown worth of the

reactor at BOEC is $6.25. The complete shutdown at BOEC accounts for zero net reactivity

when the shim rods are inserted and also $4.50 and $1.75 of total negative reactivity insertion by

the safety and ultimate shutdown systems, respectively. By inspection of Figure 4-6, Figure 4-9

and Figure 4-11, it is apparent that the total shutdown worth of the reactor is $9.75. The

complete shutdown at EOEC accounts for the total negative reactivity inserted by the safety and

ultimate shutdown systems as well as an additional $3.50 by fully inserting the shim rod system.

Thus, if an emergency shutdown is required at EOEC with all shim rods removed, the safety and

ultimate shutdown system can provide $6.25 of negative reactivity.

Reactivity Worth of Boron versus Technetium

It is important to compare the reactivity worth of metallic Tc-99 with the more

conventional SFR reactor poison: B4C. As noted earlier, the atom concentration of metallic Tc-

99 is approximately 3.5 times that of B-10 in enriched B4C. However, the unresolved resonance

absorption cross section of B-10 is roughly three times greater than that of Tc-99 for most of the

fast spectrum. Figure 4-12 contrasts the reactivity worth of the shim rod bank if natural boron or

enriched boron is used in the form of B4C as opposed to metallic Tc-99.

It is apparent from Figure 4-12 that enriched B4C has a greater reactivity worth in the

AHFTR than Tc-99. However, the use of B4C would only reduce the length of control needed to










shut the core down by approximately 10 cm. Therefore, the technetium rod is considered as a

viable alternative to enriched B4C as a control material in the AHFTR.


4.00
-*-Tc-99 metal
3.50 -B4C (natural)
4 -1 B4C (90% enriched)
) 3.00 -

0 2.50

2.00

c 1.50

2 1.00 -----
C,
0.50

0.00
0 10 20 30 40 50 60
Rod Tip Position from Bottom of Active Core (cm)


Figure 4-12. Shim control rod bank reactivity worth at BOEC for metallic Tc-99 compared to
natural and enriched boron in the form of B4C (DIF3D)

Top versus Bottom Inserted Shim Rods

One of the key design features of the primary control assembly was to have the shim rods

inserted through the bottom of the core ensure that they did not steal neutrons away from the

axial targets. For the sake of comparison, the volume average radial flux profile of the active

core and targets has been plotted for the case where all shim rods are inserted through the bottom

of the core versus all shim rods being inserted through the top of the core. These flux plots are

given in Figure 4-13.

Notice that the flux in the targets is slightly less for the top inserted shim rods than it is for

the bottom inserted shim rods. The reduction of flux indicates that the transmutation efficiency

of the axial targets would be reduced if the shim rods were inserted through the top of the core.

Unlike the safety or ultimate shutdown rods, the shim rods require being in the core at all times.

Hence, the neutron absorbing effect of a top inserted control rod would have a constant impact










on the targets region's transmutation efficiency, regardless of its gradual removal from the active

core. Thus it is more feasible to have a bottom inserted shim rod cluster design as was discussed

in the primary control assembly proposed for the AHFTR.


4.5E+15

4.0E+15

3.5E+15

n 3.0E+15

E 2.5E+15

x 2.0E+15 .

I 1.5E+15
I- Target-Bottom Loaded Shim Rod
1.0E+15 -r Active Core-Bottom Loaded Shim Rod
5.OE+14 --Target-Top Loaded Shim Rod
"- Active Core-Top Loaded Shim Rod
0.0E+00
0 20 40 60 80 100 120 140
Radial Distance from Core Center (cm)
Figure 4-13. BOEC radial flux distributions of the active core (volume averaged over core
height) and axial target fluxes for shim rods being inserted through: the top or the
bottom of the core (DIF3D)

Axial Power Tilt and Shim Rod Insertion

Figure 4-14 shows the axial power distribution for various shim rod positions in the

AHFTR core. As discussed previously, the large mean-free-path of neutrons in the core

combined with the relatively gray neutron absorbing worth of technetium reduces the affect of

localized flux depressions on the overall core flux distribution. Therefore, the axial flux

distribution is virtually identical regardless of the length of control rod inserted into the active

core region.

Other Reactivity Feedbacks

As indicated previously, the AHFTR, like the ABR and many other SFR designs has a

positive void coefficient. The positive reactivity feedback is a direct consequence of neutron











energy spectrum hardening resulting from the loss of the slight moderation provided by the

sodium coolant. When the spectrum hardens, the lack of down scattering causes the number of

neutrons above the fission threshold of fertile isotopes (U-238 and MAs) to increase. An

increase in above-threshold fissions causes an increase in the neutron multiplication contribution

of these fertile isotopes. Figure 4-15 gives the neutron spectrum of the AHFTR inner core region

during normal steady state operation and a scenario where all of the sodium coolant is voided

from the core.


130%

120%

110%

100%

S90%

to 80% -o-6.7125
eO -13.425
= 70% 20.1375
7i S --26.85
60% --33.5625
-+40.275
50% 46.9875
53.7
40%
0 10 20 30 40 50 60 70 80 90 100
Axial Distance from Bottom of Active Core (cm)
Figure 4-14. Axial power distribution for increasing shim rod length into the active core (inner
enrichment zone) region (DIF3D)

This spectrum hardening can be unfavorable if there is no other competing feedback

mechanism that can negate the void induced positive reactivity insertion. For SFRs, some

negative reactivity feedback comes in the form of Doppler resonance broadening of capture cross

sections as the fuel temperature increases. Additionally, because SFRs typically exhibit a high

degree of leakage, they can rely on feedback mechanisms that increase leakage as the fuel and

structural materials increase in temperature. The negative reactivity feedback caused by this fuel

expansion is most pronounced in metallic fuels and was a key control aspect of EBR-II [35].










1.2E+15


1.0E+15


8.OE+14


.l

I 4.0E+14
X
UJ-
2.0E+14


O.OE+O0 ** -* *-
1E-05 1E-04 1E-03 1E-02 1E-01 1E+00 1E+01
Energy (MeV)
Figure 4-15. Neutron spectrum at BOEC of the inner core region for steady state operation
versus a complete loss of sodium coolant (DIF3D)

The void and Doppler reactivity worth for the AHFTR are given in Table 4-3. Also given

is the negative reactivity provided by leakage feedback when fuel expansion changes the radius

and height of the active core. Leakage feedback is also provided by the GEM. The void effect

of the helium bubble expansion in the GEM is also given in the table.

As can be seen from Table 4-3, if the sodium coolant is completely lost from the active

core (coolant remains in the gas plenum, axial and radial reflectors), a positive reactivity

insertion of about $8.60 occurs. However, if thermal expansion effects cause both the height and

diameter of the core to increase by only 2.5%, the core will lose approximately six dollars worth

of reactivity. If both the height and diameter increase by 5.0%, the expansion would remove

approximately $13.30, which is sufficient reactivity to completely negate the positive reactivity

insertion by the void. Also, an additional dollar of negative reactivity can be supplied by the

combined contribution of the GEM and Doppler feedbacks. Assuming a loss of primary coolant

flow, the GEMs provides can provide $0.52 of negative reactivity. The Doppler feedback can

provide an additional $0.0014 per degree increase in fuel temperature.









Table 4-3. Core reactivity worth for independently separate reactivity feedback effects
Difference
BOEC Nr EOEC Difference from Normal
from Normal
Normal Steady State $
$3.79 -- $0.0 --
(all rods out)
All Coolant Voided
$12.13 -$8.34 $8.60 -$8.60
(all rods out)
Doppler (+100K) $3.65 $0.14 -$0.15 $0.15
(all rods out)
Normal Steady State $0.0 n/a n/a
$0.0 -- n/a n/a
(shim rods in)
All Coolant Voided
Al C t V d $8.62 -8.62 n/a n/a
(shim rods in)
Doppler (+100K) -$0.14 $0.14 n/a n/a
(shim rods in)
Values Below This Point Taken with All Rods Out
GEM Voided $3.27 $0.52 -$0.50 $0.50
Radial Expansion (2.5%) -$0.35 $4.14 -$4.17 $4.17
Radial Expansion (5.0%) -$5.01 $8.80 -$8.88 $8.88
Axial Expansion (2.5%) $1.82 $1.97 -$2.01 $2.01
Axial Expansion (5.0%) -$0.71 $4.50 -$4.58 $4.58

Axial Fuel Expansion

During the IFR program, axial fuel expansion of Pu-U-Zr (and U-Zr) EBR-II fuels were

tested at the Transient Reactor Test (TREAT) facility. The results of these tests are summarized

by Rhodes et al [68]. For these tests, fuel pins were first irradiated to varying burnups in the

EBR-II. These irradiated fuel pins were then subjected to transient overpower (TOP) at TREAT.

These power levels were in excess of four times their nominal power level in EBR-II. All

overpower transients caused extensive melting in the test fuel which amounted to one-half of the

original fuel inventory. The pre-failure data from these tests indicated that the metallic fuel slugs

expanded axially in excess of the 1% attributed to purely thermal expansion. Rhodes et al

attributed this additional expansion to swelling caused by dissolved fission product gasses being

liberated by the partial melting of the metallic fuel. Post test examination of the fuel pins that

did not exhibit cladding rupture showed large bubbles in the fuel that formed from fission gasses

which coalesced into voids as the fuel partially melted and then expanded. It is this expansion of









coalesced gas bubbles that caused the fuel to expand during the TOP. The level of axial

expansion was later quantified as a function of:

* The amount of molten fuel created during the overpower
* The concentration of fission gas made available when the fuel melted
* The initial bubble size due to surface tension effects
* The pressure of gas in the fuel pin gas plenum resisting the expansion

The results obtained by Rhodes et al showed that one of the test pins achieved an axial

elongation of 3.7%. Also, the axial expansion could be well predicted using a simplified model

where only gas trapped in the bubbles was available for expansion.

The results of these tests are promising from the standpoint of reactor control. If these

models can be developed to a higher fidelity, they might be used to optimize the axial expansion

to meet the TOP scenarios for a given SFR design with a safety margin to cladding breach.

Thermal expansion and fuel assembly bowing was used to validate the safety case for

future ALMRs during the IFR program. This passive feedback capability was demonstrated in

the Shutdown and Heat Removal Test (SHRT) conducted at EBR-II [69,70]. The SHRT tests

demonstrated passive reactor shut down using fuel expansion and natural circulation during two

LOCA scenarios: an Unprotected Loss of forced circulation Flow (ULOF) as well as an

Unprotected Loss of Heat Sink (LOHS) [71,72].

Radial Fuel Bowing

Radial expansion of the core through fuel assembly bowing is a phenomenon that is highly

coupled between the core physics of the TOP reactivity insertion (either by LOCA or other

accident initiators) and the thermal-hydraulic and thermo-mechanical behavior of the sodium

coolant and fuel assemblies. The bowing of fuel assemblies in the radial direction is caused by

lateral temperature gradients across the cross-sectional area of the fuel assembly. The

temperature gradient across the fuel assembly is related to the power gradient in the reactor









during the transient. The temperature gradient causes differences in the axial thermal expansion

of the assembly's HT-9 duct wall from one side of the fuel assembly to the other. It is this

difference that induces a bending moment on the length of the fuel assembly. Assuming the core

temperature gradient during the transient is positive for decreasing radius (core hotter at the

center), the bending moment pushes the fuel assemblies outwards. At a fundamental level, the

amount of bowing created by these temperature gradients is roughly approximated by treating

the assembly as a tubular beam with an applied bending moment. Varying complexities of

calculations of this type have been implemented in computer codes since the 1970's to try to

accurately predict fuel bowing. However, these codes to date have all assumed steady state

conditions and were typically benchmarked against critical pile simulations (also at steady-state)

using perturbation theory [73,74,75,76].

Fuel bowing was used to enhance the passive safety of the EBR-II. The EBR-II fuel

assembly was located in the core by two closely placed lower grid plates. To minimize radial

expansion below the mid-plane of the core, a metal button was incorporated onto the surface of

each six of the hexagonal sides approximately half-way up the length of the fuel assembly.

These buttons held the fuel assemblies tightly in place below the mid-plane. EBR-II did not

have an upper grid plate. The combination of the buttons and the two lower grid plates

effectively cantilevered the fuel assembly below the core mid-plane. If a strong bending moment

were to be induced on the fuel assembly, the lack of an upper grid plate would allow the fuel to

blossom (analogous to a flower) at the top of the core. This blossoming effect was an integral

feature of the passive safety attributes of the EBR-II. The EBR-II passive reactivity feedback by

expansion forces was demonstrated by the SHRT tests. Similar negative reactivity feedbacks for









fuel assembly bowing have also been verified in critical pile tests performed at steady state

conditions by various groups [73].

Despite the success of these tests, the transients initiated were demonstrated on fairly long

time periods compared to the neutron lifetime in a fast reactor. Due to the small delayed neutron

fraction in fast reactors compared to thermal reactors, the neutron lifetime is much shorter and is

in the range of tenths of micro-seconds. If a void were instantaneously introduced into the core

for some hypothetical situation, the reactivity insertion and resulting TOP could possibly occur

on a timescale much faster than fuel bowing could occur.

However, it is important to realize that the total core void worth calculations in Table 4-3

were performed assuming that sodium was only voided in the core and not the surrounding

reflectors. The likelihood of such a hypothetical reactivity insertion from void formation in the

fueled region alone may not be probable. If this is the case, this hypothetical void insertion may

be considered in the realm of core disruptive accidents (CDA) and may or may not require a

robust reactor design for compensating for this remote possibility [77]. If such a robust design is

required, then it is outside the realm of this dissertation work due to the fact that the ABR

designs currently under consideration also have large positive void induced reactivity insertions

for this accident event (Table 3-13).

It should be briefly mentioned that the coolant outlet temperature of the current ABR

design is approximately 780 K. The boiling point of sodium (at atmospheric pressure) is 1156 K.

If a pool type reactor with the bulk sodium at or just above atmospheric pressure is assumed, the

temperature margin before sodium boiling occurs is roughly 370K. Sodium has a large thermal

conductivity (60 W/m-K) and fairly large heat capacity (1,250 J/kg-K). Therefore, a pool type

design could offer a large thermal momentum which minimizes the rate of increase in the core









coolant temperature. If a loop type design is used, the pressurization of the sodium could

increase the margin to boiling. However, a loop design would possibly have to sacrifice the

large thermal momentum offered by a large sodium pool. The decision for pool or loop type

SFR reactor cooling strategies is currently the subject of much debate in the SFR community.

Both methods have their own pros and cons. The choice for a pool or loop type reactor does not

play a role in the transmutation performance of the reactor and does not enter into the

calculations of this dissertation. For the purpose of discussion in the fuel performance discussion

in Chapter 6, a pool type design is assumed.

Technetium Transmutation Rate

When Tc-99 absorbs a neutron, it transmutes into Tc-100 which decays with a half-life of

15.86 seconds by beta particle emission into a stable atom of Ru-100, ruthenium. Because, Ru-

100 is followed by two more stable isotopes of ruthenium on the chart of the nuclides, it is

unlikely that successive neutron capture from the original Tc-99 transmutation will produce a

significant mass of radioactive material. Therefore, it can be assumed that Tc-99 transmutation

by neutron capture will remove the radiotoxicity associated with Tc-99 from the fuel cycle.

Hence, the technetium shim rod, though not a true transmutation target such as that proposed by

Yang et al, performs a secondary purpose as a Tc-99 burner.

Given that some Tc-99 material is inserted into the reactor to some extent in the form of

the shim rods, it is expected that some Tc-99 will be destroyed as a function of the amount of rod

that is inserted into the core. To evaluate the rate at which the shim rod Tc-99 is destroyed, a

REBUS calculation was performed for a single cycle (non-equilibrium depletion mode) with the

shim rods fully inserted into the core throughout the entire irradiation. Using this simulation, the

overall depletion rate of Tc-99 was computed.









This depletion rate was then divided by the entire length of the shim rod to find the average

Tc-99 depletion rate per length of rod. Thus, the average Tc-99 depletion rate per length of

control rod is: 1.4E-6 kg/MWD per cm or shim rod. If it can be assumed that the amount of

excess reactivity is linear as a function of burnup, than it can also be assumed that the rate that

the shim rod is removed is also a linear function of irradiation time. For a shim rod length of 55

cm and a cycle length of 214 days, the shim rod removal rate is: 0.26 cm/EFPD. Integrating the

Tc-99 depletion rate per unit length of rod over the entire cycle length then gives the total Tc-99

consumption rate of the core accounting for the shim rod movement. Thus, the Tc-99

consumption rate of the shim rods is: 0.0143/MWY. A summary of results on Tc-99

consumption by the AHFTR is given in Table 4-4.

Table 4-4. Summary of Tc-99 consumption by the AHFTR
Tc-99 Depletion Rate per Unit Length of Shim Rods (kg/MWY/cm) 1.4E-6
Shim Rod Removal Rate (cm/EFPD) 0.26
Cycle Length (EFPD) 214
Tc-99 Consumption Rate by Shim Rods (kg/MWY) 0.0143
Tc-99 Production Rate by Fissions in the Fuel (kg/MWY) 0.0053
Tc-99 Net Consumption Rate by the AHFTR (kg/MWY) 0.0089
Tc-99 Production Rate of reference PWR fuel (kg/MWY) 0.0074

Using the TRITON calculation of the reference UOX PWR fuel assembly, which was used

to generate the isotopic vector of the ABR and AHFTR external feed, the rate of Tc-99 produced

by the LWR fleet was found to be: 0.0074 kg/MWY. Therefore, the AHFTR shim rods can

consume approximately 1.9 times more Tc-99 than that produced by a PWR for equal amounts

of energy produced by each reactor type. In other words, the shim rods associated with one

megawatt of installed AHFTR capacity can consume the Tc-99 produced by 1.9 megawatts of

installed PWR capacity.

Despite the attractiveness of burning Tc-99 in the shim rod system of the AHFTR, these

calculations up till now have not reflect the Tc-99 produced by the ABR or the AHFTR in the









fast reactor fleet. This is partly because, unlike the UREX+ aqueous reprocessing technology,

pyroprocessing does not currently offer a separation strategy for Tc-99.

However, if Tc-99 separation were possible with pyroprocessing the net Tc-99

consumption of the entire AHFTR core can be found. The AHFTR fuel produces approximately

0.0053 kg/MWY of its own Tc-99 through fissions of fuel atoms. Therefore, the net destruction

of Tc-99 by the AHFTR is only: 0.0143 kg/MWY 0.0053 kg/MWY = 0.0089 kg/MWY.

Hence, if it were possible to recover the Tc-99 produced by fission, the support ratio of PWRs to

AHFTRs would be reduced to 1.2.









CHAPTER 5
DIFFUSION VERSUS TRANSPORT BENCHMARKS

A benchmarking effort is conducted to evaluate the validity of the diffusion theory

calculation used for the core physics and transport analysis. The Variational Anisotropic Nodal

Transport (VARIANT) code is an "add on" module that comes with the current publicly

available version of the DIF3D code [78]. The VARIANT spherical harmonics treatment can be

applied to the hexagonal-z nodal discretization which describes the AHFTR geometry.

VARIANT can be invoked with minimal changes to the DIF3D input. Therefore, the

equilibrium cycle calculation performed by REBUS is re-evaluated using the VARIANT option

of the DIF3D code.

Due to the inherent archaic memory allocation structures available in the DIF3D/REBUS

system, only three moments of the angular flux (P3) were calculated. Also, because of these

memory restrictions, the number of energy groups was reduced from 33 to eight. Therefore, the

amount of energy information is reduced in lieu of increased angular information. Because of

these limitations, the MCNP code was used to evaluate the axial and radial spatial flux profiles

as well as the driver and target energy spectrums. A short FORTRAN processing code was

written for copying the batched homogenized isotopic number densities from each core region in

DIF3D/REBUS model into the MCNP model. The corresponding hexagonal-z nodal geometry

scheme used in the DIF3D model was also preserved in the MCNP model.

Also tested in this chapter is the spatial and energy shielding treatment used in the MC2-2

cross section collapsing algorithm. The collapsed cross sections (group constants) used for the

deterministic calculations are generated by completely homogenizing the pin and fuel assembly

geometry detail into a representative zero-dimensional infinitely dilute mixture. One of these

mixtures is defined for every region in the core sharing a similar isotopic composition and,









hence, neutron spectrum. These regions include the: inner, middle, outer enrichment zones and

the axial target region above the active core. Also, a mixture is created for the shield and

reflector compositions. Using a critical buckling search calculation, the group constant library is

generated for each zero-dimensional mixture in separate calculations performed by MC2-2.

Therefore, the spatial shielding and the region-to-region neutron shadowing effects in the reactor

core are not accounted for during the cross section generation. This task would generally be

performed in a lattice calculation at the pin or assembly level for most thermal spectrum

applications before using the group constant set in a core simulator.

For fast reactors, the mean-free-path is significantly longer than in thermal spectrums

allowing for the homogenization over the fuel assembly. However, the axial targets are slightly

moderated to an epithermal or hard-epithermal spectrum. Thus, it is important to test whether or

not spatial shielding occurs within the target fuel rod placed adjacent to a moderator rod. A unit

cell calculation of one zirconium hydride rod surrounded by six targets with reflective boundary

conditions is modeled using the MCNP code. The flux for one of these six targets is evaluated as

a function of radius going into the fuel slug. This radial distribution is determined by dividing

the fuel slug into five equal volume zones and tallying the neutron flux entering each of these

zones.

The region-to-region shadowing effect is also evaluated using the MCNP code. The six

pins from the cell calculation are homogenized into an annular region of equal volume

surrounding the zirconium hydride pin. Shadowing is determined by tallying the neutron flux as

a function of radial distance away from the center of the annulus.

DIF3D, VARIANT and MCNP Methods

VARIANT solves the multigroup steady-state neutron diffusion and transport equations in

two and three dimensional Cartesian and hexagonal geometries using variational nodal methods.









Anisotropic scattering is treated in these calculations. However, for the coupling of VARIANT

with REBUS, it was found out that the REBUS code is ill-equipped to accept the higher order

scattering matrices in the microscopic cross section data file format (ISOTXS) provided by the

MC2-2 code. ISOTXS is a binary data file for transferring microscopic cross section data

between codes by different authors which uses the standardized format adopted by the

Committee on Computer Code Coordination (CCCC). The difficulty of REBUS to process the

higher order scattering data in ISOTXS is caused by the method in which REBUS homogenizes

fuel isotopic number densities over each region of the core (i.e., Inner, Middle, Outer Core and

Targets).

Recalling Figure 2-1, REBUS homogenizes fuel compositions of all fuel batches

independent of depletion stage within a user defined "same spectrum" region. This new region

averaged number density is then used in the physics calculation to generate the flux in that

region. This flux is then multiplied by the original un-homogenized number densities at each

region to get the reaction rate at that location in the core. These reaction rates are then applied to

the original isotopic compositions of each fuel batch at each stage within that region to perform

the depletion process.

Therefore, the exact position of each individual fuel assembly within the core is not

tracked. It is only necessary to track the mass of each batch as it moves from region to region as

a function of cycle or stage number of its depletion. For many fast reactor core designs (ABR,

AHFTR, etc.), the fuel is not shuffled. Instead, a fuel assembly typically resides in the same

location in which it was originally loaded from BOL through EOL. Because the fuel is not

shuffled, the core can be divided into separate enrichment zones throughout the core.









For this reason, even though the MC2-2 code provides the higher order scattering data,

REBUS does not have the data storage structures to receive it. The creators of VARIANT made

substantial modifications in DIF3D to handle the input of anisotropic cross sections. However,

the VARIANT module expects this data to be input in macroscopic data file format (COMPXS).

Very little formatting changes were made for the absorption and fission "principal" cross

sections, which can be in ISOTXS or COMPXS form. Nevertheless, REBUS can only accept the

ISOTXS format due to the homogenization routine. Logistically, it is impossible to input

COMPXS formatted data into REBUS because the nature of the equilibrium search manipulates

the atom densities of the fresh fuel composition (i.e., TRU enrichment) in order to meet the

constraints of equilibrium cycle length and burnup. Because of these limitations, the higher

order approximation of the scattering cross section had to be dropped for the coupling between

VARIANT with REBUS.

Flux Spectrum Analysis

A P3 approximation of the flux and current are performed by VARIANT but with zero

order scattering in the PN equations. To this end, MCNP is used to compare the BOEC radial

and axial flux distributions as well as neutron energy spectrum with the DIF3D/REBUS and

VARIANT/REBUS results. Figure 5-1 shows the flux spectrum for the AHFTR mid-plane at the

inner most row of fuel.

It is important to note the close agreement between the MCNP and the DIF3D spectrum.

The major observable difference between the two curves is the general magnitude of the flux

which can be seen to be centered on 0.25 MeV. The flux depression at 0.25 MeV is a result of

several well resolved but overlapping sodium resonances in that energy range. Note the flux

depression at 30 keV which corresponds to the well resolved iron capture resonance at that

energy. The VARIANT spectrum does not capture the true shape of the energy distribution due











to the eight energy group resolution. Figure 5-2 shows the flux spectrum for the target zone for


the innermost row of fuel.


1.4E+15


1.2E+15


1.0E+15


8.0E+14


6.0E+14


4.0E+14


2.0E+14


O.OE+00 -
1E-05


1E-01 1E+00 1E+01


Figure 5-1. Neutron energy spectrum for the AHFTR at the core mid-plane for the inner most
row of driver fuel (taken at BOEC)


1E-04 1E-03 1E-02
Energy (MeV)


1E-01 1E+00 1E+01


Figure 5-2. Neutron energy spectrum for the AHFTR at the target region located directly above
the inner most row of driver fuel (taken at BOEC)


1E-04 1E-03 1E-02
Energy (MeV)


6.0E+14


5.0E+14


4.0E+14


3.0E+14


2.0E+14


1.0E+14


O.OE+00
1E-05









Here again, the DIF3D spectrum shows good agreement with the MCNP spectrum. In fact,

the difference in overall magnitude is less than in the driver fuel. Note that because of the

epithermal spectrum, the sodium resonance at 3 keV is much more pronounced than in Figure 5-

1. However, much of the flux is still in the epithermal-to-fast energy range such that heavy

metal resonances dominating from 0.1 eV through 100 eV have very little importance. For a

later discussion, it appears that the resonance shielding of the lighter metal elements (comprising

the structure and coolant), with resolved resonances between 1 keV and 1 MeV, have the most

affect on the flux.

The difference in total flux magnitude between the three calculation methods may be

attributed to an increasing accuracy of the angular treatment provided by VARIANT and

REBUS. VARIANT provides a better approximation of the angular distribution of the flux and

transport cross section. MCNP provides full anisotropic treatment of scattering as well as the

flux gradient as opposed to the diffusion approximation made by DIF3D or the isotropic

scattering assumption made with VARIANT.

Spatial Flux Analysis

The diffusion approximation under predicts the curvature of the axial and radial flux

gradient [79]. This is observed in the axial and radial flux profile for the AHFTR. The axial flux

profile for the inner most row of fuel is given in Figure 5-3. The error bars for two standard

deviations of the flux in the MCNP calculation are shown in the plot but are of the same size as

the plotted data point. Notice the peak flux occurring at the mid-plane is less for the DIF3D

calculation than the other two curves. The total flux calculated by VARIANT is actually much

closer to the MCNP result than the DIF3D result. This indicates that the fewer energy groups

used in the VARIANT calculation was a valid simplification even though the fine spectral












resolution of the flux energy spectrum is lost. Figure 5-4 shows the radial distribution of the flux


at the core mid-plane.


5.0E+15


4.5E+15


w 4.0E+15

E
.3.5E+15
x
LL.
S3.0E+15
0
F-
2.5E+15


2.0E+15




Figure 5-3. Comparison o


5.0E+15


4.5E+15


S4.0E+15
in

3.5E+15
E
o
3.0E+15
LIL

0 2.5E+15
I-

2.0E+15


1.5E+15


o DIF3D
x VARIANT
O MCNP


0 10 20 30 40 50 60 70 80 9(
Axial Distance from Core Bottom (cm)

fthe total BOEC flux axial profile for the inner most r

















o DIF3D
X VARIANT
oMCNP


0 20 40 60 80 100 120 140
Radial Distance from Core Center (cm)

Figure 5-4. Comparison of the total BOEC flux radial profile calculated at the core mid-plane


The differences in the mid-plane radial flux distribution between calculation methods are


negligible for the inner regions of the core. However, for increasing radial and axial distance


from the center, some differences in the flux gradient become observable. The target region


radial distribution (Figure 5-5) shows notably more difference for regions nearest the center of


0



ow of fuel











the core. These differences may be explained by the differences in angular treatment between


the three methods. It is interesting to note that the target flux calculated by VARIANT is greater


than the MCNP calculation for the inner rows of targets. This may be attributable to the lack of


anisotropic expansion of neutron scattering used for this calculation. The anisotropic scattering


off of target hydrogen could cause more neutrons to pass through the target without being


absorbed. Not having this anisotropic effect in the VARIANT calculation causes some of the


neutrons to be reflected back into the target and active core regions.


Radial Flux Distribution in the Targets

2.5E+15
2.4E+15 -
X X
2.3E+15 -; _-----
I X
2.2E+15
<' 2.1E+15-
E x
2. 2.0E+15 -
1.9E+15
LL.
i 1.8E+15
o DIF3D
I- 1.7E+15 --n
x VARIANT
1.6E+15 o MCNP-
1.5E+15
0 20 40 60 80 100
Radial Distance from Core Center (cm)
Figure 5-5. Comparison of the total BOEC flux radial profile calculated for the axial target
region

The DIF3D curve under predicts the flux given by both MCNP and VARIANT.


Therefore, it is expected that the MA destruction rate should be on the conservative side of the


full transport flux evaluation. This result can be seen in Table 5-1. As expected, the rate of


transmutation is higher for VARIANT than it is for the DIF3D calculation. However, all of the


values in the comparisons made by Table 5-1 are in close agreement with each other. Therefore,


the use of the diffusion theory is deemed as an acceptable solution to the transport equation for


fast reactor core and epithermal target physics analysis. It is important to make the connection










that there is very little impact on the fuel cycle performance evaluation when using diffusion or

transport theory to calculate the AHFTR's reactor and transmutation performance.

Table 5-1. Select reactor parameters for the AHFTR given by DIF3D and VARIANT.
DIF3D VARIANT
BOEC k-eff 1.012377 1.013683
tCR 0.718088 0.715677
Cycle Length (EFPD) 213.957 212.078
Inner Core Enrichment 20.77% 20.70%
Am-241 Transmutation Half-Life (EFPY) 2.31 2.26
Am-241 Consumption Rate (kg/EFPY) 2.12E+01 2.14E+01
MA Consumption Rate (kg/EFPY) 4.00E+01 4.04E+01

Differences between BOEC and EOEC Fluxes

As discussed in the Chapter 2, the relative or percent change in the initial fissile

concentration in the fuel as a function of burnup is small. This is the fissile requirement of the

core is high due to the high leakage (i.e., geometric buckling). Therefore, it should be expected

that the neutron spectrum in the core is fairly insensitive to the fissile atom depletion. Figure 5-

6, shows the neutron spectrum at the mid-plane and target level in the first row of fuel in the

AHFTR.

1.4E+15
BOEC Active Core Region
1.2E+15 -e- EOEC Active Core Region f',
5 -*- BOEC Target Region
M1.OE+15 -- EOEC Target Region

8.0E+14

6.0E+14
E
x 4.0E+14

2.OE+14

O.OE+OO
1E-05 1E-04 1E-03 1E-02 1E-01 1E+00 1E+01
Energy (MeV)
Figure 5-6. Comparison between the BOEC and EOEC neutron spectrums for the inner most
row of fuel (DIF3D)











Notice that there is a slight change in the magnitude of the flux between BOEC and EOEC.

This difference is attributed to the slight change in the average fissile concentration in the core

between BOEC and EOEC from the fuel burnup. It should be expected that the average flux in

the core should generally increase as the fissile material is being decreased by depletion due to

the fact that generally number density and flux are inversely proportional to each other. To

illustrate this point, the radial flux distribution at the core mid-plane is plotted for both BOEC

and EOEC in Figure 5-7. The axial flux profile is plotted for the innermost row of fuel for both

BOEC and EOEC in Figure 5-8.

-*- BOEC Active Core Mid-plane -e- EOEC Active Core Mid-plane
-- BOEC Target Region --- EOEC Target Region
5.0E+15

4.5E+15 -

< 4.0E+15

S3.5E+15

x 3.0E+15
LJ.
o 2.5E+15
I- *-- ---.t-----
2.0E+15

1.5E+15
0 20 40 60 80 100 120 140
Radial Distance from Core Center (cm)
Figure 5-7. Comparison between the BOEC and EOEC radial flux profiles (DIF3D)

Depletion Test

To check the affect of the difference in flux values on the results of the REBUS depletion

calculation, a benchmark with MONTEBURNS was performed to compare the k-eff as a

function of irradiation time.

In equilibrium-mode, the REBUS code performs a minimum of two DIF3D calculations:

BOEC and at EOEC. The BOEC calculation is performed to determine the flux values needed to

deplete the fuel. The EOEC calculation is performed to determine whether or not the










uncontrolled k-eff has converged to one, indicating that the code has reached the end of the

irradiation (Figure 2-1). A middle-of-equilibrium-cycle (MOEC) calculation is performed to

update the flux values during the depletion to ensure that any spectrum changes, though small,

will be reflected in the depletion. Due to the small change in flux spectrum and intensity

throughout the irradiation (Figure 5-6, Figure 5-7 and Figure 5-8) this three-point check is

usually acceptable. The limitations of the REBUS code in equilibrium-mode, only allows a

maximum of two DIF3D calculations to be performed between BOEC and EOEC. For the many

core design calculations performed in this dissertation, using the equilibrium-mode, only one

MOEC calculation was performed between BOEC and EOEC. This decision was made in order

to minimize computational time.

-*- BOEC First Row of Fuel -- EOEC First Row fo Fuel

5.0E+15

4.5E+15 -

3 4.0E+15 -
E


S3.0E+15
I- \
2.5E+15 -

2.0E+15 -
0 10 20 30 40 50 60 70 80 90
Radial Distance from Core Center (cm)
Figure 5-8. Comparison between BOEC and EOEC axial flux profiles (taken at the inner-most
row of fuel) (DIF3D)

To check the validity of this three-point check, a non-equilibrium-mode REBUS

calculation was performed using the geometry and atom density data extracted from the

equilibrium-mode output. In non-equilibrium mode the REBUS code allows an arbitrary number

of DIF3D calculations to be performed between BOC and EOC. Therefore, the BOEC atom










density and geometry used in the previous standalone DIF3D calculations (Figure 5-1 through

Figure 5-8) of this chapter was also used for BOC values in a representative non-equilibrium

coupled DIF3D/REBUS calculation. In a similar manner, the atom density and geometry data

used in the previous standalone MCNP calculations of this chapter was applied to a coupled

MCNP/MONTEBURNS calculation. Figure 5-9 shows the k-eff results from the non-

equilibrium DIF3D/REBUS calculation and the MCNP/MONTEBURNS calculation.

1.05
a REBUS
Note: Error Bars have two sigma confidenceREBUS
1.045 o MONTEBURNS

1.04

1.035

1.03

9 1.025

1.02

1.015

1.01 -

1.005 -

1 .
0 50 100 150 200 250
Effective Full Power Day
Figure 5-9. Reactivity curve comparison between REBUS and MONTEBURNS

There is a bias of approximately 30 milli-k-eff between the REBUS and MONTEBURNS

calculations. Part of this difference can be attributed to the difference in the diffusion

approximation made by DIF3D and the transport calculation performed by MCNP. However,

much of the difference exists due to a discrepancy between fission product masses at BOC in the

DIF3D and MCNP models. This discrepancy is caused by the MC2-2 cross section library (used

by DIF3D and REBUS) having more fission product isotopes than the standard ENDF libraries

available to MCNP and MONTEBURNS. MC2-2 also uses ENDF libraries to produce the group









constants used in DIF3D/REBUS. However, the reason MC2-2 has more ENDF isotopes than

MCNP is because almost all of the fission products in the comprehensive ENDF database are

important to fast reactor calculations. In general, all fission products have approximately the

same cross section value in the unresolved resonance range. However, not all of these fission

products have appreciable value in the thermal range. Therefore, for thermal reactor calculations

these extra fission product isotopes are not needed and do not come standard with the publicly

available release of MCNP.

There is also a slight difference in the slope of the reactivity curve between REBUS and

MONTEBURNS. The smaller MONTEBURNS slope is explained by the difference in fission

product isotopes, which are generated by the ORIGEN part of MONTEBURNS, which do not

exist in the MCNP cross section library. When these isotopes are generated by the ORIGEN

code but can not be recognized by the MCNP code, MONTEBURNS simply decides not to

continue tracking these isotopes in the depletion calculation. Therefore, the neutron absorption

of these dropped fission products are not reflected in the MONTEBURNS reactivity curve

causing the shallower slope.

Before REBUS was adopted for fuel cycle analysis in this work, a benchmark calculation

was performed between REBUS and MONTEBURNS using the Advanced Burner Test Reactor

(ABTR) design as a reference for fuel composition and geometry [80]. The ABTR is a smaller

prototypic version of the larger ABR design, proposed by ANL as a proof of concept reactor

with a primary function for materials testing. The ABTR benchmark started with a fresh fuel

composition and not one obtained from an equilibrium cycle. The results of this "fresh core"

benchmark calculation are given in Figure 5-10.











Reactivity curve


1.28000


1.27500


1.27000


1.26500


1.26000


1.25500


0 18 36 54 72 90 108 126 144 162 180
Cycle Lenght (days) -- REBUS-BOC
-- REBUS-MOC
-- REBUS-EOC
note: Error bars have two sigma confidence X Monteburns
I Linear (Monteburns)
Figure 5-10. Advanced Burner Test Reactor benchmark using "fresh core" fuel composition
showing the gradual departure of MONTEBURNS from REBUS1

This benchmark was conducted "internally" by the members of the INL fuel cycle analysis

team (which the author is a member of) to identify the practicality of using REBUS (an ANL

code) for fast reactor calculations. The non-equilibrium REBUS and MONTEBURNS

calculations in the ABTR benchmark showed very good agreement with each other at BOC.

However, as fission products were dropped from the MONTEBURNS calculation, the EOC

result by MONTEBURNS diverged from the REBUS calculation.

Spatial Self-Shielding Test

The method by which the group constants are generated in MC2-2 is consistent with zero

dimensional slowing down theory techniques that were in common use at the time the MC2-2

code was developed. The use of such methods is generally considered an acceptable practice for

fast reactor core simulation due to the fast neutron mean-free-path being generally larger than the



1 The non-equilibrium REBUS calculations were performed using BOC, MOC "Middle-of-
Cycle" and EOC isotopics to generate the MC2-2 cross section library.









pin-cell. There is little change in the neutron spectrum over the local space domain due to this

long mean-free-path. Therefore, the use of corrections of the flux at node, cell or the boundaries

between different fuel compositions such as discontinuity factors, is not applied in the DIF3D or

VARIANT codes. However, incorporation of moderating pin-cells within the target geometry

raises the question of the applicability of such generic assumptions.

An MCNP sub-lattice model of one zirconium hydride pin surrounded by six target pins

was created to represent an equivalent lattice calculation for this arrangement (Figure 5-11). The

neutron flux was tallied in five equal volume zones within the zirconium hydride pin. Similarly,

the fuel slug of one of the six targets was sub-divided into five equal volume tally zones. The 33

energy bin group structure used in the MC2-2 calculation was used to tally these fluxes as a

function of energy. The height of this MCNP model is 20 cm. Reflective boundary conditions

were used on all sides to simulate a repeating lattice of this geometry.

It should be noted that the geometry shown is Figure 5-11 is not a true lattice calculation.

This is because the geometry can not be simply folded over at the problem boundary condition

into a hexagonal lattice. Five Equal Volume
Tally Zones
















A B
Figure 5-11. MCNP sub-lattice representation (A) of a repeating pin-cell arrangement (B)









The picture in the right-hand-side of Figure 5-11 shows how the seven pin arrangement

would actually look within a hexagonal lattice. Nevertheless, it will be seen that the geometry

shown can be used as a valuable analysis tool for evaluating the neutron spectrum variations

within the repeating lattice. Figure 5-12 and Figure 5-13 shows the 33 group energy spectrum as

a function of the five equal volume tally zones for both the zirconium hydride and target slugs

respectively. These plots show a similar spectrum shape and magnitude to that shown in Figure

5-2. Indeed both of these plots are very similar to each other. The small change in neutron

spectrum as a function of penetrating depth into the moderator is similar to that found by

Konashi et al [81]. Therefore, there is virtually zero spatial shielding as a function of fuel slug

radius. Because of this uniform irradiation, the target slug should not experience a non-uniform

radial burnup distribution or "rim-effect".

The uniform irradiation across the fuel slug radius can be attributed to the epithermal

mean-free-path in this target spectrum to be greater than that of the pin diameter in the seven pin

model of Figure 5-11. The hexagonal flat-to-flat dimension of the model is 2.31 cm, whereas the

neutron mean-free-path (for all reaction types including scattering) is 2.97 cm. Note the fuel

slug outer diameter is 0.557 cm. As a comparison, the mean-free-path calculated for a generic

PWR IMF fuel assembly (Table 1-2) is 2.34 cm, whereas the lattice pitch and pellet diameter is

1.26 and 0.82 cm respectively.

Spatial Shadowing Test

Now that the irradiation distribution in the pellet has been analyzed, there is still one final

check in the applicability of the infinitely homogeneous approximation used by MC2-2. The

region-to-region neutron shadowing effect is evaluated by modifying the seven pin sub-lattice

model shown in Figure 5-11. A new MCNP unit-cell model is created by homogenizing the six

target pins into an annulus that surrounds the zirconium hydride pin (Figure 5-14).















6E+14
5E+14
4E+14
3E+14 E
a 2E+14
1E+14
OE+00

Energy C, do
(MeV)
Radius (cm)
Figure 5-12. Neutron spectrum as a function of energy and the zirconium hydride slug radius
(MCNP)






6E+14
5E+14


3E+14 -
5 2E+14
1E+14

Energy O E+00


Radius (cm)
Figure 5-13. Neutron spectrum as a function of energy and the target slug radius (MCNP)

The inner and outer bounds of this annulus are the nearest and furthest points of the

original six targets from the zirconium hydride pin. The HT-9 cladding and sodium bond and

coolant are also homogenized with the fuel slug into the annulus' composition. A "white

albedo" boundary condition was used at the annulus outer perimeter. The white boundary

condition replaces the incident angular dependence with an isotropic distribution. This is


rrlr~cn~


I









different from a reflective boundary condition which reflects neutron flux with a reflection of the

angular distribution that is incident upon it as if it were a perfect mirror.



0.727 1.265









0.279
cm













Figure 5-14. MCNP unit-cell model with homogenized fuel annulus: zirconium hydride
(turquoise), sodium bonded gap (olive), HT-9 cladding (yellow), sodium coolant
(red), homogenized fuel annulus (blue)

The white boundary condition was used for this problem to give the annulus a

representative flux at the boundary to that of a hexagon despite using the circular shape. Similar

to the sub-lattice geometry from Figure 5-11, the annulus is divided into five equal volume zones

for tallying the neutron flux. Figure 5-15 shows the 33 group neutron spectrum as a function of

radius for these flux tally zones.

As can be seen from Figure 5-15, there is very little difference in the neutron spectrum flux

magnitude as a function of radial distance from the zirconium hydride pin. This is expected

because the epithermal mean-free-path discussed in the previous section is on the dimensional

level of the annulus which is filled with mostly sodium. The annulus outer diameter is 2.53 cm

compared to a mean-free-path of 2.97 cm. This is compared to the LWR IMF case which has a

mean-free-path of 2.34 cm which is roughly twice the rectangular pitch of 1.26 cm. Note, the









volume fraction of water in a PWR pin-cell is approximately 65 v/o, whereas the volume fraction

of ZrH1.6 in the annular pin cell is approximately 5 v/o.






6E+14
5E+14

O,5 4E+14
-n






from the unit-cell origin (MCNP)
6, 0q 1E+14









It is important to distinguish between the mean-free-path between total interactions versus

the mean-free-path between fission events. Table 5-2 indicates the mean-free-path for total
interactions versus that for only scatter, capture or ssion reactions.
(MeV) tk* c 0 LO W

Radius (cm)
Figure 5-15. Neutron spectrum within the homogenized annulus as a function of radial distance
from the unit-cell origin (MCNP)

It is important to distinguish between the mean-free-path between total interactions versus

the mean-free-path between fission events. Table 5-2 indicates the mean-free-path for total

interactions versus that for only scatter, capture or fission reactions.

Table 5-2. Comparison of mean-free-paths of different reaction types for various irradiation
regions having different spectrums (MCNP)
Total Interaction (cm) Scatter (cm) Capture (cm) Fission (cm)
ABR (CR=0.5) 4.15 4.44 386.42 457.24
AHFTR Active Core 3.75 4.04 339.86 390.53
AHFTR Target Region 2.97 3.11 166.07 451.81
LWR IMF 2.34 2.97 21.16 27.89

The loss of energy per collision in sodium is significantly less than that of hydrogen in

water. Therefore, without slowing down to thermal energies, neutrons are more likely to fission

at fast energies, where the probability (i.e., cross section) is small compared to thermal energies.

Therefore, the average length of travel between fission (or capture) reactions for a SFR neutron









spectrum is much greater than for an LWR spectrum. This is also true of the "slightly"

moderated target region of the AHFTR as shown in Table 5-2.

Calculation Validation Remarks

Despite the lack of higher order scattering in the VARIANT/REBUS calculation, close

agreement in the neutron spectrum and spatial flux distribution was found between DIF3D,

VARIANT and MCNP. The DIF3D/REBUS and VARIANT/REBUS BOEC k-effective was

1.012377 and 1.013683 respectively. The MCNP calculation with BOEC isotopic number

densities based on the DIF3D/REBUS calculation was 1.04071 with a standard deviation of

0.00024. As is seen in the benchmark analysis given in the VARIANT user's manual, the full

transport evaluation gives a k-eff that is higher than for diffusion by roughly 25 milli-k-eff [78].

Naturally, this bias is not constant for any core geometry. It is expected that for increasing core

size, the difference between the core average fluxes of the full transport evaluation and the

diffusion approximation would decrease as a result of spatial heterogeneities becoming less

significant compared to the overall governing core physics. The much higher k-eff obtained by

MCNP is a result of the lack of fission product isotopes that were available to the MCNP code.

Despite the lack of these isotopes, the MCNP calculations showed similar trends with regards to

flux values and a loosely representative reactivity curve in the MONTEBURNS calculation.

Due to the close agreement in neutron spectrum and spatial distribution between all three

calculation methods, little difference was observed in the MA transmutation rate. Therefore, the

diffusion approximation is considered an acceptable method for evaluating the fuel cycle and

reactor performance analysis of the AHFTR core design.

It was also found that the arrangement of moderating and target rods in the AHFTR

demonstrate negligible resolved resonance spatial self-shielding and shadowing effects. Though

neutrons are being slowed down by scatters in the zirconium hydride rod, the neutron energies









are still to fast to be considered a true thermal spectrum. Most of the neutrons in the target

region's neutron spectrum have energies above 100 eV which is at the upper end (or above) the

resolved resonance range for essentially all heavy metals. At these energies, the resolved

resonances of heavy metals are poorly resolved. The flux depressions observed in the neutron

spectrum of the AHFTR are, in fact, a result of the lighter sodium and iron atoms with resolved

resonances in the fast range. The nearly fast epithermal spectrum has a mean-free-path

sufficiently greater than the dimensions of the fuel slug; allowing neutrons to pass through it

without being absorbed non-uniformly in the slug periphery which would give a "rim-effect".

This uniform irradiation makes it possible to assume infinite dilution in the target region for the

group constant calculation performed by MC2-2. Also because the epithermal neutron mean-

free-path is on the dimensional level of the repeating target-and-moderating rod arrangement, the

neutron shadowing effect is negligible.









CHAPTER 6
THE AHFTR FUEL DESIGN

The AHFTR design employs flattened core geometry, epithermal upper axial targets and

metallic ternary alloy fuel. The combination of axial targets, combined with the co-

pyroprocessing approach draws from the Integral Fuel Cycle (IFC) strategy demonstrated by the

EBR-I and later by EBR-II. In the IFR program, it was envisioned that both driver blanket fuel

would be used to breed fissile plutonium [40,82,83]. Once discharged from the core, driver and

fuel assemblies (axial blankets were included as part of the driver fuel assembly) would be

chopped into small segments. Next the sodium bond occupying the fuel-to-clad gap would be

extracted and the remaining chopped cladding hulls and fuel slugs placed in perforated steel

baskets. These baskets would be taken to the electrorefiner for processing. All of these

processes were performed in a system of interconnected hot-cells at a facility adjacent to the

EBR-II primary containment building. A more detailed synopsis of metal fuel electrorefining

will be provided in a following section.

In the AHFTR design, the moderation effect in the targets suppresses the fission of

plutonium isotopes while increasing the capture rate in Am-241. Similar to the Pu-239

conversion from uranium in the IFR blankets, plutonium isotopes are generated by this Am-241

transmutation. The moderating effect enhances this transmutation conversion process as

discussed in Chapter 3. The combined effect of americium transmutation into Pu-238, and

conversion of Pu-239 from the U-238 in the target, provides a plutonium source within the

AHFTR fuel cycle analogous to the IFR blanket fuel. The co-pyroprocessing approach,

envisioned by the IFR scenario, is also applicable to the AHFTR because of this plutonium

creation. The feasibility of this scenario is validated by the fact that the MA concentration

accumulating in the fresh driver fuel can be kept to below 5 w/o (MA/HM) as a guideline









adopted from the CAPRA program. The 5 w/o limit was imposed for the foregoing AHFTR

parametric design study in Chapter 3, for the purpose of preserving the fuel irradiation

performance established within the IFR experience database for U-Pu-Zr metal alloy fuels.

The 5 w/o limit was also considered important to maintain a low MA driver fuel

concentration from the standpoint of reactor kinetics. As discussed in the introduction and

Chapter 4, MAs can create an unacceptably high void coefficient due to the resulting spectrum

hardening.

Fuel Pin Design

The neutron trap effect afforded by the moderated target region allows the neutron leakage

leaving through the top of the reactor to be captured in the targets as opposed to being lost. In

the absence of the targets, axially leaked neutrons would leak out of the top of the active core

into the gas plenum region which is mostly voided space. Placing the axial targets below, in

addition to above the driver fuel, is possible and was considered in the conceptualization of this

design. However, including a second target region of equal size to that of the upper region

would double the amount of unburned americium discharged to the pyroprocessor by the targets.

Hence, the MA content in the driver fuel would roughly double. Remember, from a reactor

kinetics and fuel reliability standpoint, it is desirable to maintain the americium content below

the 5 w/o limit adopted for this work. Also lower axial targets would act as a neutron trap below

the core thus robbing the AHFTR of the reflection necessary for optimizing the core's reactivity

requirements. Therefore, a bottom axial reflector comprised of stainless steel, which is a

traditionally common design aspect for SFR's, was adopted. For the AHFTR, S-PRISM and

ABR designs, this axial reflector constitutes an HT-9 plug approximately a meter long which

comprises the bottom end of the fuel pin. This end-plug could either be cast as an integral part of









the fuel cladding or inserted as a separate component and diffusion bonded into the pin. Figure

6-1 shows the general design configuration of the AHFTR fuel pin.

Sodium Level Axial Target or ZrH1.6 Slug Bottom Fitting





Gas Plenum Driver Slug Axial Reflector
Figure 6-1. Conceptual AHFTR fuel pin design

In Figure 6-1, the axial target slug is a zirconium metal alloy similar to the driver fuel. The

elemental constituents of this alloy are 0.5Np/9Pu/9Am/1.5Cm/40U/40Zr by weight. These

weight percent are derived from the blending of the Np+Pu, Am+Cm+Bk+Cf and U mass

streams produced by the aqueous separations plant shown in Figure 3-4. As mentioned in

Chapter 3, these streams are blended with the ratios: 10Np+Pu/10Am+Cm+Bk+Cf/40U/40Zr.

The SNF TRU isotopic vector was assumed to be that for a typical 17x17 PWR fuel assembly

having a discharge burnup of 50 MWD/kg and cooled for five years before being transported to

the aqueous separation plant. An additional decay time of two years was assumed for the time

after separation which includes: reprocessing, fuel fabrication and transportation to the AHFTR.

The SNF TRU isotopic data was generated by an infinitely repeating fuel assembly lattice

calculation using the coupled NEWT/TRITON depletion module of the SCALES.1 code system.

The isotopic vector was given previously in Table 2-1.

Target Alloy Selection and Design Considerations

The elemental composition of the target was selected, based on the irradiation experience

gained by the AFC-1B and AFC-1F tests performed at the Advanced Test Reactor (ATR) at INL

[84]. The AFC-1B tests consisted of four uranium-free metal alloy fuel samples with varying

MA, plutonium and zirconium concentrations.









* A1B1 and A1B4: 48Pu/12Am/40Zr
* A1B2: 40Pu/10Am/10Np/40Zr
* A1B3: 60Pu/40Zr
* A1B5: 40Pu/60Zr

The AFC-1F test consisted of four uranium bearing metal alloy fuel samples with varying

MA, uranium and zirconium concentration.

* A1F1: 28Pu/4Am/2Np/66U/30Zr
* A1F2: 27Pu/3Am/2Np/28/40Zr
* A1F3: 34Pu/4Am/2Np/60U/20Zr
* A1F4: 29Pu/7Am/64U

The AHFTR target composition is a hybrid compilation of these test compositions:

combining a modest amount of uranium (-40 w/o in some cases) and zirconium content (-40

w/o in some cases) and high MA (-10 w/o in most cases) content. The choice to mix fertile and

fissile materials in this composition was made with the intent to decrease the expected fission

density in the axial target fuel alloy.

Metallic fuels experience a swelling incubation period (low rate of swelling) followed by a

transition period (high rate of swelling) as a function of burnup [60,85]. The AFC-1 test Post

Irradiation Examination (PIE) performed by Hilton et al. revealed that the time of incubation is

proportional to the amount of fission damage imparted to the overall fuel matrix. Because the

zirconium content was varied in these tests, the total fission energy released per unit HM mass

(MWD/kgiHM) (iHM stands for initial heavy metal of fuel) or equivalently atom % heavy metal

destruction) could not be used as a metric to measure cumulative fission damage. As an

alternative, the time integrated cumulative missions per unit volume of fuel alloy was used

instead.

As is common in metal fuel irradiation behavior, many of the metal fuel samples

experienced fuel-to-clad contact due to swelling. This behavior was also common to metal fuels









tested at EBR-II. The samples that experienced this behavior were the ones placed in the highest

flux and also had the highest plutonium content. However, also similar to EBR-II experience,

there was no cladding breach despite this contact. Figure 6-2, borrowed from Hilton et al, shows

the optical microscopy and fission gas release for three of the AFC-1B samples.

Pu,Am-40Zr Pu,Am,Np-40Zr Pu -40Zr Pu-60Zr






4.6 at. % 6.1 at. % 6.6 at. % 7.8 at. %
3.5E20 fiss/cm3 3.9E20 fiss/cm3 6.1E20 fiss/cm3 3.8E20 fiss/cm3

Figure 6-2. Optical microscopy for three of the AFC-1B samples performed by Hilton et al
(PIE) [84]

As can be seen in Figure 6-2, the high plutonium content of the 60Pu/40Zr composition

leads to greater fuel swelling than the low plutonium 40Pu/60Zr sample at equal burnup. This is

expected because the amount of fission damage imparted to the total fuel volume (i.e., fission

density) is greatest in the 60Pu/40Zr case. It is important to note that the correlation between

atom percent burnup and fission density is roughly linear for equal zirconium content.

Therefore, for equal zirconium volume and actinide volume, the fission damage is the same.

This can be seen in the three TRU/40Zr cases. It appears that the transition swelling begins

somewhere between 3.9E20 and 6.1E20 fiss/cm3. This result is reflected in the fission gas curve

of Figure 6-3, which was borrowed from Hilton et al.

The fission gas release curve seems to also exhibit an incubation period followed by

transition period. The metal fuel alloys used at EBR-II also exhibited this connection between

swelling and fission gas release. Fuel transition swelling would occur early in the irradiation

leading to a high incidence of cladding breach for the first generation of EBR-II fuel (Mark I).










RAFC-l8 UL-xPu-lOZr

90,00%

60.00%
BO. OO% .--------------
7G' .?0%




60.00% -
a 4[H&.C'O- *- -------- /^ --------



0.00% CI -
O.OOE+00 500E+20 1. OE+21 1.50E+21 2,00E+Z1
Flss3ion renUlty, ftlcan
Figure 6-3. Fuel swelling performance for the AFC-1B samples by Hilton et al (PIE) [84]

The second generation of fuel (Mark II) used a much larger cladding-to-fuel gap distance,

which allowed the fuel to expand inside of the larger gap and delay contact with the cladding

[61]. It is important to remember that the bond material selected for metal fuels is sodium which

provides high gap conduction. When the Mark II fuel was allowed to expand, it was discovered

that the fission gas bubbles generated in the fuel would eventually interconnect after a certain

burnup was reached, as documented by Walters [60,62]. This "interconnected porosity" allowed

most of the accumulated fission gas to be released from the fuel where it would be collected in

the gas plenum at the top of the fuel pin. This sharp increase in fission gas release slowed the

rate of fuel swelling in the transition period, allowing a significantly higher burnup than the

Mark I before the limiting inter fuel-to-clad contact pressure was reached. The Mark I fuel could

only be irradiated to 2-4 at.% (19 MWD/kgiHM 37 MWD/kgiHM). Whereas, the Mark II fuel

could be irradiated to 15-18 at.% (140 MWD/kgiHM 160 MWD/kgiHM) [86]. The high

burnup and fission gas release for both metal and oxide SFR fuels stipulates a gas plenum

approximately one and one-half times the length of the driver fuel [60,61,62].









Target and Driver Fuel Burnup Criteria

For the AHFTR targets, it was desirable to delay the transition from incubation to

transition swelling as long as possible. To accomplish this, the plutonium concentration in the

AHFTR axial target alloy was reduced from that used in the AFC-1B tests and replaced by U-

238. The result is a better balance of reactivity suppression at BOL by neutron capture in both

uranium and americium. The addition of uranium and americium gives a breeding contribution

to reactivity at EOL. Table 6-1 and Table 6-2 give the peak and average bumup and cumulative

fission density, respectively, for the inner, middle and outer enrichment zones for the driver fuel

and targets.

Table 6-1. Peak burnups for the driver fuel and targets (REBUS)
urnup Pmrnup Peak Fission Density
Region (MWD/kgiHM) Peak Bumup (at. %) (fiss/cm3)
___ (MWD/kgiHM) / ___ (fiss/cm')
Inner Core Mid-plane 119.18 12.76% 4.7456E+21
Middle Core Mid-plane 122.64 13.13% 4.8836E+21
Outer Core Mid-plane 96.52 10.33% 3.8434E+21
Inner Core Targets 201.78 21.60% 3.4435E+21
Middle Core Targets 173.85 18.61% 2.9669E+21

To ensure fuel reliability with no cladding failures, an average fuel bumup of

approximately 100 MWD/kgiHM is common for both metallic and oxide fueled cores [62]. This

is the case for both S-PRISM and the ABR (CR=0.75) designs [7,19]. The average bumup for

the AHFTR is also in this range. Walters et al show that the corresponding fuel bumup for EBR-

II Mark II fuel (90U/10Zr) could reach about 15 at. % with a low probability of cladding failure

[60,61].

The AHFTR average and peak burnup stays below this limit with the exception of the

targets. However, the target peak and average fission density, in the targets, is still below that of

the driver fuel. Therefore, it is expected that the AHFTR fuel performance will be within the

design criteria established by EBR-II experience.









Table 6-2. Average burnups for the driver fuel and targets (REBUS)
urnup AmrnuAvg. Fission Density
*\ Avg. Bumup (at. %)
Region (MWD/kgiHM) A Bumup (fiss/cm3)
Inner Core Mid-plane 106.19 11.37% 4.2285E+21
Middle Core Mid-plane 102.79 11.00% 4.0928E+21
Outer Core Mid-plane 70.52 7.55% 2.8079E+21
Inner Core Targets 197.02 21.09% 3.3622E+21
Middle Core Targets 162.15 17.35% 2.7672E+21

The peak driver fuel burnups were taken at the active core mid-plane. It is important to

note that the peak actinide burnup occurs in the targets. This is due to the higher zirconium

content in the targets than in the driver fuel. In fact, because of the lower mass density of

zirconium than the metal actinides, the 40 w/o zirconium fraction occupies 60 v/o of the alloy

volume. However, the peak fission damage per volume occurs at the active core's mid-plane.

Therefore, the fuel performance of the driver fuel is the actual limiting case with respect to fuel

performance. The mid-plane also happens to be the location in the core that experiences the

highest flux, volumetric power density and specific power. Therefore, it can be expected that the

fuel at the active core mid-plane will experience the peak cladding damage, fuel swelling and

fuel center-line temperature.

Cladding Damage Criteria

The time integrated fast fluence and dpa for the HT-9 cladding is given in Table 6-3 for the

inner, middle and outer enrichment zones for the driver fuel and targets. As indicated in Chapter

5, the peak flux occurs at the active core mid-plane. Therefore, it is expected that the mid-plane

is where the highest cladding damage occurs as verified in Table 6-3.

Table 6-3. Average and peak fast fluence (E>0.1 MeV) and peak dpa (REBUS)
Region Peak Fast Fluence* (cm-2) Peak dpa
Inner Core Mid-plane 3.5360E+23 164.32
Middle Core Mid-plane 3.6125E+23 170.63
Outer Core Mid-plane 3.2986E+23 159.38
Inner Core Targets 2.0791E+23 89.04
Middle Core Targets 2.0614E+23 90.21
* Fast fluence is calculated by the REBUS code for energies greater than 0.1 MeV.









-2
A fast fluence limit of 4.0E23 cm-2 was assumed as the maximum allowable cladding

exposure for the AHFTR based on the irradiation experience gained by the FFTF reactor. HT-9

qualification at FFTF showed no elongation or cladding breach after being irradiated up to

3.9E23 cm-2 [63]. A cladding damage criterion of 200 dpa was selected as a secondary cladding

performance standard for the AHFTR. The 200 dpa limit is the approximate damage function

-2
corresponding to a fluence of 4.0E23 cm-2 in a "typical" SFR fast spectrum (such as that

observed in the HT-9 tests at FFTF). The cumulative displacement damage was calculated by

applying a 33 group displacement cascade response function library to the calculated neutron

flux. As can be seen from Table 6-3, the target atomic displacement damage for the cladding in

the target region is still less than in the active core. Also this peak cladding damage is within the

4.0E23 cm-2 (or 200 dpa) design criteria for HT-9 established by experience gained at FFTF [63].

The use of irradiation tolerant steels for cladding and structural components in a SFR is

critical to the reliable performance of the fuel. A cladding damage of at least 20 dpa,

(corresponding to a fuel burnup of about 50 MWD/kg) is typical in the zirconium based cladding

used by LWRs [87]. HT-9 would not make a favorable LWR cladding material due to its higher

thermal neutron capture cross section and poorer oxidation performance than zirconium.

However, HT-9 has been proposed for the S-PRISM's and ABR's cladding and structural

materials due to its irradiation creep resistance, high tensile strength throughout irradiation and

low irradiation induced swelling.

Because of the high fuel swelling and cladding damage, the SFR fuel pin is designed to

accommodate the anticipated fuel-to-clad interaction, >90% fission gas release, etc. To have

sufficient space to accommodate fuel expansion, the ratio of cross sectional area of the fuel slug

divided by the cross sectional area of the gap and fuel region, "fractional smear density", is









larger than for LWR fuel. The smear density adopted for the AHFTR is 0.75 as was done for the

metal fueled EBR-II, S-PRISM and ABR. This is compared to a smear density of roughly 0.95

for LWRs, as well as oxide fueled SFR's. As a result of the larger smear density, the fuel-to-clad

gap distance is larger. To achieve adequate heat conduction between the fuel and cladding, this

gap is filled with sodium instead of helium. The sodium bond provides a much higher thermal

conductivity than helium, which allows the difference in temperature between fuel surface and

the inner cladding wall to be kept within 10-15K. The thermal performance of the AHFTR

metallic fuel and target will be discussed in a later section.

Though the gap is filled with sodium, the upper region of the fuel pin, allocated for the

fission gas plenum, is filled with helium gas. Because a pool type design is assumed for the

AHFTR, the coolant pressure is near one atmosphere (not including gravitational pressure head).

Therefore, the helium in the fuel pin does not require significant, if any, internal pressurization.

Hence, the gas plenum pressure at fabrication is assumed to be one atmosphere.

Fuel Pin Thermal Performance Criterion

The overall design rationale of the AHFTR is to minimize the excess reactivity required to

achieve the lowest tCR possible. It does this by increasing the core's axial buckling which

increases axial leakage. If the sum of axial and radial buckling must be equal to the fuel material

buckling (criticality condition), increasing axial buckling will reduce the required radial

buckling. Since the AHFTR radius was increased by an additional row of fuel, the net effect

results in a flattened radial flux and radial power density (i.e., a reduced radial buckling). This

flattened radial power profile translates into a reduction in the ratio of peak-to-average power

density and LHGR over the core. Therefore, the AHFTR design rationale is intended to

inherently increase fuel performance by reducing peaking. The reduction in peaking enables

most of the fuel to be irradiated evenly to within the same performance margins. This is









different from smaller SFR core designs, such as the ABR, with a higher radial buckling. Small

cores have been proposed by Hill et al to decrease the tCR through increasing the radial buckling

[54]. An increase in radial buckling translates into a higher radial power gradient than the

flattened core. This effect was demonstrated by comparing the "tall" and "flat" AHFTR designs

in Chapter 3. For the "tall" and homogeneous "ABR" designs, all of the peaking occurred in the

inner core. Small SFR designs, such as the reference ABR design, require more enrichment in

the outer core to overcome the reactivity lost through radial leakage. Without this enrichment

splitting, only the inner core fuel could be irradiated to the limiting fuel and cladding design

constraints, because the rest of the fuel would not have the reactivity to achieve the same fission

density.

Fuel Temperature Criterion

There are two basic performance criteria. First, the power in the hottest fuel pins should be

less than that necessary to melt the fuel. Secondly, the cladding inner wall temperature must be

less than that necessary to cause eutectic melting between the fuel and cladding. As will be

shown, the high fuel thermal conductivity of the metal alloy creates a large temperature margin

between the peak fuel temperature and the solidus temperature for ternary alloy

19TRU/71U/10Zr fuel, taken to be about 1320 K (1100C) [7].

As an interesting side note, cladding breach as a result of fuel melt is not probable due to

the HT-9 melting temperature (-1500C or -1800K) being higher than that of the fuel. This

result is illustrated by experimental irradiations of EBR-II Mark IA fuel, which was fabricated

with only the bond sodium in the lower half of the fuel pins [61]. The absence of bond sodium

caused extensive fuel melting. However when the fuel melted, it relocated, thus closing the fuel-

to-bond gap, where it froze in place without melting the cladding. A small amount of eutectic

interaction with the cladding was observed with a 10% penetration into the cladding wall.









Cladding Temperature Criterion

Eutectic phase formation between SFR metal fuel and cladding is a type of failure

mechanism that can occur in SFR metallic fuels. Eutectic liquefaction is a result of metallurgical

interaction between actinides (and fission products), and the iron in the HT-9 cladding, which

produces a low melting point phase [86]. It is important to note that eutectic phase formation

does not occur in all SFR fuel pins. It occurs only after sufficient chemical interaction between

fuel and cladding has occurred to produce the low melting point phase, and if the temperature is

high enough in this phase to cause melting. This fuel-to-clad chemical interaction (FCCI) begins

as the inter-diffusion trading of fuel and cladding constituents. Unlike eutectic melting, the

FCCI is a cumulative process as lanthanides are continuously produced by fission. Pahl et al

described the results of FCCI tests, performed at EBR-II, as an inter-diffusion of lanthanides in

the fuel and cladding constituents leading to formation of brittle layers in the cladding wall that

were prone to failure [88]. Lahm et al discusses the erosion of the EBR-II Mark II cladding as a

result of this FCCI inter-diffusion [89]. As the fuel slug swelled and came into contact with the

cladding, the contact pressure led to creep damage and stress rupture later in life. Once the

contact is made, FCCI occurs, allowing iron in the cladding to be traded with lanthanide fission

product metals in the fuel. Lahm et al found that this newly formed uranium-iron phase had a

eutectic solidus temperature below the melting point of the cladding. If the fuel was operated at

temperatures above this eutectic temperature, a liquid interface between cladding and fuel would

form accelerating the FCCI diffusion. The diffusion feedback causes the cladding to thin and

lose its strength as iron in the cladding is consumed by the liquid interface. This cladding

wastage serves to increase the frequency of stress rupture in the driver fuel.

Because, the EBR-II Mark II fuel did not contain plutonium, the eutectic phase was

composed of uranium and iron. Lamb et al reports that cladding penetration was only observed









at temperatures above 985K (7150C), for the Mark II fuel [89]. However, plutonium and iron

have a binary eutectic solidus temperature near 650K (2800C) [60]. This eutectic solidus

temperature closely corresponds to the onset of FCCI at 930K (660C) for the Mark II fuel,

reported by Pahl et al [88]. The EBR-II Mark-V fuel was a 19Pu/71U/10Zr alloy with HT-9

cladding. For the AHFTR fuel pin design, the minimum value of 930K value was chosen as the

limiting temperature, taken at the inner cladding surface. A similar approach was used for the S-

PRISM design to show whether or not eutectic penetration could be tolerated during the

bounding design basis of 113% over-power at SCRAM [19]. The full thermal-hydraulic analysis

of the AHFTR over-power scenario is outside of the scope of this work which is more concerned

with the fuel cycle aspects of the heterogeneous design. However, given the similarity of the

AHFTR design to that of the S-PRISM and ABR, it is expected that the transient response issues

of MA containing fuels will be dealt by future authors.

Thermal Analysis

The peak inner cladding and fuel centerline temperature was calculated using the region

specific peak volumetric power density taken from the REBUS output. Given the poor neutronic

communication amongst individual fuel rods, as discussed in Chapter 2, local pin power peaking,

within the fuel assembly is considered to be a negligible effect. Therefore, pin power

reconstruction within the hottest fuel assembly and/or core region is deemed unnecessary.

Indeed, pin power reconstruction is not a feature provided by the DIF3D/REBUS code system.

Therefore, the average fuel pin power for the hottest fuel assembly, in the hottest enrichment

zone, is taken to also be the peak fuel pin power in the following thermal analysis.

Fuel Assembly Power Peaking

As noted previously, the enrichment zone homogenization routine in REBUS does not

permit the specific tracking of individual fuel assemblies. Instead, a mass balance is used to


200









track the mass of each batch as it is charged and discharged from its enrichment zone within the

core. Because the reactivity change of the average fuel assembly, and representative batch, does

not change significantly with burnup, the shuffling of fuel assemblies is not required. Therefore,

for the AHFTR, like the ABR and S-PRISM, the driver fuel is not shuffled during its irradiation.

A fresh fuel assembly is simply loaded, irradiated and discharged from the same location within

the core. For the sake of simplicity, the use of the term "fuel assembly", being synonymous with

the corresponding fuel batch, is used in place of the word "batch" for the remainder of this

discussion.

It is evident from Figure 6-4, that the peak power density occurs at the core mid-plane of

row five (yellow columns) which is in the middle core enrichment zone. This highlights the

importance of enrichment zoning in SFR design. The reactivity contribution of a slightly higher

enrichment for the middle and outer core provides for a flatter radial power distribution than can

be achieved by radial buckling alone. The draw on reactivity by radial leakage can be inferred

from the fact that the outer core fuel has the highest enrichment but also the lowest power

density over the entire core.

The core region specific volumetric power density for the homogeneous ABR reference

case with a tCR=0.5 is given in Figure 6-5. Observing the yellow column of Figure 6-5, between

row one and four, one can see the radial leakage affect on the curvature of the power density.

The power density for the ABR decreases until the fuel reactivity is increased by the higher

enrichment of the middle core. It can be seen from the yellow column of Figure 6-4, that the

AHFTR exhibits virtually zero curvature between row one and four. In contrast, the ABR also

has a relatively flat radial power density. This is because the ABR uses a higher gradient of

enrichment splitting than the AHFTR (Table 3-13). The ABR enrichments in the middle and












outer core are 1.25 and 1.5 times, respectively, higher than for the inner core. Whereas for the


AHFTR, the enrichments in the middle and outer core are 1.125 and 1.25 times, respectively,


higher than for the inner core.


50 1 1 0-14 cm
1 -28cm
26 42 cm
0 43-56cm

7 90 cm

Row Number from Radial Center 8


Figure 6-4. AHFTR power density profile as a function of fuel row and axial region (REBUS)


450

400

350

300

250

200

150

100

50

0


0 20 cm
20 40 cm
aJ: 60 cm
60-80 cm


S280 100cm
2 3 4 5 6 7

Row Number from Radial Center 8


Figure 6-5. ABR reference power density profile as a function of fuel row and axial region
(REBUS)


202









Because of the higher enrichment splitting, the ABR experiences a higher ratio of peak-to-

average heat generation rate. The peak-to-average LHGR, at BOEC, is 1.5 for the ABR and 1.38

for the AHFTR. The peak-to-average ratio is defined as the hottest LHGR (mid-plane) for the

hottest fuel assembly divided by the average LHGR taken over all fuel assemblies in the core.

This core wide peaking factor also takes into consideration the peak-to-average of the axial

LHGR distribution of the individual fuel assembly.

max(LHGR(z))
PFx LHGR(z)dzz (6-1)

max(( LHGR(z)dz )/z)
PFFuelAssembly max( LHGR z) (6-2)
Z((qLHGR(z)dz) z)) FA
FA FA /
PFcore = PFAxial x PFFuelAssembly (6-3)

Where: PF stands for peaking factor and FA stands for Fuel Assembly. The LHGR(z)

represents the axial LHGR distribution across the fueled portion of the fuel assembly and z is the

total length of the fueled portion.

The ratio of the mid-plane LHGR to the average LHGR of the hottest fuel assembly is 1.2

and 1.23 for the ABR and AHFTR, respectively. The peak axial fuel assembly average LHGR to

core average LHGR is 1.25 and 1.12 for the ABR and AHFTR, respectively. Multiplying the

fuel assembly peaking factor by the axial peaking factor gives the core wide peaking factors,

(i.e., 1.2x 1.25=1.5 and 1.23 x 1.12=1.38). The fuel assembly peaking factors for the ABR and

AHFTR cores are given in Table 6-4 and Table 6-5.

Note the small amount of change between the fresh and last bum of the fuel for any given

row in the core. Hence, the peaking factor is generally insensitive to the amount of exposure

received by the fuel assembly in any given row of the reactor. This is expected, given the impact


203










of leakage on the radial power distribution as noted above. Table 6-6 gives the fuel assembly


average peaking factors for the AHFTR at EOEC.


Table 6-4. ABR fuel assembly peak-to-core average LHGR ratio taken at BOEC (REBUS)
Batch Number
Row Fresh Once Twice Thrice Fourth Fifth Sixth Burned
Number Fuel Burned Burned Burned Burned Burned
One 1.19 1.15 1.12 1.09 1.06 1.03 n/a
Two 1.18 1.14 1.11 1.08 1.05 1.02 n/a
Three 1.15 1.12 1.08 1.05 1.02 0.99 n/a
Four 1.12 1.09 1.06 1.03 1.00 0.97 n/a
Five 1.26 1.18 1.12 1.07 1.02 0.97 n/a
Six 1.07 1.01 0.97 0.92 0.89 0.85 n/a
Seven 0.78 0.74 0.71 0.68 0.65 0.62 0.60

Table 6-5. AHFTR fuel assembly peak-to-core average LHGR ratio taken at BOEC (REBUS)
Batch Number
Row Fresh Once Twice Thrice Fourth Fifth Burned
Number Fuel Burned Burned Burned Burned
One 1.12 1.11 1.10 1.09 1.08 1.07
Two 1.12 1.11 1.11 1.10 1.08 1.07
Three 1.11 1.10 1.10 1.09 1.07 1.06
Four 1.10 1.09 1.08 1.07 1.06 1.04
Five 1.12 1.11 1.09 1.07 1.06 1.04
Six 1.02 1.01 0.99 0.98 0.96 0.94
Seven 0.92 0.90 0.88 0.87 0.85 0.83
Eight 0.64 0.63 0.62 0.62 0.61 0.60

Table 6-6. AHFTR Fuel Assembly Peak-to-Core Average LHGR ratio taken at EOEC (REBUS)
Batch Number
Row Fresh Once Twice Thrice Fourth Fifth Burned
Number Fuel Burned Burned Burned Burned
One 1.14 1.13 1.12 1.11 1.09 1.08
Two 1.14 1.13 1.12 1.11 1.09 1.07
Three 1.12 1.12 1.11 1.09 1.08 1.06
Four 1.11 1.10 1.09 1.07 1.06 1.04
Five 1.12 1.10 1.08 1.06 1.05 1.03
Six 1.01 1.00 0.98 0.97 0.95 0.93
Seven 0.90 0.88 0.87 0.85 0.83 0.82
Eight 0.63 0.62 0.62 0.61 0.60 0.59

Comparing Table 6-5 with Table 6-6, it can be seen that the peak fuel assembly location

shifts from Row Five in the middle core to Row One in the inner core. This can be attributed to

a greater level of plutonium breeding in the inner core than in the middle core. The tCR in each

enrichment zone is given in Table 3-13 in Chapter 3. The conversion ratio is essentially a










function of the neutron balance between the probabilities of parasitic capture by U-238 versus

neutron escape from the core through leakage. The inner core has the smallest enrichment of all

the enrichment zones. It also sees the least amount of neutron escape from radial leakage.

Therefore, the inner core stands to have the largest plutonium breeding gain and hence relative

increase in power generation to the rest of the core.

The internal breeding effect on inner core reactivity can be seen by comparing the LHGR

for the Row One (Inner Core) versus Row Five (Middle Core) (Figure 6-6 and Figure 6-7). The

decrease in LHGR with irradiation is slightly less for Row One driver fuel than it is for Row Five

driver fuel. Also important to note is the increase in LHGR for the targets. The increase in

LHGR with irradiation is slightly more for Row One targets than it is for Row Five targets.

However, the overall change in LHGR with irradiation for targets and driver fuel is relatively

slow. If the conversion ratio of the active core is reduced, than the internal breeding would also

be reduced, thus creating a greater loss rate in LHGR from BOL to EOL. For the targets, the

composition was selected to have some initial plutonium and uranium at BOL so that the

"external" breeding in the targets would maintain a fairly constant LHGR.

40

35

30

E 25

20

I 15
.J
10
-*-Midplane Row Five -l-Midplane Row One
5 -A-Target Row One Target Row Five

0 1 1 1 1
1 2 3 4 5 6
Cycle Number

Figure 6-6. BOEC LHGR for the driver fuel mid-plane and target regions (REBUS)










40

35

30

E 25

20
20

I 15

10
--Midplane Row Five --Midplane Row One
5 -Target Row One Target Row Five

0 -------------------------
1 2 3 4 5 6
Cycle Number

Figure 6-7. EOEC LHGR for the driver fuel mid-plane and target regions (REBUS)

Peak Fuel Pin Hot Channel Analysis

For the following analysis, the axial LHGR distribution for the hottest fuel assembly is

used to determine the axial fuel centerline and inner cladding temperature profile. This "hottest"

fuel assembly is located in Row Five of the AHFTR. The highest LHGR for Row Five fuel

occurs at BOL. The sodium coolant channel surrounding a "typical" fuel pin within this hottest

fuel assembly is considered. To perform sodium channel analysis, the coolant channel height is

sub-divided into a series of axial temperature nodes. Steady state conditions are applied at each

node to establish a Nusselt correlation as a function of axial height from the coolant inlet at the

bottom of the active core. Using a curve fit to the axial LHGR data and an empirical formula for

the sodium heat capacity, the coolant temperature drop across each axial node is determined.

This information is used to determine the bulk coolant temperature at each axial node. Once the

bulk coolant temperatures are known, established empirical correlations for the sodium Nusselt

number and thermal conductivity is applied to calculate the convective heat transfer coefficient

at each axial node.









Once the axial bulk coolant temperature and convective heat transfer coefficient

distribution is known, a one-dimensional radial heat transfer model is applied at the channel wall

to determine the: cladding outer and inner wall temperature and fuel surface and centerline

temperature.

REBUS generates a snapshot of the average reactor power by region in the core (Figure 6-

4) as a function of cycle number for each fuel batch (Figure 6-6 and Figure 6-7). The axial

LHGR data for fresh fuel in Row Five was tabulated from the power generation data for each

axial region in Row Five of the REBUS model. The REBUS model divides each row of fuel into

six axial regions: five axial driver regions plus one target region. REBUS generates region

powers for each of these six axial regions in each row of fuel. The code further breaks down the

regional power data, which is homogenized over all fuel batches in that region, into the

individual power contributions to that region by each fuel batch. These batch power values are

given as a function of cycle number. Because each batch is assigned its own row, the average

power produced by any given row (and axial region in that row) is the volumetric average of the

batch power contribution at each stage (i.e., cycle) of its irradiation in that region.

This REBUS generated batch power "history" was then divided by the total length of all

the fuel pins within the batch to give the LHGR in each axial region in each row as a function of

batch and cycle number. This LHGR region specific power history data for Row Five was used

to represent the "typical fuel pin" of fresh fuel occupying that region of the AHFTR. Figure 6-8

shows this axial BOL LHGR distribution. A fourth order polynomial was used to fit the REBUS

tabulation for the 71.6 cm section of driver fuel. A linear curve fit was applied for the target

region from 71.6 to 91.6 cm. A caveat must be noted here. The LHGR for each data point from

the REBUS tabulation directly corresponds to the average power for that axial region in the










computer model. Given the scoping-calculation nature of the following one-dimensional channel

Nusselt number calculation, the approximation by the "moving average" is deemed appropriate.


40

35 -

30

E 25

0
S15

10 --Curve Fit Data + Tabulated From REBUS

5

0
0 10 20 30 40 50 60 70 80 90
Distance from Bottom of Active Core (cm)

Figure 6-8. Peak pin axial BOL LHGR distribution (REBUS)

The LHGR curve fit is used to calculate the heat energy released "q" through the cladding,

or channel wall, for each node by multiplying the node LHGR at a location "z" by the intra-node

distance "Az". The node-to-node marching procession is illustrated in Figure 6-9.


q' = q'(z')xAz (6-4)

Where: q' is the LHGR at "z"

Knowing that the first node is at the core inlet temperature, the axial bulk coolant

temperature distribution is then calculated using the heat capacity relationship.


Tb = Tb + -(6-5)
q' xCp(Tb)

Where: m is the average mass flow rate of a "typical" coolant channel, q is the amount of

heat energy transferred to the channel cladding wall in Az, Cp is the heat capacity of node "i"

and Tb is the bulk coolant temperature at node "i" or "i+1".










The average channel mass flow rate was approximated by considering the core inlet

temperature and outlet temperature (after mixing above the core). Since the sodium heat

capacity does not change appreciably in this temperature range, an average sodium heat capacity

for the entire core was assumed for the channel mass flow calculation. Given the AHFTR core

power, the heat capacity relationship is used to calculate the total core mass flow rate.


Channel Exit Temnerature


ic: inner cladding
oc: outer cladding
fs: fuel surface
cl: centerline


Flow Direction









nm


0









Az
s


Core Inlet Temnerature


zN
q4
TbN



Z1
z'
q
Tb


z2
q2
Tb2


z'
q1
Tb1


Figure 6-9. Axial node procession for Nusselt analysis

This total mass flow is then simply divided by the number of fuel pins in the core. This

generic calculation gives a slightly conservative estimate of the hot channel mass flow rate.

Typically a SFR thermal-hydraulic design would introduce a flow orifice at the coolant entrance

at the bottom of the fuel assembly. Most modern SFR fuel assembly designs incorporate a

hexagonal HT-9 shroud that encompasses the fuel pin sub-assembly. The HT-9 shroud prevents

cross flow between fuel assemblies. This allows the mass flow through each fuel assembly to be









tailored specifically to meet the thermal requirements of the fuel in that assembly. Therefore, for

the AHFTR, one might envision the inner, middle and outer core to have three different levels of

flow orificing to ensure a fairly flat exit coolant temperature and average fuel temperature profile

across the core. Since the hot channel mass flow rate has not been artificially increased to

demonstrate this flow orificing, the fuel temperatures discussed in this analysis will be slightly

higher than the expected operating value. However, as this analysis will show, even without

flow orifices, the fuel still meets the thermal fuel performance criteria discussed earlier.


mh = average x (6-6)
a ..e.r..ge X (/,,,y., Tj., ) x No.FP
p erg outlet ie O

Where: Q is the thermal power of the AHFTR (1000 MW), Cpaverage is the average sodium

coolant heat capacity throughout the core, No.FP is the total number of fuel pins in the core.

By repeating the axial marching procedure for all "N" nodes, the bulk coolant temperature

is found as a function ofz. Once the bulk coolant temperature profile is known, the Schad-

Modified correlation, as recommended by Todreas and Kazimi as well as Waltar and Reynolds,

is applied to determine the Nusselt number for each node [62,90]. Once the Nusselt number is

known for each coolant channel axial node, the radial temperature profile calculation is

performed for all "N" nodes. For the remainder of the radial temperature analysis, the axial "i"

subscript is dropped.

Nu= (- 16.5 + 24.96(P/D)- 8.55(P/D)2 )Pe3 (6-7)

For: 1.1


Where: Nu is the Nusselt number at node "i", Pe is the Pecklet number at node "i", P is

the hexagonal fuel pin-cell flat-to-flat pitch and D is the fuel pin diameter. Then the convective

heat transfer coefficient for each node is as follows.









Nu x k(Tb )
h = (6-8)
DH

Where: h is the convective heat transfer coefficient, k(Tb) is the sodium thermal

conductivity evaluated at the coolant bulk temperature and DH is the fuel pin hydraulic diameter

which is equivalent to D.

Assuming one-dimensional radial heat diffusion and constant heat current, Newton's Law

of Cooling is applied to calculate the outer cladding surface temperatures at each node.


Toc = Tb + (6-9)
'rxDf xh

Where: Df is the diameter of the fuel slug

With the outer cladding surface temperature known, Forrier's Law in cylindrical

coordinates is used to calculate the inner wall temperature.

T = T+ x + In(b/ a) (6-10)
2; x kHT-9 (TIoc)

Where: kHT-9 is the HT-9 thermal conductivity at the outer cladding wall temperature.

A gap heat transfer coefficient is approximated by dividing the thermal conductivity of the

sodium bond, at the inner wall temperature, by the fuel-to-clad gap distance. This approximation

does not take into account the surface roughness of the fuel or a jump distance approximation.

Given the relatively large gap distance of 0.4 mm, compared to the allotted gap spacing for

surface roughness (-0.01 mm), and the higher thermal conductivity of sodium, compared to

helium, this is probably a reasonable approximation of the gap heat transfer coefficient. The

temperature of the fuel slug surface is calculated using the gap conductance model.


Tf = x +d (6-11)
z x Df x k(T/,)Id









Where: k is the sodium bond temperature taken at the cladding inner wall temperature and

d is the fuel-to-clad gap distance.

The assumption was made that the fuel thermal conductivity does not vary widely with

temperature. This is deemed as an acceptable "first glance" approximation by the

recommendation of Hofman et al [86]. Given the higher thermal conductivity of metal fuel, as

opposed to oxide, the radial temperature rise from the slug surface to centerline is only a few

hundred degrees, as is verified by this calculation. Hence, it can be expected that the

temperature dependent thermal conductivity will be relatively insensitive to this change in

temperature.

In fact, the thermal conductivity is highest at BOL, as is the case for the hottest AHFTR

fuel assembly, and remains high throughout most of the irradiation. Hofman et al discusses the

effect of swelling and interconnected porosity phenomenon on the thermal conductivity of

19Pu/71U/10Zr fuel irradiated at EBR-II [86]. In the first few at. % of burnup, pore formation,

from fission gas retention in the fuel, causes the fuel to swell to nearly contacting with the

cladding. During this free swelling time, the fuel thermal conductivity goes down. Also, much

of the bond sodium is displaced and forced into the gas plenum volume. Near the point of fuel-

to-clad contact the pores become mostly interconnected which releases fission gas to the plenum.

The interconnected pores are large enough to allow the bond sodium to infiltrate the fuel causing

the thermal conductivity to increase. The bond sodium infiltration was sufficient enough to

restore most of the thermal conductivity that was lost during the initial swelling [60,83].

The approximate temperature independent thermal conductivity of 20Pu/70U/10Zr metallic

fuel is discussed by Hoffman et al for the ABR parametric study on conversion ratio. This

summary is provided in Figure 6-10 which was borrowed from Hoffman et al [7]. It should be










noted that this data was not generated by Hoffman et al whose work was primarily focused in the

reactor physics and fuel cycle analysis of the ABR. Hoffman referenced the data in Figure 6-10

to internal communications with Hofman and Pelton held at ANL in the 1980's and 1990's.


1,500 ---- -- 50
Fuel SOllaSl Temllperalure
1,400 Fuel Co'i3uctivity t 2C.0 C) 45
1,300 Zr Weight Fraction 40



1.000 250 0
E 900oo-, 20 E
800. 15
700 - --- 10
600- -- ----- --- 5
5 00 ,....-------- 0
0% 20% 40% 60% 80% 100%
TRU Enrichment. TRU,'HMI
Figure 6-10. Assumed thermal properties of TRU-U-Zr metal alloy fuel borrowed from the
ANL-AFCI-177 report by Hoffman et al [7]

Assuming the thermal conductivity was constant throughout the fuel slug, the centerline

temperature was calculated by solving the heat diffusion equation with constant volumetric heat

generation.


+ d+ dT\ q'
Sdr dr) k 0 (6-12)


Using the boundary conditions of a known temperature at the fuel surface and also zero

temperature gradient when the radius is zero, a solution to the heat diffusion equation is found.


Tfc = Tf, + q (6-13)
4z x kf (T7},)


Where: kf is the average fuel thermal conductivity taken at the fuel surface temperature.

The fuel centerline, fuel surface, and cladding inner and outer surface temperatures are plotted in










Figure 6-11. It is interesting to note that the axial fuel centerline temperature profile becomes

fairly flat above the core mid-plane. The curvature of the axial LHGR distribution (Figure 6-8)

gives the lowest heat generation at the upper and lower ends of the fuel rod. This reduces the

heat input through the channel cladding, which is directly proportional to the difference in fuel

centerline and surface temperature from Equation 6-13. It is an interesting finding that the

reduction in (Tcl-Tfs) is roughly matched by the coolant temperature rise (Toc-Tb) with increasing

z.


1400



-Fuel Centerline
1200 Fuel Surface
Clad Inner Surface
Clad Outer Surface
SFuel Solidus Temp
-1100 Clad/Fuel Eutectic Temp




E 900 1

800 -- "" -

700 _. ",_ "

600
0 10 20 30 40 50 60 70 80 90
Distance from Bottom of Active Core (cm)
Figure 6-11. Axial temperature profiles for a "typical" fuel rod in the hottest fuel assembly

The resulting minuscule change in fuel centerline temperature, in the upper half of the

AHFTR fuel rod, can be considered an important phenomenon for discussing the likelihood of

atomic inter-diffusion between the target and fuel slug compositions in this region of the core.

The anticipated metallurgical interaction between the fuel alloy and target alloy composition will

be discussed in the following section.









The fuel centerline temperature is never more than 230 K greater than the fuel surface

temperature. Due to the high thermal conductivity of the fuel, the fuel temperature is always at

least 300 K less than its melting temperature. In contrast, the minimum difference between the

inner cladding temperature and the eutectic melting temperature is only 55 K. Therefore, it is

apparent that the eutectic melting temperature is the most constraining temperature limit.

However, this feature is not a special property of the AHFTR, but is fairly common to all SFRs

with metal fuel.

Hofman et al states that the most common failure mechanism for the EBR-II Mark II fuel

was a small inter-granular crack caused by FCCI at the restrainer dimples (discussed in Chapter

4) [86]. The restrainer dimples (120 degrees apart) were three sharp indentations inside the fuel

pin cladding. Their purpose was to prevent the fuel from somehow ratcheting upwards, due to

bowing effects, inside the cladding during the irradiation and then later falling back down at an

inappropriate time in the irradiation, resulting in positive reactivity insertion. These dimples

became stress risers, as the fuel swelled and the plenum gas pressure increased. The failure rate

at these high stress points was further enhanced by FCCI and a eutectic phase formation. PIE of

the Mark II fuel showed that the ratcheting effect was not as significant as originally thought.

When the dimples were eliminated from the EBR-II fuel pin design, much higher burnups could

be achieved without clad rupture.

Metallurgical Diffusion Effects

Given the intimate contact of the axial target and driver fuel slugs, it is expected that some

inter-diffusion of the alloy constituents, of the two different fuel types, should be expected. The

targets and drivers slugs are essentially composed of the same primary elemental ingredients:

plutonium, uranium and zirconium metal. They are simply mixed in different proportions. Since

the axial fuel temperature gradient is nearly zero at the target-to-driver interface (Figure 6-11), it









is expected that the axial inter-diffusion between the target and fuel slugs will be driven only by

the atom concentration gradients and not axial thermal gradients.

Inter-Diffusion Data

Inter-diffusion between these elements has been well characterized during the IFR program

and its predecessors. This work was done to increase understanding of the radial redistribution

of Pu, U and Zr induced by the thermal gradient between the metal fuel centerline and surface

temperature. Isothermal diffusion coefficients of these constituents were measured by Petri et al

[91]. Ternary inter-diffusion coefficients were calculated from the common composition

between two diffusion couples with intersecting diffusion paths. In general, Petri et al found that

the diffusion coefficients increased for increasing plutonium concentration and decreased for

increasing zirconium concentration. This result suggests that it is more likely that driver fuel

species are more likely to diffuse into the target than it is for the target species to diffuse into the

driver fuel. The results of Petri et al's experiments are given in Table 6-7.

Table 6-7. Ternary inter-diffusion coefficients measured at the common composition (matino
plane) between two diffusion couples*
Composition (a/o) Inter-diffusion Coefficients (10-12 m2/s)
Pu U Zr DUzrzr DUZrpu DUpuzr DUpuPu DPUzrzr DPUuzr
13 75 12 0.16 -0.33 -0.10 1.7 0.49 1.3
13 75 12 0.16 -0.29 -0.04 1.5 0.45 1.1
9 74 17 -0.45 1.3
15 62 23 -0.62 1.4
16 80 4 0.20 1.7
16 80 4 0.06 3.1
11 73 16 0.63 0.84
11 74 15 0.72 0.94
*Table borrowed directly from Petri et al [91]. D'jk stands for the ternary inter-diffusion
coefficient for species j diffusing into species k within the solvent i.

The composition given in the left-hand side of the Table 6-7 represents the common

composition at the matino plane between the two diffusion couple compositions. The matino









plane represents the original contact plane, between diffusion couples, where the accumulation

on one side is balanced by the depletion of a species on the other side.

Petri et al describes a negative diffusion coefficient for zirconium indicating that zirconium

inter-diffuses up a positive Pu concentration gradient to regions of higher Pu contents.

Therefore, it is expected that the high zirconium content of the targets is likely to diffuse toward

the higher plutonium concentration in the driver fuel. For plutonium, the inter-diffusion

coefficient with zirconium is more than an order of magnitude smaller than the plutonium-to-

plutonium diffusion coefficient. Hence, plutonium inter-diffusion is affected strongly by its own

concentration gradient, but is relatively uninfluenced by the Zr concentration gradient.

Therefore, it is expected that the higher Pu concentration in the driver fuel is likely to diffuse

toward the lower plutonium concentration in the targets. The inter-diffusion coefficient of

uranium into zirconium is of the same order as the plutonium-to-plutonium diffusion coefficient.

Therefore, it is expected that the higher U concentration in the driver is likely to diffuse into the

high zirconium concentration in the targets. The inter-diffusion coefficients for uranium and

plutonium are all roughly three times higher than for any of the zirconium diffusion coefficients.

Penetration Distance

To quantify the inter-diffusion impact upon the target composition, an effective diffusion

coefficient and penetration depth is offered by Petri et al.

x = 2 xtxD (6-14)

Where: t is the annealing time, Def is the average effective diffusion coefficient.

The average effective diffusion coefficient is a proxy form of the ternary inter-diffusion

coefficients that represents diffusion of a single species crossing the matino plane. Therefore, it

can be applied in an analogous manner to a binary diffusion coefficient, which is easier to









conceptualize. The maximum De" is reported by Petri et al for each species. As expected from

Table 6-7, Pu has the highest Def of 2.2E-12 m2/s.

To calculate the expected penetration depth of driver plutonium diffusing into the target,

the target-to-driver interface is assumed to be the matino plane in the AHFTR fuel rod. Given,

isothermal conditions at the target-to-driver interface from the hot channel analysis, and

assuming an annealing time equal to the fuel pin in-core exposure time, an effective penetration

depth is calculated.


S ( days 24hr 3600 s22 10 2 102
x 2x 6cyclesx221 --- --- 2.2x1012m 1002cm
x cycle day hr s m2 (6-15)
x = 2.2452cm

Therefore, it can be expected that about 11% of the length of the axial target length will be

affected by the plutonium diffusion coefficient to some varying degree. As stated in Chapter 5,

the mean-free-path of the epithermal neutrons in the targets is 2.97 cm. Therefore, it is unlikely

that the plutonium diffusion gradient will have a significant impact on the local power produced

in the area of the target-to-driver fuel interface. However, it should be pointed out that Petri's

diffusion measurements did not include MAs. Therefore, further testing should be done to

quantify the diffusion gradients including neptunium, americium and curium.

Transmutation Gas Generation and Plenum Sizing

The transmutation chain of the even neutron numbered americium isotopes, Am-241 and

Am-243 ultimately lead to the production of Pu-238 and Cm-245, as shown in Figure 1-2 and

Figure 1-3. The accumulation of these isotopes is favorable from a physics perspective because

these isotopes can be used as fuel at a later time in the fuel cycle. However, in order to produce

these fissile isotopes, a shorter lived intermediary non-fissile curium isotope has to be produced.

For Am-241, this isotope is Cm-242, whose production and decay (Tl/2=162.8 days) is










essentially in secular equilibrium with the beta decay of the neutron capture product Am-242.

For Am-243, this isotope is Cm-244, which is sufficiently long lived (T1/2=18.1 years) to

accumulate in the AHFTR fuel cycle. Both Cm-242 and Cm-244 decay by emission of an alpha

particle. The kinetic energy of this alpha particle is quickly lost by ionization interactions with

atoms of the fuel matrix. Once, the alpha particle is stopped it will have picked up two electrons,

becoming an atom of helium. These helium atoms accumulate in the fuel as a function of

burnup. When the burnup drives the fuel swelling to the point of interconnected porosity, it is

expected that these helium atoms will be released to the gas plenum in addition to the fission

product gasses. This additional gas production requires quantification in order to determine the

plenum height which in turn impacts the thermal-hydraulic design of the reactor. The major

contributions to helium production by alpha decay for all the actinides in the targets are shown in

Figure 6-12.





Am-241
0.76%
Cm-242
88.80%


SPu-238 Am-243

8.82%
1.36% 0.03%
Cm-244
8.82%8 Pu-240
0.04%
Cm-243
0.18%
Pu-239 U Am-243 D Pu-240 D Cm-243 Am-241
SPu-238 PU241 D Cm-244 U Cm-242
Figure 6-12. Percent contribution to helium production by alpha decay for target fuel actinides
during the course of irradiation (BOL to EOL)

Almost all HM actinides decay by alpha decay. Most of these isotopes, like Pu-239 have

half-lives ranging from decades to thousands of years. Therefore, most actinides have little









contribution to helium production in the fuel pin unless present in appreciably large quantities.

This is the case for Pu-238 as well as Cm-244, whose alpha decays are not in secular equilibrium

with the americium transmutation, but are present in the fuel in large enough quantities to

produce a non-trivial amount of helium.

Since helium gas production is an unavoidable result of neutron capture by americium, it is

an issue for the driver fuel as well. In fact, due to the fission threshold requirement for Am-241

and Am-243 to fission, all SFR designs with americium in the fuel, must account for

transmutation helium. Helium production is discussed by Taiwo et al for americium bearing fuel

pins for a multi-recycled heterogeneous PWR fuel assembly design [92]. Taiwo et al quoted the

helium component to be about 20% of the total gas production for that particular fuel design.

Taiwo et al also goes on to mention that there is a direct relationship between the presence of

Am-241, Pu-238 and Cm-244 and the amount of helium produced. Hence, the percent helium

per total gas production will increase as the ratio of americium to total HM loaded into the core

increases.

Helium and Fission Product Gas Calculation

Since the REBUS calculation does not report total number of alpha decays as one of its

outputs, a post processing code was developed to recreate the fuel buildup/depletion algorithm

performed by REBUS. This was done in order to reproduce the exact alpha decay history of the

fuel as a function of irradiation time. The depletion algorithm uses an exponential matrix

method, similar to that used by the ORIGEN code, to calculate the number density of each

actinide isotope as a function of discrete time steps [50]. With the depletion history of the fuel

known in much finer time steps than reported in the REBUS output, these number densities are

then converted into alpha decay activity. For the helium calculation, pure exponential decay is

assumed within the time step. This alpha decay is numerically integrated in time for each time









step. The total cumulative number of alpha disintegrations throughout the entire irradiation is

found by summing the integrated alpha decay over all time steps. The total integrated alpha

activity corresponds to the total number of helium atoms generated.

The exponential matrix method provides a simultaneous solution of the generalized burnup

equation for each time step. The generalized burnup equation is given in Equation 6-16.


N-=Y N, +Bi -N AN,- yf 0 N, (6-16)


Where: Ni is the number of atoms of daughter i, Nj is the number of atoms of parent j, k is

the relevant radioactive decay constant, 4 is the local neutron flux, ac is the capture cross section,

Ca is the total absorption (capture plus fission) cross section, g represents the 33 group neutron

flux and cross section set, Yj is the yield fraction of radioactive decay for isotope j going into i,

and Bj is the branching ratio of isotope j going into isotope i. Equation 6-16 is determined for all

combinations of i and j to form a list of first order ordinary differential equations for each

daughter i. These equations are put into the matrix form shown below in Equation 6-17.

In matrix form, the Yield Fraction and Branching Ratio take on an additional numerical

meaning. When i is not the radioactive decay daughter of parent j, then the yield fraction is zero.

Similarly when a neutron capture in j does not create i, the branching ratio forj going into i is

zero. Otherwise, when i is created by j the yield fraction or branching ratio is determined by the

physics characterized by the radioactive decay or neutron reaction, respectively. The vector

form of Equation 6-17 is expressed by Equation 6-18.

Given the non-linear behavior of buildup and depletion of isotopes throughout the entire

irradiation, the solution can not be solved for in a simple integration, as is done for the scalar

form of Equation 6-18. However, if the solution to this equation is assumed to be linear across a








small time step, then an accurate approximation can be found for each step. Therefore, the entire

radiation time is broken into T time steps. Then the linear solution to Equation 6-18 is found for

each step.

d F -AN, 9 [rY4N.+BrZ -J
dt N

dt
(6-17)

dN1\N
dt- ... I+B \0, N]



d= AN (6-18)
dt

+ = Nt[ex(t)] (6-19)

Where: Nt+1 is the Nt vector found at time t+At after t and Nt, found during the previous

time interval, and is also the initial condition to Equation 6-18 for the next times step. The

exponential matrix is defined through a Taylor series expansion valid for small incremental times

of At.

[eI(A)] =+At+ 12t2 (6-20)
2

Where: I is an identity matrix and A2 is the vector multiplication of matrix A with itself.

Once the Ni vector is found for all T time steps, a vector for decay activity is defined by

multiplying the number of isotope i atoms in each row by its alpha decay constant: k,.

Assuming that the change in Ni over At is negligible, the total number of decays, in At, can be

found by simply multiplying the activity by At.









dN'
NHe^)- = a Ax At = Z, x N, x At (6-21)
dt

However, in an effort to reduce the necessary time steps to give reasonably accurate

results, a slightly more elegant integration within the time step is applied. First, the activity

within the time step is defined.

A=dN = (Ne -(At) (6-22)
dt

Where: Ai is the activity or rate of decay of isotope i. The helium production in the time

step is found by integrating Equation 6-23 over At.

t+At
Nt,'^ 'A = T N,' Je "-' (dt N= 1 e- (At) (6-23)
t

The total helium generation over the irradiation time is found by summing the solution to

Equation 6-23 over all time steps (t). A similar technique is used to calculate the atoms of

krypton and xenon fission gas atoms produced by fission. Instead of Xk, the fission reaction rate,

in conjunction with the fission yield for krypton and xenon, is used instead: BAr((D and Bxe(Gy.

The total fission gas yield for U-Pu fuel in a SFR is about 27 % [62]. This percentage accounts

for all intermediate short lived decays, following fission, that ultimately lead to formation of a

stable Kr or Xe atom.

Transmutation and Fission Gas Analysis

This depletion algorithm and gas production calculation is then applied for all regions of

the core used in Figure 6-4. The code calculates the average number of helium and fission gas

atoms produced per fuel assembly per cycle in each region of the core. Using the ideal gas law

and the physical dimensions of the AHFTR fuel pin, the total number of He, Kr and Xe atoms,

for each fuel assembly, are converted into pressure. For the conversion of atom density into


223









pressure, an approximate sodium coolant outlet temperature was assumed (-750 K). The

dimensions used for this calculation are given in Table 6-8. The sodium coolant channel

dimensions are also given. These dimensions were also used in the hot channel analysis in the

previous section.

Table 6-8. AHFTR fuel pin dimensions
Total Fuel Pins Per Assembly 271
Pin Pitch-to-Diameter Ratio 1.1760
Pin Pitch (cm) 0.8879
Pin Diameter (cm) 0.7550
Cladding Thickness (cm) 0.0559
Cladding Inner Diameter (cm) 0.6432
Fuel Smear Density (%) 75
Fuel-to-Cladding Gap (cm) 0.0431
Fuel Slug Diameter (cm) 0.5570
Axial Reflector Height (cm) 114.6600
Active Fuel Height (cm) 71.6000
Target Fuel Height (cm) 20
Gas Plenum Height (cm) 191.1400

Fission gas release, due to interconnected porosity, is a function of fuel swelling, which in

turn is a function ofburnup for a given fuel composition. Extensive experimental data exists for

the EBR-II Mark I, II, III and IV binary alloy 90U/10Zr fuels irradiated throughout the life of

EBR-II. However, fuel qualification of the Mark V ternary 19Pu/71U/10Zr composition, of

interest to the IFR program, had just begun before the program was terminated in 1992. The

second phase of the AFC-1 program, at the ATR, is expected to achieve fuel burnups high

enough to quantify swelling and interconnected porosity in high zirconium and high MA fuels.

Therefore, without an established database correlating MA concentration with swelling, a gas

release fraction of 75 % was assumed, as recommended by Tsai et al [64]. The average partial

pressure contribution to the total plenum pressure is given for He, Kr and Xe as a function of row

number in Figure 6-13.


224











35
B Helium Krypton E Xenon
30 -- -- -- -------------

S25 -- --- -- -- ---------





520- - -- -- ----------
30

25
E

S20

2 15
a-

10

5

0
1 2 3 4 5 6 7 8
Row Number

Figure 6-13. Gas plenum pressures resulting from transmutation and fission gas production
(plotted as a function of fuel assembly row number)

Notice, the sharp fall in gas production for the outer core rows (seven and eight). The


smaller gas production in this region is characterized by the smaller power generation in the


outer core. Also, the outer core the outer core has no axial targets so the contribution to helium


gas production is less in these outer two rows. However, it is important to note the percent of the


total pressure represented by helium, even in the absence of targets. Even though the outer core


does not have axial targets, the driver fuel still has a non-trivial concentration of Pu-238, Am-


241 and Cm-244. Without the axial targets, the fraction of total helium generated by Pu-238 is


10% as opposed to the 1% shown in Figure 6-12.


These pressures are representative of plenum pressures observed at EBR-II for the Mark II


through Mark V experimental irradiations [93]. Using empirical models for the peak strain,


creep and cumulative damage fraction (CDF), a rule of thumb was adopted to allow a plenum


length approximately 1.5 times the length of the driver fuel. As with the Mark II and Mark III


fuel, the 75% smear density, for Mark V, was adopted to ensure low FCCI, respect the cladding


tensile strength, and minimize the contact pressure due to fuel-to-cladding mechanical interaction


225









(FCMI). The end result is a low probability of cladding rupture. The corresponding plenum

pressure ranged between 25 and 40 atmospheres for steady state operation depending on fission

gas release, fuel radial and axial expansion, etc. Knowing that transmuted helium gas would

create the need for additional plenum volume, a conservative estimate of two times the rod

(driver and target) height was used for the AHFTR design. The resulting plenum pressures fall

within the EBR-II database, indicating that the AHFTR fuels will experience similar tolerance of

FCCI, FCMI or creep rupture.

Fuel Design Basis Summary

Though the design criteria for the AHFTR are quite different than for existing LWR

technology, the standards applied have been proven to provide safe operating conditions for SFR

in general. The design criteria used for the AHFTR are virtually identical to that adopted by the

S-PRISM and ABR designs and is backed by over 40 year operational experience of EBR-I and

EBR-II and the 14 years of experience gained at FFTF. The AHFTR pin design was approached

in a way that could incorporate the axial targets as an integral feature of the overall driver fuel.

Therefore, it is not surprising that the fuel performance, cladding damage and temperature profile

fall within the same limitations imposed upon the driver fuel.

The similarities between target and driver fuel performance are principally related to the

similarities in fuel composition. The choice to incorporate some fissile plutonium and uranium

into the fresh target slug minimizes the power shift from BOL to EOL. The suppression of

power and transmutation of fertile material (namely U-238 and Am-241) in the axial target

region produces fuel atoms that can be used later as fuel. The transmuted material (which is

mostly plutonium) experiences a spectrum shift from epithermal to fast neutron energies when it

is re-fabricated for a second life as driver. When the transmuted plutonium isotopes are

introduced as driver fuel, their fissile worth becomes very near that of fissile Pu-239. Efficient


226









conversion, of MAs into plutonium isotopes, eliminates the need to multi-recycle the unburned

MAs back into fresh targets for additional irradiations. Only the target's initial pass through the

reactor is required to transmute the majority of the SNF MA mass into plutonium fuel.

Fuel Processing Considerations

This recycling strategy (i.e. breeding plutonium outside the active core before combining it

with recycled driver fuel) is exactly the same IFC scenario demonstrated during the IFR

program. In fact the AHFTR fuel cycle uses all of the primary components of the IFR fuel cycle.

All fuel processing and fabrication operations are expected to be performed by remote handling

in a hot-cell facility adjacent to the reactor. A similar hot-cell facility, called the Fuel Cycle

Facility (FCF), was used to develop pyroprocessing and associated process technologies

"electro-refining" for EBR-II and the IFR program. The FCF was an annular argon gas filled

hot-cell that encompassed all fuel dissolution and re-fabrication processes. Spent EBR-II fuel

assemblies were passed from disassembly, to pyroprocessing, blending, casting and assembly

fabrication in a clockwise fashion within this hot-cell.

The AHFTR "electrorefinery" would serve a similar form and function as the FCF, having

a single interlock with the outside world, which would pass casks of imported Np+Pu,

Am+Cm+Bk+Cf and uranium provided by a larger centrally located SNF aqueous separations

facility (Figure 3-4). FCF fuel handler's manipulated objects in the hot-cell using hand operated

master-slave manipulator arms. Unlike the FCF, the electrorefinery would be automated on an

industrial scale with a system of conveyors and robotic manipulators.

In the pyroprocessor, the chopped cladding hulls, fuel slugs and zirconium hydride slugs

are immersed in a eutectic solution of LiCl-KCl electrolyte. A direct current is applied with the

positive pole connected to the fuel basket and the negative pole applied to a steel cathode that is

also immersed in the electrolyte. The current that is passed through the salt bath induces electro-


227









transport of the uranium to the metal cathode. The transuranics are collected by a liquid

cadmium cathode in a ceramic crucible at the bottom of the pyroprocessor. Hydrogen from the

ZrH1.6 slugs would be extracted by passing a current between the anode basket and a palladium

cathode. Palladium metal becomes PdLiHx readily by immersion into a LiCl-KCl-LiH system.

Eutectic salt hydrogen recovery technology has been developed at the bench top level for

fabricating tritium production targets as well as hydrogen batteries for fuel cell applications

[93,94,95]. Further exploration of the hydrogen recovery technology, at the deployment level, is

necessary to bring it to the same technology readiness level as pyroprocessing of SFR fuels.

After separations, the recycled and external feeds would be brought together to form the

target and driver fuel slugs. For EBR-II and the IFR program, americium containing slugs were

melted in an induction furnace and then injected into quartz molds. This process worked well for

the 90U/10Zr and 19Pu/71U/10Zr fuels. However, when americium was added to the mix, 40 %

of it was lost due to vaporization before the fuel slug solidified [96]. The vaporization loss was

attributed to volatile contaminants and vaporization losses at the casting temperature of 1465 C.

This problem was overcome during the AFC-1 tests performed more recently (this decade) for

testing at the ATR. To create the high MA compositions for the AFC-1 fuels, an arc smelter was

used to reach much more rapid heating of the feedstock [97]. Also, the injection and cooling

times were reduced by using a vacuum assisted injection process. These factors combined

eliminated most of the americium losses. However, the arc melting technology was only applied

at the table-top level and still needs to be demonstrated at the deployment level to prove that it

can be consistently performed in an assembly line fashion.

The AHFTR electrorefinery is expected to work in conjunction with the S-PRISM power

block concept [19,98]. The AHFTR design has the same thermal power rating as S-PRISM. In


228









the AHFTR power block model, two 1000 MWth cores, each having their own primary and

secondary sodium coolant loops, share the same primary containment building, steam generator

and turbine system. The power block business model is analogous to a power utility purchasing

smaller fossil fuel boilers units to comprise a much larger plant. The nuclear utility purchases

individual power blocks and adds them to the same reactor site in a similar fashion that a coal

utility purchases additional boiler units depending on the rate of market growth for electricity

demand in the local market. A cartoon of the power block reactor site is given in Figure 6-14.

Turbine Plant



Power Block
Sodium Fast wer c
Reactor



Primary
Containment
o i Building
Cooling _
Towers


Sub-Surface Spent Fuel
Transport Tunnel and Tracks
-------------'------------------------

Receiving Bay: Rail Tracks
Internal and External
Interlock
Electrorefinery


Figure 6-14. Envisioned power block reactor plant model describing the physical relationship
between SFRs (ABR or AHFTR) and the electrorefinery


229









There is one important difference between the power block model and the boiler unit

paradigm. The primary source of fuel is derived from the electrorefinery which is built at a set

capacity. The utility must purchase the electrorefinery and co-locate it at the site where the

power blocks are to be built. This virtually ensures that in order to achieve return on investment

of the electrorefinery, the maximum amount of power blocks should be built early on. A high

fuel throughput per unit of capital footprint pays down the interest on the electrorefinery capital

investment more quickly.

For the AHFTR power block concept, it is envisioned that sufficient power blocks will be

purchased, such that core reload operations will be a continuous, instead of cyclic, operation.

The refueling of the AHFTRs would be conducted out of phase of each other, so that one reactor

is going offline as the next reactor is going online. This ensures that the electrorefinery is, at all

times, recycling the currently discharged fuel with a constant throughput. A cycle length of 220

EFPD days and a capacity factor of 0.85 gives an outage time of 36.56 days. Therefore, ten

AHFTR reactors are required to ensure that one reactor is offline at any given time of the year.

At any given time of the year, nine reactors are at full power and one is shutdown for refueling.

Assuming a thermal efficiency of 0.37%, these nine reactors produce 8,990 MWth and 3,326

MWe of electricity.

Higher Mass Actinide Considerations

The choice of pyroprocessing eliminates the possibility of MA actinide partitioning and

multi-reprocessing of the MAs in targets. Therefore, the buildup of curium and the higher mass

actinides berkelium and californium in the driver fuel must be tolerated. This problem was

encountered for multi-recycling IMF in studies of LWR recycling strategies. In IMF, each

reactor pass depletes the concentration of neptunium, plutonium and americium while

simultaneously generating curium, berkelium and californium. A similar result is observed in


230











multi-recycling all of the MAs in the homogeneous ABR [38]. The increase in decay heat,

gamma and neutron emission in the recycled ABR fuel becomes an order of magnitude greater

than if curium and the higher mass actinides are discarded.

Curium and the higher mass actinides are produced in the AHFTR targets as well as the

driver fuel. Figure 6-15 shows the relative neutron emission activity per mass of initial TRU,

after recycle and blending with the external feeds. In Figure 6-15, each reprocessing technology

represents the actinide grouping assumed for the external transuranic feed from the SNF aqueous

plant (Table 1-8).

PUREX: Np, Am, Cm, Bk and Cf are discarded
UREX+2/+3: Am,Cm, Bk and Cf are discarded
UREX+4: Cm, Bk and Cf are discarded
UREX+la: All TRU is kept in the fuel and none is discarded
UREX: All TRU is kept in the fuel and none is discarded

1E+11
CR=0.5 Oxide
F CR=0.5 Metal
1E+10 --NIMF
EMOX
SHT-SFR (Targets)
1E+09 HT-SFR (Driver)
--
L 1E+08
a.
= 1E+07
0

W 1E+06

1E+05

1E+04
PUREX UREX+2/+3 UREX+4 UREX+Ia/Pyro

Figure 6-15. Average neutron emission rate for processed initial TRU in fresh fuel

In the case of the metal fueled SFR, the neutron emission rate remains more or less

constant across the different levels of actinide partitioning. This is due to the lumped transuranic

grouping limitation of pyroprocessing. In the case of oxide, however, different aqueous

separation processes only allow specific isotopes to be fabricated into new fuel. In the oxide









fueled SFR, a UREX+ separation is assumed for the reactor discharge in addition to the SNF

feed. With the exception of the metal fueled SFR, the neutron emission rate stays low until

curium is kept in the fuel, as is the case for UREX+la and pyroprocessing. Even MOX and IMF

are found to have low emission rates when curium and the higher mass actinides are discarded.

However, it should be noted that even if curium, berkelium and californium are separated

from the fresh fuel charge, the waste stream created is highly radioactive, and requires a long

term storage solution. This is counterintuitive because the philosophy of an advanced "burner"

reactor is to destroy nuclear waste.

The AHFTR appears to exhibit a much reduced neutron source compared to the thermal

spectrum transmutation schemes. This may be attributed to the net destruction of Cm-244 in the

driver fuel. In thermal reactors, the fission-to-absorption ratios for the fertile curium isotopes

(Cm-244,246,248) are small and on the order of 10% to 20% (Table 1-2). The low fission

importance provides an open gateway to produce higher mass actinides with each neutron

capture. The AHFTR targets also have low fission-to-absorption ratios for these isotopes, thus

allowing higher mass actinide generation. However, when these isotopes are processed and

charged to the driver fuel, the much faster neutron spectrum of the active core closes the neutron

capture gateway with significantly higher fission-to-capture ratios (Table 2-3). Therefore, the

transmutation gateway towards higher mass curium, berkelium and californium is analogous to a

diode. The transmutation diode is open for neutrons below the threshold for fission and closed at

neutron energies above the threshold (one MeV).

The gamma decay energy rate per mass of initial TRU is shown in Figure 6-16. The trends

are similar to the neutron emission data. However, gamma energy emission rate is less sensitive,

than the neutron emission rate, to the separation of curium from the initial TRU. The AHFTR


232











targets show the highest gamma energy emission rate, even though this mass is processed

directly from the SNF and not multi-reprocessed. Hence, there is inadequate time for buildup

trends leading to high emission rates. The gamma decay rate from Am-241 and Cm-244

constitutes 66% and 20%, respectively, of the gamma energy produced.

Gamma decay is not the primary mode of decay for Am-241 and Cm-244. The secondary

gamma emission of the excited daughter results in an associated gamma field (e.g., the Am-241

decay into various excitations ofNp-237). The gamma emission rate of the AHFTR fuel is

roughly 1.75 times that of the metal fueled reference ABR. The emission rate is not dominated

by any single isotope as it is in the targets but rather is represented with relative equally by: Pu-

238, Am-241, Cm-242, Cm-243 and Cm-244.

1E+00
SCR=0.5 Oxide
0 CR=0.5 Metal
NIMF
SMOX
U HT-SFR (Targets)
I-
HT-SFR (Driver)
0
1 IE-01
a.


C-
1E-02
E
E



1E-03
PUREX UREX+2/+3 UREX+4 UREX+la/Pyro

Figure 6-16. Average gamma decay energy rate for processed initial TRU in fresh fuel

The alpha decay heat rate per mass of initial TRU is shown in Figure 6-17. The trends are

similar to the neutron emission data. The target and heat generation rate is dominated by the

alpha decay of Cm-244 (targets) and also Cm-242 (targets and fuel) (Figure 3-17 and Figure 3-

18). The AHFTR fuel heat generation rates are both roughly equal to the thermal recycling


schemes and roughly twice as high as the metal fueled SFR.


233










Though the AHFTR driver and target fuels exhibit higher gamma and heat emission rates,

these rates do not exceed the worst case scenario for multi-recycling in an LWR. Also, these

higher emission rates should be expected, considering the high throughput of americium and

curium feedstock in the AHFTR recycling center. Because of these high emission rates, it is

assumed that all AHFTR fuel handling processes will be performed remotely in a dedicated hot-

cell facility. However, the dedicated MA burning, performed by the AHFTR, provides for more

economic fabrication of Np+Pu fuels for other reactors. The Np+Pu feedstock is free of curium

and the higher mass actinides. Therefore, it could perceivably be handled using glove-box

accessible processes. An economic evaluation of the cost savings of this "two-tier" strategy will

be evaluated in a later section.


1000
m CR=0.5 Oxide
CR=0.5 Metal
NIMF
I MOX
U HT-SFR (Targets)













PUREX UREX+2/+3 UREX+4 UREX+la/Pyro

Figure 6-17. Average alpha decay heat rate for processed initial TRU in fresh fuel
Repository Considerations(Driver)










Much emphasis has been placed on the importance of destroying Am-241 in this work. As
100



0






PUREX UREX+2/+3 UREX+4 UREX+la/Pyro

Figure 6-17. Average alpha decay heat rate for processed initial TRU in fresh fuel

Repository Considerations

Much emphasis has been placed on the importance of destroying Am-241 in this work. As

can be seen in Figure 1-4, the heat contribution by Am-241 in UOX-SNF causes the heat

generated in the repository to peak approximately 1000 years after it is closed. Therefore,

transmuting Am-241 in the AHFTR should decrease the amount of heat generated in the


234









repository. However, as stated previously the irradiation of Am-241 results in the buildup of Pu-

238 and Cm-244 in the fuel cycle. These two isotopes decay by alpha particle emission with

half-lives ranging in the decades. As stated in the previous section, buildup of Cm-244 in the

AHFTR fuel cycle results in a fresh fuel that is thermally hot within the time frame that fuel

recycling occurs. Some of this material will be lost from the fuel cycle due to process losses at

the reprocessing and fuel fabrication stages at the electrorefinery. These losses would be

recovered as HLW and ultimately be sent to permanent geologic disposal in the repository.

Therefore, it is necessary to quantify the thermal heat trends of the fuel within the time frame of

the repository.

To do this, the ORIGEN component of the MONTEBURNS code was used. Using

MONTEBURNS, the AHFTR fuel was depleted to the EOC (representing EOEC from the

REBUS calculation) as was done in the benchmark calculation in Chapter 5. After the in-core

buildup/depletion calculation was completed, the MCNP component of MONTEBURNS was

switched off and only the ORIGEN code was used to decay the fuel out to the geologic time

frames of the repository. Because the decay was performed on the entire core inventory at EOC,

this calculation does not represent the repository performance due to the AHFTR spent fuel.

Instead, the calculation offers a scenario where the core is shut down at EOEC, defueled and all

fuel assemblies sent to the repository. Because the decay calculation represents HM in the fuel

cycle averaged over its various stages of depletion in the core, it is a better representation of the

isotopic composition of process losses than purely analyzing the spent fuel composition. The

results of the coupled MONTEBURNS/ORIGEN calculation are shown in Figure 6-18.

Note that the total decay heat in Figure 6-18 does not have a hump at 1000 years as it does

in Figure 1-4. Instead the near term repository heat is dominated by the presence of Pu-238 and


235










Cm-244 which decay away in the first few hundred years after emplacement. The heat

contribution of Am-241 still peaks after 1000 years due to the fact that its production by Pu-241

beta decay is faster than its radioactive decay. However, the overall HM heat generation is

relatively insensitive to the Am-241 peak due to the much higher concentration of Pu-238 and

Cm-244 in the AHFTR fuel cycle than in SNF.

1E+06

S1E+05
a-
1E+04 -*-Np-237
S-e-Pu-238
6 1E+03 Pu-239
S--Am-241
Am-243
1 IE+02
-1E+02 Cm-244
-Cm-245
2 1E+01 --total





1E-02
1E+00 1E+01 1E+02 1E+03 1E+04 1E+05 1E+06
Time After Irradiation (Years)
Figure 6-18. Decay heat plot for AHFTR for a scenario where the reactor is shutdown at EOEC,
defueled and all fuel assemblies sent to the repository1

Because the AHFTR heat plateaus (due to the peak heat of Pu-238) earlier in the repository

life than SNF, it can be argued that the waste stream produced by the AHFTR fuel cycle will be

easier to monitor and manage in the repository than SNF. This is because the heat plateau occurs

during the practical time that the repository can be operated and monitored by present day

institutions. Very few human organizations, except for a select few religions, have enjoyed the

longevity that would be needed to operate the repository for thousands of years. Therefore, if a


1 Note the magnitude of the decay heat in Figure 6-18 is at least two orders of magnitude greater
than Figure 1-4. To some extent, this is anticipated due to the one order of magnitude higher
concentration of TRU in fast reactor fuel than in SNF. It is also important to note that
reprocessing losses can usually be assumed to less than one percent meaning that the amount of
waste generated per energy extracted from the fuel is one to two orders less for a closed fuel
cycle than an open fuel cycle [13].


236









sudden deviation from the predicted performance of a SNF repository occurs due to the Am-241

peak, it can not be expected that a human institution will exist that can respond to the safety

implications of this change. However, in the AHFTR scenario, the heat generation peak would

occur much sooner, in the first few hundred years. Many human institutions including the

existence of the United States have subsisted in this few hundred year time scale. Therefore, the

decay heat related performance of an AHFTR repository would be much easier to control using

human influences.


237









CHAPTER 7
ECONOMICS OF THE TWO-TIER AHFTR FUEL CYCLE

Given the higher rate of MA consumption by the AHFTR compared to the ABR, the

potential waste disposal attributes to the back-end of the fuel cycle are apparent. However, it is

necessary to evaluate the overall AHFTR fuel cost in order to realize any cost reductions that

MA waste burning can bring to a closed fuel cycle scenario. The cost attractiveness of fuel

recycling versus direct-disposal of SNF has created an ongoing public debate on the subject for

several decades. Two relatively recent independent reports, by the Massachusetts Institute of

Technology (MIT) and Harvard University, state that the direct-disposal fuel cycle is currently

more economically competitive from a fuel cost standpoint than plutonium reprocessing in a de-

regulated non-subsidized nuclear market [99,100].

Economic Issues of Reprocessing

The fundamental cost driver for SNF recycling is the reprocessing cost of separating the

fissile (-1%) plutonium (or transuranic) component of the SNF from the non-fissile uranium

(-95%) and fission products (-4%). Because of the low transuranic concentration in SNF, a

large HM throughput of SNF is required to remove all of the unwanted uranium and fission

products. Reprocessing costs are further complicated by plutonium proliferation and safeguards

issues. Criticality safety and radiological protection adds additional costs to recycling and

fabricating plutonium or transuranic bearing fuel. The combined effect is a fuel cost of a

transuranic fuel cycle that is more expensive than the combined cost of mining, milling,

conversion, enrichment and fabrication of UOX-LWR fuel.

For a fast reactor, the amount of energy extracted per mass of fuel (i.e., burnup) is at least

twice as high as that attainable by irradiating MOX fuels in LWRs. This allows more revenue in

electricity sales to be created per unit of fuel cost. However, fast reactors inherently require a


238









very robust and tolerant fuel and reactor design in order to achieve these high burnups and still

operate safely. Therefore, commercial demonstration fast reactors have typically been built at a

capitol cost 10-50% higher than LWRs [99]. Also, because the spent fast reactor fuel (SFF) in

many SFR fuel cycles (including this analysis) is reprocessed at the reactor site without allowing

for fission product decay, a hot-cell is required for all fuel handling operations. This remote

handling infrastructure requirement imposes a higher cost of fuel recycling than the glove-box

type of infrastructure used for SNF plutonium recycling (i.e., MOX-LWR fuels).

Transuranic reprocessing also incurs a disposal fee for the HLW generated by the

separation process. The HLW is mostly comprised of radioactive fission products that are

removed during the separation process. In addition, a variety of low level wastes (LLW) and

intermediate level wastes (ILW) are created by process losses and equipment contamination as a

normal part of reprocessing and fuel fabrication plant operation These waste materials can take

on many forms from the mixed-waste (chemical plus radiological hazard) nitric acid solutions

produced by aqueous processes that qualify as HLW to contaminated gloves and clothing that

only qualify as LLW. These wastes comprise a non-trivial component to the front-end of the

closed fuel cycle. For this analysis, a standard "base-case" HLW disposal fee is adopted for both

aqueous reprocessing and pyroprocessing. Base-case values are also adopted for the unit costs of

the following fuel cycle services.

* Price of Uranium ore, Conversion, Enrichment and UOX fuel fabrication
* NWPA SNF disposal levee
* Onsite SNF interim cask storage fee
* Cost of aqueous reprocessing, Cost of pyroprocessing


1 It should be noted that ILW is not defined by United States laws and regulations. However,
this term is adopted in the international community and is used in this dissertation as a
nomenclature placeholder for reprocessing related waste streams such as chopped HT-9 cladding
and/or the zirconium constituent of Pu/U/Zr fuels. This type of waste could be considered as
"greater than class C LLW" in the United States as will be discussed in a later section.


239









* HLW disposal (including ILW and LLW in this cost)
* Fast reactor fuel fabrication hot handling (hot-cell) or cold handling (glove-box)

A series of sensitivity studies will be applied to these base case unit costs to draw attention

to the dominating fuel costs of the ABR and AHFTR fuel cycles.

Capitol Costs

The MIT report sites most Generation-III or III+ LWR plant overnight costs to be in the

range of $1,400/kWe to $2,000 per 1 kWe of installed capacity. These values reflect the

differences in plant infrastructure footprint, reactor vessel and component modularity, as well as

a construction learning curve between the first-of-a-kind (FOAK) and the Nth-of-a-kind (NOAK)

deployment of a particular plant design. The Harvard report mentions that the overnight costs of

"proposed" commercial SFRs have been traditionally quoted to be 10% to 50% higher than for

LWRs. This estimate is also summarized by Kochetov et al [101]. These higher capital costs are

generally quoted as NOAK costs. However, many of these "proposed" commercial designs,

such as the S-PRISM, are rooted in previous reactor experience with FOAK reactors. It is not an

overgeneralization to say that all SFR plants to date were built without a standardized design or

an established SFR regulatory framework. In addition, SFR power reactors have always had a

dual role as a design concept demonstration and/or a fast flux fuel test facility. Because of the

FOAK nature of past SFR power systems, a fair degree of over-engineering was introduced into

their designs. This over-conservatism is necessary for a FOAK reactor plant to compensate for

the lack of technological maturity. For this dissertation, no assumptions are made with regards

to the learning curve or degree of technological maturity between, FOAK and NOAK of a

commercial scale SFR fleet. However, it is assumed that most of the technological hurdles for

an "economical" commercial SFR can be overcome by evolutionary design innovation and not

limited by the fundamental physics of the SFR concept in general.


240









As an example, Konomura et al conducted a feasibility study of a conceptual commercial

scale Japanese Atomic Energy Agency SFR (JSFR) to determine where design simplifications

are necessary for reducing the construction cost [102]. Konomura et al identified the long piping

of the secondary sodium loop as one of the largest construction expenses of the JSFR design.

The secondary sodium loop transfers heat from the intermediate heat exchanger (IHX) in the

sodium pool (or primary loop if a loop type system is considered) to the steam generator. The

IHX and secondary loop sodium does not circulate through the reactor core which essentially

eliminates neutron activated sodium from leaving the reactor vessel. This feature is a necessity

in case of a leak between the reactor vessel and the steam generator that would ultimately lead to

a sodium fire.

Historically, austenitic steels have been used for these pipes for their strength at elevated

temperatures, sufficient ductility and compatibility with sodium. Ductility is a requirement for

managing thermal stresses that arise during steady-state or transient operation. However,

austenitic steels have a relatively high thermal expansion coefficient compared to other steels.

The thermal expansion necessitates long piping designs with a number of elbows. On the other

hand, high chromium content ferritic steels typically have a smaller thermal expansion than

austenitic steels. High chromium steels were not used in the past due to their poor ductility.

However, Konomura et al sites recent developments with tungsten and molybdenum alloying in

12-Cr steels in the 1990's that may give high chromium steels the advantage in ductility required

to allow their use in the secondary loop plumbing. The HT-9 cladding developed in the late

1980's to early 1990's for SFR fuel pin cladding (and in-core structures) is also a 12-Cr

ferittic/martensitic steel.









Konomura et al proposed that the shortening of piping in addition to: loop number

reduction, a compact reactor vessel and structure and integration of certain components could

significantly reduce the construction cost. These goals are also shared by the compact

modularity and shared steam loop philosophy of the S-PRISM power block approach.

Konomura et al used the Nuclear Utility Service (NUS) code of account, developed in 1969 for

LWRs, in order to evaluate the JSFR construction cost. Using NUS, the total plant construction

cost was evaluated as a sum of all facility and equipment unit costs multiplied by the mass or

quantity of each component. Using this model and appropriately discounting the construction

cost for each component, Konomura et al found that the overall JSFR cost would be in the range

of 200,000 yen/kWe or equivalently about $1,750/kWe in present United States dollars. This

value is within the range of Generation-Ill options listed by the MIT report.

Due to the apparent similarities between the AHFTR power plant design and other

proposed commercial SFR concepts (i.e., ABR, S-PRISM, JSFR, etc.), a rigorous component

cost analysis is not necessary. However, if a base case overnight construction cost of

$1,800/kWe is assumed with a discount rate of 6% and a five year construction time, then the

capital construction cost, at the time the plant goes on line, is $2,140/kWe. After accounting for

interest on the initial investment as well as taxes, insurance, etc., the capital cost contribution to

the price of electricity is about 4.0 /kWxhr(e). The economic model used to calculate the

capital cost contribution to the price of electricity is discussed in a later section. A ballpark

estimate of $1,800/kW(e) is assumed for the AHFTR overnight capital cost. This assumes zero

FOAK costs for a standardized SFR design with reactors built in succession at a preapproved

site. Also, for the purpose of this analysis, the operations and maintenance cost contribution to

the price of electricity will be assumed to be the same as for a LWR.


242









Fuel Costs

The Harvard report makes special mention of the fact that civilian plutonium separation

practices to date have not kept in line with purchase of MOX fuel assemblies, resulting in a 200

metric ton commercial plutonium world stockpile. Therefore, the Harvard report stipulates that

this stockpile could be used to constitute the "first-core" required to start the SFR and launch its

closed fuel cycle. Hence, as was done in the Harvard report, the assumption is made for the

AHFTR that this first-core at each new SFR (ABR or AHFTR) has zero fuel cost. The general

transuranic loading of both the AHFTR and the ABR is approximately three metric tons (Table

3-12). Therefore, approximately 65 reactors can be started using the commercial stockpile. This

is the equivalent of six reactor sites often SFRs per site (using the power block model) making

3,330 MWe per site. However, if more reactors are required, the commodity price of stockpiled

commercial plutonium would be driven up by an increase in demand. Thus, it may be prudent to

consider the approximate 600 metric tons of stockpiled weapons grade plutonium existing in the

world, to supplement the commercial stockpile. If the weapons plutonium stockpile is

considered, then 27 SFR power stations (using the power block model) may be constructed.

First-Core Reprocessing Cost

If the first-core plutonium must be separated from SNF, a significant capital cost is

incurred of approximately $200/kWe to $300/kWe which corresponds to a reprocessing cost of

approximately $20,000/kgiHM to $30,000/kgiHM, respectively, of separated plutonium. This

estimate is based on a cost of aqueous reprocessing of SNF to be approximately $1,000 to $1,500

per kilogram of SNF and only 1% of this SNF being plutonium. Therefore, the cost to bring the

plutonium concentration from l%/kg in SNF to the 20%/kgiHM of TRU enrichment requirement

for the SFR driver fuel is:


243









$1,000 1 kgSNF 0.20 kgPu
x x = $20,000/kgiHM
kgSNF 0.01 kgPu 1 kgiHM
or: (7-1)
$1,500 1 kgSNF 0.20 kgPu
x x = $30,00/kgiHM
kgSNF 0.01 kgPu 1 kgiHM

Assuming, a $2000/kWe overnight construction rate, the added cost of the first-core

plutonium is equivalent to a 10% to 15% increase in the overnight capital cost of the reactor

plant. Because, this "first-core" reprocessing cost is attached to the reactor's total cost before

actual operation begins, it can be considered as a capital or startup cost. This large pre-

operational reprocessing cost is analogous to the cost of supplying the initial heavy water to a

pressurized heavy water reactor (PHWR). The operational cost of resupplying heavy water to a

PHWR is minimal. However, the pre-operational cost to supply the initial heavy water for a

PHWR's plumbing can be as high as 10% of a PHWR's overnight cost [103].

Nth Core Reprocessing Cost

For all subsequent refuelings, a much smaller amount of SNF HM must be processed

externally in order to retrieve the transuranic material needed to refuel the ABR and AHFTR.

This is because much of the transuranic material needed for all Nth refuelings has already been

pre-concentrated in the form of the fast reactor fuel, or more precisely the SFF. Hence, the

throughput of HM required from SNF to reconstitute the TRU enrichment of fresh fuel is much

less for a closed SFR fuel cycle than if the SFF were not recycled. The smaller HM throughput

requirement stems from the fact that SFF needs less uranium to be subtracted in order to bring

the spent fuel transuranic concentration back to that needed for fresh fuel. It is important to note,

that as the SFR's CR approaches zero, the concentration of TRU in the SFF relative to the initial

TRU enrichment decreases, thus increasing the reprocessing cost of SNF. In the ABR fuel cycle,

SNF is reprocessed at a centrally located aqueous reprocessing facility (presumably UREX+)


244









(Figure 1-1). Therefore, the HM mass throughput, and hence reprocessing cost per mass of

product, of this aqueous facility is strongly dependent on the conversion ratio selected for the

ABR (and/or AHFTR) design. The mass throughput requirements of the electrorefinery are

similarly affected by the concentration of TRU in the SFF. However, because of the high TRU

concentration per HM in SFF compared to SNF, the electrorefinery requires less SFF HM

throughput "capacity" per unit mass of separated TRU produced.

The cost of fuel reprocessing is a function of the rate of return on investments made in its

design and construction, the day-to-day operations and maintenance costs and the mass

throughput of the SFR.

$/kgHM = (Design x p)+ (Construction x Yp)+ (O & M) (7-2)
Capacity

Where: "Design" and "Construction" are the design and construction overnight costs in

millions of dollars (M$), respectively. 4 is the carrying charge on the initial investment per year

(yr1). The carrying charge reflects the annual payment of the reprocessing facility capital cost

plus interest, insurance and taxes. O&M is the operations and maintenance cost in millions of

dollars (M$/yr). Capacity is the annual HM throughput capability of the reprocessing facility

(kg/yr).

Unfortunately, due to the technical requirements of remote fuel handling, the

electrorefinery construction cost will be capital intensive. In order to have a pyroprocessing cost

competitive with aqueous reprocessing, the design and construction costs must be reduced. To

compensate the higher construction costs in Equation 7-2, the electrorefinery should have a

modular design which could reduce the design cost of the facility to nearly zero. Because of the

construction cost, the HM throughput capacity must be maximized to a feasible limit. This fact


245









requires that the electrorefinery service multiple SFRs as is proposed in the ten reactor power

block model.

As stated before, reducing the SFR's CR will increase the HM throughput capacity.

However, the reprocessing requirements of the SFR will also increase which would necessitate a

larger electrorefinery be constructed in order to meet this demand. Therefore, the benefit of

decreasing the CR to decrease the reprocessing service cost would be washed out by the cost to

construct the reprocessing plants needed to keep up with fuel demand.

Operations and maintenance of the reprocessing facility is an unavoidable expense which

is necessary for safe handling and protection of separated plutonium and transuranic material.

The French La Hague plant employs 6,000 to 8,000 highly trained individuals that oversee

operations and maintenance of equipment and processes [99]. A report by Smith et al and

another by Kim, Kazimi et al details the expected fuel reprocessing plant costs for an Accelerator

Transmutation of Waste (ATW) system and an ABR respectively [104,105]. Both reports

indicate that the O&M cost is expected to be between 23% and 33% of the total reprocessing

cost per HM throughput.

For the base case of this analysis, the reprocessing cost was allowed to vary from

$1,500/kgHM for the minimum aqueous processing cost to $3,000/kgHM for the maximum

electrorefinery cost. The $1,500/kgHM is a common estimate of large scale commercial aqueous

reprocessing costs for the United Kingdom's Thermal Oxide Reprocessing Plant (THORP) and

France's UP2 and UP3 facilities in La Hague [99]. This number reflects the cost of reprocessing

under government enforced pay-ahead contracts that required utilities to essentially pay off the

entire cost of these plants over a ten year base load period with no required return to investors.

The Harvard report estimated the reprocessing cost to be $1,350/kgHM for a government owned


246









facility capable of borrowing money at low risk-free government rates and amortization of

capital over a 30 year plant lifetime. The Harvard report also calculated the reprocessing cost to

be greater than $2,000/kgHM for a privately owned facility with a government guaranteed rate of

return. If no rate of return is guaranteed, the reprocessing cost may be as high as $3,000/kgHM.

Since, it is considered likely that the electrorefinery will be privately owned and operated in a

consortium with the reactor utility; this value is assumed as the maximum cost for

pyroprocessing. The $1,500/kgHM number (2003 dollars) was used for the fuel cycle

calculations in both the MIT and Harvard studies. Despite the ATW reports that give a much

more optimistic estimate, the MIT and Harvard value is considered for aqueous reprocessing for

the purpose of this study. This value also closely matches a 1994 report by the French Nuclear

Energy Agency of $720 Euro/kgHM (1994 currency). Assuming a 3% rate of inflation and the

present day conversion rate, this corresponds to approximately $1,515/kgHM in 2007 US dollars.

High Level Waste Disposal Cost

As a byproduct of aqueous reprocessing HLW, ILW and LLW are produced. Low level

waste generally consists of gloves, protective clothing and tools that have been contaminated

with low levels of short lived radioisotopes. The definition of ILW includes resins, chemical

sludge and metal reactor fuel cladding, as well as contaminated materials from equipment

decommissioning [106]. It should be noted that ILW is not defined by United States laws and

regulations. However, this term is adopted in the international community and is used in this

dissertation as a nomenclature placeholder for reprocessing related waste streams such as

chopped HT-9 cladding and/or the zirconium constituent of Pu/U/Zr fuels. If ILW contains

transuranics, it may be considered transuranic waste which is a HLW that requires repository

storage. It is common in Europe to cement ILW inside metal containers that are destined for

geologic disposal. If this is the case, this type of waste could be considered as "greater than class


247









C LLW" in the United States [107]. However, as a general rule, non-transuranic containing ILW

such as fuel cladding hulls can be disposed of at surface burial sites with LLW. HLW consists of

the separated fission products as well as pragmatic actinide losses that can not be fully recovered

from the acid or molten salt solutions. The NEA study assumed that for aqueous reprocessing,

these wastes would be vitrified into a glass waste form a few years after their creation and then

held at the reprocessing site for 50 years.

Disposal practices for ILW and HLW have been demonstrated for pyroprocessing by the

treatment of EBR-II spent fuel. For these treatment operations, EBR-II SFF is pyroprocessed to

separate the uranium from the NaCl-LiCl eutectic solution. The transuranics and fission

products are separated from this solution and then mixed with zeolite to form a ceramic waste

form. The noble metals and cladding hulls from this process are added to zirconium and then

melted down into a metal waste form.

The MIT and Harvard reports both assessed a fee of $300/kgHM for disposing of these

HLW, ILW and LLW for aqueous reprocessing. In this dissertation, this value is adopted for the

electrorefinery as well.

Fuel Fabrication Cost

Similar to reprocessing, transuranic fuel fabrication also requires large capital-intensive

facilities with a large workforce of highly skilled and trained personnel. The Harvard report

suggested that the fuel fabrication cost of such a facility would be in the range of $1,010/kgHM

for a government operated facility; in the range of $1,460/kgHM for a privately owned facility

with a guaranteed rate of return; and approximately $2,140/kgHM for a privately owned facility

with no guaranteed rate of return. The Harvard report also mentions that a cost of $120/kgHM

should be assessed for transporting transuranic fuel from the fuel fabrication facility to the


248









reactor site. The MIT report assumed a value of $1,500/kgHM for the cost of fabricating Pu-

only MOX-LWR fuel.

Similar to the ABR and AHFTR reactor plants, the electrorefinery is expected to be

privately owned with no guaranteed rate of return. Because, fuel fabrication is carried out in the

same circuit as pyroprocessing, most of the transportation costs can be avoided at the ABR or

AHFTR electrorefinery. However, some transportation cost is incurred in order to bring the

make-up transuranic materials from the aqueous plant to the electrorefinery. Because of the

additional shielding requirements (hot-cell environment) brought on by high gamma and neutron

radiation fields, and also the additional complexity of target fabrication, a conservative fuel

fabrication estimate of $2,500/kgHM is for the reprocessing fee of the electrorefinery. This

estimate is used for both the ABR and the AHFTR.

Due to the use of Lightly Enriched Uranium (LEU), the fuel fabrication safety

requirements for LEU-UOX fueled LWRs are less than for transuranic fueled reactors. A

fabrication cost of $250/kgHM is the current average market price of LEU UOX-LWR fuel

fabrication. This is the value used for the ABR and AHFTR fuel cycle study. The MIT report

assumed a value of $275/kgHM. The Harvard report assumed a value of $250/kgHM.

Front-End Uranium Costs

The price of uranium ore has historically given a small contribution to fuel costs. In fact

the primary reason why the LWR has become the work horse of the world's nuclear electricity

industry is because uranium was found to be in much greater supply then originally expected

during the early days of the industry. For the past 20 years, the price of U308 has stayed below

$20/lb U308 ($40/kgU in the form of U308). However, this price has increased to as high as

$140/lb U308 between 2006 and 2007. The issue of the abundance of the world's "minable"

uranium supplies, and hence price, is the center of much industry debate in recent years. Though


249









currently the price of uranium ore is in a state of flux, it is assumed for the purpose of this

economic analysis that future prospecting will discover additional reserves. Therefore, the pre-

2006 price of $20/lb U308 is used for the ABR and AHFTR fuel cycle analysis.

Conversion of the U308 ore into UF6 gas has been a relatively minor cost to the front-end

of the fuel cycle. However, the price of conversion has also been in a relative state of flux over

the past 10 years, varying from $2/kgU in 2001 to $11/kgU in 2006. For the following fuel cycle

analysis, a value of $11/kgU (in the form of UF6) is assumed for the cost of conversion.

Uranium enrichment is the process of increasing the isotopic concentration of the fissile U-

235 isotope in uranium above what is found in nature. The enrichment process produces a

product stream, with a higher U-235 concentration, and a depleted uranium "tails" stream, with a

lower U-235 concentration than that given by its natural abundance. The enrichment separative

work unit (SWU) is defined as a metric of the cost of energy required by the enrichment process

to achieve a desired U-235 concentration (i.e., enrichment). The more SWU that is applied to the

natural uranium feedstock, the higher the U-235 enrichment in the product. Also, the more SWU

applied leaves less U-235 concentration in the depleted uranium tails.

The price of the SWU has steadily risen since 2001 from a value of $105/SWU to a value

of $145/SWU in 2006. The price increase of enrichment is driven by the price increase in

uranium ore. As the price of uranium ore increases, fuel purchasers have specified a smaller U-

235 concentration in the depleted uranium. The less amount ofU-235 wasted in depleted

uranium translates into more U-235 being extracted from the original uranium ore purchase.

Therefore, the fuel purchaser does not need to buy as much uranium to arrive at the specified

enrichment if more SWUs are purchased. This strategy is known as underfeedingg". Because of

rising uranium ore prices, underfeeding has been performed by the United States Enrichment


250









Corporation (USEC) since 2003 [108]. The increase in SWU demand has caused the price of

enrichment to increase. For the following ABR and AHFTR fuel cycle analysis, an enrichment

cost of $145/SWU is assumed.

Back-End Uranium Costs

The NWPA was established to create a comprehensive national program for the safe long-

term disposal of highly radioactive wastes. The NWPA directed the Department of Energy

(DOE) to study suitable sites for a geologic repository for this long-term storage. The repository

envisioned by the NWPA is an engineered disposal facility located deep underground that can

store 70,000 metric tons of SNF and HLW. In 2002, Congress and the President approved the

development of a geologic repository at Yucca Mountain, Nevada. The Yucca Mountain

Environmental Impact Statement (YM-EIS) designated that the Yucca Mountain repository

would receive 63,000 MTHM (metric ton heavy metal) of commercially generated SNF, 2,333

MTHM of DOE generated SNF and 4,667 MTHM of DOE generated HLW.

To finance the disposal of commercially generated SNF in the repository, the NWPA

created the Nuclear Waste Fund (NWF) which imposes a flat fee of 1 mil/kWxhr(e) (1 mil is

equal to ten percent of one cent) of electricity produced by nuclear fuel. This back-end fuel cost

has been imposed on all nuclear electricity sold in the United States after the NWPA was passed

in 1982.

Under the provisions of the NWPA, the DOE was required to accept SNF from the

commercial industry for geologic disposal no later then January 31, 1998. This has not occurred

due to logistical, legal and legislative delays. The delays have been driven by the Yucca

Mountain site evaluation requiring a comprehensive understanding of its long term geologic

behavior. Because of these delays and the fact that wet storage space provided by the spent fuel

pool for most reactor plants is a finite premium, most nuclear utilities have been forced to









acquire interim storage capability in the form of dry storage casks. The cost of establishing this

interim storage capacity has been estimated by the MIT and Harvard reports to be in the range of

$100/kgHM and is therefore used for the economics analysis of the ABR and AHFTR fuel cycle.

This cost is determined by the cost to establish the storage facility (usually consisting of a

concrete pad, fences, surveillance equipment, etc.) and purchase of the casks themselves. Since

very little needs to be done with the casks once they are loaded on the pad, the operations,

maintenance and surveillance costs are usually considered negligible and typically lumped in

with the O&M cost of the power plant. It should be noted that if dry cask storage is needed for

SNF at a reactor site that has been decommissioned, the cost of pad space for all of the reactor's

legacy SNF can be much higher and in the range of $300/kgHM because these costs would no

longer be able to be rolled into the O&M costs of the reactor plant.

The assumption is made that it is probable that initial reprocessing capacity will not be able

to meet the rate in which SNF is generated by LWRs. Therefore, it is foreseeable that interim

storage will be necessary as a buffer between the rate of SNF generation and the rate at which it

can be accepted by the reprocessing company. It is also assumed that these back-end costs will

be credited to the reprocessing company as a fee for taking ownership of the SNF. The incentive

for the LWR utility to pay the reprocessing company for SNF removal ($100/kgHM), as opposed

to waiting for repository disposal, comes from the avoidance of the $300/kgHM dry cask storage

fee in the event fuel is not removed before the plant is decommissioned.

Discounting and Financing

The cost of commercially generated electricity seen by customer ratepayers is broken down

into four primary components: cost of reactor design, construction capital, fuel purchase, and

operations and maintenance.


252










Celec = cde + Ccap + Cfuel + CO&M (7-3)

For the ABR and AHFTR economics analysis, only the NOAK scenario is evaluated.

Hence, the cost of the reactor design is appreciably reduced and assumed to be zero.

Operations and Maintenance

For the sake of this analysis, the operations and maintenance cost per electricity generated

is considered the same for SFRs and LWRs. Also the cost of decommissioning is rolled into this

amount.

Corn + CddFdd
cOaM C += (7-4)
8766 x P x / x

Where: Com is the average operations and maintenance costs associated with payroll,

equipment maintenance and update, etc. ($). Cdd is the cost to decommission, dismantle and

remediate nuclear materials from the reactor site at the end of its operational life ($). Fdd is the

amount of money paid each year by the utility into an annuity account for decommissioning

activities. P is the reactor thermal power rating (kW). q is the reactor plants thermal-to-electric

efficiency. s is the availability of actual total heat energy produced in a year divided by the

theoretical amount of heat energy produced if the reactor ran at full power throughout the same

year. 8766 is the total number of hours in a year. The annuity factor (Fdd) is defined by the

future value of a series of uniform payments into an account.


Fdd dd (7-5)
S(1 + idd )n 1

Where: idd is the rate of return on the decommissioning annuity fund. n is the number of

years allotted for making payments into the annuity. For the ABR and AHFTR analysis the

annuity is assumed to be paid over the five year construction time of the plant and idd is assumed

to be 5.83 %/yr, which gives a value of 17.8% for Fdd.


253









Therefore, the Com cost of electricity is $0.0122/kW-hr(e), if Com/(Pxrxs) and

Cdd/(Pxqrxs) are $80/kW(e) and $150/kW(e), respectively.

Construction Capital

Like the operations and maintenance model, the capital construction cost financing model

of the ABR and/or AHFTR is borrowed from the Harvard report.

Ccp (I + ldc )( + Fpreop Xi + F7,,, )
C cap dc preop nt (Fc +F +E ) (7-6)
8766 xPx x q x

Where: cap is the costs paid annually for the construction of the reactor ($/kWxhr(e)).

This is essentially the construction capital contribution to the cost of electricity. Ccap is the total

overnight construction cost ($). Fide and Fpreop are factors that account for interest during

construction and other pre-operational costs. Foont is a contingency factor to provide for cost

overruns and other unforeseen costs. For is the "fixed charge rate" which is the fraction of the

total investment which must be repaid each year including interest rate and return on investment

(yr-'). Ftax and Fins are the annual charges on property tax and insurance (yr-'), respectively. For

simplicity, it is assumed that Fpreop and Fcont both equal 10%. Also, it assumed that the fixed

charge rate, tax and insurance rates total to 10 %/yr.

The interest rate factor during construction (Fidc) accounts for the interest charges collected

during construction. The MIT and Harvard reports employ a special curve fit such as sinusoidal

or binomial to generalize the amount of money borrowed for each year of construction. The

amount of money borrowed at each of these years would then be discounted to the year the

reactor goes online. For the ABR and AHFTR analysis, a flat distribution is assumed where the

amount borrowed is the same for each year of construction before being discounted.

1
F;dC =-L+ i:1 + id)" (7-7)
k=l


254










Where: n is the total number of years required for construction, k is the integer year of

construction starting from ground breaking. idc is the interest rate on moneys borrowed during

construction (yr-1). For the ABR and AHFTR analysis, the construction time is assumed to be

five years and the idc is assumed to be 5.83 %/yr, which gives a value of 19% for Fide.

The fixed charge rate (For) is the fraction of the initial investment that is paid each year to

pay off the principal money borrowed with a return on investment. It is assumed that a fixed

interest rate is applied over the lifetime of the reactor. Hence, the fixed charge rate may be

determined by a simple "capital recovery formula".

Fcr u i(l+i)N (7-8)
For-E (7-8)
p (1+)N -1

Where: u is the yearly required payment. p is the present worth of the capital investment. i

is the discount rate for the return on the investment (yr-'). N is the reactor lifetime in which the

principal is paid off (yr). If a guaranteed discount rate of five percent and a reactor lifetime of 40

years are assumed, the fixed charge rate is 5.83 %/yr. If the sum of Fcr, Ftax and Fins is 10%, the

tax and insurance portion is 4.17 %/yr.

The discount rate i is treated in Equation 7-8 in a similar fashion to how interest is

borrowed from a bank. However, in a broader sense, it is the "effective cost of money" for the

investment made into the utility company by bonds and stockholders. This investment may be

broken down into the percent of stocks and bonds invested in the utility and the taxes payable on

the bonds.

i= f;i + (- r)fi, (7-9)

Where: fs is the fraction of the investment owned by stocks. fb is the fraction of the

investment owned by bonds (fb=l-fs). is is the stock rate of return (yr-'). ib is the bond rate of


255









return(yr1). c is the effective percentage of the investment paid to federal, state and local taxes.

If the tax rate is 45%, the stock rate of return is 10 %/yr, the bond rate of return is 8 %/yr and

90% of the investment is owned by stocks, the discount rate would be 5 %/yr.

If the SFR overnight capital construction cost is "Ccap/(Pxr) = $1,800/kW(e)" then the cap

capital cost of electricity from Equation 7-6 is $0.0427/kW(e).

Fuel Investment

A simplified discounting model, similar to that used by the MIT report, is used to address

the levelized cost of electricity corresponding to the purchase of fuel. Fuel costs are discounted

in a similar manner to how interest is accrued during the construction capital investment.

However, instead of reflecting just the discount rate, which is essentially another way of

expressing interest or the investor rate of return, a broader "carrying charge" definition is applied

for fuel costs. For this discounting model, the carrying charge is considered to be the total cost

of money, taxes and insurance on funds borrowed at the time a fuel service is purchased. This

carrying charge is used to discount the cost from the time the service is purchased to the

midpoint of the irradiation of the fuel in the reactor.

1000 1
Cfrel =1 x I x M,, M (7-10)
2Cfuel q (7-10)

Where: Mi is the mass of HM processed at stage i per megawatt-day of energy that that

mass produces in the core (kgHM/MWD). ci is the cost of the service to process the material at

stage i ($/kgHM). 4 is the carrying charge on the investment to purchase the service i, At is the

discounting time elapsed from the moment the service is purchased to the mid-point of the fuel

irradiation in the reactor (days). 1000/24 is the unit conversion required to convert megawatts to

kilowatts and days into hours. r is the reactor plants thermal-to-electric efficiency.


256









The carrying charge may be determined by summing the discount rate, tax rate and

insurance rate of the fuel. For simplicity, these values are assumed to total to 10 %/yr. This is

also the value used in the MIT report. This "lumped" carrying charge method is slightly

different from the way that fuel purchases have historically been discounted with the carrying

charge. In many fuel purchase arrangements, the fuel purchase including carrying charge is

amortized in step with fuel burnup until the fuel is fully amortized when it is discharged from the

reactor. However, for a general and straight forward comparison of fuel costs between SFRs and

LWRs, the "lumped" carrying charge analysis is used as was done by the MIT report.

Fuel Cycle Base Cases

The choice of reactor type (or types) and their relationship with recycling centers has been

the focus of much scientific debate since the late 1990's when it became apparent that a second

repository would be likely needed if nuclear electric generation was to expand without reducing

the volume of SNF. This debate has caused the DOE to change the naming acronym of its fuel

cycle research activity four times since the late 1990's when research restarted after the IFR

program was terminated (e.g., Accelerator Transmutation of Waste (ATW), Advanced

Accelerator Applications (AAA), Advanced Fuel Cycle Initiative (AFCI), Global Nuclear

Energy Partnership). The timeline of these research activities are not always in series to each

other. Currently, the AFCI and GNEP programs are a parallel effort.

Each of these programs was highly focused on a single reactor technology (i.e. Large

accelerator driven systems, water moderated reactors, fast reactors and graphite reactors) to solve

all of the problems of the once-through fuel cycle in a single tier. A tier is referred to as a

stepping off point in the fuel cycle where one reactor type accepts, as fuel, the nuclear waste

generated by another reactor type. Each of these reactor concepts, and the fuel technologies

developed for waste incineration, has demonstrated transmutation strengths and drawbacks for


257









each individual isotope within the SNF. As discussed throughout this dissertation, transmutation

is highly sensitive to the neutron energies and flux intensities in which the fuel is exposed. The

elemental partitioning technologies developed under the envelope of the UREX+ process has

been demonstrated as a viable technology for partitioning SNF into specialized waste streams for

individualized transmutation options and/or disposal options. This enables the feasibility for

heterogeneous arrangements within the fuel and/or reactor.

The following analyses will demonstrate how the above cost calculations are employed to

re-produce some of the fuel cycle scenarios and conclusions addressed by the MIT and Harvard

reports. Then the symbiotic two-tier fuel cycle, as described in Figure 1-8 and Figure 3-4, with a

heterogeneous fleet of ABRs and AHFTRs is analyzed.

The MOX Fuel Cycle

The MIT report focuses primarily on the cost competitiveness between the once-through

irradiation of LWR UOX versus recycling SNF to create LWR-MOX fuel. This pseudo-closed

MOX fuel cycle assumes that the UOX fuel is irradiated in an LWR and then recycled to create

MOX fuel for a second reactor-pass in a LWR and then disposed of as MOX-SNF. This

scenario, which is currently in practice in Europe, Russia and Japan, is not truly closed because it

only reprocesses the SNF once, which still creates spent fuel assemblies that require repository

storage. Such a fuel cycle, where fuel is only recycled once before requiring direct disposal, is

considered an open fuel cycle with a single-tier recycling scenario. In theory, because additional

energy is extracted from the fuel by the second reactor pass, such a fuel cycle reduces the mass

of SNF destined for repository storage per uranium mined. However, the front-end cost of

reprocessing and fabricating plutonium bearing fuels has historically made the MOX fuel cycle

more expensive than the UOX fuel cycle. Figure 7-1 shows the mass flow relationships of the

open one-tier MOX fuel cycle.


258









For the following fuel cycle calculations, the fresh UOX fuel is assumed to have a uranium

enrichment of 4.5 w/o and is irradiated to 50 MWD/kg during three 500 EFPD cycles in a

standard 17x17 PWR fuel assembly. For these calculations, the coupled NEWT-ORIGEN code,

TRITON, is used to perform the buildup/depletion calculations for LWR fuel. Also, the

TRITON code was used to decay the fuel in the time elapsed from UOX-LWR discharge to

MOX-LWR fueling.

Mining Conversion Enrichment Fabrication LWR
(U308) (UF6) (UOX) (UOX-FA) (UOX-SNF)

0.231 kgU/MWD 0.230kgU/MWD 0.143 kgSWU/MWD 0.020 kgU/MWD


0.156kgHM/MWD = s 2 ii l' kgU/MWD
"collected for MOX" disc luigcd as UOX-SNF"


Reprocessing Fabrication LWR Repository
(MOX) (MOX-) (MOX (MOX-SNF)
0.020 kgHM/MWD 0.020 kgiHM/MWD 0.019 kgHM/MWD
"discharged as MOX-SNF"
HLW, ILW, LLW
.................................. .................................
Repository and
Surface Disposal

Figure 7-1. The open single-tier MOX fuel cycle

For this open single-tier fuel cycle, a theoretical Np+Pu-MOX is considered. Here, the

Am+Cm+Bk+Cf stream is discarded as transuranic HLW and destined for the repository. This

Np+Pu-MOX has a transuranic enrichment of 10% TRU per HM. The total fissile enrichment,

including U-235, Pu-239 and Pu-241 is 6.8 w/o. Using the TRITON code, it was found that this

composition could be irradiated for three 500 EFPD cycles to a burnup of 50 MWD/kg in a

standard 17x17 PWR fuel assembly.


259









It is important to note the relatively large amount of UOX-SNF needed in order to produce

a small amount of MOX. Performing a simple mass balance calculation shows that 8.2 times

more SNF HM is needed at the MOX fuel cycle front-end than is produced per megawatt-day at

the back-end of the UOX fuel cycle (Figure 7-1). Equivalently, roughly 8.2 UOX-SNF

assemblies need to be reprocessed in order to create one Np+Pu-MOX assembly. The

significance of this large front-end throughput becomes evident in the Np+Pu-MOX cost

analysis. For the Np+Pu cost analysis, the following assumptions are made.

* Fresh UOX fuel enrichment: 4.5 w/o
* Uranium tails depletion: 0.3
* Process Losses: 1%
* LWR Fuel Burnup: 50 MWD/kg
* LWR Thermal Efficiency: 33%
* Reprocessed UOX-SNF uranium is used for the uranium in the Np+Pu-MOX

The unit costs for fuel cycle services and fees are listed in Table 7-1. Detailed descriptions

of the background and assumptions of these unit costs are discussed earlier in this Chapter.

Using these assumptions and unit costs, the following fuel costs are calculated for the zero-tier or

UOX-LWR component of the fuel cycle. These costs are tabulated in Table 7-2.

Table 7-1. Unit costs and fees for the open single-tier MOX fuel cycle
Item Amount Unit
Carrying Charge Rate: 4 10 % -
Mined U $20.00 $/lb U308
Mined U $37.40 $/kgU as U308
Conversion $11.00 $/kgU as UF6
Separative Work $145.00 $/kg-SWU
Aqueous Reprocessing* $1,000.00 $/kgHM Feed
HLW Disposal $300.00 $/kgHM Feed
Fabrication UOX $250.00 $/kgU Product
Fabrication MOX** $1,460.00 $/kgHM Product
NWPA Levee $0.001 $/kw-hr(e)
Interim Cask Storage $100.00 $/kgHM Waste
*Note: For the Np+Pu-MOX it is assumed that no extra processing steps are required to extract
a pure Am+Cm+Bk+Cf stream from fission product waste streams (e.g., UREX+2). **Note:
Because Am+Cm+Bk+Cf is not involved in fuel fabrication, the requirement for a hot-cell
facility is excluded and the cost of MOX fabrication is assumed to be performed in a glove-box
operation at a cost of $1,460/kgHM instead of $2,500/kgHM].


260









Table 7-2. Fuel costs for the UOX-LWR tier, equivalently the once through fuel cycle
$/MW Sticker Price Borrowing Time Carrying Charge
($/kgU) (Yr) ($/kgU)
Mining & Milling $7.18 $360.55 4.25 $153.23
Conversion $2.10 $105.53 4.25 $44.85
Enrichment $16.35 $821.60 3.25 $267.02
UOX Fabrication $4.98 $250.00 2.75 $68.75
NWPA Levee $8.16 $410.04 -2.25 -$92.26
Interim Cask Storage $1.89 $94.76 -2.25 -$21.32

Combining the sticker price and carrying charges of the fuel purchase gives a total fuel

cost of $2,463/kgU. Dividing by the fuel burnup of 50 MWD/kg gives a fuel cost of electricity

of $0.0062/kW-hr(e). Repeating this calculation for the Np+Pu MOX-LWR tier gives the

following fuel cycle costs shown in Table 7-3.

Table 7-3. Fuel costs for the MOX-LWR tier of the open single-tier MOX fuel cycle
$/MWSticker Price Borrowing Time Carrying Charge
MW ($/kgiHM) (Yr) ($/kgiHM)
Credit for UOX-SNF -$78.50 -$3,944.84 4.25 -$1,676.56
UREX+2 $155.52 $7,814.71 4.25 $3,321.25
HLW Disposal $46.66 $2,344.41 3.25 $761.93
MOX Fabrication $29.35 $1,474.60 3.25 $479.25
NWPA Levee $8.16 $410.04 -2.25 -$92.26
Interim Cask Storage $0.00 $0.00 -2.25 $0.00

Combining the sticker price and carrying charge of the fuel purchase gives a total fuel cost

of $10,892/kgiHM. Dividing by the fuel burnup of 50 MWD/kgiHM gives a fuel cost of

electricity of $0.027/kW-hr(e). Therefore, the fuel cost of producing Np+Pu-MOX electricity is

more than four times the cost of UOX electricity. The much higher Np+Pu-MOX cost is

primarily due to the large sticker price ($155.52/MWD $7,814.71/kgiHM 50 MWD/kgiHM) of

reprocessing, which is a result of the reprocessing plants large HM throughput (0.156

kgiHM/MWD) and the price of reprocessing ($1000/kg).

Notice the influence of the UOX-SNF credit. This is the fee that the UOX utility pays to

the aqueous reprocessing company for assuming responsibility for the UOX-SNF. It is

equivalent to the cost of interim storage at the aqueous reprocessing plant plus the NWPA levee









imposed on the UOX-SNF. Because, the aqueous reprocessing company assumes responsibility,

it has the option of charging a removal fee equal to what the DOE would have charged to store

the SNF in a repository. Also because the aqueous reprocessing company is providing the

interim storage service as part of its regular operations, the reprocessing plant stores the dry

storage casks and passes this cost along to the UOX-LWR utility. These credits are a significant

source of income to the aqueous reprocessing company and alleviate some of the reprocessing

cost. However, these credits are not sufficient to offset the total cost of reprocessing and the fuel

fabrication cost. Note that the cost of MOX fabrication is more than five times as high as UOX

fabrication.

It is an interesting finding of the MIT report, that because of the UOX-SNF credit, if the

NWPA levee would have been set at approximately 3.0 mil/kW-hr(e), the fuel cost of MOX

would be less than for UOX. In the calculations in Table 7-3, the 3.0 mil/kW-hr(e) would

increase the UOX fuel cost to $0.0078/kW-hr(e) and reduce the Np+Pu-MOX fuel cost to

$0.0060/kW-hr(e).

The ABR Fuel Cycle

The Harvard report offers a comparison of the once-through LWR fuel cycle with a SFR

closed fuel cycle. Because the Harvard report considers the recycling of SFF, the system

analyzed by the report is considered a truly closed fuel cycle. In such a fuel cycle, all fuel is

recovered by reprocessing and only the HLWs due to fission product and/or MA separation

requires a long term disposal solution. The Harvard report only considered SFRs with a

conversion ratio equal to one or higher. Therefore, the only reprocessing cost considered was

that for SFF. However for the ABR, two types of reprocessing plants are required: an aqueous

facility for SNF and a pyroprocess facility for SFF. Therefore, the aqueous reprocessing cost

associated with a continuous external supply of SNF TRU to the ABR must be considered.


262









Because ABRs, in this scenario, are the final destination of the TRU produced by LWRs, the

ABR fuel cycle is considered closed with a single tier. Figure 7-2 shows the mass flow

relationships of the closed one-tier ABR fuel cycle.

Mining Conversion Enrichment Fabrication LWR
(U308) (UF6) (UOX) (UOX-FA) (UOX-SNF)

0.231 kgU/MWD 0.230 kgU/MWD 0.143 kgSWU/MWD 0.020 kgU/MWD


0.0346 kgHM/MWD = IS I l11 kgU/MWD
"collected for MOX" di,,cluigcd as UOX-SNF"


Aqueous Processing Repository and
(processed SNF) S Surface Disposal
H L W IL W L L W .......................................................................
0.00102 kgHM/MWD (including all TRU)



Pyroprocessing Fabrication ABR
(processed SFF) (fuel assembly) (fuel assembly)
S0.00592 kgHM/MWD 0.00687 kgiHM/MWD



......... 0.00598 kgHM/MWD
.................................................................... R ep o sito ry an d
Surface Disposal
H L W IL W L L W ........................................................................

Figure 7-2. The closed single-tier ABR fuel cycle

Performing the mass balance calculation shows that only 1.8 times more SNF HM is

needed at the ABR fuel cycle front-end than is produced per megawatt-day at the back-end of the

UOX fuel cycle. For the ABR cost analysis, the following additional assumptions are made.

* All TRU, including Np+Pu+Am+Cm, is recycled at both reprocessing plants
* Metal fueled ABR with tCR=0.5
* Average ABR Fuel Burnup: 135 MWD/kgiHM
* ABR Thermal Efficiency: 37%


263









* Reprocessed UOX-SNF transuranics are used for the make-up TRU to the ABR
* Reprocessed UOX-SNF uranium is used for the make-up uranium to the ABR

The unit costs for the ABR fuel cycle services and fees for the closed single-tier fuel cycle

are listed in Table 7-4. Using these assumptions and unit costs, the following fuel costs are

calculated for the ABR tier of the fuel cycle. These costs are tabulated in Table 7-5.

Table 7-4. Unit costs and fees for the closed single-tier ABR fuel cycle
Item Amount Unit
Aqueous reprocessing $1,500.00 $/kgHM Feed
Pyroprocessing $3,000.00 $/kgHM Feed
HLW Disposal* $275.00 $/kgHM Feed
Fabrication ABR $2,500.00 $/kgHM Product
*Note: Since Am+Cm+Bk+Cf are not present in large quantities in the HLW, it may be assumed
that vitrified HLW will not require deep geologic repository storage and could be stored in a
surface repository at a cost of $275/kgHM instead of $300/kgHM

Table 7-5. Fuel costs for the ABR tier of the closed single-tier ABR fuel cycle
D Sticker Price Borrowing Time Carrying Charge
($/kgiHM) (Yr) ($/kgiHM)
Credit for UOX SNF -$17.47 -$2,340.62 3.74 -$875.39
UREX+la $51.90 $6,955.14 3.74 $2,601.22
Pyroprocess $17.94 $2,404.30 3.74 $899.21
HLW Disposal-Aqueous $9.52 $1,275.11 3.74 $476.89
HLW Disposal-Pyro $1.64 $220.39 2.86 $63.03
ABR Fabrication* $2.54 $340.17 2.86 $97.29
ABR Fabrication** $14.80 $1,983.55 2.86 $567.29
NWPA Levee $8.16 $1,093.44 2.86 $312.72
*Aside: Cost of fuel fabrication corresponding to the HM taken from SNF, **Aside: Cost of
fuel fabrication corresponding to the HM taken from SFF.

Combining the sticker price and carrying charge of the fuel purchase gives a total fuel cost

of $16,073/kgiHM. Dividing by the fuel burnup of 135 MWD/kgiHM gives a fuel cost of

electricity of $0.014/kW-hr(e). Therefore, the ABRs fuel contribution to the cost of electricity is

a little more than two times the fuel cost of UOX electricity and about one-half the fuel cost of

MOX electricity. Similar to MOX, the cost of reprocessing SNF is principally responsible for

the higher ABR fuel cost. However, the ABR can achieve 2.7 times the fuel burnup of the

Np+Pu-MOX case. Therefore, even though the ABR fuel is about 1.5 times more expensive per


264









mass of fuel purchased than the Np+Pu-MOX it extracts more fission energy per fuel purchased.

However, this higher energy extraction per mass of fuel purchased is still not sufficient to reduce

the ABR fuel cost to less than that of the once-through UOX fuel cycle.

The Combined AHFTR and ABR Fuel Cycle

The AHFTR has essentially two fuel suppliers. It receives an external make-up supply of

Np+Pu for supplementing the mass exhausted by fission in each cycle. It also receives an

external make-up supply of Am+Cm+Bk+Cf for fabricating fresh targets. Both of these mass

flows have origins in LWR SNF. However, the AHFTR receives a greater proportion of

Am+Cm+Bk+Cf per total transuranic mass than exists in the initial SNF. This extra material is

being diverted from ABRs which are receiving only Np+Pu based fuel. If the Am+Cm+Bk+Cf

mass flow were not diverted to the AHFTR, it would be considered as part of the HLW stream

created by the aqueous recycling center. Hence it would require a long term disposal solution

and add to the HLW disposal cost. Conversely, if either the AHFTRs or ABRs accept this

Am+Cm+Bk+Cf mass stream as part of its fuel, then the MA contribution of the HLW cost is

avoided. However, the purpose of the AHFTR concept is to accept this unwanted material in

order to alleviate material handling expenses for the ABR. Because, the AHFTR provides this

service to the ABRs as well as consuming Np+Pu from LWRs, its fuel cycle is considered closed

but having a two-tier recycling scenario. Figure 7-3 shows the mass flow relationships of the

closed two-tier AHFTR fuel cycle.

Performing the mass balance calculation shows that only 0.63 kg of SNF is needed to

produce one kg of fuel at the front-end of the AHFTR fuel cycle. The larger reprocessed fuel

output than the SNF input is due to the contribution of MAs provided by the SNF partitioned for

the ABR's fuel. This property of the AHFTR fuel cycle will be used to quantify the fuel savings

afforded by recycling Am+Cm+Bk+Cf in the AHFTR as opposed to the ABR.


265










Mining Conversion Enrichment Fabrication LWR
(U308) (UF6) (UOX) (UOX-FA) (UOX-SNF)


0.231 kgU/MWD


0.230 kgU/MWD 0.143 kgSWU/MWD


0.020 kgU/MWD


0.0346 kgiHM/MWD
"collected for MOX"



0.0119 kc,.HM l\\'I
"collected fol MR-ON


Aqueous Processing
(processed SNF)


HLW, ILW, LLW...............................
HLW, ILW, LLW


I s5


Cold-Fabrication
(fuel assembly)


Repository and
Surface Disposal


II Ill kgU/MWD
discuiggcd as UOX-SNF"


II. IlI. kJl-i IlWD
Ilcliard Is Li(O)X-SNF"






0.00104 kgHM/MWD (w/o
Am+Cm+Bk+Cf)


Pyroprocessing Hot-Fabrication ABR
(processed SFF) (fuel assembly) (fuel assembly)

0.00649 kgHM/MWD 0.00642 kgiHM/MWD


HL\\ IL\\ LL\\


Repository and
Surface Disposal


0.00656 kgHM/MWD


0.00104 kgHM/MWD (including Am+Cm+Bk+Cf)


0.00987 kgHM/MWD

HLW, ILW, LLW


0.00977 kgiHM/MWD


0.00997 kgHM/MWD


Figure 7-3. The closed double-tier AHFTR fuel cycle


It is important to note that the external supply of HM is identical for both the ABR and the

AHFTR (0.00104 kgHM/MWD). These external feeds demonstrate the fact that one gram of

HM is converted by fission into one MWD of energy. When the fuel cycle is evaluated from a


266


I I I









pure mass flow perspective, as it is shown in Figure 7-3, it is apparent that it is irrelevant whether

or not this material is all plutonium, all MAs, all uranium or a mix in-between.

The rate that HM is discharged from the core and returned to the pyroprocessor is

determined by how massive the core is and also the refueling rate. The AHFTR and ABR have

approximately the same cycle length and batch-fraction. Therefore, the refueling rate is about

the same (Table 3-13). However, the AHFTR has a HM mass approximately 1.3 times that of

the ABR (Table 3-12). The larger core, explains the larger mass flow rate through the AHFTR

pyroprocessor. For the AHFTR cost analysis, the following additional assumptions are made.

* The ABR and AHFTR draw HM from an infinite SNF reservoir at their own independent
rates

* The external supply of TRU from the aqueous reprocessing plant, which is destined for
ABR and AHFTR driver fuels, is only Np+Pu.

* All Am+Cm+Bk+Cf separated from SNF is diverted from HLW and sent to the AHFTR
for target fabrication.

* Some of the ABR driver fuel can be fabricated in a glove-box facility collocated at the
aqueous reprocessing plant using the Np+Pu separated from SNF.

* Metal fuel AHFTR with tCR=0.7

* Average AHFTR Fuel Burnup: 93 MWD/kgiHM

* AHFTR Thermal Efficiency: 37%

The unit costs for the ABR component of the closed double-tier AHFTR fuel cycle

services and fees are listed in Table 7-6. Using these assumptions and unit costs, the following

fuel costs are calculated for the ABR tier of the fuel cycle. The ABR costs are tabulated in Table

7-7.

Combining the sticker price and carrying charge of the fuel purchase gives a total fuel cost

of $16,821/kgiHM. Dividing by the fuel burnup of 135 MWD/kgiHM gives a fuel cost of


267









electricity of $0.014/kW-hr(e). As one recalls, the ABR fuel cost of electricity with

Am+Cm+Bk+Cf in the driver fuel, was also $0.014/kW-hr(e).

Table 7-6. Unit costs and fees for the closed double-tier AHFTR fuel cycle
Item Amount Unit
Aqueous reprocessing $1,500.00 $/kgHM Feed
Pyroprocessing $3,000.00 $/kgHM Feed
HLW Disposal $275.00 $/kgHM Feed
Fabrication ABR Cold Handling $1,500.00 $/kgHM Product
Fabrication ABR Hot Handling $2,500.00 $/kgHM Product
Fabrication AHFTR $2,500.00 $/kgHM Product

Table 7-7. Fuel costs for the ABR tier of the closed double-tier AHFTR fuel cycle
$/MW Sticker Price Borrowing Time Carrying Charge
($/kgiHM) (Yr) ($/kgiHM)
Credit for UOX SNF -$17.78 -$2,384.83 3.74 -$891.92
UREX+3 $52.83 $7,086.50 3.74 $2,650.35
Pyroprocessing $19.67 $2,637.87 3.74 $986.56
HLW Disposal-Aqueous $10.57 $1,417.30 3.74 $530.07
HLW Disposal-Pyro $1.82 $243.73 2.86 $69.71
ABR Fabrication* $1.56 $209.38 2.86 $59.88
ABR Fabrication** $16.22 $2,176.25 2.86 $622.41
NWPA Levee $8.16 $1,094.51 2.86 $313.03
*Aside: Cost of fuel fabrication corresponding to the HM taken from SNF (assuming a glove-
box facility for cold-fuel handling), **Aside: Cost of fuel fabrication corresponding to the HM
taken from SFF (assuming a hot-cell facility for hot-fuel handling)

Therefore, reduction in the fuel fabrication costs using the glove-box facility had negligible

impact on the fuel cost of electricity. This result indicates that the ABR fuel cost is fairly

insensitive to whether or not some of it can be fabricated at the centrally located aqueous

reprocessing plant. The negligible cost savings is a direct result of most of the HM being

pyroprocessed as SFF at the electrorefinery. Repeating this calculation for the AHFTR tier gives

the following fuel cycle costs. The AHFTR fuel costs are shown in Table 7-8.

Combining the sticker price and carrying charge of the fuel purchase gives a total fuel cost

of $8,529/kgiHM. Dividing by the fuel burnup of 93 MWD/kgiHM gives a fuel cost of

electricity of $0.0104/kW-hr(e). Therefore, the fuel cost of the AHFTR is only 1.6 times the fuel

cost of UOX electricity and is 73% of the fuel cost of the ABR. The reason for the cheaper


268









AHFTR fuel cost is the credit that the AHFTR receives for taking responsibility for the separated

Am+Cm+Bk+Cf from the aqueous reprocessing company. Similar to the closed single-tier ABR

case, the assumption is made that the cost of HLW disposal can be reduced if none of the SNF

transuranic wastes, with the exception of small process losses, require geologic repository

disposal.

Table 7-8. Fuel costs for the AHFTR tier of the closed double-tier AHFTR fuel cycle
$/MWD Sticker Price Borrowing Time Carrying Charge
($/kgiHM) (Yr) ($/kgiHM)
Credit for UOX SNF -$6.00 -$555.73 3.74 -$207.84
Credit for ABR HLW -$13.70 -$1,269.28 3.74 -$474.71
UREX+3 $17.83 $1,651.35 3.74 $617.60
Pyroprocessing $29.91 $2,770.15 3.74 $1,036.04
HLW Disposal-Aqueous $3.27 $302.75 3.74 $113.23
HLW Disposal-Pryo $2.74 $253.93 2.86 $72.62
AHFTR Fabrication* $2.59 $239.86 2.86 $68.60
AHFTR Fabrication** $24.67 $2,285.37 2.86 $653.62
NWPA Levee $8.16 $755.87 2.86 $216.18
*Aside: Cost of fuel fabrication corresponding to the HM taken from SNF (assuming a hot-cell
facility for hot-fuel handling), **Aside: Cost of fuel fabrication corresponding to the HM taken
from SFF (assuming a hot-cell facility for hot-fuel handling)

Also, the AHFTR has a smaller aqueous reprocessing cost than the ABR because the

AHFTR draws less fuel from SNF for its external feed rate than the ABR (0.63 versus 1.85 in

Figure 7-3, also see Table 3-14). The AHFTR requires less externally supplied make-up TRU

from SNF because, it receives an additional mass stream from the pyroprocessor which would

not be available if MAs were not diverted from the HLW and irradiated in targets.

The ABR in the closed double-tier case is being charged for the HLW disposal because it

is not assuming responsibility for the Am+Cm+Bk+Cf wastes generated in the front end of its

fuel cycle. The AHFTR receives a credit ($300/kg-$275/kg=$25/kg) for taking these wastes.

This credit is equal to the difference in the cost to dispose of the Am+Cm+Bk+Cf stream in the

HLW waste ($300/kgHM) and the disposal fee without Am+Cm+Bk+Cf ($275/kgHM). This


269









credit combined with the UOX-SNF credit fully compensates for the cost of aqueous

reprocessing.

The assumption that the HLW disposal fee can be reduced from $300/kgHM to

$275/kgHM is not sufficient to reduce the AHFTR fuel cost to below the cost of the once-

through fuel cycle. However, assuming that a new HLW disposal fee of $225/kgHM can be

achieved if all the MAs are eliminated from the HLW mass stream, the fuel cost of electricity for

the AHFTR becomes equivalent to the fuel cost of UOX-LWRs.

Sensitivity Analysis

Though it is apparent that the HLW disposal fee can be arbitrarily adjusted to make, the

AHFTR fuel costs comparable to UOX-LWR fuel costs, the overall closed two-tier fuel cycle at

this stage in the analysis is not cost competitive to one-through. This is because the majority of

the SFRs in the fast rector fleet are ABRs with a non-competitive fuel cost. Therefore, a

comparative analysis is needed to show how much fuel cycle unit costs need to change in order

to make the overall closed two-tier fuel cycle cost competitive. First, the cost of uranium ore,

conversion, enrichment and UOX fabrication will be adjusted to raise the once-through fuel cost

to meet the ABR or AHFTR fuel cost. Next, the cost of aqueous reprocessing, pyroprocessing,

HLW disposal and hot-fuel fabrication will be adjusted to lower the ABR and AHFTR fuel costs

to the breakeven point with the once-through cost. Finally, a theoretical SFR utility operating a

mix of ABRs and AHFTRs is evaluated to see what conditions need to apply in order to make

the average cost of fuel to the utility equivalent to the once-through cost.

Breakeven Unit Costs for ABRs and AHFTRs

The breakeven fuel costs are calculated by holding all values of the above three base cases

constant and evaluating the perturbation of each individual unit cost required to equate the

overall fuel cost of electricity between the once-through and closed two-tier fuel cycle. This is


270









done by increasing the fuel cycle service costs of the once-through fuel cycle until the once-

through option is as costly as the closed two-tier base case calculation. Conversely, the unit

costs of reprocessing, HLW disposal and fuel fabrication can be reduced until the fuel costs of

the closed two-tier option is as cheap as the once-through base case.

Table 7-9 shows the breakeven unit costs necessary to equate the ABR fuel costs to the

UOX-LWR once-through fuel cycle or visa versa. The negative unit costs indicate that it is

physically impossible to equate the fuel costs unless an external additional cash flow "fringe

benefit" of revenue could be created by the given process in addition to the fuel cycle service

provided by that process. Obviously, these cash flows do not exist in reality but are simply a

numerical artifact of the unit cost perturbation. It is interesting to note that the cost of uranium

would have to increase to $143/lb of U308 in order to make fuel recycling in an ABR an

attractive fuel cycle alternative to the once-through UOX-LWR fuel cycle. Coincidently, the

Uranium Exchange Consulting Company (UxC) quoted the price of uranium to be $135/lb U308

in the summer of 2007. The UxC is an industry recognized uranium fuel cycle consulting

company that continuously publishes the current market price for fuel cycle services including

uranium ore, conversion and enrichment.

The breakeven unit costs necessary to equate the AHFTR fuel costs to the UOX-LWR

once-through fuel cycle, or visa versa, is shown in Table 7-10. Remember, the AHFTR receives

a credit that is equivalent to some fraction of the market HLW disposal costs which represents

the money saved by the aqueous reprocessing company from not having to dispose

Am+Cm+Bk+Cf transuranic waste. It is assumed that because pyroprocessing does not allow

for removal of Am+Cm+Bk+Cf, the pyroprocessor HLW disposal cost is also equal to the

reduced value of this cost. Therefore, the breakeven unit cost of the pyroprocessor's HLW









disposal fee in Table 7-10 is equivalent to the reduced value of the HLW disposal fee extended

to both aqueous reprocessor and pyroprocessor for not having to dispose of Am+Cm+Bk+Cf

transuranic waste.

Table 7-9. Breakeven unit costs for equating the once-through fuel cost of electricity to that of
the ABR in the closed two-tier fuel cycle
Fuel Service Base Case Unit Cost Breakeven Unit Cost Unit
UOX-LWR Fuel Service Costs Increased/ABR Fuel Service Costs Held Constant
Mined Uranium $20 $143 $/lb U308
Conversion $11 $242 $/kgU as UF6
Separative Work $145 $566 $/kg-SWU
Fabrication UOX $250 $2,727 $/kgU Product
NWPA Levee $0.001 $0.005 $/kw-hr(e)
Interim Cask Storage $100 $1,232 $/kgHM Waste
ABR Fuel Service Costs Decreased/UOX LWR Fuel Service Costs Held Constant
Aqueous Reprocessing $1,500 $44 $/kgHM Feed
Pyroprocessing $3,000 -$4,822 $/kgHM Feed
HLW Disposal (Pyro)* $275 -$8,080 $/kgHM Feed
HLW Disposal ABR (Aqueous)* $300 -$1,156 $/kgHM Feed
Fabrication Cold Handling $1,500 -$51,144 $/kgHM Product
Fabrication -Hot Handling $2,500 -$5,942 $/kgHM Product
*Note: The aqueous process HLW disposal fee is higher than the pyroprocessing fee because in
this scenario the ABR does not take responsibility for the Am+Cm+Bk+Cf forcing the
reprocessing company to either dispose of it or pay for its incineration in an AHFTR.

Table 7-10. Breakeven unit costs for equating the once-through fuel cost of electricity to that of
the AHFTR in the closed two-tier fuel cycle
Fuel Service Base Case Unit Cost Breakeven Unit Cost Unit
UOX-LWR Fuel Service Costs Increased/AHFTR Fuel Service Costs Held Constant
Mined Uranium $20 $84.78 $/lb U308
Conversion $11 $132.73 $/kgU as UF6
Separative Work $145 $366.66 $/kg-SWU
Fabrication UOX $250 $1,555.20 $/kgU Product
NWPA Levee $0.001 $0.012 $/kw-hr(e)
Interim Cask Storage $100 $1,265.49 $/kgHM Waste
AHFTR Fuel Service Costs Decreased/UOX LWR Fuel Service Costs Held Constant
Aqueous Reprocessing $1,500 -$878.68 $/kgHM Feed
Pyroprocessing $3,000 $164.03 $/kgHM Feed
HLW Disposal (Pyro) $275 $227.53 $/kgHM Feed
HLW Disposal ABR (Aqueous) $300 $349.30 $/kgHM Feed
Fabrication -Hot Handling $2,500 -$147.84 $/kgHM Product

Notice that this (Am+Cm+Bk+Cf free) (free) HLW disposal cost only needs to be less than

one third of the currently estimated HLW disposal fee in order to make the AHFTR cost


272









competitive with the once-through UOX-LWR fuel cycle. Also it is noteworthy to point out that

the cost of uranium ore required to make the once-through fuel costs as expensive as the AHFTR

fuel costs is only $85/lb U308. The UxC quoted price of uranium ore was $90/lb U308 in the

December of 2007 at the time that this dissertation was being written.

It has been shown that the cost of uranium or the HLW disposal fee can be arbitrarily

altered to show the cost competitiveness of either the ABR or the AHFTR independently.

However, in reality the closed two-tier fuel cycle is intended to allow these two rector types to

operate in parallel. Therefore, it is necessary to determine the support ratio of AHFTRs needed

to burn all of the Am+Cm+Bk+Cf produced by the aqueous reprocessing plant in support of both

reactors.

The ABR requires the separation of 0.0352 kg SNF/MWD of SNF in order to produce the

Np+Pu needed to satisfy the external feed demand of its driver fuel. In order to meet the

Am+Cm+Bk+Cf external feed requirements of the targets, the AHFTR requires the separation of

0.1160 kgHM/MWD of SNF in addition to the SNF separation requirements of its driver fuel

(0.0119 kg SNF/MWD). The ratio of the AHFTR processing demand on SNF to get the

Am+Cm+Bk+Cf from ABRs over the rate that Am+Cm+Bk+Cf can be produced by the ABR

reprocessing demand is effectively the support ratio of ABRs to AHFTRs in the fast reactor fleet.

Hence, a support ratio (ABR per AHFTR 0.1160/0.0352 = 3.3) is required to burn, in

AHFTRs, all of the Am+Cm+Bk+Cf separated at the aqueous reprocessing plant.

Using this support ratio, a utility is envisioned that operates a fleet of SFRs with one

AHFTR for every 3.3 ABRs. In order to evaluate the cost competiveness of the SFR utility, the

average fuel cost between the ABRs and AHFTRs is evaluated, as opposed to evaluating for each


273









reactor type individually. The breakeven unit costs necessary to equate the utility's average fuel

costs to the UOX-LWR once-through fuel cycle, or visa versa, is shown in Table 7-11.

Table 7-11. Breakeven unit costs for equating the once-through fuel cost of electricity to that of
the average fuel cost of a combined fleet of ABRs and AHFTRs
Fuel Service Base Case Unit Cost Breakeven Unit Cost Unit
UOX-LWR Fuel Service Costs Increased/Combined Fleet Fuel Service Costs Held Constant
Mined Uranium $20 $129 $/lb U308
Conversion $11 $217 $/kgU as UF6
Separative Work $145 $519 $/kg-SWU
Fabrication UOX $250 $2,454 $/kgU Product
NWPA Levee $0.001 $0.006 $/kw-hr(e)
Interim Cask Storage $100 $1,236 $/kgHM Waste
Combined Fleet Fuel Service Costs Decreased/UOX LWR Fuel Service Costs Held Constant
Aqueous Reprocessing $1,500 -$42 $/kgHM Feed
Pyroprocessing $3,000 -$3,247 $/kgHM Feed
HLW Disposal (Pyro) $275 -$69 $/kgHM Feed
HLW Disposal ABR (Aqueous) $300 $753 $/kgHM Feed
Fabrication -Hot Handling $2,500 -$3,985 $/kgHM Product

It is apparent from Table 7-11 that the cost of uranium must reach $130/lb U308 in order

for the overall closed two-tier fuel cycle to reach cost competitiveness with the once-through

UOX-LWR fuel cycle. It is also interesting to note that if the cost of uranium is only $20/lb

U308 then the HLW disposal credit will not be sufficient to achieve cost competitiveness. This

is specified by the negative HLW disposal fee of a transuranic free HLW stream (i.e.,

equivalently the pyroprocessing HLW disposal fee). However if the cost of the once-through

fuel cycle was slightly more expensive as a result of a higher uranium ore cost, then a HLW

disposal credit might be found that could reduce the average SFR fuel cost to meet the higher

once-through fuel cost.

Cost Sensitivities for a Combined ABR and AHFTR Fleet

In this section, a SFR utility is envisioned that operates a mix of ABRs and AHFTRs in

order to burn the TRU generated by the UOX-LWR fleet. Based on the mass flows from Table

3-14, the support ratio (normalized per megawatt of installed reactor capacity) between ABRs,

AHFTRs and LWRs is:


274









3.3 ABRs per AHFTR
0.6 SFRs per LWR

In order to make it more likely that the SFR fleet becomes cost competitive with the UOX-

LWRs, the cost for the price of uranium is set to $80/lb U308 as opposed to $20/lb U308. The

higher uranium price increases the once-through UOX-LWR fuel cost of electricity to $0.01/kW-

hr(e). Next, a sensitivity analysis is conducted on the SFR utility where the unit cost of aqueous

reprocessing and pyroprocessing is each individually varied from $1,000/kgHM to

$3,000/kgHM. For each of these perturbations, the "transuranic-free" HLW disposal cost is

adjusted until the average fuel cost to the SFR utility becomes equal to the $0.01/kW-hr(e) fuel

cost of the once-through UOX-LWR fuel cycle.

Figure 7-4 shows the ABR and AHFTR fuel costs for this breakeven scenario. Notice that

the fuel costs for the ABR increase with increasing reprocessing cost. This trend is affected by

the fact that the ABR is held accountable for MAs that it does not bum. Therefore, the ABR's

fuel cost does not receive the HLW disposal credit and must effectively reimburse the aqueous

reprocessing company the full market price ($300/kgHM) of transuranic HLW disposal. The

fuel cost increases almost linearly because the aqueous reprocessing unit cost contribution to the

total fuel cost is being linearly increased from $1,000 to $2,000.

The ABR and AHFTR fuel costs for a varying pyroprocessing costs are given in Figure 7-

5. It is important to note that the ABR fuel cost is more sensitive to the cost of aqueous

reprocessing than it is to pyroprocessing. This is because of the much higher SNF HM

throughput to the aqueous reprocessing plant (0.0346 kgHM/MWD) than the pyroprocessor

(0.00656 kgHM/MWD) (Figure 7-3). The higher throughput of the aqueous plant is due to the

TRU content of SNF being so much less than that of SFF. Even though the cost of


275












pyroprocessing can be twice as high as aqueous reprocessing, the large aqueous plant throughput


causes it to be much more sensitive to the cost of reprocessing.


1.8
0-
]1.6

1.4

1.2
I--
o
a
0.8

' 0.6

;0.4
0
Co
S0.2
U.
0.0


MABR Fuel Cost
*AHFTR Fuel Cost
OAverage Fuel Cost to SFR Utillity


$1,000 $1,500 $2,000
Cost of Aqueous Reprocessing ($/kg HM)

Figure 7-4. Fuel cycle costs for the ABR and AHFTR for varying costs of aqueous reprocessing
adjusted by the HLW disposal credit for a breakeven average fuel cost


1.6
SABR Fuel Cost
S1.4 AHFTR Fuel Cost








-0.4 --------------------------
D Average Fuel Cost to Utillity
1.2

1.0
I-"
0.8

0.6







S $1,000 $1,500 $2,000 $ 0 $ 0
-0.2
u_

-0.4
Cost of Pyroprocessing ($/kg HM)

Figure 7-5. Fuel cycle costs for the ABR and AHFTR for varying costs of pyroprocessing
adjusted by the HLW disposal credit for a breakeven fuel cost


The negative fuel cost for the AHFTR corresponds to a HLW disposal credit that becomes


large enough to pay for all of the AHFTR fuel costs and eventually debits back to the SFR utility


276









some of the fuel expenditures paid on the ABR fuel. Figure 4-15 shows the impact of the

perceived price versus cost of HLW disposal.

Note that if the actual cost of HLW disposal is not at all dependent on the presence of

transuranic waste than it can not be varied as it is done in this study. This would be the case if

deep geologic disposal is required for HLW even if Am+Cm+Bk+Cf are not present in the

waste. However, if the Am+Cm+Bk+Cf is a large cost driver then its removal could greatly

reduce the cost of HLW disposal. This may be the case if the cost of surface disposal of HLW is

possible and this disposal is significantly cheaper then putting this waste in a deep geologic

repository. If it is possible to have a HLW disposal fee so small that the closed fuel cycle is

cheaper than the once-through fuel cycle, then the reprocessing company has the option to

charge the most expensive market price for fuel services of the once-through competition.

Therefore, the HLW disposal fee is used in Figure 7-5 to control the market expense or debit of

AHFTR fuel in order to make the average fuel cost to the SFR utility equivalent to the once-

through fuel cycle.

Figure 7-6 gives the transuranic-free HLW disposal fee as a function of the varying unit

costs of reprocessing. These values are used to calculate the HLW disposal credit necessary to

create the fuel costs of Figure 7-4 and Figure 7-5. The resulting HLW disposal credit is

calculated in the same way as was done in Table 7-8. This credit is equal to the current market

price ($300/kgHM) minus the transuranic-free price from Figure 7-6.

The reason that the aqueous reprocessing cost in Figure 7-4 stops at $2,000/kgHM is

because the transuranic-free HLW disposal fee drops to zero below this value. Hence, the cost

saving afforded by a reduced HLW disposal cost is insufficient to force the fuel cycle to become


277










cost competitive with the once-through fuel cycle if the aqueous reprocessing cost rises above

$2,000/kgHM.


$250
variMn Aqueous Reprocessing
"t i Pyroprocessing
$200
LLr
h f
.. $150


S$100


2 $50
-)
C

$0
$1,000 $1,500 $2,000 $2,500 $3,000
Unit Cost of Fuel Service ($/kg HM)
Figure 7-6. Transuranic-free HLW disposal fee adjusted until the fuel cost breakeven occurs for
varying costs of reprocessing

It is apparent from Figure 7-5 and Figure 7-6 cost 7-6 that even if the cost of pyroprocessing is as

high as $3,000/kgHM, the mixed-fleet fuel cycle can become cost competitive if the transuranic-

free HLW disposal cost can be reduced to $125/kgHM. Therefore, if the unit costs from Table

7-1, Table 7-4 and Table 7-6 are adopted, but with a $80/lb U308 uranium ore price and a

$125/kgHM HLW disposal fee, then the SFR mixed-fleet utility will be cost competitive to the

once-through fuel cycle.

Converting Neutrons into Dollars

If the mixed-fleet scenario from the previous section is considered with a $125/kgHM

HLW disposal fee and a $80/lb U308 ore price with all other unit costs the same as in Table 7-1,

Table 7-4 and Table 7-6, then the fuel costs to the ABR and ATHFTR will be $0.0140/kW-hr(e)

and -$0.0024/kW-hr(e), respectively. Combining these costs with the 3.3 ABR per AHFTR

support ratio gives an average fuel cost to the mixed-fleet utility of $0.0101/kW-hr(e). This


278









$0.0101/kW-hr(e) value is the cost of the once-through LWR fuel cycle assuming $80/lb U308.

Therefore, for these assumptions in unit costs, the mixed-fleet fuel cost breaks even with that of

the once-through fuel cycle.

It is interesting to note that if the ABR (CR=0.5) of the closed single-tier fuel cycle is

allowed to pay this cheaper HLW disposal fee, its fuel cost is reduced to only $0.0139/kW-hr(e).

In fact, using the $125/kgHM HLW disposal assumption with all the other unit costs of Table 7-

1 and Table 7-4, the breakeven price of uranium would be approximately $125/lb U308. As an

aside, if one assumes the ABR (CR=0.75) design as opposed to the ABR (CR=0.5) design, the

breakeven uranium price is $105/lb U308. Therefore, there is a breakeven uranium price

difference of at least $25/lb U308 between the mixed-fleet fuel costs and the single-tier ABR fuel

costs. The reason for the cost savings can be explained by a better neutron utilization afforded

by incorporating the axial targets. Because of the MA in the targets becomes a source of

plutonium in the AHFTR fuel cycle (Table 3-14), the overall mixed-fleet aqueous reprocessing

requirements are actually less than both the ABR (CR=0.75) and ABR (CR=0.5) designs.

At face value, the HLW disposal credit that the AHFTR receives for taking ownership of

the Am+Cm+Bk+Cf mass stream could be misinterpreted as a placeholder for redistributing the

HLW disposal fee between the ABR and the AHFTR. This is actually not the case. The AHFTR

considers the Am+Cm+Bk+Cf material in the fuel cycle as a commodity that is necessary for

breeding plutonium. Hence, the AHFTR regards Am+Cm+Bk+Cf as a fuel and not a waste.

Remembering back to Figure 7-3, the rate that HM is fed to both the ABR and AHFTR is exactly

the same (1.04 gram iHM/MWD). This external feed is equivalently the rate that the external

supply of mass is converted into fission energy. Remembering back to Table 3-15, 10% of this

external mass stream (0.10 gram iHM/MWD), for the AHFTR, is Am+Cm+Bk+Cf Also, the


279









Np+Pu supply is 0.15 gram iHM/MWD and the uranium supply is 0.78 gram iHM/MWD. On

the other hand, the ABR (CR=0.5) does not receive any Am+Cm+Bk+Cf and has a Np+Pu

supply of 0.45 gram iHM/MWD and a uranium supply of 0.59 gram iHM/MWD. Based on the

fact that the AHFTR requires less Np+Pu feed than the ABR, it can be inferred that the aqueous

reprocessing costs will be significantly less. The difference in the ABR and AHFTR fuel

aqueous reprocessing cost can be seen by comparing Table 7-5 with Table 7-8.

So, the AHFTR derives 10% of its fission energy from the fission of Am+Cm+Bk+Cf, or

its eventual transmutation into a fissile isotope that later undergo fission. The creation of fission

energy through Am+Cm+Bk+Cf and uranium conversion reduces the AHFTR demand for

externally supplied Np+Pu. Therefore, the AHFTR incurs less of an aqueous separations cost

than the ABR (Table 7-5, Table 7-8). The aqueous reprocessing costs of the mixed-fleet and

single-tier ABR scenarios are given in Table 7-12 for comparison purposes. Note that the

external feed Np+Pu requirement of the AHFTR is less than for the ABR cores. Therefore, the

aqueous reprocessing costs are less for the AHFTR than the ABRs. For completeness, the

pyroprocessing costs are also given in Table 7-12.

Table 7-12. Inter-comparison of aqueous reprocessing costs for single-tier ABR scenarios and
the double-tier combined ABR/AHFTR mixed-fleet
Single-Tier Double-Tier (Mixed-Fleet)
ABR ABR
Np+Pu ABR AHFTR
(CR=0.5) (CR=0.75) Np+Pu ABR AHFTR
TRU External Feed (g iHM/MWD) 0.46 0.20 0.45 0.25
Aqueous Reprocessing Cost ($/kgiHM) $6,821.38 $3,037.21 $7,086.50 $1,651.35
Pyroprocessing Cost ($/kgiHM) $2,633.36 $3,630.82 $2,637.87 $2,770.15

The conversion of waste into fuel is made possible by the conservation of neutrons in the

axial targets. This neutron conservation allows the fCR to be high by creating plutonium fuel

while allowing the tCR to be closer to the ABR by destroying Am+Cm+Bk+Cf transuranics

(Table 3-13). The ABR leaks excess neutrons out of the core in order to achieve a low tCR. For


280









the ABR, the tCR and fCR is closer in value than the AHFTR because the plutonium economy of

the fuel cycle is directly linked to the reactors neutron economy (i.e., leakage versus parasitic

absorption in U-238). Unlike the ABR, the AHFTR uses excess neutrons to create fuel out of

transuranic waste (i.e., Am+Cm+Bk+Cf). Hence, the fCR of the AHFTR is actually very near

that of the ABR (CR=0.75) design which has a calculated fCR of 0.84 and a tCR value of 0.77.

Yet the AHFTR's tCR is only 0.72. The difference in the tCR conversion ratios is created by the

credit given to the creation of plutonium fuel from MAs (which are trans-uranium) instead of

from uranium.

Because the AHFTR is consuming waste by converting it into fissile material, it provides a

valuable service to the aqueous reprocessing company and indirectly to the ABR. This is the

source of the HLW disposal credit. If all of the Am+Cm+Bk+Cf were evenly distributed

throughout a fleet of ABRs with no AHFTR, this mass stream would eventually get converted

into fission energy. However, more SNF would have to be separated at the aqueous reprocessing

plant to provide plutonium to the ABR's, than if some of the plutonium was created by

Am+Cm+Bk+Cf transmutation which is allowed by neutron recovery in axial targets. This is

why the fuel costs are higher for a utility operating an ABR in a single-tier fuel cycle than the

average fuel costs to a utility operating a mixed-fleet in a double-tier fuel cycle. In fact, the

sensitivity analysis indicates that if a pyroprocessing cost of $3,000/kgHM and a $125/kgHM

transuranic-free HLW disposal cost are possible, then the average fuel cost to the mixed-fleet

SFR utility would be competitive with the cost of the once-through fuel cycle assuming the

current-day uranium ore price of $80/lb U308.









CHAPTER 8
SUMMARY AND CONCLUSIONS

In the past several years there has been a renewed interest in sodium fast reactor (SFR)

technology for the purpose of destroying transuranic waste (TRU) which is produced by light

water reactors (LWR). The main driver for this decision comes from the fact that higher neutron

energies allow all of the actinides, including the minor actinides (MA), to contribute to fission.

Though MAs constitute only a tenth of one percent of all LWR spent nuclear fuel (SNF) mass,

they are the most radiologically hazardous and constraining factor in the design of a geologic

repository. The higher SFR neutron economy over thermal spectrum reactors makes possible a

sustainable recycling strategy that can continuously irradiate the initial MA content of SNF,

including all of their transmutation products. This sustainable burner strategy is the focus of the

current Advanced Burner Reactor (ABR) designs intended to consume the TRU present in the

SNF produced by LWRs.

A measure of TRU destruction is the SFR's conversion ratio (CR). The definition of the

CR in recent years used for fuel cycle analysis has taken on the meaning of the net TRU

produced divided by the net TRU destroyed by the reactor system. This definition is different

from the traditional fissile definition (fCR) which is more based on the neutron balance between

fissile produced divided by fissile destroyed. The slight difference between these two definitions

is the fact that the transuranic (tCR) definition treats all TRU as if they were all of the same

fissile quality. In reality this is not the case because most of the long-lived SNF MAs are not

fissile by definition due to their neutron-to-proton pairing. Essentially, all even neutron

numbered actinides exhibit a fission threshold, at about one MeV, due to the fact that additional

kinetic energy is needed in order to overcome the critical energy for fission. The average

neutron spectrum of a SFR exists in the energy range of this fission threshold. However, for


282










most real SFRs there is sufficient down-scattering to ensure that only a fraction of the neutrons in

the core exist at energies high enough to induce fission in the MAs (Figure 8-1).

1E+06



1E+04

-c






X


1.E-07 1.E-05 1.E-03 1.E-1E0201 1.E+01
c.






Figure 8. EN-Am-241 Capture -Am-241 Fission


1E-04
1.E-07 1.E-05 1.E-03 1.E-01 1.E+01
Energy (MeV)
Figure 8-1. ENDF/B-VI cross section data for fission and neutron capture of Am-241, Pu-238
and Pu-239 plotted against the neutron spectrum of a metallic fueled sodium fast
reactor

Nevertheless, transmutation of these MAs, in particular Am-241, leads to plutonium

isotopes with significantly higher fission cross sections. Given that the capture cross section for

Am-241 below its fission threshold is as high as the fission cross section of Pu-239, slight

moderation of the SFR neutron flux can yield an energy spectrum conducive to Am-241

transmutation instead of Pu-239 fission. This fact is capitalized upon in this dissertation by

adding an axial blanket to the ABR design that contains a combination of plutonium, americium

and moderator which enhances neutron capture in Am-241. This hybrid ABR design is referred

to in this dissertation as the Axial Heterogeneous Fast Transmutation Reactor (AHFTR). Since,

the Am-241 neutron capture ultimately produces the plutonium isotope Pu-238, the axial target

region becomes a source of plutonium in the fuel cycle similar to a uranium blanket in a Pu-239


283









breeder reactor. Pu-238 is itself an even neutron numbered actinide. However, its sub-threshold

fission cross section is significantly greater than the initial Am-241 (Figure 8-1). This fact

allows the AHFTR to have an fCR value similar to that of an equivalent ABR design but with a

tCR value that is less.

Method of Reactor Design

Due to the large mean-free-path of fast neutrons compared to thermal neutrons, the effect

of geometric buckling (i.e., neutron leakage) is a stronger contributor to the SFR's: CR, power

distribution and coolant void response. To meet the criticality condition of a high leakage SFR

core, the fissile concentration in the fuel is necessarily high. In order to decrease the tCR, and

hence increase TRU burning performance, the concentration of fertile uranium in the fuel must

be decreased. The ABR and AHFTR both accomplish this by simultaneously increasing the

TRU concentration while decreasing the U-238 concentration. This increase in the fissile

inventory of the core necessitates that leakage be increased in order to meet the criticality

requirement that the core's geometric and materials buckling be equated. In the AHFTR design,

the radius of the core was increased by one row of fuel over that of the reference ABR design.

The greater AHFTR core radius decreases the radial geometric buckling or curvature of the flux

in the radial direction. This has the effect of decreasing the peak-to-average power ratio in the

core which enables the AHFTR fuel to be irradiated more evenly than in the ABR (Figure 8-2).

It also relaxes the need to smooth the radial power profile using higher enriched fuel in the outer

region of the core.

Because of the reduced emphasis on higher enriched fuel, the AHFTR can draw upon the

already existing experience established by past SFR's such as the Experimental Breeder

Reactor's I and II (EBR-I and EBR-II). Thus, the fuel composition and assembly design of the


284











AHFTR closely matches that previously established by real reactors. Hence, with the exception

of the target material, little fuels development work should be required to implement its design.

4.5E+02

4.0E+02



_3.5E+02


S2.OE+02

1.5E+02 --20%
O 1.0E+02
0. "-*-40%
1.OE+02 ---60%
5.0E+01 80%
--100%
0.OE+00 I I
1 2 3 4 5 6 7 8
Radial Fuel Assembly Row Number
A

4.5E+02

4.0E+02

c 3.5E+02
E
O3.0E+02

_2.5E+02

C 2.OE+02

1.5E+02 -20%
S-1--40%
1.0E+02 -60%
80%
5.0E+01 -*-100%
+--120%
0.OE+00
1 2 3 4 5 6 7 8
Radial Fuel Assembly Row Number
B
Figure 8-2. Radial power distribution for six axial slices through the core for the ABR (At) and
AHFTR (B) (REBUS)

Because the AHFTR radial buckling was reduced, the geometric buckling in the axial

direction had to be slightly increased. The resulting increase in axial leakage favors

transmutation in the targets because more neutrons are lost from the active driver core and

invested in neutron captures in the target region. In the axial target region, some of the fuel pins


285









contain blank rods of zirconium hydride moderator. These moderating rods shift the fast flux of

the active core region to a more epithermal flux where the capture cross section of Am-241 is

high compared to the Pu-239 cross section. This spectrum modulation causes the neutron mean-

free-path to shorten while simultaneously increasing the importance of neutron capture in Am-

241 relative to fission in Pu-239. Therefore, the leakage (and hence buckling) is reduced locally

in the target region. This neutron trap effect enables enhanced utilization of neutrons in the

AHFTR's fuel cycle. Note that the Pu-238 generated by the Am-241 is exposed to epithermal

neutrons in the target spectrum and therefore has little fissile value. However when the bred Pu-

238 is recycled and placed in the faster spectrum of the active core region, its fissile value

becomes much higher. This spectrum shift, as a result of the fuel recycling process, is the

distinguishing factor that gives the AHFTR an enhancement in fCR without a corresponding

increase in the tCR.

Reactor Control and Safety Features

Some debate has existed in the transmutation community over whether or not Tc-99 can be

effectively burned in a SFR. Tc-99 is a fission product with a concentration in SNF roughly

equal to that of the MAs. Also, due to its long life and radiotoxicity, it is an isotope that must be

considered in the long term design of a geologic repository. Tc-99 has approximately the same

capture cross section of that ofU-238 in the fast spectrum. However, the neutron capture cross

section of Tc-99 is roughly three times less than that of Am-241. Therefore, moderated target

studies by previous authors have shown that this isotope is fairly difficult to destroy in SFRs.

Because the transmutation rate of Tc-99 is relatively small, in this dissertation it is used as the

neutron poison in control rods for reactivity shim purposes only.

It was found that using the control assembly pattern used by the reference ABR design, Tc-

99 could be used for shim and safety rods in the AHFTR. The effectiveness of Tc-99 as a


286









control rod poison lends itself to the relatively high atom concentration of Tc-99 atoms in the

form of metallic technetium. The atomic density of metallic technetium is approximately three

times higher than the B-10 concentration in 90% enriched boron carbide. Enriched boron

carbide is a typical neutron absorber in many SFR designs due to the large unresolved cross

section value of B-10. The neutron capture cross section of B-10 is approximately three and a

half times that of Tc-99 in the fast neutron energy range. Therefore, Tc-99's reactivity shim

worth in the SFR is almost but not quite as significant as enriched boron carbide.

As a side benefit, it was also found that the Tc-99 destruction rate in the shim rods during

their withdrawal cycle from the AHFTR active core was approximately equal to the Tc-99

production rate in a typical pressurized water reactor. Therefore, though it may not be possible

to "deep-bur" Tc-99 in dedicated targets as was proposed in previous transmutation studies, it

may be feasible to modify the control rod design of SFRs (such as was down with the AHFTR)

to have a net destruction benefit of technetium in the fuel cycle.

The AHFTR's void and Doppler coefficients were also calculated and compared to the

ABR. Because special effort was taken to minimize the MA concentration in the driver fuel, the

AHFTR's void and Doppler coefficients are comparable to the ABR design. The issue of a

positive void coefficient resulting from spectrum hardening during a complete loss of sodium

coolant was also addressed. This positive whole-core void worth also exists for the ABR and is

typical of many SFR designs. In the past EBR-II experience, this positive void coefficient was

countered by thermal expansion of the fuel. Fuel expansion decreases the density of the fuel and

increases leakage. The net effect of these expansion feedbacks was an overall negative power

coefficient. The AHFTR's radial and axial expansion coefficients were calculated to be negative

and showed that the positive reactivity insertion of a whole core void event could be countered


287









by a reasonable amount of thermal expansion. The AHFTR Doppler coefficient is negative and

comparable to the reference ABR case.

Code Validation

The majority of the fuel cycle calculations were performed using a suite of fast reactor

analysis codes developed over the past three decades by Argonne National Laboratory: MC2-2,

DIF3D and REBUS. Since these codes are adapted to fast reactor analysis, they make the

assumption that the neutron spectrum and flux magnitude do not vary over large "same

spectrum" regions of the core. Therefore, the cross section group constants generated by MC2-2

are done so using slowing down theory with a zero-dimensional critical buckling search. Core

flux calculations are made by DIF3D using diffusion theory and homogenizing "smearing" fuel

assemblies over radial and axial regions of the core in a hexagonal-z nodal discretization.

The accuracy of the moderated axial target transmutation analysis depends primarily on the

change in the mean-free-path between the fast neutrons of the active core and the moderated

neutrons of the target regions. Benchmark calculations were performed between DIF3D, and the

variational transport code VARIANT which uses spherical harmonics, and the transport code

MCNP which uses the Monte Carlo method. Good agreement was found between these three

different codes for both the neutron spectrum and flux distributions of the AHFTR core. Further

pin-lattice calculations, using MCNP, of the repeating moderator/target rod pattern showed that

spatial self-shielding and shadowing effects in the target region were negligible. This acceptable

result is due to the fact that only sufficient moderation was provided to shift the fast spectrum to

epithermal energies below the one MeV fission threshold of the MAs. This softened fast

spectrum is sufficient to achieve a moderate enhancement in the Am-241 transmutation rate

without decreasing the neutron energies to the range of well resolved cross section resonances.

Also, the epithermal neutron mean-free-path in the target region was sufficiently larger than the


288









fuel pin diameter and pitch dimensions. Thus, these target region neutrons do not have a high

visibility to the local heterogeneities created by the moderating zirconium hydride pins.

Transmutation Target Fuel Design

The general philosophy of the AHFTR design is to demonstrate that a MA burner core

could be constructed using existing or near term achievable technology. Therefore, the reactor

design was tailored in such a way as to draw upon the existing fabrication and irradiation

experience with metallic Pu/U/Zr fuels. Therefore, the composition of the AHFTR driver fuel

was limited to approximately 20TRU/70U/10Zr which is almost identical to the ternary metal

alloys tested at EBR-II. Also for feasibility purposes, the transmutation target's composition was

based on recent irradiation testing of high MA content metal alloy fuels, which was performed at

the Idaho National Laboratory's Advanced Test Reactor.

A set of fuel design criteria for the AHFTR was established to ensure that the

heterogeneous core design does not produce peak power levels in the driver fuel and targets that

could cause the fuel to fail. These design criteria were based on the experience with metal alloy

SFR fuels gained by the operation of EBR-II and the Fast Flux Test Facility (FFTF).

First, all fuel assemblies in the core were restricted to a fast fluence limit of 4x 1023 cm-2

(E>0.1 MeV) which is the operational limit, established at FFTF, for HT-9 fast reactor grade

steel. HT-9 was used for cladding and structural components at both EBR-II and FFTF. Second,

the gas plenum pressure in the fuel pins was not allowed to exceed that typical of EBR-II which

is in the range of 30 atm to 50 atm. The gas plenum pressures calculated for the AHFTR were

found to be within the operational experience of the EBR-II fuel pin data. In fact, the gas

plenum length of the reference ABR design is sufficiently long to provide adequate space for the

production of fission gasses and transmutation helium. Third, the temperature of the inner

cladding wall was not allowed to exceed 650C to prevent or minimize the possibility of fuel-to-


289









cladding chemical or eutectic interaction which is a unavoidable characteristic of metallic alloy

fuel. Finally, all assemblies were limited to pellet centerline temperatures equal to or less than

those of the homogeneous reference ABR which are also representative of EBR-II. Also, based

on the EBR-II experience, the burnup of the AHFTR fuel was kept to within approximately 100

MWD/kg. It may be possible to increase this burnup limit as might be seen in some of the ABR

designs. However, much of the EBR-II experience dictates that the likelihood of cladding failure

became an issue at burnups higher than 100 MWD/kg due to swelling and mechanical and

chemical interaction between cladding and the metal fuel. Because, it was desirable to establish

the feasibility of a MA burner it was felt that a low probability of fuel failure was desirable. This

is especially true when one considers the various volatility and radiological hazards of

neptunium, americium, curium and californium which is present in non-trivial quantities in the

AHFTR driver fuel and targets.

Fuel Processing and Repository Considerations

Much emphasis has been put on the importance of transmuting MAs, in particular Am-241,

in this work. The impact of this transmutation on the radioactivity and the thermal heat

production of the transmuted fuel have been addressed. It was determined that the driver fuel

and transmutation targets both had specific neutron dose rates (per kg TRU) very similar to the

reference ABR. This is an indication that the higher mass actinides created during the target

irradiation, which are a source of spontaneous fission neutrons, can be burned effectively in the

fast spectrum of the active core region. Though the AHFTR driver and target fuels exhibit

higher gamma and heat emission rates than the ABR case, these rates do not exceed the worst

case scenario for multi-recycling in an LWR. Also, these higher emission rates should be

expected, considering the high throughput of americium and curium feedstock in the AHFTR

recycling center.


290









The decay heat produced by fuel recycling process losses that might be seen by a geologic

repository was also addressed. It was found that the increased concentration of Pu-238 and Cm-

244 in the AHFTR fuel cycle caused a heat generation rate plateau in the repository after a few

hundred years after actinide wastes were placed in the repository. This trend is different than

exhibited by SNF because the higher Am-241 in this LWR waste causes the heat plateau to occur

no sooner than 1000 years. Because the AHFTR heat plateau occurs earlier in the repository life,

it can be argued that it is more likely that human institutions will be able to manage the

repository performance during this peak heat time than if the repository were filled with SNF.

Fuel Cycle Economic Analysis

The economics of a mixed ABR and AHFTR fleet was considered. A SFR utility was

envisioned that operates a mix of ABRs and AHFTRs in order to burn the TRU generated by the

LWR fleet. In the mixed SFR fleet, the SNF neptunium and plutonium would be the feedstock

for refueling the ABR's. Some of this neptunium and plutonium and all of the SNF americium,

curium, berkelium and californium would be diverted to fuel the AHFTR. In this capacity, the

AHFTR becomes a dedicated MA burner in the mix of reactors. Based on the mass balance

between TRU production and destruction rates for each reactor, a support ratio of 3.3 MWth of

ABRs per one MWth of AHFTRs was found. Also, 0.6 MWth of SFRs are needed per one

MWth of LWRs.

This economics analysis shows that the conversion of americium into plutonium reduces

the reprocessing demand for separating plutonium from SNF. This reduces the reprocessing

costs to the reactor. Also because, the reactor consumes MAs and Tc-99, which are high level

wastes (HLW), the reactor provides a valuable service to the SFR utility. The fact that the

AHFTR saves fuel costs by transmuting waste atoms into fuel atoms is a direct example of how









the reactor attains an observably high fissile conversion ratio while at the same time being a

transuranic burner.

A sensitivity analysis of fuel costs on reprocessing unit costs indicates that if the cost of

HLW disposal can be reduced to $125/kgHM due to the removal of MAs (and Tc-99), then the

average fuel cost to the mixed-fleet SFR utility would be competitive with the cost of the once-

through fuel cycle assuming the current-day uranium ore price of $80/lb U308.

Concluding Remarks

It is recommended that if SFRs and LWRs are to exist in a symbiotic fuel relationship, then

their waste management practices be balanced in such a way as to deliver the most amount of

reactivity per neutron spectrum requirements. Sodium fast "breeder" reactors were envisioned at

the beginning of the commercial nuclear industry because at that time it was commonly believed

that uranium resources would become scarce. This was not the case and today LWRs use the

fissile U-235 atom in enriched uranium to drive the chain reaction in those reactors. However,

thermal spectrums as a general rule, with the exception of some thorium reactors, are deficient in

the reactivity needed to sustainably bombard non-fissile material. This, in fact, is the nature of

the breeder reactor. Breeder reactors possess the neutron balance between fissile production

sources and sinks to continuously transmute non-fissile (i.e., fertile) isotopes into fissile isotopes.

Due to their fission threshold nature and the fact that their transmutation leads to plutonium

isotopes, MAs are by definition fertile. Therefore, investing neutrons into their transmutation

into fissile material is the key element for extracting their maximum fissile worth. It is true that

a high leakage core will reduce the net transuranic production. However, creating a high leakage

core without a blanket (radial or axial) of fertile material will waste neutrons that could

otherwise be used for transmutation. If a blanket of fertile MAs is used in place of uranium for


292









the purpose of breeding plutonium, then by definition a fissile breeder is created by enhancing

transuranic minor actinide burning.


293













APPENDIX A

REBUS INPUT DECK


Key components of the REBUS input deck used to model the final AHFTR design is given


in this appendix.







REBUSINPUTAPPENOX.txt
.:'.,S M i:p r l i Ui|, ) *:-r C':METi aB B tainaer l B.B ean nai-t Birt aea ea
(:.:raTECIL A SEM L> i l.ii
30 CR01A ... 0 .. OC, OO .E. ,0 4 EF'fi ,..,-1
30 CROi 1 0 0 4 '4.)1 ) 1'"R0E-,)
30 CR01C 1 0 0 '..li'80.01 1.14660E+02
30 CROID 1 0 0 1.Vl'-,:.'"- 1.28980E+02
30 CR01E 1 0 0 1.I0,i6-i0; 1.13.00E-07
30 CR01F 1 0 0 1.1 OEli-02 1 12 i.;.'
30 CR01G 1 0 0 1 *;'.c-i..02 l.-.'0l-i.,
30 CR01H 1 0 0 1. -: O .- !i r,;.- -i.
30 '.,1,H 1 0 0 1 I't lltO .' 2.,:'.6 : : -. :
30 R,)Il 1 0 0 2., t2,i'.'E,2 2.7 2 0zGO.
30 CR013 1 0 0 2 3. 2t.''2 2."'JlC_.-i
30 CR01K 1 0 0 1 tb'E.? 3.9'0l-0.
**INNER CORE/ ICO1 / ROW 01 FUIE' I.iEMALL.S'
30 1,:,l4 2 0 0 WO(-000 E-O i 4 S 4'. E
30 1 1B 2 0 0 4 ;* 4i Et.'.! '1 ,'.i4 ..i
30 IC01C 2 0 0 9 1'1E. 01 1 146tliEii2
30 ICOlo 2 0 0 1 l E. 1 _',: ',IE.
30 IC01E 2 0 0 1 1 >0(EI 02
30 IC01F 2 0 0 1 itL.:. I i0E
30 IC01G 2 0 0 1. 'l 20._ 1 'A 1 f.0
30 IC01H 2 0 0 1 "l-'iE ~6 Si 660E.
30 TGO1H 2 0 0 1.5 l: i :U .V 6,U0 ;-)
30 ICOl1 2 0 0 .0La;rjEj 2 i _-O
30 IcO13 2 0 0 .2. l 2 2. ?-I .,:
30 ICOlK 2 0 0 .'iE.u ; 'iif,
**INNER CORE / T102 / ROW 02 FUEL ASSEMBLIES**
30 ICO2A 3 0 0 O.OOOOOE+00 .. '86 l)COi-.)1
30 I...:J 3 0 04 s55 i 'E,'il i l. i-;e l
30 it:02C 3 0 0 c9 1'50E.1I 1. ibEi, Wi' 2
30 TC00r 3 0 0 1 4u.E-2' I ')RO i,2
30 i'0',E 3 0 0 1 jS)'it( .) 1.43300E+02
30 IC02F 3 0 0 I 4;.:"E.l? 1 5-,70E.02
30 Tr0?G 3 0 0 1 'r:C>E ,? "I 'i1 E'2
30 1i'.',21 3 0 0 1 "I'1- E .2 n i l ~?i:r .
30 TGO2H 3 0 0 1 .-;E, ,.i' 06OE,(. i.
30 T'2T? 3 0 0 2 '56 26 :E-
l1 '.ii' 3 0 0 i .' 2 E-
30 ICO2K 3 0 0 2.74295E602 1 '9-00E-*0
**INNER CORE / IC03 / ROW 03 FUEL ASSEMBLIES**
30 I!,0_ 4 0 0 0.00OC E.-0O 4.58640E+01
30 I.i 4 0 4 0 0 I4 ri-4,)L 9.17280E+01
30 IC03C 4 0 0 9.17280Et01 1 l-(t.,...,'
30 IC03D 4 0 0 I i-iFrE .rJ2 1 I ','. O .
30 ICO3E 4 0 0 i.L SS,.i:E. 2 1 7F C.i)0)
30 IC03F 4 0 0 1 ?O410E.02 I :'.,Ei,2
30 I.-',"I 4 0 0 1 "r-iE."'" 1.71940E+02
30 icO'i' 4 0 0 1 'lil.E .i0? .!. rrOE.o0
30 TG03H 4 0 0 1.86260E+02 2 i't.it)f E'
30 IC03I 4 0 0 ?2 06C Oe.M2 ? ii2"-iE-.')
30 IC033 4 0 0 2 2]c.)E'.j 2.74295E+02
30 IC03K 4 0 0 2 "1p'95.-0; '400E+02
t'INNER CORE / TrC4 / ROW 0- FlJrL 4 iSi:'RLIE i**
30 TC1-4, S 0 0 : u0O,-MQE-00 4.58640E+01
30 IC' 5 0 0 4 7~,:iii -01 9 !-?SOE-0C
3', IC04C 5 0 0 9.17280E+01 1 14mI+.0_
30 IC04D 5 0 0 1 -I6r-0c E-1.; i .'9,E.'?
30 IC04E 5 0 0 1 ,9,'."'--2 1 43OnE-n?
30 IC04F 5 0 0 1.43300E+02 1 .'60 'iu,'l
30 IC04G 5 0 0 1. 5-Gr0E-02 i ,"iiEL r,
30 IC04H 5 O 0 1.'1' ,'t-(0? I '*'-.Or'
Page 1







Figure A-1. REBUS A.NIP Type 30 Geometry Cards and A.BURN Type 11 In-Core Fuel

Management Cards


294


















REBUSINPUTAPPENDX.txt
30 CRO1E 5 17 0 1.28980E+02 1.43300E+02
30 CR01F 5 17 0 1.43300E+02 1.57620E+02
30 CR01G 5 17 0 1.57620E+02 1.71940E+02
30 CRO1H 5 17 0 1.71940E+02 1.86260E+02
30 CROHH 5 17 0 1.86260E+02 2.06260E+02
30 CR01I 5 17 0 2.06260E+02 2.33260E+02
30 CR013 5 17 0 2.33260E+02 2.74295E+02
30 CRO1K 5 17 0 2.74295E+02 3.97400E+02
"*CONTROL ASSEMBLY (5,19) **
30 CR01A 5 19 0 0.00000E+00 4.58640E+01
30 CROl1 5 19 0 4.58640E+01 9.17230E+01
30 CRO1C 5 19 0 9.17280E+01 1.14660EE+0
30 CRO1D 5 19 0 1.14660E+02 1.28980E+02
30 CR01E 5 19 0 1.28980E+02 1.43300E+02
30 CR01F 5 19 0 1.43300E+02 1.57620E+02
30 CR01G 5 19 0 1.57620E+02 1.71940E+02
30 CRO1H 5 19 0 1.71940E+02 1.86260E+02
30 CROHH 5 19 0 1.86260E+02 2.06260E+02
30 CR01I 5 19 0 2.06260E+02 2.33260E+02
30 CR013 5 19 0 2.33260E+02 2.74295E+02
30 CRO1K 5 19 0 2.74295E+02 3.97400E+02
**CONTROL ASSEMBLY (5,21) **
30 CR01A 5 21 0 0.00000E+00 4,58640E+01
30 CR01B 5 21 0 4.58640E+01 9.17280E+01
30 CRO1C 5 21 0 9.17280E+01 1.14660E+02
30 CR010 5 21 0 1.14660E+02 1.28980E+02
30 CRO1E 5 21 0 1.28980E+02 1.43300E+02
30 CROIF 5 21 0 1.43300E+02 1.57620E+02
30 CR01G 5 21 0 1.57620E+02 1.71940E+02
30 CROIH 5 21 0 1.71940E+02 1.86260E+02
30 CROHH 5 21 0 1.86260E+02 2.06260E+02
30 CR01I 5 21 0 2.06260E+02 2.33260E+02
30 CR013 5 21 0 2.33260E+02 2.74295E+02
30 CR01K 5 21 0 2.74295E+02 3.97400E+02
"*CONTROL ASSEMBLY (5,23) **
30 CR01A S 23 0 0.00000E+00 4.58640E+01
30 CR01B 5 23 0 4.58640E+01 9.17280E+01
30 CRO1C 5 23 0 9.17280E+01 1.14660E+02
30 CR01D 5 23 0 1.14660E+02 1.28980E+02
30 CRO1E 5 23 0 1.28980E+02 1.43300E+02
30 CROIF 5 23 0 1.43300E+02 1.57620E+02
30 CR01G 5 23 0 1.57620E+02 1.71940E+02
30 CRO1H 5 23 0 1.71940E+02 1.86260E+02
30 CROHH 5 23 0 1.86260E+02 2.06260E+02
30 CR011 5 23 0 2.06260E+02 2.33260E+02
30 CR013 5 23 0 2.33260E+02 2.74295E+02
30 CRO1K 5 23 0 2.74295E+02 3.97400E+02
#################################################################################COM
POSITIONS MAPPED INTO REGIONS WITHIN REACTOR##########################
"**FUEL MANAGEMENT PATH INNER CORE****
11 CPL01 0 1ICSC ICO1D 2ICSC ICOlD
11 CPLO1 0 3ICSC IC01D 4ICSC IC010
11 CPLO1 0 SICSC IC01D 6ICSC IC01O
11 CPLO1 0 7MELTR
**"*FUEL MANAGEMENT PATH INNER CORE`***
11 CPLO2 0 1ICSC ICO1E 2ICSC ICOlE
11 CPLO2 0 3ICSC ICO1E 41CSC ICO1E
11 CPLO2 0 5ICSC IC01E 6ICSC ICOlE
11 CPLO2 0 7MELTR
****FUEL MANAGEMENT PATH INNER CORE'***
11 CPL03 O IICSC IC01F 2IC5C ICO1F
11 CPL03 0 3ICSC ICO1F 4ICSC ICOIF
Page 6








Figure A-1. Continued.


295










APPENDIX B
PARAMETRIC DESIGN ANALYSIS

The full results of the parametric analysis performed in Chapter 3 are given in this

appendix. The main parameters varied in this analysis were the core height and pin pitch-to-

diameter ratio (Table 3-1 and Table 3-2). The reactor thermal power of the reactor core was held

constant at 1000 MWth for all of these cases. The burnup in the first row of fuel at the core mid-

plane was restricted to 18 at. %. These analyses gave an indication of the affects of core and fuel

pin geometry on the: TRU enrichment, excess reactivity and cycle length. The down selections

to the "tall" and "flat versions of the AHFTR in Chapter 3 were made based on the results of this

parametric analysis.



4.0%

3.500

3.000



0. 2.00o
1.500
LU. 1.00o

0.500 1357
0.0O0 1.293
1.176 p/d
1016 91.6 81.6 1.1
81.6 71.6 616 51.6
61.6
51.6
Active Core Height (cm)

Figure B-1. Excess Reactivity of the core design in Figure 3-1 "tall" for varying core height and
p/d (REBUS)


296






















In I
Z

^ti



x
w


3.5000o

3.0000

2.5000

2.0000

1.5000

1.0000

0.5000

0.0000


r 1.357
1.293
1.176 1


61.6 51.6
51.6
Active Core Height (cm)

Figure B-2. Excess Reactivity of the core design in Figure 3-14 "flat" for varying core height
and p/d (REBUS)


1 2500

S2000
I-
1500

1000


1.357
1.293
1.176 p
1.1


Active Core Height (cm)


Figure B-3. TRU enrichment of the core given in Figure 3-1 "tall" for varying core height and
p/d (REBUS)


297






















I-

I-


1.357
000 1.293
1.176 p/d
91.6 81.6 1 1.1

61.6
51.6
Active Core Height (cm)

Figure B-4. TRU enrichment of the core design in Figure 3-14 "flat" for varying core height and
p/d (REBUS)


350

300

250

200

150

100

50


61.6
1.176 51.6
1 9293


101.6


Active Core
Height (cm)


p/d 1.357

Figure B-5. Cycle length of the core given in Figure 3-1 "tall" for varying core height and p/d
(REBUS)


298


















O 300
a-
I..
250

S200
-J
150

0 100
101.6
50

0
61.6 Active Core
1 1.176 19 51"6 Height (cm)
1.293
p/d 1.357

Figure B-6. Cycle length of the core design in Figure 3-14 "flat" for varying core height and p/d
(REBUS)


299









LIST OF REFERENCES


1. US Department of Energy, "The History of Nuclear Energy," US Department of Energy,
Office of Nuclear Energy, Science and Technology, DOE/NE-0088, Washington, DC.

2. E. BLAKE, "US Capacity Factors: A Small Gain to an Allready Large Number," Nuclear
News, 27-32, La Grange Park, Illinois, (May 2007).

3. R. COCHRAN and N. TSOULFANIDIS, The Nuclear Fuel Cycle: Analysis and
Management, 2nd ed, American Nuclear Society, La Grange Park, Illinois, (1999).

4. "The Nuclear Waste Policy Act: As Ammended," US Department of Energy, Office of
Civillian Radioactive Waste Management, Washington, DC, (2004).

5. "The Future Path for Advanced Spent Fuel Treatment and Transmutation Research," US
Department of Energy, Office of Nuclear Energy, Science and Technology, Report to Congress
on Advanced Fuel Cycle Initiative, Washington, DC, (2003).

6. "Nuclear Waste Fund Fee Adequacy: An Assessment," US Department of Energy, Office
of Civillian Radioactive Waste Management, DOE/RW-0534, Washington, DC, (2001).

7. E. HOFFMAN, R. HILL, and M. WILLIAMSON, "Preliminary Core Design Studies for
the Advanced Burner Reactor Over a Range of Conversion Ratios," ANL-AFCI-177, Argonne
National Laboratory, Argonne, Illinois, (2006).

8. E. SHWAGERAUS, P. HEJZLAR, and M. KAZIMI, "A Combined Non-Fertile and UO2
PWR Fuel Assembly for Actinide Waste Minimization," Nuclear Technology, 149, 281-303
(2005).

9. W. CARMACK, R. FIELDING, P. MEDVEDEV, M. MEYER, M. TODOSOW, H.
HAMILTON, J. NINO, S. PHILPOT, and J. TULENKO, "AECL/US INEERI-Development of
Inert Matrix Fuels for Plutonium and Minor Actinide Management in Power Reactors-Fuel
Requirements and Down Select Report," INL/EXT-05-00436, Idaho National Laboratory, Idaho
Falls, Idaho, (2005).

10. V. BERTHOU, C. DEGUELDRE, and J. MAGILL, "Transmutation Characteristics in
Thermal and Fast Neutron Spectra: Application to Americium," Journal ofNuclear Materials,
320, 156-162 (2003).

11. F. MANN and R. SCHENTER, "Calculated Neutron Capture Cross Sections to the
Americium Ground and Isomeric States," Nuclear Science and Engineering, 63, 242-249 (1977).

12. "Yucca Mountain Science and Engineering Report, Rev. 1," US Department of Energy,
Office of Civillian Radioactive Waste Management, DOE/RW-539-1, Washington, DC, (2002).

13. R. WIGELAND and T. BAUER, "Repository Benefits of AFCI Options", ANL-AFCI-129,
Argonne National Laboratory, Argonne, Illinois, (2004).


300










14. K. WIRTZ, Lectures on Fast Reactors, American Nuclear Society, La Grange Park,
Illinois, (1982)

15. H CHOI and T. DOWNAR, "A Liquid-Metal Reactor for Burning Minor Actinides of
Spent Light Water Reactor Fuel I," Nuclear Science and Engineering, 133, 1-22 (1999).

16. A. STANCULESCU. J. GARNIER, J. ROUAULT, E. KIEFHABER, and R.
SUNDERLAND, "Plutonium Burning and Minor Actinides Transmutation in a Fast Reactors:
First Results Obtained within the Frame of the CAPRA Program," Proc. International Nuclear
Congress: Atoms for Energy, Oct 2-6, 1994, Lyon, France, (1994).

17. A. VASILE,G. RIMPAULT, J. TOMMASI, C. DE SAINT JEAN, M. DELPECH,
K.HESKETH, H, BEUAMONT, R. SUNDERLAND, T. NEWTON, P. SMITH, W. MASCHEK,
D. HAAS, C. RAEDT, G. VAMBENEPE, and J. LEFEVRE, "Fast Reactors Fuel Cycle Core
Physics Results from the CAPRA-CADRA Programme," Proc. ofGlobal01, September 13-17,
2001, Paris, France, (2001).

18. E. GLUEKLER, "US. Advanced Liquid Metal Reactor (ALMR)," Progress in Nuclear
Energy, 31, No. 1-2, 43-61 (1997).

19. A. DUBBERLY, K. YOSHIDA, C. BOARDMAN, and T. WU, "Superprism Oxide and
Metal Fuel Core Designs," Proc. of ICONE8: International Conference on Nuclear Engineering,
April 2-6, 2000, Baltimore, Maryland, (2000).

20. E. MORRIS and M. SMITH, "Development of Low Conversion Ratio Fast Reactors for
Transmutation," ANL-AAA-057, Argonne National Laboratory, Argonne, Illinois, (2002).

21. L. BUIRON, F. VARAINE, D. LORENZO, H. PALANCHER, and B. VALENTIN,
"Minor Actinides Transmutation in SFR Depleted Uranium Radial Blanket, Neutronic and
Thermal Hydraulic Evaluation," Proc. of Global07, September 9-13, 2007, Boise, Idaho, (2007).

22. K. FUJIMURA, T. SANDA, M. MAYUMI, S. MORO, M. SAITO, and H. SEKIMOTO,
"Feasibility Study of Large Mox Fueled FBR Core Aimed At the Self-Consistent Nuclear
Energy System," Progress in Nuclear Energy, 37, No. 1-4, 177-185, (2000).

23. K FUJIMURA, T. SANDA, M. YAMAWAKI, and K. KNASHI, "Enhancement of
Transmutation Characteristics of the Minor Actinide Burning Fast Reactor Core Concept Using
Hydride Fuel Targets and Its Introduction Scenario," Journal ofNuclear Science and Technology,
38, No. 10, 879-886 (2001).

24. M. YAMAWAKI, H. SUWARNO, T. YAMAMATO, T. SANDA, K. FUJIMURA, and K.
KAWASHIMA, "Concept of Hyrdide Fuel Target Subassemblies in a Fast Reactor Core for
Effective Transmutation of MA," Journal ofAlloys and Compounds, 271-273, 530-533 (1998).









25. V. ELISEEV and E. POPLAVSKAYA, "Possibilities of Deep Burnup of Americium and
Neptunium in the Core of a Fast Reactor," Atomic Energy, 96, No. 3, 170-176 (2004).

26. M. ROME, A. HARISLUR, E. HERNIOU, A. ZAETTA, and J. TOMMASI, "Use of Fast
Reactors to Burn Long-Life Actinides, Especially Am, Produced by Current Reactors, "Proc. of
Physor96, September 16-20, 1996, Mito, Japan, (1996).

27. M. ROME A. HARISLUR, and E. HERNIOU, Study of Once-Trhough Americium
Matrices Strategy in Fast Reactors," Proc. oflCONE5, May 26-30, 1997, Nice, France, (1997).

28. T. SANDA, K. FUJIMURA, K. KOBAYASHI, K. KAWASHIMA, M. YAMAWAKI, and
K. KONASHI, "Fast Reactor Core Concepts for Minor Actinide Transmutation Using
HydrideFuel Targets," Journal ofNuclear Science and Technology, 37, No. 4, 335-343 (2000).

29. International Atomic Energy Agency, "Fast Reactor Database", Last Visited: March 1,
2008, http://www.iaea.org/inisnkm/nkm/aws/frdb/index.html, (2008).

30. H KURAISHI, T. SAWADA,H. NINOKATA, and H. ENDO, Inherent and Passive Safety
Sodium-Cooled Fast Reactor Core Design with Minor Actinide and Fission Product Incineration,
Nuclear Science andEngineering, 138, 205-222 (2001).

31. K. ARIE, M. SUZUKI, M. KAWASHIMA, R. FUJITA, H. ENDO, A. TOUKURA, M.
SAITO, and Y. FUJII-E, "Long-Lived Fission Product Burning Based on the Actinide Recycle
Metal Fuel Core," Progress in Nuclear Energy, 32, No. 3-4, 665-672 (1998).

32. Y. KIM, S. LEE, J. KIM, and W. PARK, "Feasibility of Using Burnable Poisons for
Reduction of Coolant Void Reactivity in LMR for TRU Transmutation," Proc. ofPhysor04,
April 25-29, 2004, Chicago, Illinois, (2004).

33. M. MESSAOUDI, and J. TOMMASI, "Fast Burner Reactor Devoted to Minor Actinide
Incineration," Nuclear Technology, 137, 84-96 (2002).

34. R. RICE, R. CURRAN, F. MCGINNIS, M NOVICK, and F. THALGOTT, "EBR-I, Mark
Ill-Design Report," ANL-5836, Idaho National Laboratory, Idaho Falls, Idaho, (1958).

35. L. KOCH, EBR-II: Experimental Breeder Reactor, ANL Internal Publication, Argonne
National Laboratory, Argonne, Illinois, (2004).

36. Y. CHANG, R. HILL, E. FUJITA, D. WADE, Y. KUMAOKA, M. SUZUKI, M.
KUWASHIMA, and H. NAKAGAWa, Core Concepts for "Zero-Sodium-Void-Worth Core in
Metal Fueled Fast Reactor," Proc. of International Conference on Fast Reactors and Related
Fuel Cycles, Oct. 28 Nov. 1, 1991, Kyoto, Japan, (1991).

37. G. VANDEGRIFT, M. REGALBUTO, S. AASE, A. BAKEL, T. BATTISTI, D.
BOWERS, J. BYRNES, M. CLARK, J. EMERY, J. FALKENBERG, A. GELIS, C. PEREIRA,
L. HAFENRICHTER, Y. TSAI, K. QUIGLEY and M. VANDER POL, "Designing and


302









Demonstration of the UREX+ Process Using Spent Nuclear Fuel," Proc. ofA TA TLANTE 2004:
Advances for Future Nuclear Fuel Cycles, June 21-24, 2004, Nimes, France, (2004).

38. R. FERRER, M. ASGARI, S. BAYS, and B. FORGET, "Computational Neutronics
Methods and Transmutation Performance Analysis for Fast Reactors," INL/EXT-07-12466,
Idaho National Laboratory, Idaho, (2007).

39. S. PILLON, J. SOMERS, S. GRANDJEAN, and J. LACQUEMENT, "Aspects of
Fabrication of Curium-Based Fuels and Targets," Journal ofNuclear Materials, 320, 36-43
(2003).

40. J. ACKERMAN, T. JOHNSON, and J. LAIDLER, "Waste Removal in Pyrochemical Fuel
Processing for the Integral Fast Reactor," Proc. of The Interational Symposium on Actinides:
Processing & Materials, February 27-March 3, 1994, San Francisco, California, (1994).

41. H. HENRYSON II, B. TOPPEL, and C. STENBERG, "MC2-2: A Code to Calculate Fast
Neutron, Spectra and Multi-Group Cross-Sections," ANL-8144, Argonne National Laboratory,
Argonne, Illinois, (1976).

42. K. DERSTINE, DIF3D: "A Code to Solve One-, Two-, and Three-Dimensional Finite-
Difference Diffusion Theory Problems," ANL-82-64, Argonne National Laboratory, Argonne,
Illinois, (1984).

43. B. TOPPEL, "A User's Guide to the REBUS-3 Fuel Cycle Analysis Capability," ANL-83-
2, Argonne National Laboratory, Argonne, Illinois, (1983).

44. R. MCKNIGHT, J. STILLMAN, B. TOPPEL, and H. KHALIL, "Validationand
Applicationof a Physics Database for Fast Reactor Fuel Cycle Analysis," Proc. ofANS: Topical
Meeting on Advances in Reactor Physics, April 11-15, 1994, Knoxville, Tennessee, (1994).

45. J. DUDERSTADT and L. HAMILTON, Nuclear Reactor Analysis, 1st ed., John Wiley &
Sons, Inc., New York, (1976).

46. J. HOOVER, G. LEAF, D. MENELEY, and P. WALKER, "The Fuel Cycle Analysis
System, REBUS Reactor Burnup System," Nuclear Science and Engineering, 45, 52-65 (1971).

47. D. HAMRIN, SCALE: "A Modular Code System for Performing Standard Computer
Analysis for Licensing Evaluation," ONRNL/TM-2005/39, Oak Ridge National Laboratory, Oak
Ridge, Tennessee, (2006).

48. Los Alamos National Laboratory: X-5 Monte Carlo Team, "MCNP A General Monte
Carlo N-Particle Transport Code: Version 5," LA-UR-03-1987, Los Alamos National Laboratory,
Los Alamos, New Mexico, (2003).


303









49. D. POSTON and H. TRELLUE, "MONTEBURNS 2.0: An Automated, Multi-Step Monte
Carlo Burnup Code System," LA-UR-99-4999, Los Alamos National Laboratory, Los Alamos,
New Mexico, (1999).

50. A. CROFF, "User's Manual for ORIGEN2 Computer Code," ORNL/TM-7175, Oak Ridge
National Laboratory, Oak Ridge, Tennessee, (1980).

51. D. PELOWITZ (editor), "MCNPX User's Manual: Version 2.5.0," LA-CP-05-0369, Los
Alamos National Laboratory, Los Alamos, New Mexico, (2005).

52. R. HILL, D. WADE, E. FUJITA, and H. KHALIL, "Physics Studies of Higher Actinide
Consumption in an LMR," Proc. of the International Conerence on thePhysics ofReactors, April
23-27, 1990, Marseille, France, (1990).

53. M. SMITH, E. MORRIS, and R. HILL, "Physics and Safety Studies of a Low Conversion
Ratio Sodium Cooled Fast Reactor," Proc. of Global03, November 16-20, 2003, New Orleans,
Louisanna, (2003).

54. J. CAHALAN, M. SMITH, R. HILL, and F. DUNN, "Physics and Safety Studies of a Low
Conversion Ratio Sodium Cooled Fast Reactor," Proc. ofPhysor04, April 25-29, 2004, Chicago,
Illinois, (2004).

55. T. KIM and W. YANG, "A Metal Fuel Core Concept for 1000 MWt Advanced Burner
Reactor (Interim Report)," ANL-AFCI-185, Argonne National Laboratory, Argonne, Illinois,
(2007).

56. W. YANG, T. KIM, and C. GRANDY, "A Metal Fuel Core Concept for 1000 MWt
Advanced Burner Reactor," Proc. of Global07, September 9-13, 2007, Boise, Idaho, (2007).

57. R. HOSTENY, "The ARC System Fuel Cycle Analysis Capability: REBUS-2," ANL-
7721, Argonne National Laboratory, Argonne, Illinois, (1978).

58. "Advanced Fuel Cycle Initiative (AFCI) Comparison Report, FY 2006 Update," US
Department of Energy, Office of Nuclear Energy, Science and Technology, Report to Congress
on Advanced Fuel Cycle Initiative, Washington, DC, (2006).

59. K. MOORE and W. YOUNG, "Phase Studies of the Zr-H System at High Hydrogen
Concentrations," Journal ofNuclear Materials, 27, 316-324 (1968).

60. L. WALTERS, and J. KITTLE, "Development and Performance of Metal Fuel Elements
for Fast Breeder Reactors," Nuclear Technology, 48, 273-280 (1980).

61. L. WALTERS, and B. SEIDEL, "Performance of Metallic Fuels and Blankets in Liquid-
Metal Fast Breeder Reactors," Nuclear Technology, 65, 179-231 (1984).


304









62. A. WALTAR and A. REYNOLDS, Fast Breeder Reactors, Pergamon Press Inc., Maxwell
House, Fairview Park, Elmsford, New York, (1981).

63. D. GELLES, "Microstructural Development in Reduced Activation Ferritic Alloys
Irradiated to 200 dpa at 420C, Journal ofNuclear Materials, 212-215, 714-719 (1994).

64. H. TSAI, A. COHEN, M. BILLONE, and L. NEIMARK, "Irradiation Performance of U-
Pu-Zr Metal Fuels for Liquid Metal Cooled Reactors," Proc. oflCONE3, April 23-27, 1995,
Kyoto, Japan, (1995).

65. R. MEULEKAMP and S. VAN DER MARCK, "Calculating the Effective Delayed
Neutron Fraction with Monte Carlo," Nuclear Science and Engineering, 152, 142-148 (2006).

66. W. YANG, Y. KIM, R. HILL, T. TAIWO, and H. KHALIL, "Long-Lived Fission Product
Transmutation Studies," Nuclear Science and Engineering, 146, 291-318 (2004).

67. R. HILL, D. WADE, J. LIAW, and E. FUJITA, "Physics Studies of Weapons Plutonium
Disposition in the Integral Fast Reactor Closed Fuel Cycle," Nuclear Science andEngineering,
121, 17-31 (1995).

68. RHODES, T. BAUER, G. STANFORD, J. REGIS, and C. DICKERMAN, "Fuel Motion
in Overpower Tests of Metallic Integral Fast Reactor Fuel," Nuclear Technology, 98, 91-99
(1992).

69. H. PLANCHON, R. SINGER, D. MOHR, and E. FELDMAN, "The Experimental Breeder
Reactor II Inherent Shutdown and Heat Removal Tests Results and Analysis," Nuclear
Engineering and Design, 91, 287-296 (1986).

70. G. GOLDEN, H. PLANCHON, J. SACKETT and R. SINGER, "Evolution of Thermal-
Hydraulics Testing in EBR-II," Nuclear Engineering andDesign, 101, 3-12 (1987).

71. D. MOHR, L.CHANG, E. FELDMAN, P. BETTEN, and H. PLANCHON, "Loss-of-
Primary-Flow-Without-Scram Tests: Prestest Predictions and Preliminary Results," Nuclear
Engineering andDesign, 101, 45-46 (1987).

72. J. HERZOG, T. AOYAMA, H. TSAI, and E. DEAN, "A Transient Overpower Experiment
in EBR-II", Proc. ofARS94: International Meeting:, Advanced Reactor Safety, April 17-21,
1994, Pittsburgh, Pensyvania, (1994).

73. H. OIGAWA, S. IIJIMA, and M. BANDO, "Experiments and Analysis on Fuel Expansion
and Bowing Reactivity Worth in Mock-up Cores of Metallic Fueld Fast Reactors at FCA,"
Journal ofNuclear Science and Technology, 36, No. 10, 902-913 (1999).

74. OHMAE, A. MORINO, N. NAKAO, and S. HIRAO, "Channel Deformation Analysis for
Fast Reactor Fuel Assemblies Undergoing Swelling and Thermal Bowing," Nuclear Engineering
andDesign, 23, 309-320 (1972).


305










75. P. REINHALL, K. PARK, and R. ALBRECHT, "Analysis of Mechanical Bowing
Phenomena of Fuel Assemblies in Passively Safe Advanced Liquid-Metal Reactors," Nuclear
Technology, 83, 197-204 (1988).

76. S. KAMAL and Y. ORECHWA, "Bowing of Core Assemblies in Advanced Liquid Metal
Fast Reactors," Proc. of Joint ASME ANS Nuclear Power Conference, July 20-23, 1986,
Philadelphia, Pensylvania, (1986).

77. "Preapplication Safety Evaluation Report for the Power Reactor Innovative Small Module
(PRISM) Liquid-Metal Reactor Final Report," Office of Nuclear Reactor Regulation, U.S.
Nuclear Regulatory Commission, NUREG-1368, Rockville, Maryland (1994).

78. G. PALMIOTTI, E.LEWIS, and C. CARRICO, "VARIANT: VARIational Anisotropic
Nodal Transport for Multidimensional Cartesian and Hexagonal Geometry Calculation," ANL-
95/40, Argonne National Laboratory, Argonne, Illinois, (1995).

79. S. GLASSTONE and A. SESONSKE, Nuclear Reactor Engineering: 3rd Edition, Van
Nostrand Reinhold Company, New York, New York, (1981).

80. Y. CHANG, P. FINCK and C. GRANDY, "Advanced Burner Test Reactor Preconceptual
Design Report," ANL-ABR-1, Argonne National Laboratory, Argonne, Illinois, (2006).

81. K. KONASHI, B. TSUCHIYA, M. YAMAWAKI, K. FUJIMURA, and T. SANDA,
"Development of Actinide-Hydride Target for Transmutation of Nuclear Waste," Proc. of
GlobalO, September 13-17, 2001, Paris, France, (2001).

82. C. TILL, Y. CHANG, and W. HANNUM, "The Integral Fast Reactor An Overview,"
Progress in Nuclear Energy, 31, No. 1-2, 3-11 (1997).

83. D. WADE and R. HILL,"The Design Rationale of the IFR," Progress in Nuclear Energy,
31, No 1-2, 13-42 (1997).

84. B. HILTION, D. PORTER, and S. HAYES, "AFC-1 Transmutation Fuels Post Irradiation
Hot Cell Examination 4 to 8 at.% Final Report," INL/EXT-05-00785, Idaho National Laboratory,
Idaho Falls, Idaho, (2006).

85. D. OLANDER, Fundamental Aspects ofNuclear Fuel Ellements, National Technical
Information Center, Office of Public Affairs Energy Research and Development Administration,
Springfield, Virginia, TID-26711-P1, (1976).

86. G. HOFMAN, L. WALTERS, and T. BAUER, "Metallic Fast Reactor Fuels," Progress in
Nuclear Energy, 31, No. 1-2, 83-110 (1997).

87. K. EHRLICH, J. KONYS, and L. HEIKINHEIMO, "Materials for High Performance Light
Water Reactors," Journal ofNuclear Materials, 327, 140-147 (2004).


306










88. R. PAHL, C. LAHM, and S. HYES, "Performance ofHT9 Clad Metallic Fuel at High
Temperatures," Journal ofNuclear Materials, 204, 141-147 (1993).

89. C. LAHM, J. KOENIG, P. BETTEN, J. BOTTCHER, W. LEHTO, and B. SEIDEL,
"EBR-II Driver Fuel Qualification for Loss-of-Flow and Loss-of-Heat-Sink Tests without
Scram," Nuclear Engineering andDesign, 101, 25-34 (1987).

90. TODREAS and M. KAZIMI, Nuclear Systems: Vol. 1, Thermal Hydraulic Fundamentals,
Hemisphere, New York, (1990).

91. PETRI, and M. DAYANANDA, "Isothermal Diffusion in Uranium-Plutonium-Zirconium
Alloys," Journal ofNuclear Materials, 240, 131-143 (1997).

92. T. TAIWO, T. KIM, J. STILLLMAN, R. HILL, M. SALVATORES, and P. FINCK,
"Assessment of a Heterogeneous PWR Assembly for Plutonium and Minor Actinide Recycle,"
Nuclear Technology, 155, 34-54 (2006).

93. J. PLAMBECK, and J. EIDER, H. "Laitinen, Electrochemisty of the Lithium Hydride
Cell," Journal of the Electrochemical Society, 113, No. 9, 931-937 (1966).

94. T. NOHIRA and Y. ITA, "Electrochemical Hydrogen Absorbing Behavior of Pd and Pd-Li
Alloys in a Molten LiCl-KCI-LiH System," Journal of the Electrochemical Society, 144, No. 7,
2290-2295 (1997).

95. H. ITOZ and Y. HASEGAWA, "Electrode Behavior of Hydrogen Reduction in LiCl-KCl
Melt: Voltammetric Analysis," Journal of the Electrochemical Society, 147, No. 1, 285-295
(2000).

96. C. TRYBUS, J. SANECKI, and S. HENSLEE, "Casting of Metallic Fuel Containing
Minor Actinide Additions," Journal ofNuclear Materials, 204, 50-55 (1993).

97. B. HILTON, S. HAYES, D. PORTER, M. MEYER, J. KENNEDY, G. CHANG, D.
UTTERBECK, and R. AMBROSEK, "Irradiation Performance of Fertile-Free Metallic Alloys
for Actinide Transmutation," IMF 1: Workshop on the Research, Development and
Implementation ofFissile-Free Nuclear Fuels, October 10-12, 2006, Park City, Utah, (2006).

98. A. DUBBERLY, C. BOARDMAN, D. CARROLL, C. EHRMAN, and C. WALTER, "S-
PRISM Fuel Cycle Study," Proc. oflCAPP03, May 4-7, 2003, Cordoba, Spain, (2003).

99. M. BUNN, S. FETTER, J. HOLDREN, and B. VAN DER ZWAAN, "The Economics of
Reprocessing vs. Direct Disposal of Spent Nuclear Fuel," Harvard University/DE-FG26-
99FT4028, Cambridge, Massachusetts, (2003).

100. E. BECKJORD, Executive Director, "The Future of Nuclear Power," Massachusetts
Institute of Technology, Cambridge, Massachusetts, (2003).


307










101. L. KOCHETKOV, A. MALENKOV, V. POPLAVSKII, O. SARAEV,K. SUKNEV, and A.
KIRYUSHIN, "Achieving Competitiveness in Fast Reactors with Sodium Coolant," Atomic
Energy, 83, No. 6, 863-867 (1997).

102. M. KONOMURA and D M. ICHIMIYA, "Design Challenges for Sodium Cooled Fast
Reactors," Journal ofNuclear Materials, 371, 250-269 (2007).

103. Email Communication with Dr. Alistair Miller of Atomic Energy of Canada Limitted,
D20 Prices, millera@aecl.ca, (Febrary 14, 2005).

104. R. SMITH, M. SHAY, S. SHORT, C. EHRMAN, and T. MYERs, "Estimated Cost of an
ATW System," PNNL-13018, Pacific Northwest National Laboratory, Richland, Washington,
(1999).

105. D. KIM, M. KAZIMI, M. DRISCOLL, and N. TODREAS, "The Economics of
Tranmutation Fuel Cycles, Core Physics and Engineering Aspects of Emerging Nuclear Energy
Systems for Energy Generation and Transmutation," November 28 December 1, 2000, Argonne,
Illinois, (2000).

106. D. GROOM, Expert Group and Editorial Group Chairman "The Economics of the Nuclear
Fuel Cycle," Nuclear Energy Agency: Organization for Economic Co-operation and
Development, Paris, France, (1994).

107. League of Women Voters, The Nuclear Waste Primer: A Handbookfor Citizens, 2nd ed.,
Lyons Press, New York, New York, (1986).

108. United States Enrichment Corporation, "2006 Annual Report: Investing in Nuclear's
Future," http://www.usec.com/, Bethesda, Maryland, (2006).


308









BIOGRAPHICAL SKETCH

Samuel E. Bays was born the son of Gerald and Marva Bays in Topeka, Kansas on July 14,

1979. He was raised on a small family farm in the northeastern part of the state and graduated

from Mission Valley High School in May 1997. He attended Kansas State University where he

received his Bachelor of Science in Mechanical Engineering with an emphasis in Nuclear

Engineering in May 2002. He later received a Master of Science degree in Nuclear Engineering

at the University of Florida in August 2007. He now works in the reactor physics department at

Idaho National Laboratory. His research interests include all aspects of reactor design, nuclear

waste transmutation and hybrid nuclear energy systems.


309





PAGE 1

1 HETEROGENEOUS SODIUM FAST REACTOR DESIGNED FOR TRANSMUTING MINOR ACTINIDE WASTE ISOTOPES INTO PLUTONIUM FUEL By SAMUEL EUGENE BAYS A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2008

PAGE 2

2 2008 Samuel Eugene Bays

PAGE 3

3 To Nikki.

PAGE 4

4 ACKNOWLEDGMENTS I would like to thank m y facu lty advisor, James Tulenko, for his constructive comments and guidance of this work. I would also like to thank the other members of my advisory panel, Edward Dugan, Samim Anghaie, Ronald Bane y and Steve Herring for their sensible recommendations. I would like to especially thank and recognize Steve Herring for his technical guidance and support of this work. I greatly than k and appreciate Idaho National Laboratory for its financial support of this diss ertation project. Thanks go to Douglas Crawford and David Nigg for their management of projects and activities in which this work supported. I thank Doug Porter, Mitch Meyer, Steve Hayes, Jon Carmack and Rory Kennedy for their technical feedback and discussions on transmutation and fast reacto r fuels. Special thanks are given to Pavel Medvedev for his input on establishi ng the fuel performance criteria of the transmutation targets. Special thanks are given to Bevi n Brush and Thomas Johnson for th eir technical insight of fuel handling and separations operations at the EBR-II fuel cycle facilit y. Special thanks are given to Steve Piet, Gretchen Matthern and David Shropshire for their technical insight into the system dynamics and economics of closing the domestic and international fuel cycl es. Special thanks are given to Roald Wigeland, Giuseppe Palmiotti and Massimo Salvatores for their technical insight into reactor physics and design methodologie s of sodium fast reactors. Many thanks are given to my friends and colleagues of the INL fu el cycle analysis team: Mehdi Asgari, Rodolfo Ferrer, Benoit Forget and Michael Pope for the excellent studies and results that we have produced over the last two years.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........9 LIST OF FIGURES.......................................................................................................................12 ABSTRACT...................................................................................................................................22 CHAP TER 1 INTRODUCTION..................................................................................................................24 Motivation and Objectives...................................................................................................... 26 Transmutation Physics.......................................................................................................... ..31 Neutron Spectrum Influence on Transm utation Behavior............................................... 32 Transmutation and Nuclear Stability............................................................................... 36 Isotopic Aspects of Repository Impacts.......................................................................... 39 Background on Previously Proposed TRU Burning SFRs..................................................... 41 The Advanced Burner Reactor Design Concept............................................................. 43 Transmutation Target Designs: Radial Blankets and Moderated Targets ...................... 45 Transmutation Target Designs: Ax ial Blankets and Axial T argets................................ 48 Transmutation Based Reactivity Control Concept.......................................................... 49 Design Rationale of an Axial Heterogeneous Fast Transm utation Reactor........................... 52 Compensation for Inherent Positi ve Void Reactivity Feedback ..................................... 53 Benefits of Axial Targets for a De dicated MA Burner Core Design .............................. 54 Technology Compatibilities and Synergies............................................................................ 58 Actinide Partitioning : PUREX and UREX ..................................................................... 59 Pyroprocessing and the Integral Fuel Cycle .................................................................... 61 Assumptions for Using Transm utation Targets in SFRs ................................................. 63 2 COMPUTATIONAL METHODS AND FAST REACTOR PHYSICS................................ 66 Calculations and Fuel Cycle Modeling................................................................................... 66 Fast Reactor Equilibrium Fuel Cycle Calculations Using the REBUS Code................. 68 Light Water Reactor Spent Fuel Calculations Using the TRITON Code....................... 72 Scoping Calculations and Benchm ar king Using the MCNP Code................................. 73 Physics of the Reference Metal Fu eled Advanced Burner Reactor ........................................74 Conversion Ratio and High Leakage Cores.................................................................... 75 Conversion Ratio and High TRU Enriched Fuels........................................................... 80 Physics of the Axially Heterogene o us Fast Transmutation Reactor...................................... 82 Axial Targets and Axial Leakage Recovery.................................................................... 82 Axial Targets and Minor Actinide Conversion............................................................... 85 Combining Leakage and Capture Effects............................................................................... 90

PAGE 6

6 3 AXIAL TARGET DESIGN ANALYSIS............................................................................... 93 Waste Management Philosophies and Conversion Ratio Definition...................................... 93 Transmutation Based Reactor Design.................................................................................... 96 Transmutation Targets and Accom panying Fuel Cycle.................................................. 99 Transmutation Target Physics....................................................................................... 103 Parametric Study............................................................................................................... ....106 Effects of Pin Diameter and Core Height...................................................................... 107 Effects of Moderating Pins............................................................................................111 Tall and Flattened Axial Heterogeneous Core Designs........................................................ 115 Radial and Axial Power Profiles................................................................................... 116 Reactivity Feedbacks..................................................................................................... 123 Fuel Performance Indicators.......................................................................................... 125 Final Down-Selection: The AHFTR Design....................................................................... 128 Fuel Cycle Performance of the Final AHFTR Design.................................................. 129 Reactor Performance Characterist ics of the Final AHFTR Des ign............................... 132 Transmutation Analysis of the Final AHFTR Des ign................................................... 135 4 REACTOR REACTIVITY CONTROL STRATEGY......................................................... 139 Tc-99 versus B-10 as a C ontrol Rod Neutron Poison ...........................................................139 Control Assembly Design..................................................................................................... 143 Traditional SFR Control Assembly Design: B4C Ultimate Shutdown Assembly........ 143 Technetium Based Primary Control Assembly Design................................................. 144 Gas Expansion Module..................................................................................................148 Control Rod Worth.............................................................................................................. .150 Reactivity Worth of Bor on versus T echnetium............................................................. 155 Top versus Bottom Inserted Shim Rods........................................................................ 156 Axial Power Tilt and Shim Rod Insertion.....................................................................157 Other Reactivity Feedbacks..................................................................................................157 Axial Fuel Expansion.................................................................................................... 160 Radial Fuel Bowing....................................................................................................... 161 Technetium Transmutation Rate........................................................................................... 164 5 DIFFUSION VERSUS TRANSPORT BENCHMARKS.................................................... 167 Diffusion and Transport Methods......................................................................................... 168 Flux Spectrum Analysis................................................................................................ 170 Spatial Flux Analysis..................................................................................................... 172 Differences between BOEC and EOEC Fluxes.............................................................175 Depletion Test................................................................................................................. ......176 Spatial Self-Shielding Test................................................................................................... 180 Spatial Shadowing Test........................................................................................................182 Calculation Validation Remarks........................................................................................... 186 6 THE AHFTR FUEL DESIGN.............................................................................................. 188

PAGE 7

7 Fuel Pin Design.....................................................................................................................189 Target Alloy Selection a nd Design Considerations ....................................................... 190 Target and Driver Fu el Burnup Criteria ........................................................................ 194 Cladding Damage Criteria.............................................................................................195 Fuel Pin Thermal Performance Criterion............................................................................. 197 Fuel Temperature Criterion...........................................................................................198 Cladding Temperature Criterion.................................................................................... 199 Thermal Analysis..................................................................................................................200 Fuel Assembly Power Peaking...................................................................................... 200 Peak Fuel Pin Hot Channel Analysis............................................................................. 206 Metallurgical Diffusion Effects............................................................................................215 Inter-Diffusion Data......................................................................................................216 Penetration Distance...................................................................................................... 217 Transmutation Gas Genera tion and Plenum Sizing.............................................................. 218 Helium and Fission Product Gas Calculation................................................................ 220 Transmutation and Fission Gas Analysis...................................................................... 223 Fuel Design Basis Summary................................................................................................. 226 Fuel Processing Considerations.....................................................................................227 Higher Mass Actinide Considerations...........................................................................230 Repository Considerations............................................................................................. 234 7 ECONOMICS OF THE TWO-TIER AHFTR FUEL CYCLE............................................ 238 Economic Issues of Reprocessing........................................................................................ 238 Capitol Costs.........................................................................................................................240 Fuel Costs.............................................................................................................................243 First-Core Reprocessing Cost........................................................................................ 243 Subsequent Core Reprocessing Cost............................................................................. 244 High Level Waste Disposal Cost................................................................................... 247 Fuel Fabrication Cost....................................................................................................248 Front-End Uranium Costs.............................................................................................. 249 Back-End Uranium Costs.............................................................................................. 251 Discounting and Financing...................................................................................................252 Operations and Maintenance......................................................................................... 253 Construction Capital......................................................................................................254 Fuel Investment.............................................................................................................256 Fuel Cycle Base Cases.......................................................................................................... 257 The MOX Fuel Cycle....................................................................................................258 The ABR Fuel Cycle.....................................................................................................262 The Combined Mixed-Fleet AHFTR and ABR Fuel Cycle...................................... 265 Sensitivity Analysis ........................................................................................................... ...270 Breakeven Unit Costs for ABRs and AHFTRs.............................................................270 Cost Sensitivities for a Combined ABR and AHFTR Fleet.......................................... 274 Converting Neutrons into Dollars......................................................................................... 278 8 SUMMARY AND CONCLUSIONS...................................................................................282

PAGE 8

8 Method of Reactor Design.................................................................................................... 284 Reactor Control and Safety Features.................................................................................... 286 Code Validation....................................................................................................................288 Transmutation Target Fuel Design....................................................................................... 289 Fuel Processing and Repository Considerations................................................................... 290 Fuel Cycle Economic Analysis............................................................................................. 291 Concluding Remarks............................................................................................................ 292 APPENDIX A REBUS INPUT DECK......................................................................................................... 294 B PARAMETRIC DESIGN ANALYSIS................................................................................ 296 LIST OF REFERENCES.............................................................................................................300 BIOGRAPHICAL SKETCH.......................................................................................................309

PAGE 9

9 LIST OF TABLES Table page 1-1 ABR fission and capture cross section data ....................................................................... 35 1-2 IMF fission and capture cross section data ........................................................................ 36 1-3 Fuel assembly and pin dimensions for the S-PRIS M and ABR designs........................... 46 1-4 Summary of relative studie s on transmutation targets .......................................................48 1-5 Examples of real fast reactor plants that have incorpor ated axial blankets .......................49 1-6 Fuel assembly design of the AHFTR compared to ABR designs...................................... 58 1-7 Core design for the AHFTR co m pared to similar ABR designs....................................... 59 1-8 Waste stream partitioning afforded by various reproces sing technologies ....................... 60 1-9 Technology compatibility assumptions used for fuel cycle anlaysis ................................. 65 2-1 Isotopic composition for UOX SNF..................................................................................73 2-2 Atom densities and one-group microsc opic cross sections for Pu-239 and U-238 ........... 76 2-3 AHFTR fission and capture cross sections data................................................................89 3-1 Reactor thermal power and core heig hts evaluated in param etric analysis....................... 97 3-2 Reference fuel assembly design fo r varying height-to-diam eter ratio............................... 97 3-3 Transmutation utilization factor....................................................................................... 105 3-4 Single group cross section ratio....................................................................................... 105 3-5 Transmutation half-lives for a preliminary AHFTR design ............................................ 108 3-6 Transmutation half-life of Am-241 fo r varying num ber of moderating rods.................. 111 3-7 Core design summary for the refere nce ABR with tall and flat AHFTR ........................ 116 3-8 Fuel cycle comparison for the reference ABR with tall and flat AHFTR.......................122 3-9 Physics comparison for the refere nce ABR with tall and flat AHFTR ............................ 125 3-10 Fuel performance comparison for the reference ABR with tall and flat AHFTR ........... 127 3-11 Mass flow analysis of the downsellected AHFTR........................................................... 130

PAGE 10

10 3-12 Target and active core region fuel inventory of the downsellected AHFTR ...................131 3-13 Initial reactor physics and fuel perform ance of the downsellected AHFTR................... 134 3-14 Mass production and destruction ra tes per installed m egawatt per year......................... 135 3-15 Isotopic contribution of external supply versus contributing to fission ........................... 137 4-1 Atomic concentrations of absorber atoms in B4C versus technetium.............................. 142 4-2 Control assembly types and dimensions.......................................................................... 143 4-3 Core reactivity worth for independently separate reactivity feedback effects ................. 160 4-4 Summary of Tc-99 consumption by the AHFTR............................................................ 165 5-1 Select reactor parameters for th e AHFTR given by DIF3D and VARIANT. .................175 5-2 Comparison of mean-free-path s of different reaction types ............................................185 6-1 Peak burnups for the driver fuel and targets.................................................................... 194 6-2 Average burnups for the driver fuel and targets.............................................................. 195 6-3 Average and peak fast fluence (E>0.1 MeV) and peak dpa............................................ 195 6-4 ABR fuel assembly peak-to-core average LHGR ratio taken at BOEC .......................... 204 6-5 Beginning of Equilibrium Cycle AHFTR fuel assembly peak-to-core average LHGR.. 204 6-6 End of Equilibrium Cycle AHFTR fuel assem bly peak-to-core average LHGR............ 204 6-7 Ternary inter-diffusion coefficients................................................................................. 216 6-8 AHFTR fuel pin dimensions............................................................................................224 7-1 Unit costs and fees for the ope n single-tier MOX fuel cycle ...........................................260 7-2 Uranium Oxide once through fuel cycle fuel costs.......................................................... 261 7-3 Mixed Oxide fuel cycle fuel costs................................................................................... 261 7-4 Unit costs and fees for the cl osed single-tier ABR fuel cycle .........................................264 7-5 Advanced Burner Reactor fuel cycle fuel costs...............................................................264 7-6 Unit costs and fees for the clos ed double-tier AHFTR fuel cycle ...................................268 7-7 Advanced Burner Reactor fuel costs in a mixed ABR and AHFTR fleet....................... 268

PAGE 11

11 7-8 Fuel costs for the AHFTR tier of the m ixed ABR and AHFTR fleet..............................269 7-9 Advanced Burner Reactor two-tier f uel cycle breakeven unit costs................................ 272 7-10 Axial Heterogeneous Fast Transmutati on Reactor fuel cycle breakeven costs ...............272 7-11 Mixed fleet fuel cycle breakeven costs............................................................................ 274 7-12 Inter-comparison of aqueous reprocessing costs for various scenarios ........................... 280

PAGE 12

12 LIST OF FIGURES Figure page 1-1 Synergy between GNEP and IF R fuel recycling strategies ............................................... 30 1-2 Important transmutation reactions of am ericium and neptunium into plutonium............. 33 1-3 Periodic relationship between fissile and fertile transuranic isotopes. .............................. 38 1-4 Decay heat plot for pres surized water reactor SN F........................................................... 40 1-5 Core layouts for the S-PRISM and ABR designs..............................................................44 1-6 Cross section plots of U-238 and Tc -99 total absorpti on cross sections ........................... 52 1-7 Core design of the Axial Heter ogeneo us Fast Transmutation Reactor.............................. 56 1-8 Modification to the ABR fuel cycle by the axial targets ................................................... 57 1-9 Partitioning and transmutation scenar io of an ABR and AHFTR m ixed-fleet.................. 63 2-1 Coupled DIF3D core physics a nd R EBUS fuel cycle algorithm.......................................69 2-2 Flow diagram of data transfer between the MC2-2 and REBUS codes.............................71 2-3 Material buckling of a s imple bare hom ogeneous SFR.....................................................77 2-4 Critical radius required to equate ge om etric buckling with material buckling.................. 77 2-5 Generalized conversion ratio of sim ple bare homogeneous SFR...................................... 79 2-6 Mean-free-path of neutron travel between interactions ..................................................... 79 2-7 Leakage fraction of a simple bare homogeneous SFR....................................................... 80 2-8 Conversion ratio and the corresponding TRU enrichm ent for a fixed core size............... 81 2-9 Axial current distributio n of the AHFTR and ABR...........................................................84 2-10 Axial flux distribution of the AHFTR and ABR............................................................... 85 2-11 Moderated target region binned m icroscopic reaction rate spectra ................................ 86 2-12 Active core region binned microscopic reaction rate spectra......................................... 87 2-13 Comparison of capture and fission cro ss sections between Am-241 and Pu-238 ............. 88 2-14 Comparison of flux spectrums between LWR IMF target region and active core............ 91

PAGE 13

13 3-1 Preliminary AHFTR core design used in param etric analyses.......................................... 98 3-2 Zirconium hydride phase diagra m for varying hydrogen content.....................................99 3-3 Representation of the axial target pin-la ttice showing the orie ntation of targets ............100 3-4 AHFTR fuel cycle scenario for the REBUS calculation................................................. 101 3-5 Americium and plutonium cross sect ion plots versus a SFR neutron flux ...................... 104 3-6 Spectrum comparison between inner, m iddle, outer core and targets............................. 106 3-7 Change in target isotope masse s as a function irradiation tim e....................................... 107 3-8 Active core radial power density profile for a prelim inary AHFTR............................... 110 3-9 Inner core axial power density profile for a preliminary AHFTR................................... 110 3-10 Lin.-Log Scale: Neutron flux in the ta rget region for varying moderating rods ............. 112 3-11 Log.-Log Scale: Neutron flux in the ta rget region for varying moderating rods ............ 112 3-12 Percent of the initial Am-241 ma ss rem aining in the target rod...................................... 114 3-13 Percent of the initial Pu-238 m ass created in the target rod............................................ 115 3-14 Flat preliminary AHFTR design with eight row s of fuel instead of seven.................. 117 3-15 Radial power density profile fo r six axial slices through the core ................................... 118 3-16 Axial power density profile for each row of fuel............................................................. 119 3-17 Target region buil dup and depletion curve ...................................................................... 131 3-18 Active core region build up and depletion curve ..............................................................132 4-1 Total neutron absorption cross section plots for select absorber m aterials..................... 142 4-2 Ultimate shutdown control assembly configuration........................................................ 145 4-3 Primary control asse m bly configuration.......................................................................... 147 4-4 Gas expansion module assembly configuration............................................................... 149 4-5 Shim control rod bank reactivity worth at BOEC............................................................ 150 4-6 Shim control rod bank reactivity worth at EOEC............................................................ 151 4-7 Safety control rod bank reactivity worth at BOEC ..........................................................152

PAGE 14

14 4-8 Neutron spectrum in the targets as a function of safety rod insertion ............................. 152 4-9 Safety control rod bank reactivity worth at EOEC ..........................................................153 4-10 Ultimate shutdown system worth at BOEC..................................................................... 154 4-11 Ultimate shutdown system worth at EOEC..................................................................... 154 4-12 Shim control rod bank reactivity wo rth for m etallic Tc-99 compared to B4C.................156 4-13 Radial flux distributions for shim rods inserted from the top or the bottom................... 157 4-14 Axial power distribution for increasing shim rod insertion into the active core............. 158 4-15 Neutron spectrum for steady state oper ation versus a com plete loss of sodium............. 159 5-1 Core midplane energy spectrum for code validation comparison................................... 171 5-2 Target region energy spectrum for code validation com parison..................................... 171 5-3 Cod validation comparison of the flux axial profile........................................................ 173 5-4 Cod validation comparison of the flux ra dial prof ile at the core mid-plane.................... 173 5-5 Cod validation comparison of the flux radi al prof ile at the axial target region............... 174 5-6 Comparison between the BOEC and EOEC neutron spectrum s..................................... 175 5-7 Comparison between the BOEC and EOEC radial flux profiles ..................................... 176 5-8 Comparison between BOEC and EOEC axial flux profiles............................................ 177 5-9 Reactivity curve comparison between REBUS and MONTEBURNS............................ 178 5-10 Advanced Burner Test Reactor benc h mark using fresh core composition.................. 180 5-11 MCNP sub-lattice model.................................................................................................. 181 5-12 Neutron spectrum as a function of the zirconium hydride slug radius............................ 183 5-13 Neutron spectrum as a functi on of the target slug radius .................................................183 5-14 MCNP unit-cell model with homogenized fuel annulus.................................................. 184 5-15 Neutron spectrum within the homogenized annulus........................................................ 185 6-1 Conceptual AHFTR Fuel Pin Design.............................................................................. 190 6-2 Optical microscopy for three of the AFC-1B samples..................................................... 192

PAGE 15

15 6-3 Fuel swelling performance for the AFC-1B samples...................................................... 193 6-4 AHFTR power density profile as a f unction of fuel row and axial region ...................... 202 6-5 ABR reference power density profile as a function of fuel row and axial region ........... 202 6-6 BOEC LHGR for the driver fuel m id-plane and target regions....................................... 205 6-7 EOEC LHGR for the driver fuel m id-plane and target regions....................................... 206 6-8 Peak pin axial BOL LHGR distribution...........................................................................208 6-9 Axial node procession for Nusselt analysis..................................................................... 209 6-10 Assumed thermal properties of TRU-U-Zr m etal alloy fuel............................................ 213 6-11 Axial temperature profiles for a typical fuel rod in the hottest fuel assembly ............. 214 6-12 Percent contribution to helium production by alpha decay............................................. 219 6-13 Gas plenum pressures resulting from transm utation and fission gas production............. 225 6-14 Envisioned power block reactor plant model................................................................... 229 6-15 Average neutron emission rate for processed initial TRU in fresh fuel .......................... 231 6-16 Average gamma decay energy rate for processed initial TRU in fresh fuel.................... 233 6-17 Average alpha decay heat rate for processed initial TRU in fresh fuel........................... 234 6-18 Decay heat plot for AHFTR fu el cycle lo sses in a repository......................................... 236 7-1 The open single-tier MOX fuel cycle..............................................................................259 7-2 The closed single-tier ABR fuel cycle............................................................................. 263 7-3 The closed double-tier AHFTR fuel cycle....................................................................... 266 7-4 Fuel cycle costs for the ABR and AHFTR for varying aqueous unit costs.....................276 7-5 Fuel cycle costs for the ABR and AH FTR for varying pyroprocessing unit costs .......... 276 7-6 Transuranic-free HLW disposal fee adjustm ent for breakeven....................................... 278 8-1 Cross section data for Am-241, Pu-238 and Pu-239........................................................ 283 8-2 Radial power distribution for the ABR versus the AHFTR.............................................285 A-1 REBUS input cards..........................................................................................................294

PAGE 16

16 B-1 Excess Reactivity of the core design in tall for varying core height and p/d ...............296 B-2 Excess Reactivity of the core design in flat for varying co re height and p/d ...............297 B-3 TRU enrichment of the core given in tall for varying core height and p/d .................. 297 B-4 TRU enrichment of the core design in flat for varying co re height and p/d ...............298 B-5 Cycle length of the core given in t all for varying core height and p/d .........................298 B-6 Cycle length of the core design in flat for vary ing core height and p/d .......................299

PAGE 17

17 LIST OF ABBREVIATI ONS AND TERMINOLOGY AAA Advanced Accelerator Applications ABR Advanced Burner Reactor ABTR Advanced Burner Test Reactor AFCI Advanced Fuel Cycle Initiative AHFTR Axial Heterogeneous Fa st Transmutation Reactor ALMR Advanced Liquid Metal Reactor ANL Argonne National Laboratory ARR Advanced Recycling Reactor ATR Advanced Test Reactor ATW Accelerator Transmutation of Waste B4C Boron Carbide BOC Beginning of Cycle BOEC Beginning of Equilibrium Cycle BOL Beginning of Life CAPRA Consommation Amliore du Plut onium dans les Racteurs Avancs CCCC Committee on Computer Code Coordination CDA Core Disruptive Accident CDF Cumulative Damage Fraction COMPX CCCC format binary macroscopic cross section data file Critical Radius Effective core radius required to meet the reactivity criticality requirement by equating geometric buckli ng with material buckling CRGT Control Rod Guide Tube D-9 Austenitic steel that can be used for SFR fuel cladding and structural components

PAGE 18

18 DIF3D Multigroup neutron diffusion code developed for modeling steady-state reactor physics and other critical systems developed by ANL DOE Department of Energy dpa Displacements per Atom EBR-I Experimental Breeder Reactor I EBR-II Experimental Breeder Reactor II EFPD Effective Full Power Day EFR European Fast Reactor ENDF Evaluated Nuclear Data File Enrichment Zone Same TRU enrichme nt/composition regions within a SFR EOC End of Cycle EOEC End of Equilibrium Cycle EOL End of Life FCCI Fuel-to-Cladding Chemical Interaction FCF Fuel Cycle Facillity FCMI Fuel-to-Cladding Mechanical Interaction FFTF Fast Flux Test Facillity FOAK First of a Kind GEM Gas Expansion Module GNEP Global Nuclear Energy Partnership HLW High Level Waste HM Heavy Metal HT-9 Ferritic/Martensitic high chromium steel used for SFR fuel cladding and structural materials IBA Integral Burnable Absorber IFBA Integral Fuel Burnable Absorber

PAGE 19

19 IFC Integral Fuel Cycle IFR Integral Fast Reactor iHM Initial Heavy Metal or fresh fuel going into the reactor IHX Intermediate Heat Exchanger ILW Intermediate Level Waste IMF Inert Matrix Fuel INL Idaho National Laboratory ISOTXS CCCC format binary micros copic cross section data file JSFR Japanese Atomic Energy Agency SFR KALIMER Korean Advanced Liquid Metal Cooled Reactor LANL Los Alamos National Laboratory LEU Light Enriched Uranium LHGR Linear Heat Generation Rate LLW Low Level Waste LOCA Loss of Coolant Accident LOHS Unprotected Loss of Heat Sink LWR Light Water Reactor MA Minor Actinide Matino Plane Contact surface between two inter-diffusing compositions MC2-2 Multigroup cross section generatio n code by solution of the neutron slowing down equations developed by ANL MCNP Monte Carlo N-Particle general geometry, continuous energy and angle Monte Carlo transport co de developed by LANL mil one-tenth of one cent MIT Massachusetts Institute of Technology Mixed waste Chemically and radi ologically hazardous waste

PAGE 20

20 MOEC Middle of Equilibrium Cycle MONTEBURNS Coupling code for MCNP and ORIGEN developed by LANL MOX Mixed Oxide MWD Megawatt-day MWY Megawatt-year NOAK Nth of a Kind NRC Nuclear Regulatory Commission NUS Nuclear Utility Service code of account NWF Nuclear Waste Fund NWPA Nuclear Waste Policy Act OCRWM Office of Civilian Ra dioactive Waste Management ORIGEN General use isotope buildup, depletion and decay code for solving the Bateman equations by the exponential matrix method developed by ORNL ORNL Oak Ridge National Laboratory p/d Pitch-to-diameter ratio PHWR Pressurized Heavy Water Reactor PIE Post Irradiation Examination PUREX Plutonium Uranium Redox Extraction PWR Pressurized Water Reactor Pyroprocess Metal fuel reprocessing where by uranium and transuranics are separated by electro-deposition on cathodes immersed in a eutectic LiCl-KCl bath REBUS REactor BUrnup System Fuel cy cle analysis code developed by ANL SCNES Self-Consistent Nuclear Energy System SFF Spent Fast reactor Fuel SFR Sodium Fast Reactor SHRT Shutdown and Heat Removal Test

PAGE 21

21 SNF Spent LWR Nuclear Fuel S-PRISM Super Power Reactor Innovative Small Module Support Ratio Mass balance ratio of installed thermal capacity for the mass consuming reactor per installed thermal capacity for the mass producing reactor SWU Separative Work Unit THORP Thermal Oxide Reprocessing Plant, Britain TOP Transient Overpower TREAT Transient Reactor Test facility TRU Transuranic TRU Enrichment Concentration by volume of TRU over HM in SFR or MOX-LWR fuels ULOF Unprotected Loss of forced circulation Flow UOX Uranium Oxide UP1 & UP2 Reprocessing plants at LaHague, France UREX Uranium Extraction aque ous reprocessing technology UREX+ Uranium Extraction Plus In cluding waste stre am partitioning USEC United States Enrichment Corporation UxC Uranium Exchange Consulting Company VARIANT Variational Anisotropic Noda l Transport Code developed by ANL YM-EIS Yucca Mountain Environmental Impact Statement ZrH1.6 Zirconium Hydride

PAGE 22

22 Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy HETEROGENEOUS SODIUM FAST REACTOR DESIGNED FOR TRANSMUTING MINOR ACTINIDE WASTE ISOTOPES INTO PLUTONIUM FUEL By Samuel Eugene Bays May 2008 Chair: James Tulenko Major: Nuclear Engineering Sciences In the past several years ther e has been a renewed interest in sodium fast reactor (SFR) technology for the purpose of destroying tran suranic waste (TRU) produced by light water reactors (LWR). The utility of SFRs as waste burners is due to the fact that higher neutron energies allow all of th e actinides, including the minor actinid es (MA), to contribute to fission. It is well understood that many of the design issues of LWR spent nuclear fuel (SNF) disposal in a geologic repository are linked to MAs. Because the probability of fission for essentially all the non-fissile MAs is nearly zero at low neutron energies, these is otopes act as a neutron capture sink in most thermal reactor systems. Furthermore, because most of the isotopes produced by these capture reactions are also non-fissile, they too are neutron sinks in most thermal reactor systems. Conversely, with high neutron energies the MAs can produce neut rons by fast fission. Additionally, capture reactions transmute the MA s into mostly plutonium isotopes, which can fission more readily at any ener gy. The transmutation of non-fissi le into fissile atoms is the premise of the plutonium breeder reactor. In a breeder reacto r, not only does the non-fissile fertile U-238 atom contribute fast fission neutrons, but also transmutes into fissile Pu-239. The fissile value of the plutonium produced by MA transmutation can only be realized in fast neutron spectra. This is due to the fact that the pr edominate isotope produced by MA

PAGE 23

23 transmutation, Pu-238, is itself not fissile. However, the Pu-238 fission cross section is significantly larger than the or iginal transmutation parent, pr edominately: Np-237 and Am-241, in the fast energy range. Also, Pu-238s fissi on cross section and fissi on-to-capture ratio is almost as high as that of fissile Pu-239 in the fast neutron spectrum. It is also important to note that a neutron absorption in Pu-238, that does not cause fission, will instead produce fissile Pu239. Given this fast fissile quality and also the f act that Pu-238 is transmuted from Np-237 and Am-241, these MAs are regarded as fertile material in the SFR design proposed by this dissertation. This dissertation demonstrates a SFR design which is dedicated to plutonium breeding by targeting Am-241 transmutation. This SFR design uses a moderated axial transmutation target that functions primarily as a pseudo-blanket fuel, which is reprocessed with the active driver fuel in an integrated recycling strategy. This work demonstrates the cost and feasibility advantages of plutonium breedi ng via MA transmutation by adopting reactor, reprocessing and fuel technologies previously demonstrated for tr aditional breeder reactors. The fuel cycle proposed seeks to find a harmony be tween the waste management advantages of transuranic burning SFRs and the resource sustainabi lity of traditional plutonium breeder SFRs. As a result, the enhanced plutonium conversion fro m MAs decreases the burner SFRs fuel costs, by extracting more fissile value from the in itial TRU purchased through SNF reprocessing.

PAGE 24

24 CHAPTER 1 INTRODUCTION Sodium Fast Reactors (SFRs) are currently be ing evaluated as a mean s of eliminating the long-lived transuranic (TRU) waste produced by Li ght Water Reactors (LWR). Historically, the inherent neutron surplus, necessary for overcomi ng neutron losses by leakage, has been used to breed fissile Pu-239 by transmuting U-238 in a bla nket. In the context of a transuranic-burning SFR, these excess neutrons are ap plied to destroying, by fission, th e transuranic atoms of Spent Nuclear Fuel (SNF) produced by LWRs. However, not all transuranic isotopes can be de stroyed by fission with the same efficiency. This is because many of the isotopes in SN F TRU can only fission above a given threshold neutron energy. Unfortunately, th ese isotopes, such as Np-237 and Am-241, present in SNF, pose the largest decay heat and radiotoxicity issues for permanent disposal of this waste in a geologic repository. The repository space benefit stems principally from the removal of Am-241 from the fuel cycle. This is because the re positorys waste emplacement drift spacing is limited by the maximum rock temperature between drifts. The mid-drift temperature is principally a function of the decay heat produced by Am-241. Additionally, alpha decay of Am-241 is chiefly responsible for the buildup of Np -237 in the repository many years after it is closed. The Am-241 cross section for fission is orders of magnitude less than its capture cross section at neutron energies less than one MeV. Because the average neutron flux for most nuclear reactors, including fast reactors, exists at energies below one MeV, fission can not be the primary mechanism for removing americium from the fuel cycle. In fact, this fission threshold property for both Np-237 and Am-241 is very pro nounced when compared to most other heavy metal (HM) actinides. Therefore, it is useful to convert these isotopes, especially Am-241, into more fissionable plutonium isotopes, using specia lized transmutation targets. Irradiating Am-

PAGE 25

25 241 in an epithermal or thermal spectrum (below one MeV) maximizes its destruction by neutron capture. This neutron capture and its subsequent decay chain leads to the breeding of the even mass number plutonium isotope: Pu-238. This even-plutonium isotope, though not truly fissile, has a larger fission cross secti on below one MeV compared with the initial Am-241 atom. The fast reactor core design proposed in this disse rtation uses the plutoni um breeding fertile property of Am-241 to maximize the available fi ssile worth that can be derived from this otherwise non-fissile transuranic waste. The relative increase in fissile worth cont ributed by the transmut ed Pu-238 can only be realized in fast neutron spectra. This is b ecause Pu-238 has a fission threshold of its own. However, the Pu-238 fission threshold is simply less sharply defined at one MeV than Am-241 (or Np-237). The fissile worth of Pu-238 is only comparable to that of odd mass number fissile Pu-239 and Pu-241 in the unresolved resonance ra nge of its fission cross section. At thermal spectrum energies, the resolved capture re sonances of Pu-238 render it effectively non-fissile with no reactivity benefit. Even though it is po ssible that a thermal spectrum reactor could be used to transmute Am-241 into Pu-238, a fast re actor would still be necessary to fully benefit from this conversion. This scenario would re quire that both thermal and fast reactors be collocated with all the necessary fu el recycling facilities at the sa me location. This concentration of large infrastructure facilities is considered unfeasible for this work. Instead, this dissertation will demonstrate a fast reactor design which is dedicated to plutonium breeding via americium transmutation. This fast reactor design uses a m oderated axial transmutation target that functions as a pseudo-blanket fuel for breeding Pu-238. The transmuted Pu-238 is then co-reprocessed along with the SFRs active dr iver fuel in an integrat ed recycling strategy.

PAGE 26

26 Motivation and Objectives The United States comm ercial nuclear power industry currently pr oduces approximately 20% of the nations electricity. LWRs have beco me the industry work horse for producing this power since the first fully commercial LWR was brought online in the United States at Shippingport Pennsylvania in 1957. Orders for new nuclear power plants grew steadily during the 1960s and continued through the 1970s. In 1979, a partial reactor meltdown accident at Three Mile Island, Pennsylvania caused investor confidence in nuclear reactor safety to essentially disappear. The accident prompted the Nuclear Regulatory Commission (NRC) to impose stricter safety standards for all existing and future commercial reactors. Though no new nuclear power plants were orde red after the Three Mile Island Accident, completion of existing construction projects since then continued to expand the industrys elec trical capacity. By 1991, 22% of the nations electricity was produced by 111 nuclear power plants [1]. Since then, several nuclear power plants have been deco mm issioned before the end of their licensed lifetimes. Despite the loss of thes e reactors, the current share of nuclear in the nations electric resources has stabilized at 20%. The sustainability of electric generation, even with increasing electric ity demands, is due in part to an overall technological maturity afforded by years of sa fe operational experience since the Three Mile Island accident. Gradual improvements in the irradi ation integrity of the uranium fuel have allowed more fission energy to be ex tracted per initial mass (i.e., burnup) than ever before. Enhancements in computer simulation have enabled an increas ed understanding of the nuclear, mechanical, hydraulic and materials physics within LWRs which allows them to be operated safely without sacrificing performance. Also, increased attention to human factors has enhanced operational safety, which in turn, has generally reciprocated pu blic trust in nuclear energy.

PAGE 27

27 These improvements have allowed nuclear power plant operators to minimize the time the reactor is shutdown for refueling and maintenance activities. Therefore, the capacity for LWRs to operate at full power has steadily increase d from approximately 55% in the 1960s to over 90% in the 21st century [ 2]. This capacity is equivalent to saying that nuclear power plants produce 90% of the energy that could be generated if they were operated at full power all of the tim e. The remaining 10% roughly accounts for the time required to discharge old fuel and reload fresh fuel. Therefore, it is arguable that LWRs have achieved their prac tical limit in nuclear electricity generating capacity. Further improvements in LWR fuel technol ogy may allow slightly higher operating capacities without adding additional fuel costs. However, this is unlikely because the burnup of enriched fuel is generally linearly proportional to the level of uranium enrichment of the fissile isotope, U-235, in the fuel [ 3]. Currently, the NRC enforces a cap on enrichm ent at five percent to protect fuel fabrication workers from possible criticality accidents. Due to the criticality safety implications of exceeding this enrichment limit, it is assumed that additional safety measures will be required by enrichment and fabricators for achieving higher burnups in LWRs. These additional safety measures could incr ease LWR fuel costs Unless technology becomes available that makes the process of enrichment significantly cheaper, the cost of additional criticality safety practices will cause the cost of LWR fuel to increase in order to achieve higher burnups. These practical limitations extend essentially to the front-end of the nuclear fuel cycle. This part of the commercial industry is respon sible for all processes from mining the uranium from the ground through its irradiatio n in the reactor. This is al so the sector of the commercial industry in which the private sector has th e most control over electricity costs.

PAGE 28

28 The back-end of the fuel cycle, and its associ ated cost, is currentl y the responsibility of the federal government in accordance with the Nuclear Waste Policy Act (NWPA). As required by the NWPA, the nuclear industry pays 0.1 per every kW-h of electricity produced by this fuel to the federal government as payment toward s its eventual permanent disposal in a geologic repository [ 4]. The LWR fleet currently produces approxim ately 2,000 metric tones of SNF each year [ 5]. Assuming that no additional reactors are bui lt, the adequacy of this levee to pay for geologic SN F disposal has been assessed as ad equate by the Office of Civilian Radioactive Waste Management (OCRWM) [ 6]. However if new LWRs ar e constructed to m eet rising electricity demands, then it becomes possible that several repositories will be required. This could substantially change the economic compatibility of the NW PA levee with future nuclear forecasts. However, if the number of repositories can be limited to one, the NWPA revenue collected from the existing LWR fleet for the remainder of the lifetime of those plants will be adequate to pay the full cost of the repository [ 7]. The utilization of the stor age spac e within the repository design is critical in determining its effective capacity to store radioactive wastes. The NWPA stipulates that the repository be designed to not only dispose of SNF, but also for disposing of separated High Level Wastes (HLW). HLW is comp rised of fission product and transuranic isotopes present in SNF. It is the HLW compone nt of SNF that represents almost the entire radioactivity produced in the nuclear fuel cycle. In act uality, roughly 90% of the SNF mass is the original uranium that was mined from the ground. Therefore, substantial improvements in repository utilization can be achieved by storing only the volume of HLW within SNF. The cost competitiveness of this option hinges upon the economic advantage gained by separating HLW versus storing unaltered SNF. Separating transura nic and fission product

PAGE 29

29 isotopes from SNF imply that tech nologies such as that used fo r nuclear fuel reprocessing are required. Reprocessing is an integral step in nuclear fuel recycling which is the process of separating the TRU isotopes from SN F to create new reactor fuel. TRU comprises only about 1% of SNF. Additio nally, TRU can be further categorized into major actinides (i.e., plutonium) and minor actinid es (MA) (i.e. neptunium, americium, curium, berkelium and californium) which represent only about 10% of TRU. Therefore, it is the approximate 0.1% of the total mass of the SNF th at strongly controls re pository utilization. Because most of TRU is fissile plutonium, it is arguable that it could be recycled in LWRs. Nevertheless, it is the MAs that create the diffi culties for the repository. These isotopes are much more difficult to destroy in LWRs due to th e fact that they have no fissile value in the LWRs thermal neutron spectrum. It is because of the improved fissile value of MAs in a fast neutron spectrum that the United States has renewed its interest in SFR technology. The current Global Nuclear Energy Partnership (GNEP) program is currently investigating a symb iosis between SFRs and the SNF generated by TRU (Figure 1-1). Unlike TRU recycling in LWRs, which only recycles TRU once (limited by fissile concentration in Pu), the GNEP scenario would continuously recycle TRU in SFRs in a closed fuel cycle afte r the initial separation from SN F. Because the GNEP scope of the SFR is tailored for TRU destruction, it is dubbed the Advanced Burner Reactor (ABR) by this program. An important note should be made here that the exact GNEP nomenclature refers to the ABR as a demonstration prototyp e for proving the technology of building a fleet of TRU burner SFRs called Advanced Recycling Reactors (ARR) [7]. However, the term advanced recycling reactor is to general descriptor for this dissertation. The term ABR is used regularly in recent

PAGE 30

30 SFR literature to mean specifically a transu ranic burning SFR design with a homogenous core configuration that excludes the use of transmutation targets or uranium blankets [ 7]. This definition is adopted by this dissertation. Figure 1-1. Synergy between GNEP and IFR fuel recycling strategies: .P urple represents the basic IFR closed fuel cycle scenario. Bl ue represents the LW R input into the IFR closed fuel cycle for the GNEP recycling scen ario. The striped orange represents the repla cement of fertile uranium blankets with MA transmutation targets.1 Reprocessing, fuel fabrication and reactor ir radiation in the GNEP closed fuel cycle, creates a role for the private industry to mini mize the back-end costs by using LWR and SFR spent fuel in the front-end. The further extraction of fission energy from the recycled fuel 1 The original IFR scenario would not require a constant supply of exte rnally reprocessed SNF TRU because the SFR would create its own plutoniu m from the external s upply of uranium. In the current GNEP scenario, the uranium blankets are removed and TRU is continuously recycled in the active core driver fuel without targets. LWR SFR LWR Reprocessing SFR Reprocessing GNEP Recycling Fertile Blanket Material (MAs replace Uranium) Spent Targets Join Driver Fuel for IFR Type Reprocessing SNF TRU (Mostly Pu)

PAGE 31

31 enhances the energy sustainability of nuclear power by making maximum use of the initial uranium mined. Currently, the GNEP fuel cycle (F igure 1-1) does not have a targeted solution for MAs. Research and development by the GNEP program has not currently revealed a conclusive decision to include MA within the clos ed fuel cycle because the americium in MAs is easily transmuted (even in the fast reactor) into the higher mass actinides of curium, berkelium and californium. These higher mass actinides, wi th relatively shorter half-lives, are highly radiotoxic and sometimes thermally hot. Add itionally, these isotopes do not significantly contribute fissile worth to the SFR compared to plutonium. If the MAs are not burned in SFRs then they mu st be discharged from the fuel cycle as transuranic HLW. The HLW disposal at the repository is one of the cost contributors of the closed fuel cycle that hinders economic competitive ness with the option to di rectly dispose of the SNF without reprocessing. However, as discus sed above, it is possible to transmute plutonium isotopes from the neptunium and americium in the MAs. Approximately two-thirds of SNF MAs are Np-237 and Am-241. Therefore, it is nece ssary to design the SFR in such a way that is favorable to producing plutonium isotopes as opposed to the higher mass isotopes. Closed SFR fuel cycles have been technologi cally demonstrated during the United States Integral Fast Reactor (IFR) program (~1984 1994) In the IFR scenario, the recycling of the SFRs uranium blanket and driver fuel was integrated into the sa me reprocessing step (Figure 11). It is proposed in this di ssertation that the MA transmuta tion can be achieved by irradiating MA targets in lieu of blankets in an IFR amendment to the GNEP scenario. Transmutation Physics Transm utation means literally th e conversion of one thing into another and is rooted in the Latin word trans-mutare which is to change. The modern definition was coined by 17th century alchemists to define the process of converting baser metals into gold. If the meaning of

PAGE 32

32 transmutation is to transform one isotope into another, then transmutation is accomplished naturally by radioactive deca y. It is the relatively long decay half-life of certain actinide and fission product isotopes that are th e essence of the repository storage problem. This is because their decay heat, radiological and chemical t oxicity are present for many hundreds to thousands of years. Transmutation can be accomplished arti ficially by adding neutrons to the long lived atomic nucleus until a less stable nucleus is cr eated with a disproportionate neutron-to-proton ratio, which is less stable, thus having a shortened half-life. There are two primary processes for converting long lived isotopes into shorter lived ones. These are successive neutron captu re or fission. Neut ron energy determines the extent that fission can play in the transmutation of certain ac tinide isotopes. This is because the fission-toabsorption ratio for MAs is dominated by the neutron flux available for absorption above the one MeV fission threshold. Generally, the reactor type and correspondi ng transmutation behavior can be categorized by the energy rang e dominated by the neutron ener gy spectrum. These are fast and thermal reactors. Neutron Spectrum Influence on Transmutation Behavior Because of the larg e MA neutron capture cro ss sections at thermal spectrum energies, thermal reactors have a high efficiency at transmuting MA isotopes by neutron capture. However, at the same time thermal reactors accumula te other actinides within the fuel cycle that are equally hazardous and long live d. A primary example of this is the transmutation of Am241, which is the principle americium isotope in SNF (Figure 1-2). Am-241 is generated from the decay of the plutonium isotope, Pu-241, with a half-life of 14.35 years. Am-241 itself has a half-life of 432.2 years and decays by alpha particle emission. It is the kinetic energy deposition of this alpha particle in SNF, which dominates the heat generation in th e repository for the first 1000 years. Am-241s well resolved thermal neut ron capture cross section resonances, in the

PAGE 33

33 thermal to epithermal energy range, expedites transmuting it into the shorter lived isotope Am242. Am-242 in turn beta decays into Cm-242 with a yield fraction of 83%. Cm-242 alpha decays with a half-life of 163 days into Pu-238. Also, the other 17% of the transmuted Am-242 decays by electron capture into Pu-242 which th rough successive capture reactions followed by decay results in the production of Am-243 and eventually Cm-244. Figure 1-2. Important transmutation and subsequent decay of americium and neptunium into Pu238, Pu-242 and curium and the higher mass act inides. The percentages shown in red are the starting concentrations of these isotopes in LWR SNF TRU. Proposed thermal spectrum transmutation fuels such as LWR Inert Matrix Fuel (IMF) are very efficient at destroying americium isotopes by transmuting them into Pu-238 and Pu-242 but also generate Cm-244 to a large extent. Cm244 has an 18 year alpha decay half-life that dominates the near term or first century heat generation for spent IMF. Successive neutron captures from Cm-244 causes the accumula tion of higher mass curium, berkelium and californium isotopes. These isotopes have very l ittle repository impact because their half-lives are short compared to the exp ected repository lifetime given in the hundreds-of-thousands of years. However, the gamma and neutron radiation fields associated with these isotopes dominate the near term radiological hazard that IMF fuel handlers would need protection from if spent IMF were recycled and re-irradiated in multiple reactor passes [ 8]. Np-237 6.36% Np-238 -, 2.1d Pu-238 3.44% Pu-239 42.98% Pu-240 21.41% Pu-241 10.63% A m-241 3.34% A m-24216.2hr Pu-242 8.51% Cm-242 163d A m-242m I.C., 141a n, 30% 15% 85% -, 82.7% E.C., 17.3% Pu-243 -, 4.956hr A m-243 2.59% Am-244 -, 10.1hr Cm-244 18.1a Cm-245 8500a -, 14.35a

PAGE 34

34 Alternatively, a SFR has a highe r efficiency at destroying TR U isotopes by fission because a greater share of the neutron flux exists above th e threshold fission energy. Essentially all of the transuranics are more fissionable in a fast rather than a thermal reactor. This means that all TRU atoms are supporting the chain reaction by contributing to fast fission. Conversely, in the thermal spectrum, the transmutation process in MA s acts as a capture sink for neutrons generated by the fissile isotopes. The single group fission and capture cross sections for a metallic fueled SFR as well as for a representative IMF assembly are tabulated in Table 1-1 and Table 1-2 respectively. The neutron physics simulation code Monte Carlo N-Pa rticle was used to generate the single group cross sections by tallying over all energies within the fuel region of the ABR. This code is used periodically throughout this disser tation for analysis and verification of results produced by other codes and is described in more detail in the next chapter. The SFR driver fuel is based on the ABR with a metallic alloy consisting of 30TRU/ 60DU/10Zr metallic alloy by weight percent. The IMF fuel is an 8 v/o weapons grade PuO2 mixed with 2 v/o AmO2 in a Nd2Zr2O7 matrix [ 9]. Notice th at the fission and capture cross secti ons are orders of magnitude less for the SFR than the IMF. It is also inte resting to note that the fission-to-capture ratios for the SFR odd-Pu isotopes are only slightly greater than that for IMF. Howeve r, the fission-to-capture ratios for the SFR MAs are an order of magn itude greater than the IMF MAs. The fission-to-capture ratio is a strong indicator of whether fission or captu re is the primary mode for removing an isotope from the fuel cycle. This is especially true when the capture cros s section for any given MA is comparable in magnitude to that of Pu-239. For example, the ratio of Am-241 capture to Pu-239 fission in the IMF fuel is 1.64 versus 0.83 for the SFR. This indicates that Am-241 capture is a stronger

PAGE 35

35 competitor against Pu-239 fission for neutrons in the IMF case as opposed to the SFR case. Furthermore, the ratio of total absorption in Am-241 over total absorptio n in Pu-239 is 1.1 for IMF versus 0.81 for the SFR. Table 1-1. SFR fission and capture cross section data based on the Argonne National Laboratory Advanced Burner Reactor design with a conversion ratio of 0.5. Fission (barns) Capture (barns) Fission per Capture Fission per Absorption Capture per (Pu-239) Fission Absorption per (Pu-239) Absorption U-234 0.340.500.670.400.29 0.40 U-235 1.730.473.710.790.27 1.05 U-236 0.100.370.260.210.22 0.22 U-238 0.040.240.160.140.14 0.13 Np-237 0.321.310.250.200.77 0.78 Np-238 3.800.1232.760.970.07 1.87 Pu-236 3.470.536.500.870.31 1.91 Pu-238 1.100.621.780.640.36 0.82 Pu-239 1.710.384.450.820.22 1.00 Pu-240 0.380.410.910.480.24 0.38 Pu-241 2.280.376.190.860.22 1.27 Pu-242 0.260.360.730.420.21 0.30 Am-241 0.271.420.190.160.83 0.81 Am-242m 3.550.3111.510.920.18 1.84 Am-243 0.201.270.150.130.74 0.70 Cm-242 0.160.250.630.390.15 0.19 Cm-243 2.370.2012.140.920.11 1.23 Cm-244 0.420.720.580.370.42 0.55 Cm-245 2.140.277.870.890.16 1.15 Cm-246 0.260.191.360.580.11 0.22 Cm-247 1.930.277.120.880.16 1.05 Cm-248 0.300.201.520.600.12 0.24 Bk-249 0.161.090.150.130.64 0.60 Cf-249 2.430.584.160.810.34 1.44 Cf-250 1.150.333.520.780.19 0.71 Cf-251 2.210.278.220.890.16 1.19 Cf-252 0.630.252.490.710.15 0.42 Combined 0.050.060.850.460.03 0.05 Considering these two facts, americium is a stronger competitor against Pu-239 for neutrons in the IMF case as opposed to the SFR. Therefore, the thermal spectrum allows more neutrons to be absorbed into americium rather than plutonium because the thermal americium capture cross section is larger than the plutonium fission cross s ection. A higher absorption rate

PAGE 36

36 facilitates greater removal from the fuel cycle and in the case of IMF is caused by nuetron capture reactions rather than fission. Table 1-2. IMF fission and capture cross section data based on a typical 17x17 ressurized Water Reactor fuel assembly design fueled with a TRU-O2/Nd2O2ZrO7 Matrix. Fission (barns) Capture (barns) Fission per Capture Fission per Absorption Capture per (Pu-239) Fission Absorption per (Pu-239) Absorption U-234 0.6219.090.030.030.99 0.67 U-235 12.554.842.590.720.25 0.59 U-236 0.367.680.050.050.40 0.27 U-238 0.138.160.020.020.42 0.28 Np-237 0.6419.990.030.031.04 0.70 Np-238 54.012.2923.630.960.12 1.91 Pu-236 30.318.333.640.780.43 1.31 Pu-238 1.878.390.220.180.44 0.35 Pu-239 19.2310.271.870.650.53 1.00 Pu-240 0.7049.270.010.012.56 1.69 Pu-241 28.058.743.210.760.45 1.25 Pu-242 0.5433.070.020.021.72 1.14 Am-241 0.8231.540.030.031.64 1.10 Am-242m 127.1423.685.370.841.23 5.11 Am-243 0.5440.020.010.012.08 1.37 Cm-242 0.483.570.130.120.19 0.14 Cm-243 55.596.858.120.890.36 2.12 Cm-244 1.0413.160.080.070.68 0.48 Cm-245 36.234.897.410.880.25 1.39 Cm-246 0.712.170.330.250.11 0.10 Cm-247 16.5010.291.600.620.54 0.91 Cm-248 0.907.360.120.110.38 0.28 Bk-249 0.8571.050.010.013.69 2.44 Cf-249 60.4315.133.990.800.79 2.56 Cf-250 1.12208.030.010.0110.82 7.09 Cf-251 140.5154.982.560.722.86 6.63 Cf-252 4.591.662.770.730.09 0.21 Combined 0.440.580.760.430.03 0.03 Transmutation and Nuclear Stability The actin ide cross sections in the SFR neutron energy spectrum are almost completely in the unresolved resonance range. Therefore, the resolved resonance cross sect ion of one isotope self-shielding the reaction rate of another does no t play a significant role. Hence, the ratio of plutonium versus americium fission is more related to the number density of each isotope and the percent of neutrons above the threshold energy. In fact, th e magnitude of the fission-to-

PAGE 37

37 absorption ratio in the epithermal to fast energy range is strongly tied to the relative magnitudes of the unresolved capture cross section below a nd the fission cross section above the one MeV threshold. For example, the capture cross sec tion for Am-241, Am-243 and Cm-244 are at least two orders of magnitudes greater than fission at energies between one keV and one MeV (Figure 1-3). In addition, the magnitude of the capture cr oss sections is roughly the same as that for the Pu-239 fission cross section. If the neutron spectrum for the target is made epithermal, a greater americium destruction rate can be achieved by the neutron ca pture mechanism below the one MeV threshold. A softer spectrum would enhan ce the neutron capture importance relative to fission in the competition for neutrons in this en ergy range. The general increase in the total absorption probability shifts the competiti on in favor of the fertile MA isotopes. As discussed in the previous section, rem oving MAs by neutron capture transmutes them into higher mass actinides. However, similar to IMF, the transmutation of Am-241 into Am-242 is followed by a decay path into curium and ev entually plutonium. A similar transmutation behavior is exhibited by Am-243. Am-243 is tr ansmuted by neutron capture into Am-244 which in turn beta decays into Cm-244 with a half-life of 10.1 hours. Cm-244 decays by alpha particle emission with an 18.1 year half-life into the even plutonium isotope Pu-240. As identified earlier, Am-241 has a long half-lif e and has a fission threshold of one MeV. Transmuting Am-241 produces the fissile Am-242,242m atom. By fissile, it is meant that virtually zero kinetic energy is re quired to overcome the energy barrier for fission to occur. This fissile characteristic is exhibi ted by essentially all actinides with an odd neutron number. Eighty-five percent of this transmuted Am-242 is the ground state which is not long lived enough to be used as fuel in the reactor [ 10,11]. Am-242s primary daughter by beta emission is Cm -242 whose short lived alpha d ecay results in Pu-238 (Figure 13). This Pu-238 has a larger

PAGE 38

38 fission-to-capture ratio in the fast spectrum th an the starting Am-241. However, Pu-238 is not truly fissile because some kineti c energy is required to cause it to fission. But, the amount of added kinetic energy needed for Pu-238 to fission is much less than that needed for Am-241. This explains why the Pu-238 fission cross secti on is so much higher than Am-241 at energies below one MeV. In addition to being more fi ssionable than the starti ng Am-241, Pu-238 is only one neutron capture away from becoming the long lived Pu-239 atom which is fissile. Figure 1-3. Periodic relationship between fissile and fertile transuranic isotopes This circular behavior is al so exhibited by Am-243 (Figure 1-3). The long-lived Am-243 nucleus is transmuted into Am-244 (fissile) wh ich decays by beta emission into Cm-244. This Cm-244 has a short half-life compared to the st arting Am-243. Its 18.1 year half-life alpha decay produces Pu-240 which is more fissionable (but no t truly fissile) than the starting Am-243 and is only one neutron capture away from th e fissile Pu-241 which is fissile. Green = Am-241 Capture Red=Am-243 Capture Gold=Cm-244 capture Blue=Am-241 Fission Purple=Cm-244 Fission Teal=Am-243 Fission Cross Section (barns) Incident Energy (MeV) 1010 105 1 1 10 5 Capture Fission Np-237 2.14E6yr 0.25 0.03 Np-238 2.117d 32.76 23.63 Pu-238 87.7yr 1.78 0.22 Pu-239 2.410E4yr 4.45 1.87 Pu-240 6.56E3yr 0.91 0.01 Pu-241 14.4yr 6.19 3.21 Pu-242 3.75E5yr 0.73 0.02 Pu-243 4.956hr f/c fast f/c therm A m-241 432.7yr 0.19 0.03 Am-242 16.02hr Am-242m 11.51 5.37 A m-243 3.73E3yr 0.15 0.01 A m-244 10.1hr f/c fast f/c therm Cm-242 162.8d 0.63 0.13 Cm-243 28.1yr 12.14 8.12 Cm-244 18.1yr 0.58 0.08 Cm-245 8.5E3 7.87 7.41 83% 17% Isotope T1/2 f/c fast f/c therm.

PAGE 39

39 Isotopic Aspects of Repository Impacts For this d issertation work the motivation behi nd transmutation of Am-241 (and Np-237) is rooted in the site selection and design constraints of the proposed repository site at Yucca Mountain in Nevada. Yucca Mountain is a ridge line located in the dese rt region of Amargosa Valley approximately 90 miles from the city of Las Vegas. The mountain ridge is formed by layers of volcanic rock called tuff. This ro ck was deposited by falling ash from successive eruptions of nearby super-volcanoes many milli ons of years ago. These volcanoes are now extinct. The heat generation rate produced by the alpha decay by Am-241 is the main governing parameter that determines the separation distance between drifts. The drift spacing is determined by the thermal heat rate produced by the SNF wast e packages. The spacing between drifts is the minimum separation needed to keep the mid-drif t rock temperature beneath the boiling point of water (96oC) at the elevation of the Yucca Mountain repository site [ 12]. The reason for the prevention of com plete dry-out between drifts is to ensure that rain water, which is transported via fractures in the tuff layers, is allowed to fl ow freely through the repos itory to the water table below it. The second heat generation limit is the drift tunn el wall surface temperature. The drift wall surface temperature limit is established to be below 200oC to prevent crystalline alteration of the rock. Wigeland et al indicated that this limit is not strongly influenced by the presence of Am241 but rather from the barium and yttrium decay products of the cesium and strontium fission products [ 13]. Cs-137 (T1/2=30.07) and Sr-90 (T1/2=28.78) both have half-lives less than 100 years. Therefore, similar to Wigeland, it is assu med for this work that these isotopes could be separated and diverted from repository storage during the fuel cycle [ 13]. Figure 1-4 shows the decay heat contribu tion for the principal heat generating SNF isotopes.

PAGE 40

40 1E-04 1E-03 1E-02 1E-01 1E+00 1E+01 1E+02 1E+03 1E+041E+001E+011E+021E+031E+041E+051E+06 Time After Irradiation (Years)Decay Heat (Watts/MTU) Np-237 Pu-239 Pu-241 Am-241 Cm-245 total Figure 1-4. Decay heat plot for pressurized wate r reactor SNF (5 w/o starting enrichment and 45 MWD/kg discharge burnup) There is a third repository design consid eration. Because of its long half-life, radiotoxicity, high solubility and low sorption in Yucca Mountain tuffs, Np-237 is the principle environmental concern to the biosphere if wa ter does come into contact with the SNF [ 12]. The natural and engineered waste package barriers are designed to m i nimize the likelihood that water may contact Np-237 for at least 10,000 years. The other soluble SNF isotope s of interest are: Tc-99, I-129 and U-234. However, the radiation dos es from these isotopes are not expected to be as significant as from Np-237. This is prin cipally due to the fact that Np-237 is the alpha decay product of Am-241. Therefore, Am-241 destruction in the SFR is not only important to the emplacement drift spacing but also important for controlling the Np-237 accumulation in the repository. Three repository design benefits are sought for SNF isotope transmutation within the heterogeneous SFR design evaluated in this diss ertation. The first is the removal of Am-241 by

PAGE 41

41 neutron capture in a heterogeneous target. The second is the remova l of Np-237 by neutron capture and fission within the driver fuel. Als o, Tc-99 is evaluated as a control rod neutron poison for reactor performance purposes with the adde d benefit that this isotope is not sent to the repository. Background on Previously Proposed TRU Burning SFRs The net TRU destruction of the SFR increa ses for decreasing conversion ratio. The conversion ratio is commonly define d in fuel cycle term s as the ratio of the TRU production rate divided by the TRU destruction rate averaged over the reactor and irradiation cycle. The most straight forward way to decrease the conversion is by decreasing the parasiti c capture of neutrons in U-238. This prevents the breeding of Pu-239 by transmutation from U-238. Decreasing parasitic capture can be done by simply removing ur anium from the fuel. It can also be achieved by shifting the neutron balance between parasi tic capture and neutron losses by leakage. Generally, for low conversion ra tio SFRs with high MA loaded fuels, there is a tradeoff between the optimal Doppler and void coeffi cients and the attainable TRU destruction efficiency. The two most basic SFR reactivity feedback mechanisms are the Doppler feedback provided by U-238 in the fuel and blankets; and the increase in axia l and radial neutron streaming that occurs during coolant voiding. The tradeoff st ems from the fact that the mechanisms commonly used to remove neutrons fr om the reactor during transients also remove them in steady state operation. For example, enha ncing axial streaming with a pancake geometry or a reduced fuel pin diameter (l arge sodium fraction) makes the void coefficient more negative. But, the increased overall leakage reduces th e available excess reac tivity. Alternatively, increasing U-238 increases resonance capture and ma kes the Doppler coefficient more negative. However, this produces further TRU from captures in the same resonances that provide the beneficial feedback. In both cases, more fissile plutonium is required to compensate for

PAGE 42

42 reactivity lost by the modifying strategy. The ad ded plutonium and lack of fertile material increases the rate that the reac tor losses reactivity as a function of the burnup of the TRU. During coolant voiding, the lack of neutron dow n scattering in sodium causes the neutron spectrum to harden. This hardening causes an in crease in the fission-to-cap ture ratios and hence an increase in multiplication [ 14]. SFR designs typically do not have a passive sodium void negative reactivity feedback. In case of a sodium density reduc tion, passive negative reactivity feedback is norm ally achieved by thermal expans ion of the fuel. Engineered and inherent negative leakage reactivity feedback mechanisms will be explained in a later discussion on SFR reactor control. Most SFR designs are fairly insensitive to ch anges in the sodium density. However, the spectrum hardening effect, caused by a complete void of sodium coolant, is more pronounced for higher MA loadings. The increase in void worth is because more neutrons are absorbed at energies greater than the threshold fission energy which causes an increased multiplication feedback. Past parametric studies in the literatu re showed that the void coefficient increased too much when the MA content exceeded roughly 10% (MA per mass of HM) [ 15]. This is because the m ultiplication increase given by Np-237 and Am -241 creates more neutrons than the Doppler broadening feedback from U-238 can absorb. In so me instances the constraint on MA density is as low as 5% or 2.5% such as the Consomma tion Amliore du Plutonium dans les Racteurs Avancs (CAPRA) type core [ 16,17]. Because the MA loading (in th e fuel) is lim ited by a void coefficient constraint, the attainable MA destruction rate for that design suffers. Several pathways to optimize total TRU destruction, and hence MA destruction, while main taining acceptable reactivity coefficients and reactivity swing have been considered. First and most importantly, the rate of total TRU

PAGE 43

43 destruction by fission is fixed by th e fission reaction rate of the core. This occurs because the rate at which the overall mass is destroyed by fiss ion is fixed by the power rating of the reactor. On average, the amount of energy produced by one megawatt of power produced in one day Megawatt-day (MWD) requi res approximately one gram of HM to undergo fission. Therefore, reducing the TRU generation is the only possibility for maximizing the net TRU destruction, which can only be reduced by eliminating Pu-239 creation from U-238. Enhanced leakage is accomplishe d by altering the geometric buc kling. Axial buckling can be increased by reducing the core he ight with respect to its diam eter giving a configuration known as a pancake core. Parasitic neutron capture is increased by the a ddition of an alternate resonance absorber. This raises the unique possibility of using fi ssion products such as Tc-99 as an alternate epithermal absorber for replaci ng U-238, thus reducing th e conversion ratio. In any case, the reduced poorer neutron population in the core caused by these modifying strategies ultimately results, in most cases, a higher fissile loading to compensate for neutron losses. Therefore, the burnup reactivity swing increases as a result or the refueling cycle length being reduced. The first option has the dr awback of requiring an unrealistic control rod worth to manage the excess reactivity. The second option has the drawback of limiting the irradiation time allotted for burning MAs. Th is second option is important because the probability for absorption in the MA is much less th an that for plutonium (Table 1-1), even in the fast spectrum, meaning that more neutrons are being consumed to fission plutonium in the fuel instead of transmuting MA in the fuel. The Advanced Burner Reactor Design Concept The ABR fuel, core internals, and power plan t ar e essentially identical to the Advanced Liquid Metal Reactor (ALMR) and Super Powe r Reactor Innovative Small Module (S-PRISM) designs that were produced during the IFR program [ 7,18,19]. In order to achieve a lo w

PAGE 44

44 conversion ratio, the ABR draws upon several of the modifications discussed above. First, to reduce parasitic capture in U-238, the internal rows of uranium blankets were removed from the S-PRISM design (Figure 1-5). These blankets were an integral component to the plutonium self-sus tainability of the IFR fuel cycle but are now contradictory to the TRU burning philosophy of the ABR. Second, the ABR core height-to-diameter ra tio is only 0.5 to enhance axia l buckling. Typically, in the absence of internal rows of blankets, the active core (driver fuel only) of a SFR breeder would need a height-to-diameter ratio closer to unity [ 14]. Figure 1-5. Core layouts for the S-PRISM and ABR designs2 2 The ABR uses the same driver fuel assembly design throughout the core but with different TRU enrichments for Inner, Middle and Outer co re regions. The three different enrichment zones are necessary for radial power profile flattening. The reflecto r and shield are both dummy assemblies filled with steel rods to reflect some of the neutrons back into the core and also to protect the rector vessel from neutron damage by fast neut rons. Some variations of the shield include neutron absorbing material. Th e Gas Expansion Module is a special reflector assembly designed to provide nega tive reactivity feedb ack by increasing leakage in the event of loss of coolant flow S-PRISM Shield Assembl y Reflector Assembl y Outer Enrichment Drive r Middle Enrichment Drive r Driver Fuel Assembl y Blanket Assembl y ABR Ultimate Shutdown Rod Assembl y Primar y Control Rod Assembl y

PAGE 45

45 Parametric studies performed by Morris et al showed that further reductions in height (from one meter) do not yield signifi cant reductions in conversion ratio [ 20]. Therefore, the ABR core height is the sam e as the one meter he ight of the S-PRISM. However, the removal of the blankets allowed the radius to be reduced from 2.7 to 2.2 meters. Finally, the ABR fuel pin pitch-to-diameter ratio was reduced from that of the S-PRISM. This had the attribute of adding more sodium volume to the core which also allows for more neutrons to stream out of the core. The volume of uranium in the pin was reduced in step with the reduction in fuel pin volume. Hence, the loading of TRU is roughly kept constant across the S-PRISM and ABR designs but the volume of uranium is decreased. This had the effect of increasing the fuel concentration of TRU per total HM in the core, also known as the TRU enrichment The increase in TRU enrichment dictates a significantl y different fuel assembly design in terms of fuel pin diameter, pin spacers, etc. than the S-PRISM. The fuel assembly and pin dimensions of the S-PRISM and the ABR reference core design used for comparisons in this dissertation are given in Table 1-3. As stated by Hoffman et al, the reduction in pin diameter and the corresponding increase in TRU enrichment cause the fuel excess reactivity to increase and the irra diation cycle length to decrease [ 7]. Because, of cycle length reduction, la rge excess reactivity an d large enrichment, the ABR has a correspondingly different refuelin g interval, control rod worth requirement and fuel composition from the current experience database based on past SFR experience. Transmutation Target Designs: Radi al Blankets and Moderated Targets The principal drawback of having an enhanced leakage SFR core design, such as the A BR, is that the reactor purposely wastes neutrons through leakage inst ead of re-investi ng them in the fuel. This is contrary to prev ious plutonium breeders, which used uranium blankets to recover the neutrons leaked from the active driver core.

PAGE 46

46 Table 1-3. Fuel assembly and pin dime nsions for the S-PRISM and ABR designs Fuel Type S-PRISM Driver S-PRISM Blanket Ref. ABR (CR=0.5)* Approximate core conve rsion ratio 1.00 0.50 Fuel type Metal Metal Metal Fuel alloy composition Pu/U/10-Zr U/10-Zr TRU/U/20-Zr TRU Enrichment (%) 20 0 35 Assembly pitch, cm 16.142 16.142 16.142 Inter-assembly gap, cm 0.432 0.432 0.432 Duct outside flat-to-flat, cm 15.710 15.710 15.710 Duct material HT-9 SteelHT-9 Steel HT-9 Steel Duct thickness, cm 0.394 0.394 0.394 Pins per assembly 271 127 324 (7 support) Fuel pin spacer mechanism Helical wi re wrap Helical wire wrap Grid Spacer wire wrap diameter, cm 0.142 0.094 n/a Bond material in cladding gap Na Na Na Overall fuel pin length, cm 407.040 407.040 407.040 Top fission gas plenum height, cm 191.140 191.140 191.140 Middle Active fuel core height, cm 101.600 101.600 101.600 Bottom Axial reflector height, cm 114.300 114.300 114.300 Fuel smeared/ fabrication density, % TD 75/100 85/100 75/100 Pin outer diameter, cm 0.744 1.201 0.6230 Cladding thickness, cm 0.0559 0.0559 0.0559 Pin pitch-to-diameter ratio 1.191 1.078 1.293 *The ABR reference design is currently not yet a finalized point design in the SFR community. The pin diameter, spacer type, fuel composition, etc. vary depending on the desired conversion ratio of the reactor design [ 7]. In general, the long m ean-free-path of fast neutrons causes SFRs to have a high neutron escape probability. In fact, the active driver fu eled region of most plutonium breeders purposely had a high leakage. However, the high leakag e was created only to maximize the amount of neutrons invested in the uran ium blankets. Therefore, the overall leakage, including the blankets, would be less than the active core by itself. The use of the blankets increases the overall utilization of neutrons to produ ce new fuel in the SFR fuel cycle. Remembering back to Table 1-1 and Table 1-2, though the fission-to-a bsorption ratio of the even neutron number MAs (i.e., Am-241, Am-243, Cm-244) are greater in the fast spectrum than in the thermal spectrum, they are still much less than that of the plutonium isotopes. In fact, the fast spectrum fission-to-absorpti on ratio of most of these MAs is actually closer to that of U-

PAGE 47

47 238. Considering that the transmut ation path of most of the MAs leads directly to much more fissionable material (i.e., Pu238, Pu-242 and Cm-245) within one to two-neutron captures, it is possible that they could be a suit able fertile material for replacing the uranium in the blankets. This heterogeneous approach has been proposed by previous authors studying fuel cycles similar to that of the ABR. Buiron et al studied wrapping a radial external blanket of unmoderated target assemblies around the active core region of the European Fast Reactor (EFR) [ 21]. Similarly, Fujimura et al and Sanda et al proposed a SFR core where moderated targe t assemblies containing neptunium and americium we re scattered throughout the active core and also in a radial blanket surroundi ng the active core of the Self-C onsistent Nuclear Energy System (SCNES) [ 22, 23,24]. The m oderation was provided by a U/MA/Zr/H di spersion matrix. These target assemblies were first heterogeneously scatte red throughout the center core regi on and eventually shuffled to the first row of the radial reflector. The shuff ling was done to ensure th at the plutonium atoms generated from neptunium and americium transm utation did not create an unacceptable power peak at the end of the one year irradiation. Th is study demonstrated that the overall destruction efficiency is enhanced by includ ing moderation to the target design. It also demonstrated a fuel cycle scenario where the transmuted plutonium fr om the targets was recycled into the driver assemblies. Similar results were observed separately by Eliseev et al and Ro me et al. Eliseev et al proposed a target assembly having the outer two rows of the assembly comprised of neptunium and americium target pins with all in side rows being moderator rods filled with zirconium hydride [ 25]. Rome et al also proposed a hete rogeneous target assem bly where the moderating rods and transmutati on target rods were equally di stributed throughout the assembly [26,27].

PAGE 48

48 A table of most relevant transmutation target design studies is offered in Table 1-4. The transmutation target design proposed in this work is a hybrid compilation of: Transmutation targets located on the top ax ial top core periphery to have a close proximity of the target and activ e core regions for flux sharing A heterogeneous lattice of MA pins and zi rconium hydride moderating pins to take advantage of existing metal fuel technol ogy for the target transmutation matrix A slight uranium content in the MA axial blanket/target region of the core (i.e., in the MA loaded pins) for minimizing power swing in the targets during irradiation In all of the radial heterogeneous core designs with target assemblies, the moderating effect of neutrons leaving the target and entering the driver fuel assembly caused localized power peaking in the neighboring fuel pins. To compen sate, all the moderated target assembly designs, found in the literature, used a thermal neutron fi lter that encompassed the target assembly to ensure that thermal neutrons did not leave the target. The filter used by Fujimura et al replaced the peripheral ring of rods within the target assembly with Tc-99 be aring rods. Eliseev et al used a cadmium laced fuel assembly shroud as the thermal filter. Table 1-4. Summary of relative studies on transmutation target assembly: matrix compositions and moderating strategies Author Reactor Type Target Location Target Composition Moderator Reference Buiron et al EFR Outer radial blanket MA-O2/UOX n/a 21 Fujimura et al SCNES Both active core and radial blanket U/MA/Zr/H Hydride Matrix 22,23, 24 Sanda et al SCNES Both active core and radial blanket U/MA/Zr/H Hydride Matrix 28 Eliseev et al BN-800 Outer core MA-O2 rocklike ZrH2 Pins 25 Rom e et al EFR Internal and outer radial blankets MA-O2 CaH2 or ZrH2 Pins 26,27 Transmutation Target Designs: Axial Blankets and Axial Targets The issue of localized po wer peaking may be a voided altogether by taki ng advantage of the large spatial gradient typical of most SFR radi al and axial power profiles. A large flux and power profile gradient is typical of most SFR core designs because of the large geometric

PAGE 49

49 buckling caused by high leakage. Given that the power on the periphery is significantly less than the peak power at the cores ax ial and radial center localized power peak ing from moderation can actually be favorable as a means to flatten the power profile across the co re. In fact, Rome et al found that the issue of power peaking was minimized when the moderated targets were placed exclusively on the core periphery and not in the core center. In this work a moderated axial target/blanket is proposed for cap turing the neutrons leaked from the active core. As it will be shown in Chapter 3, the affect of this moderation actually decreases the axial power gradient in the target region. Axial blankets are actually not a new concep t in SFR design. Tabl e 1-5 gives a list of examples of past reactor core s that incorporated axial blankets into their designs [ 29]. Axial targe t designs are also not a completely foreign concept for transmutation purposes. Kuraishi et al and Arie et al proposed re placing radial and axial blanke ts with fission product targets [30,31]. Table 1-5. Exam ples of real fast reactor plants where axial blanke ts have been incorporated into the SFR core design Reactor Location EBR-I (United States) Above and Below EBR-II (United States) Above FERMI (United States) Above and Below Clinch River (United States) Above and Below JOYO (Japan) Above and Below MONJU (Japan) Above and Below Dounrey (Great Britain) Above and Below Rapsodie (France) Below Phenix (France) Above and Below Super-Phenix (France) Above and Below BN-350/600/800 (Russia) Above and Below Transmutation Based Reactivity Control Concept Due to the fast neutrons long m ean-free-pat h, spatial heterogeneities on the dimensional level of a fuel pin virtually have zero impact on the local neutron flux. Therefore, the major geometry features of SFR are on the dimensi onal level of the fuel assembly. The fuel

PAGE 50

50 enrichment zoning in the ABR is possible due to this smearing effect. This smearing effect over the fuel assembly is also why the SFR has whole primary control and ultimate shutdown control assemblies as shown in Figure 1-5. This smearing principle of neutron absorbers in a SFR makes introduction of burnable poison ma terials problematic unless they can be implemented in standalone fuel assembly structures. Kim et al explored the ho mogeneous pin distribution and heterogeneous assembly distribution of burnable poisons in the Korean Advanced Liquid Metal Cooled Reactor (KALIMER) [ 32]. The burnable poison used was boron carbide (B4C) which was enriched to 90 w/o in the highly absorbing B-10 isotope. Th e purpose of the KALIMER study was not only to determine the B4Cs potential excess reactivit y suppression benefit, but also to determine the affect that the neutron po ison would have on the void and Doppler feedback. The first case studied by Kim was a homogene ous loading where 30 of the 271 fuel pins were replaced with burnable poison rods. The second case was a parametric study to determine the optimal location for burnable poison assemblies instead of homogeneously placed rods. Homogeneous and heterogeneous burnable pois on options were compared to a non-poisoned base case. It was found that the homogeneous ca se actually made the sodium void worth more positive than the base case. This is because th e U-238 fast fission contribution increases in the harder spectrum attained during voiding. The sodium void worth increases because the thermal (resonance) absorption cross sec tions of B-10 (and also the actin ides) are significantly greater than U-238. Therefore, the neutron spectrum b ecomes harder during voiding. Thus, undesirable fast fission neutron multiplication surpasses th e desirable neutron consumption by parasitic capture. The heterogeneous loadings had th e opposite effect because the burnable poison assembly is closer in physical dimensions to the fast neutron m ean-free-path. Since the burnable

PAGE 51

51 poison becomes more visible to fast neutrons, the achievement of lower sodium void worth is primarily owing to the parasitic capture in the burnable poison a ssembly. However, just as in most typical SFR designs the void co efficient was still positive. This concept of burnable control assemblies is applied to the primary control assembly design discussed in Chapter 4. However, the fiss ion product technetium (Tc-99) is used instead of B4C. Metallic Tc-99 is chosen as for reactivity control shim because: The atom density of Tc-99 is highe r than that of B-10 in enriched-B4C. Tc-99 has an unresolved cross section structure in the fast neutron energy range similar to U-238 making it suitable for a gray absorber rod in a SFR. The combination of long fast neutron mean-f ree-path with the gray absorbing Tc-99 in discrete primary control assemblies is id eal for reactivity shim control. The magnitude of the Tc-99 unresolved cross section resonances falls off sharply at energies above one MeV which could promote the above-threshold fission of MAs. Transmuting Tc-99 reduces the amount of this radiotoxic and long lived SNF (or HLW) isotope that would otherwise be dest ined for the Yucca Mountain repository. Messaoudi et al investigated burning Tc-99 as a potential re placement for all of the U-238 in the CAPRA core design [ 33]. The CAPRA design achieve d a low conversion ratio by replacing fuel pins from the driver fuel assemb lies with dummy dilution pins of stainless steel (or Tc-99 as proposed by Messaoudi et al). Wit hout changing the loading of TRU in the CAPRA core, the dilution rods displaced uranium from th e core. This strategy was similar to what was done for the ABR but without changi ng the fuel pin diameters. However, in the Messaoudi et al CAPRA design all U-238 was displaced from the core and was replaced by Tc-99 as a potential resolved cap ture resonance surrogate. It is important to note that Tc-99 has some resolved and unresolved resonances similar to U-238 (Figure 1-6). Messaoudi et al first loaded Tc -99 homogenously in all fuel pi ns. It was found that this homogeneous loading increased the void coefficient compared to the base case with U-238

PAGE 52

52 present. Similar to the results achieved by Kim et al, the increase was explained by the fast fission multiplication feedback becoming greater than the compensation of the Doppler broadening feedback. Unlike Kim et al, Messaoudi et al chose not to concentrate the Tc-99 in specialized standalone assemblies. Instead, a calcium hydride moderator was placed in dilution pins of discrete fuel assembly locations th roughout the core. For the CAPRA design these dilution assemblies comprised every third row of fuel assemblies in the core. The moderation allowed more neutrons to be down-scattered to lower energies where the Tc-99 resolved resonances could have more effect which reduced the magnitude of the positive void coefficient. 1E-07 1E-06 1E-05 1E-04 1E-03 1E-02 1E-01 1E+00 1E+01 1E+02 1E+03 1E+04 1E-071E-061E-051E-041E-031 E-021E-011E+001E+011E+02Energy (MeV)Cross Section (barns) U-238 Tc-99 Figure 1-6. ENDF-VI plot of U238 (blue) and Tc-99 (red) total absorption cross sections Design Rationale of an Axial Heterogeneous Fast Transmutation Reactor Because the SFRs reson ance feedback is highly dependent on the presence of sodium, it is important to have a design that can compensate with alternative neutron sinks during the event of a loss-of-coolant-accident (LOCA). The large mean -free-path of fast neut rons allows feedback by neutron streaming as a common solution to this problem in many SFR designs. However, designing core geometry for high st reaming neutron losses during LOCAs also requires a high neutron leakage during steady-state operation. Therefore, implementing axial

PAGE 53

53 targets will have a dual role: (1) MA conversio n into more fissionable plutonium isotopes, and (2) recovering the axial leakage lost by the active core during stea dy-state operation. Compensation for Inherent Positive Void Reactivity Feedback It was discovered in the ope rational experience of the Expe rim ental Breeder Reactor I (EBR-I) that the SFR cores reactivity was highly sensitive to core geometry. During overpower tests, the EBR-I exhibited oscillatory power characteristics and periods of prompt positive (increasing with increasing powe r) reactivity feedback. Operat ing the reactor in order to quantify these prompt reactivity feedbacks caused a partial meltdow n of the EBR-I Core-II fuel. It was later found that the positive power coefficient was due to thermal-mechanical bowing of the fuel pins towards the core center. The bowing was caused by axial and lateral temperature differentials across the fuel tubes. The inwa rd bowing made the reactor more compact thus causing the reactivity increase [ 34]. The positive power feedback was resolved in the Experim ental Breeder Reactor II (EBR -II) by removal of the upper grid plate [ 35]. The lack of a latera l constraint at the top of the core allowed the fuel to bow outwards, instead of inwards, under thermal gradients. The increase in outwa rd bowing with increasing temperature thus enhanced neutron streaming. Hence, the bowing and subsequent enhancement in neutron losses during a LOCA provided an inhere nt negative reactivity feedback which was demonstrated in tests performed at EBR-II. The inclusion or concentrati on of MAs in the central regi on of the core can actually increase the positive void feedback. Compared to U-238, most MAs do not possess resolved resonances at energies high enough in the SFR fa st flux spectrum to provide useful amounts of negative reactivity feedback. Also, because the fast flux energy spectrum becomes harder during a LOCA, the above-threshold multiplication cont ribution from MAs further complicates and increases the positive reactivity void feedback by the fuel.

PAGE 54

54 The positive reactivity feedback by the MAs crea tes a need to further enhance the negative streaming feedback of the core design. Finding a core geometry that possesses an enhanced leakage property during a LOCA without removi ng too much reactivity from the core during steady-state operation is essentia lly the Holy Grail of SFR desi gn. Traditionally, flattening the core height-to-diameter ratio into a pancake ge ometry has been the most straightforward route for achieving this higher leakage. Remember b ack to the previous discussion on core geometry modifications for attaining a TRU burner, a pancake geometry has also been proposed for achieving a very low conversion ratio. Howe ver, as discussed ear lier, the high leakage necessitates a high TRU enrichment and a shortened cycle length [ 36]. Therefore, there is a practical limit on how flat the pancake core can be m ade. In this work, a hybrid pancake core design is sought that slightly reduces the height of the ABR design while preserving the volume of its active core. The flatter core design: Enhances neutron losses during a LOCA compared to a taller core Reduces the conversion ratio of the active core compared to the ABR Invests the neutron leakage from the activ e core during steady-state operation into transmutation of new plutonium in the axial targets Benefits of Axial Targets for a Dedicated MA Burner Core Design Earlier studies indicate that a low conversion ratio, attained by a high leakage core design, is necessary to destroy th e undesired transuranics waste isotopes found in SNF. From a physics standpoint, a low conversion ratio is ideal for reducing the production of plutonium and MAs by reducing parasitic capture. This work makes a fundamental change in philosophy regarding MA waste management in the SFR. Previous repositor y studies establish the limitation that the Am241 isotope puts on the quantity of SNF th at can be stored in Yucca Mountain [ 13]. Because this isotope is not fissile, its rem oval from the fu el cycle can best be achieved by increasing its

PAGE 55

55 transmutation rate by neutron capture. Because this neutron capture in americium leads to a transmutation path ending in Pu-238 (and the othe r plutonium isotopes to a lesser extent), it can be used as a fertile blanket material. The reactor design proposed a nd analyzed in this dissert ation uses both a moderated epithermal transmutation target fuel and an un-mode rated fast spectrum driver fuel to achieve the maximum MA transmutation rate. Therefore, th e design has a heterogeneous fuel configuration where excess neutrons from the fast driver zone are used to transmute fertile MA in the epithermal target zone. For the remainder of this dissertation, this hybrid MA-to-plutonium converter reactor is given the name: Axially Heterogeneous Fast Transmutation Reactor (AHFTR) This breeding of transmuted isotopes can be accomplished with a MA fast flux trap at the axial top periphery of the active core. The goal of such a flux trap is to recover the fast flux leakage from the active core and moderate it to a softer spectrum which enhances neutron capture in the MA rich target. Th e relationship of this flux trap (t arget) to the active core (driver) is shown in Figure 1-7. This symbiotic arrangem ent allows the driver fuel to have a low MA content while the MA concentration in the targets is significantly higher. The moderation in the targets increases the capture cros s section magnitude (especially in Am-241) relative to fission (especially in Pu-239). Thus, the active core ne utron flux entering the axial targets is invested into capture reactions leading to transmutati on as opposed to being lost to leakage.

PAGE 56

56 Figure 1-7. Core design of the Axial Heterogeneous Fast Transmutation Reactor Middle Enrichment Zone Driver Ultimate Shutdown Rod Assembly Primary Control Rod Assembly Inner Enrichment Zone Driver Outer Enrichment Zone Driver Radial/Axial Reflectors Shield Assembly Axial Target Portion of Driver Assembly Fuel Rod Gas Plenum Outer Core Fuel Assembly (no targets) r z Gas Expansion Module Inner and Middle Core Assembly Configuration Handling Socket Coolant Inlet Nozzle

PAGE 57

57 When the transmuted product from the targets is recycled and used later as driver fuel, the Pu-238 (and other transmuted fissile material such as Pu-238 and Cm-245) will be exposed to the fully fast flux of the active core. There, the transmuted Pu-238 will have higher fast spectrum fissile worth than the initially loaded MAs (Table 1-1). To take advantage of this fast fissile improvement, the spent axial targets are co-reproce ssed with the spent driver fuel and included in the fresh driver fuel fabrication for the next reactor pass (Figure 1-1 and Figure 1-8). Figure 1-8. Modification to the A BR fuel cycle by the axial target s: Solid lines represent the reference ABR fuel cycle. Dashed line represents modifications to include axial targets. The key technical attribute of converting MAs into plutonium isotopes is the relaxation of the requirement to have a low conversion ratio. As will be discussed in the next chapter, homogenous core designs, such as pancake desi gns or the ABR, require a high TRU enrichment in order to achieve a low conversion ratio. The AHFTR axial blanket converts MAs into plutonium. Because, one type of TRU is being trad ed for another, the targets have a negligible impact on the overall core conversion ratio. Ho wever because the MAs are preconditioned into Target Fuel Fabrication Driver Fuel Fabrication Fuel Assembly Manufacturing AHFTR LWR Reprocessing SFR Reprocessing TRU Am+Cm+Bk+Cf Np+Pu TRU

PAGE 58

58 plutonium before being applied to the active core the fissile worth of TRU used by the driver fuel is increased. This effect allows the AHFTR fuel composition to be more comparable to IFR fuel testing experience than ABR designs of equa l conversion ratio. The fuel assembly and pin dimensions of the AHFTR and two ABR core designs are given in Table 1-6. Table 1-6. Fuel assembly design of the AHFTR compared to similar ABR designs Fuel Type AHFTR ABR (CR=0.75) Ref. ABR (CR=0.5) Approximate core convers ion ratio 0.700.75 0.50 Fuel type Metal Metal Metal Driver fuel alloy composition TRU/U/10-ZrTRU/U/10-Zr TRU/U/20-Zr Axial target alloy composition MA/Pu/U/40-Zr n/a n/a TRU Enrichment (%) 20 25 35 Assembly pitch, cm 16.14216.142 16.142 Inter-assembly gap, cm 0.432 0.432 0.432 Duct outside flat-to-flat, cm 15.71015.710 15.710 Duct material HT-9 SteelHT-9 Steel HT-9 Steel Duct thickness, cm 0.394 0.394 0.394 Pins per assembly 271 271 324 (7 support) Fuel pin spacer mechanism Helical wire wrap Helical wire wrap Grid Spacer wire wrap diameter, cm 0.13290.1329 n/a Bond material in cladding gap Na Na Na Overall fuel pin length, cm 407.040407.040 407.040 Top fission gas plenum height, cm 191.140191.140 191.140 Middle Active fuel core height, cm 91.600101.600 101.600 Bottom Axial reflector height, cm 114.300114.300 114.300 Fuel smeared/ fabrication density, % TD 75/10075/100 75/100 Pin outer diameter, cm* 0.755 0.755 0.623 Cladding thickness, cm 0.05590.0559 0.0559 Pin pitch-to-diameter ratio 1.176 1.176 1.293 *The pin diameter varies dependi ng on the desired conversion ratio A more practical driver fuel composition allo ws the fuel assembly design to be more similar to what was proposed for the S-PRISM. Table 1-7 compares the main differences between the AHFTR core design and the ABR designs. Technology Compatibilities and Synergies One of the technology goals of the GNEP progr am is to establish a higher level of accountability for HLW mass stream s than previous fuel cycles Aqueous reprocessing (using

PAGE 59

59 liquid organic solvents) of SNF has been performed in Europe, Russia and Japan on a commercial scale using the Plutonium-Uranium Redox Extraction (PUREX) process. Table 1-7. Core design for the AHFTR compared to similar ABR designs Fuel Type AHFTR ABR (CR=0.75) ABR (CR=0.5) Driver Assemblies 192 144 144 Inner (Lowest Enrichment) 42 30 42 Middle (Medium Enrichment) 66 42 66 Outer (Highest Enrichment) 84 72 36 Primary Control Assemblies 16 16 16 Ultimate Shutdown Assemblies 3 3 3 Gas Expansion Modules 6 0 0 Reflector Assemblies 96 90 90 Shield Assemblies 66 60 60 Core Diameter (m) 2.6 2.3 2.3 Reactor Diameter (m) 3.2 3.0 3.0 Active Core Height (m) 71.6 101.6 101.6 Total Core Height (m) 91.6 101.6 101.6 However, this technology extracts relative ly pure uranium and plutonium products and leaves behind a highly radioactive mixed waste of fission products and MAs which are also considered HLW. From the discussion on the isotopic aspects of the repository design, it is evident that not all of these isotopes carry th e same importance for needing disposal in the repository. Therefore, the GNEP program and its predecessors have developed a suite of addition separations steps to aqueous reproce ssing, called Uranium Extraction Plus (UREX+) [ 37]. The UREX+ reprocessing technology, and its associated sub-steps, divides the HLW stream into individualized mass streams. Uran ium Extraction (UREX) by itself is an aqueous partitioning technology designed only for creating a highly pure uranium stream from SNF. UREX is expanded into UREX+ by adding add itional extraction steps for partitioning: technetium, iodine and Cesium/Strontium. Actinide Partitioning: PUREX and UREX An addition al advantage of the UREX or UREX+ suite is the ability to control the level of elemental actinide partitioning. The only partitioning possible by the PUREX process is the

PAGE 60

60 creation of a plutonium and uranium product. The production of pure plutonium raises some debate over the nuclear weapons proliferation resistan ce of this technolog y. The UREX+ suite was designed to provide elemental partitioning such that plutonium is always diluted in another actinide. For example, the UR EX+1a process separate s all the TRU together as one product so that plutonium is diluted over all the transuranics As seen in Table 1-8, additional separations steps to the UREX+1a process creates an increasing number of product streams. Table 1-8. Waste stream pa rtitioning afforded by various reprocessing technologies* Process Prod. 1 Prod. 2 Prod. 3 Prod. 4 Prod. 5 Prod. 6 Prod. 7 PUREX U Pu MA+all FP UREX+1 U Tc Cs/Sr TRU+Ln FP UREX+1a U Tc Cs/Sr TRU FP+Ln UREX+2 U Tc Cs/Sr Pu+Np Am+Cm +Ln FP UREX+3 U Tc Cs/Sr Pu+Np Am+Cm FP+Lanth UREX+4 U Tc Cs/Sr Pu+Np Am Cm Cm+Ln Pyroproc. U TRU All FP *Prod. is product. Tc is technetium. Cs is cesium. Sr is strontium. Ln is lanthanides. FP is fission products Of these options, the UREX+1a process has the most resemblance to PUREX. The primary difference is that pluton ium is diluted by all of the MAs found with it in SNF. As mentioned in the motivations and objections sect ion, inclusion of these MAs in the SFR driver fuel can lead to an accumulation of curium, berkelium and californium [ 38]. The associated gamma and neutron radioactivity, as well as therm al heat, associated with decay of these actinides may significantly complicate fuel handling a nd fabrication of recycled fast reactor fuel. These high radiation fields rais e the possibility that expensiv e hot-cell facilities would be necessary for all fuel handling operations. Hotcells are large monolithic shielded facilities where radioactive material is handled remotely us ing mechanical master-slave manipulator arms. Because of their large size and re lative complexity, it is expected that hot-cell facilities are significantly more expensive than glove-box facilities. The PUREX fuel cycle allows fabrication in glove-box environments because the MAs and higher mass actinides are removed during

PAGE 61

61 reprocessing. Also, as a general rule, the commer cial PUREX fuel cycles to date have not dealt with the multi-recycling scenario. Hence, th e accumulation of highly radioactive curium, berkelium and californium has not been dealt with on a commercial level. To avoid the fuel handling penalty, it has been proposed by Pillon et al that the complexities of MA management in hot-cells could be limited to a small fraction of the fuel cycle infrastructure [ 39]. Because the MAs constitute only a sm all fraction of TRU, the associated hot-cell infrastructure (MA) can be made significantly smaller than the glove-box (Pu) infrastructure. By constantly partitioning the MAs from the plutonium driver fuel, after each reloading cycle, and continuously recycling them in targets the driver fu el fabrication process could be performed with less difficulty. To achieve the MA partitioning, the UREX+3 pr ocess from Table 1-8 is selected. This option provides that plutonium is not separate d by itself but diluted by neptunium. Also the MAs, of highest importance to repository and hot-cell criteria (Am+Cm+Bk+Cf), are partitioned and sent to target fabrication in Figure 1-8. The fu el cycle in Figure 1-8 is slightly different than that proposed by Pillon et al. Instead of continuously separating MAs from the driver fuel and sending them to targets, it is assumed that the transmutation conve rsion efficiency of americium into plutonium is sufficient enough to not require th at the MAs be multi-recycled in the targets. Therefore, the MAs separated from SNF are irradi ated only once in the targets before this mass is co-reprocessed with the driver fuel into the next batch of driver fuel. Special attention is given to the sizing of the target regi on in the AHFTR to ensure that the concentration of MAs in the driver fuel is small. Pyroprocessing and the Integral Fuel Cycle The sm all concentration of MAs in the driver fuel and the single-pass target irradiation makes feasible the choice of metallic fuel and metallic fuel reproces sing. Metal alloys of

PAGE 62

62 uranium, plutonium and zirconium were successf ully used as reactor fuel at both EBR-I and EBR-II for over 40 years. Metal fuel reprocessi ng was demonstrated at EBRR-II during the IFR program using the pyrometallurgical process in volving metal alkaline salts (Table 1-8) [ 40]. The benefits of applying pyroprocessi ng to any SFR fuel cycle are: Pyroprocess ing does not require water, avoiding many criticality safety issues. Pyroprocessing is performed at high temperatur es and does not use organic solvents. From the PUREX experience, organic solvents decompose when expos ed to the radiation fields of SNF/HLW. The reprocessing machinery of pyroprocessing is considerably more compact than aqueous reprocessing which makes it more suitable for deployment in hot-cells. Similar to UREX+1a, pyroprocessing does not allow for actinide partitioning which increases its general pro liferation resistance. Because of these attributes, pyroprocessing sufficiently meets the requirements for reducing the infrastructure dealing with MAs. The primary difference from the approach taken by Pillon et al, is that the entire AHFTR and its associated IFR st yle co-reprocessing (targets plus driver) perform the task of de dicated MA burning. Hence, the AHFTR and its associated reprocessing machinery constitute a small fr action of the overall fuel cycle and ABR/AHFTR mixed-fleet. In this fuel cycle, the AHFTRs and ABRs would work in parallel to accomplish the overall directive to burn SNF TR U. The relationship between A BRs and AHFTRs in this hybrid partitioning and transmutation stra tegy is given in Figure 1-9. In Figure 1-9, the AHFTRs are the only reactors committed to destroying the MAs separated from SNF. For the purpo se of this dissertation, the ABRs in the mixed-fleet are also assumed to use pyroprocessing t echnology which may or may not re quire hot-cell facilities due to an eventual buildup of highe r mass actinides (Cm+Bk+Cf). The rate of this buildup for various partitioning strategies is currently the topic of ongoing studies in the transmutation analysis field. Given the suitabil ity of metal fuel with pyroproce ssing, it is assumed that if these

PAGE 63

63 technologies are adopted for the AHFTR they will reach economical maturity and would ultimately be adopted for the ABRs as well. Figure 1-9. Partitioning and tr ansmutation scenario of a combined ABR and AHFTR mixedfleet Assumptions for Using Transmutation Targets in SFRs It is im portant to mention that the choice to use pyroprocessing in hot-cells which are collocated with the ABR or AHFTR is based on the f act that the need to transport spent SFR fuel to a centralized reprocessing plant is eliminat ed. Avoiding the need for a large centralized Target Fuel Fabrication Driver Fuel Fabrication Fuel Assembly Manufacturing AHFTR UREX+3 Aqueous Processing Pyroprocessing TRU Am+Cm+Bk+Cf Np+Pu TRU Driver Fuel Fabrication Fuel Assembly Manufacturing Pyroprocessing Np+Pu ABR LWR SNF TRU Reference ABR Recycling Strategy Proposed Addition to the ABR Strategy

PAGE 64

64 reprocessing plant may have a si gnificant economic advantage beca use the cost of transportation is eliminated. Also, it can be argued that the fast reactor fuel cycle services will require collocation with the reactor in or der to simplify the handling issues associated with hot fuel. Though centralized reprocessing and fuel fabrication facilities be nefit from economy of scale, they have not been demonstrated on a large scale with the high fiss ile concentrations and possible radiation fields inherent with fast reactor fuels. LWR SNF is stored for a period no less than approximately five years in order to mini mize the radiation fields from fission products (and Cm+Bk+Cf). When spent fuel (LWR or SFR) comes out of the reactor it is thermally hot due to the decay energy of these fission products. To cool the SNF during the decay time, LWR operators store the spent fuel assemblies in large pools of wate r before transportation them offsite. Fast reactor operators wi ll not have the luxury of using wate r to cool the fuel because of the high fissile content in the driver assemblies. Th erefore, the cost to cool spent fast reactor fuel could be a foreseeable additional non-trivial cost if an interim decay period is required before it can be reprocessed. These assumptions are stated in Table 1-9. Using Table 1-9 as a guide, fast reactors w ith heterogeneous targets are chosen because similar core configurations have be en used in the past explicitly for the purpose of transmutation. Historically, the main transmutation process in SFRs was the conversion of U-238 into fissile Pu-239 atoms. However, regardless of whether or not the transmutati on product is fissile, the inherent neutron economy of fast reactors enab les bombardment of hete rogeneous targets using excess neutrons. Furthermore, fast-fissi on by all actinides, in cluding MAs and their transmutation products, contributes to reactivity which gives th em a neutronic advantage over LWRs where transmutation is concerned. As disc ussed earlier, moderati ng pins in the target region is adopted to provide a spectrum bene fit that can enhance transmutation. Finally,

PAGE 65

65 pyroprocessing of metallic fuel at a hot-cell facility which is co-located with the SFR is chosen. These technologies have been demonstrated during the IFR program to have an industrial synergy when implemented together. Table 1-9. Technology compatibility assumptions with Pros and Cons (Technology options indicated by represent technologies adopt ed for this dissertation)* Pro Con MA targets in LWRs Using existing reactors utilizes existing technology. Thermal capture cross sections are high for most SNF MAs MAs and most of their transmutation daughters are neutron sinks. Conventional LWRs do not possess the surplus of neutrons required to continuously irradiate MAs (and transmuted products ) once the initial fissile materi al has been exhausted. MA targets in SFRs Fast reactors derive all of its reactivity from fast-fission enabling all actinides to be a neutron source (as opposed to a sink) Implementation of SFRs require s reprocessing technology in order to fully utilize the reactivity investment made by transmuting fertile isotopes including the MAs. Moderated targets Utilization of neutron capture enhances conversion into even-plutonium isotopes Moderated targets requi re placement on the core periphery to minimize adverse power peaking and kinetics feedbacks. Less flux on periphery. Un-moderated targets Simplifies materials science and reactor physics issues No spectrum benefits that could potentially enhance transmutation. Once-Through Deep Burn of Targets Eliminates recycling/handling of higher mass actinides Achieving a net-zero MA mass balance with deep-burn is difficult in any real reactor system. Material science does not currently exist to withstand ultra-long irradiations. MultiRecycling of Targets Ultra-long irradiations are not necessary. Closed fuel cycle allows a net-zero MA mass balance Shielding requirements of highe r mass actinides essentially requires a hot-cell infrastructure for MA recycling. Aqueous reprocessing with SFRs Economic advantages due to the economy of scale Spent fast reactor cooling will be considerably more expensive than in LWRs because cooling methods other than spent fuel pools water will be required for criticality safety. Pyroprocessing with SFRs Pyroprocessing has been demonstrated with SFRs and is more adaptable to handling thermally hot spent fast reactor fuels Pyroprocessing does not allow for elemental transuranic partitioning which requires that the targets and driver fuel be reprocessed together thus prev enting complete segregation of MA and plutonium mass streams within the fuel cycle Co-location of reprocessing with SFRs Transportation of spent fast reactor fuel is eliminated Fuel reprocessing and fabrication facilities for every reactor site adds extra capital cost Centralized Reprocessing Economic advantage due to the economy of scale Transportation of fuel is required which also necessitates a spent fast reactor fuel cooling facility Oxide fuel and targets Oxide fuels are well recognized in LWR commercial experience The higher burnup and temperatures of SFRs mutes many advantages such as fission gas retention and pellet integrity Metal fuel and targets Metal fuels have been demonstrated as compatible with SFRs but not demonstrated on a commercial scale Metal fuels are not widely recognized by the commercial nuclear industry *Further advances in any of these technologies ma y change the validity of the assumptions made.

PAGE 66

66 CHAPTER 2 COMPUTATIONAL METHODS AND FAST REACTOR PHYSICS Many of the m ethods for cross section prep aration and flux calculations used to study LWR core physics today, have their origins in the analysis of fast reacto rs. This is partly because fast reactors were essentially one of the fi rst nuclear energy systems. It is also due to the fact that the spatial smearing approximations ma de to simplify lattice and pin-cell calculations for these early methods were usually sufficient to achieve acceptable accuracy in fast reactor applications. Due to the lack of SFR commerc ialization, the demand to update fast reactor simulation and computational methods has been low. Furthermore, due to the general adequacy of smearing fuel pin and assembly heterogeneitie s over large core regions, fast reactor analysis methods have not evolved to the extent of t hose used for LWR analysis over the past two decades. Several calculation code s are used throughout this dissertation. Figure s and Table s will be labeled according to the calculation met hod used. Sometimes the results produced by one code are used as the input data for a secondary co de. If this is the case special mention of the coupling process is given in notes below the graphic or within the nearby text. Calculations and Fuel Cycle Modeling The Argonne National Laborato ry (ANL) fast reactor codes MC2-2, DIF3D and REBUS are used for the reactor physics and fuel cycle calculations [ 41,42,43]. These codes have been developed together by ANL fo r SFR design since the mid-1970s and are well benchmarked against SFR operational data [ 44]. The MC2-2 code was used to gene rate region dependent, 33 energy-group cross sections at hot fuel, cladding and coolant temper atures based on an ultra-fine group cross section library. The MC2-2 ultra-fine libraries were pre-processed at ANL from evaluated nuclear data files (ENDF) and usi ng a standard fast spectrum distribution for appropriately weighting the fine group generation. Starting with the ultra-fine group libraries,

PAGE 67

67 MC2-2 creates a collapsed coarse group cross section set by performing a zero dimensional infinite dilution critical buckling s earch using the extended P1 method [ 41]. The MC2-2 code also performs a resolved res onance broadening treatment, to account for energy shielding only, at user defined material and fuel temperatures. A continuous slowing down calculation is performed at lower energies to handle elastic scattering [ 45]. A direct multigroup spectrum calcu lation accounting for inelastic and anisotr opic scattering and upscatte ring is performed at higher energies. It should be noted that the cal culations performed by the publica lly available release of this code are limited to using an older ENDF/B-V.2 release of cross section data because the publicly available release of MC2-2 does not come standard with the preprocessed ultra-fine group libraries generated from newer ENDF/B-VII data. Also, it should be noted that there are two additional code packages (SDX and DB2), which are typically used by ANL in concert with MC2-2 and REBUS, but are not publi cly available, and thus were not used in the calculations performed by this work, which was performed at the Idaho National Labora tory (INL). The first of these codes, SDX, accounts for the spatial eff ect of heterogeneity introduced by pins in an assembly, which is modeled initially in MC2-2 as a homogenized region. The second, DB2, takes the fission energy spectrum from areas on the periphery of the core and collapses the cross sections for the neighboring control rod, reflect or, and shield regions based on the leakage spectrum. The approach taken by this dissertation (and parallel ABR core physics studies performed by INL) was to disregard the spatial heterogeneity introduced by pins in the assembly. This zero-dimensional approach in the cross s ection collapsing does not account for spatial resonance shielding effects between the variou s core regions. However, for fa st reactor calculations this is

PAGE 68

68 generally sufficient due to the long neutron mean-f ree-path. The fast flux is almost entirely in the unresolved resonance range, thus making unresolved energy shielding the dominating effects in the group collapsing. This assumption is s upported by the good agr eement between the INL benchmarking effort (INL-EXT-12466) of the A BR and the ANL scoping calculations for the ABR (ANL-AFCI-177) [7,38]. Also for this work and the INL benchmark, the reflector and shield region cross sections were collapsed, without DB2, using a generic Pu-239 fission energy spectrum in MC2-2. These 33-group cross section sets produced by MC2-2 are then used by the DIF3D code to perform the actual core physics and criticality calculation. The DIF3D diffusion code was used to solve the multigroup steady state neutron diffusion equation using a hexagonal-z nodal coordinate system [ 42]. In the nodal discre tizatio n, each hexagonal node in the lateral direction represents a fuel assembly. Because the mean-fr ee-path is on the dimensional level of the fuel assembly, the individual fuel rods ar e homogenized across this hexagon. Fast Reactor Equilibrium Fuel Cycle Calcula tions Using the REBUS Code The actual fuel depletion and fuel cycle modeling, including the mass balance between reactors and reprocessing plan ts, is performed by the REactor BUrnup System (REBUS) code [43,46]. REBUS uses DIF3D to gene rate reactio n rate and flux in formation at each time step burn step in its fuel depleti on algorithm. Once the reactor p hysics calculation is completed, the fluxes from DIF3D are fed to a depletion solver routine. This solver uses the exponential matrix method to solve the Bateman equations for isotopic buildup and decay within the discretized burn step. REBUS also performs the in-core fuel management and out-of-core cooling, reprocessing and re-fabricating for each r eactor cycle. The REBUS depletion and fuel management algorithm is given in Figure 2-1.

PAGE 69

69 Figure 2-1. Coupled DIF3D core physic s and REBUS fuel cycle algorithm In the REBUS fuel cycle model, individual fuel assemblies are homogenized into like neutron spectrum representative regions. Theref ore, independent batches of fuel are tracked Homogenize batches across user defined re g ions DIF3D Calculation Deplete Each Batch Does k-eff = 1 at End of Cycle? Modify the TRU enrichment of the fresh fuel batch Create macroscopic cross sections Has the End of Cycle been reached according to the pre-defined burnup limit? (taken to be 18 at. %, or nominally 215 EFPD) Microscopic cross section libraries Reactor and Fuel Cycle Inputs Yes No Yes Create Output File Perform in-core fuel management and out-core fuel cycle activities No No Yes Has the cycle length converged within the convergence criteria from the previous cycle?

PAGE 70

70 within the external fuel cycle bu t not explicitly spatia lly represented in the physics calculation. These regions may represent all of the batches of fuel having common transuranic enrichment, defined as an enrichment zone. However, it is not essential to divide the same-spectrum regions according to the enrichment zoning. The AHFTR fuel management model assumes that each row in the core is a same-spectrum region. Becaus e fuel is typically not shuffled in SFRs, the assumption is made that each row in the AHFTR core is filled with fuel assemblies of the exact same fresh fuel specifications but having different levels of de pletion. Therefore, the average fuel composition within each row (or more genera lly region) is the volumetric average of the same fuel assembly at the diffe rent stages of its depletion. The in-core fuel management and out-of-core fuel cycle activities ar e carried out until the equilibrium cycle was achieved. The equilibrium cycle is defined as the equilibrium or steadystate condition of the fuel cycle when the reactor performance properties (i.e., k-eff, enrichment, cycle length, etc.) become invariant from cycle to cycle. This equilibrium mode calculation was performed for both the ABR and the AHFTR. In a typical REBUS equilibrium fuel cycle calculation, the fuel management operations ar e carried out until the excess reactivity and equilibrium cycle length are f ound. A maximum fuel burnup of 18 a/o after 6 reactor cycles (seven cycles for the outer core of the ABR) was used to constrain the search procedure to determine the equilibrium cycle length. This burnup constraint nominally gives an equilibrium cycle length of approximately 215 Effective Full Power Days (EFPD) for both the ABR reference and the AHFTR. This burnup closely correlates to the maximum exposure limitations of SFR metallic fuel and cladding integrity. These limitations will be discussed in Chapter 6. To account for the spectral changes during fuel depletion, the 33-group cross section library was updated using a coupled corrector-predictor loop between MC2-2 and REBUS. This

PAGE 71

71 loop is initialized by a first gue ss of the region and depletion av eraged fuel composition in the core. This guess is made by using a SNF TR U composition, depleted uranium composition and approximating the fresh fuel TRU enrichment This composition data is used by MC2-2 to generate a guessed cross section set. Then RE BUS uses the guessed cross section set to perform a guessed fuel cycle calculation. Finally, the region and depletion averaged fuel composition is extracted from the REBUS output and imported into a new MC2-2 calculation. An automated scripting system is used to continuously re-calculate the cross sections for each enrichment zone based on that zones fuel inventory at equilibrium. Figure 2-2 shows the flow of cross section and isotopic composition data between MC2-2 and REBUS. Figure 2-2. Flow diagram of data transfer between the MC2-2 and REBUS codes Once this equilibrium fuel cycle was attained, the Beginning-of-Equilibrium-Cycle (BOEC) and End-of Equilibrium-Cycle (EOEC) total core void and Doppler coefficients were calculated. The total core void worth was attain ed by taking the BOEC nu mber densities from Initial guess of fuel enrichment and com p osition MC2-2 Cross Section Generation Cross Section Set REBUS Equilibrium Fuel Cycle Calculation Rebus Output Check: k-eff, cycle length, average fuel enrichment, etc. Y es/No Final Rebus Output Region dependent batch and time averaged com p ositions MC2-2 Cross Section Generation Cross Section Set REBUS Equilibrium Fuel Cycle Calculation Rebus Output

PAGE 72

72 each isotope in each axial and radial region from the REBUS output file into a new DIF3D (no depletion) input file. The number density for sodium for each region was reduced by the coolant volume fraction leaving only the bond sodium in the fuel rod gap with the coolant sodium voided. The sodium fraction was al so reduced for the corresponding MC2-2 calculation. The Doppler coefficient was calculated in a similar procedure without the change in sodium number density. Instead, a new MC2-2 calculation was performed with al l cross sections broadened at a temperature 100 K greater than that for the RE BUS calculation. Then a new DIF3D calculation was performed with the BOEC number densit ies unchanged from the REBUS output file. Light Water Reactor Spent Fuel Calc ulat ions Using the TRITON Code Since REBUS only deals with the closed portion of the fuel cycle involving fast reactors, the external supply of TRU to this closed fuel cycle must be generated externally. To perform this task, the Oak Ridge National Laboratory (ORNL) code package, Scale 5.1, was used to generate the composition of SNF by performing a single fuel assembly depletion and decay calculation [ 47]. Using the geometry of a (Pre ssu rized Water Reactor) PWR 17x17 pin bundle design, a two-dimensional lattice ca lculation was performed to repr esent the in-core irradiation. The SCALE5.1 depletion code, TRITON, was us ed to perform the physics and depletion calculations for this single assembly model. TRITON uses several other scale codes within SCALE5.1 to perform the cross s ection generation and depletion of the fuel in this single assembly calculation. BONAMI was used to apply Bonderenko factors to correct for energy shielding between un-resolved resonances NITAWL was used to apply the Nordheim inte gral treatment for spatial shielding between resolved resonances NEWT is a two-dimensional discrete-ordinance code used to solve the neutron transport equation and create reac tion rate and flux data.

PAGE 73

73 ORIGEN-S is a fuel depletion solver used to carry out the isotope buildup and decay process by solving the Bateman equations fo r each burn step and also for the postirradiation decay period Using TRITON simulation, a uranium oxide ( UOX) fuel composition, enriched to 4.5%, was irradiated to a burnup of 50 MWD/kg. After th is irradiation, the fuel was allowed to decay for five years to represent the spent fuel pool coo ling time. It is assumed that the fuel will be partitioned into the Np+Pu and Am+Cm+Bk+Cf st reams after this cooling off period. After partitioning, an additional decay time of two years was assumed for the time after separation which includes: reprocessing, fuel fabrication and transportation to the AHFTR. The isotopic composition of the SNF that is used as the exte rnal feed to the SFR is given in Table 2-1. Table 2-1. Isotopic composition in weight percent for UOX SNF Spent Fuel Pool and Transportation to Reprocessing Center UREX+3 Reprocessing Discharged from LWR After Five Year Decay Np+Pu Am+Cm Np-237 5.46% 5.54% 5.88% Pu-238 2.47% 2.55% 2.66% Pu-239 45.79% 46.16%49.01% Pu-240 22.61% 22.58%24.01% Pu-241 13.18% 10.28% 9.94% Pu-242 7.05% 7.01% 7.44% Am-241 0.50% 3.28% 1.01% 55.75% Am-242m 0.01% 0.01% 0.17% Am-243 1.92% 1.91% 32.37% Cm-242 0.18% 0.00% 0.00% Cm-243 0.01% 0.01% 0.09% Cm-244 0.78% 0.64% 10.89% Cm-245 0.04% 0.04% 0.65% Cm-246 0.00% 0.00% 0.08% Scoping Calculations and Benchmarking Using the MCNP Code Given the dissim ilarities between the core physic s of LWRs and SFRs, it is prudent to test the computational methods available for fast reacto r analysis using state-of -the-art tools that can simulate both reactor types. The Los Alamos National Laboratory (LANL) code Monte Carlo NParticle (MCNP) is a general purpose physics simulation tool th at uses the Monte Carlo method to recreate the exact particle physics of neut rons (and photons and electr ons) in any arbitrary

PAGE 74

74 three-dimensional geometry ove r a continuous energy range [ 48]. All scoping calculations and benchm arking analysis in this dissertation are performed with the MCNP code. MCNP allows the user to tally the neutron flux, and energy sp ectrum, as well as fiss ion and capture reaction rates, in any region of the core. This neutron tally f eature is used to indicate certain core physics parameters such as leakage and transmutation performance in the axial targets. The MCNP code is also used to benchmark the accuracy of the deterministic diffusion method used by DIF3D to model the heterogeneity between the active core (fast spectrum) and the axial targets (epithermal spectrum). To mo del the accuracy of the coupled reactor physics and depletion al gorithm of the MC2-2/DIF3D/REBUS scripting system, the LANL depletion code package MONTEBURNS was us ed in conjunction with MCNP [ 49]. MONTEBURNS uses MCNP tally data to produce single group fission, capture and n,2n cross sections and fluxes. These cross sections are used for neutron flux de term ination at a particular time step. These fluxes and cross sections are then fed into th e ORNL fuel depletion code ORIGEN2 for burnup and decay calculations to the next time step [ 50, 51]. Then MONTEBUR NS feeds the updated isotopic composition back to MCNP and the process begins anew. Physics of the Reference Metal Fueled Advanced Burner Reactor In the introduction, it was esta blished by references to lite rature that a SF R conversion ratio is essentially a function of parasitic captu re by uranium and neutron escape by leakage. A SFRs conversion ratio can be decreased by increasing axial leakage, as in a pancake design. The conversion ratio can also be decreased by decreasing th e ratio of TRU to uranium loaded in the core (i.e., TRU enri chment) as was done by the ABR. In order to establish the motivation for using a heterogeneous design, such as axial targets, it is important to first qualify these statements. The computational tools described in this chapter are used to explore these hetero geneous qualities and their affect on axial leakage. However, the

PAGE 75

75 ABR, unlike the S-PRISM or the axial target design proposed here, exploits very little heterogeneity (no blankets or transmutation targets). Because the homogeneous ABR design does not take advantage of blanke ts or transmutation targets, d ecreases in conversion ratio are made by changes in the neutron economy (bal ance between neutron losses and neutrons contributing to fission or fissile production). Conversion Ratio and High Leakage Cores Consider th e criticality conditi on that the geometric and materi al buckling must be equated in order for a reactor co re to be critical [ 45]. 22 mgBB (2-1) Where : Bg 2 is the flux curvatures geometric buckling and Bm 2 is the material buckling. For a bare right circular cylinder, the geometric buckling (without reflec tion) is defined by Equation 2-2: 2 2 222405.2 o o zrgzHzR BBB (2-2) Where : Br 2 and Bz 2 are the radial and axial compone nt of the geometric buckling, respectively. R and H are the critical radius a nd height of the core re spectively taking into account the extrapolation distance: stotottot sot tr tr oN N z 1 71.0 1 71.0 1 71.071. 0 (2-3) Where : zo is the extrapolation distance. tr is the transport mean-free-path. tr, t and s are the macroscopic cross sections for transport, total interaction and scattering, respectively. Ntot is the total number of atoms in the homogenized composition. o is the average cosine of the neutron scattering angle assuming elastic isotropic scatter in the lab system. The material buckling of a homogeneous mixture is defined as:

PAGE 76

76 sot af tr af af mD B 3 1 3 12 (2-4) Where : is the average number of neutrons pr oduced per fission. D is the diffusion coefficient. f and a are the macroscopic fission and abso rption cross sections, respectively. For the purpose of this discussion, consider only the Pu-239 and U-238 isotopes. Therefore, the macroscopic cross sections in Equation 2-4 can be represented in terms of the enrichment and the total HM atom content by: 238, 239, 238, 239, 238, 239, 238, 239, 21 1 31 1 1stot TRU stotTRUo ttot TRU ttotTRU atot TRU atotTRU ftot TRU ftotTRU mNr Nr Nr Nr Nr Nr Nr Nr B (2-5) Where : rTRU is the TRU enrichment which for this sample calculation is equal to Pu-239 divided by the sum of Pu-239 and U-238 atoms. f,239, f,238, a,239, a,238, t,239, t238, s,239 and s,238 are the fission, absorption, total interaction and scattering microscopic cross sections for Pu-239 and Pu-238, respectively. The microscopic cross sections for Equati on 2-5 were generate d by tallying the ABR neutron flux and capture and fi ssion reaction rates over all energies and colla psing using MCNP. These tallies were used to crea te representative one-group cross sections for U-238 and Pu-239. These cross sections and the combined number density Ntot of U-238 and Pu-239 are given in Table 2-2. For a constant total HM atom density increasing the TRU enrichment also increases the material buckling, as shown in Figure 2-3. Table 2-2. Reference ABR atom densities and one-group microscopic cross sections for: Pu239 and U-238 N (atom/b*cm) of ref. ABR One-group Microscopic Cross Sections (barns) Total 0.0062 Fission Scatter Absorption Fission Pu-239 0.0007 11.27 8.28 2.09 1.71 U-238 0.0054 10.88 9.53 0.28 0.04

PAGE 77

77 0.0E+00 5.0E-04 1.0E-03 1.5E-03 2.0E-03 2.5E-03 3.0E-03 3.5E-03 4.0E-03 0%20%40%60%80%100%TRU EnrichmentMaterial Buckling "Bm^2)" (cm^-1) Figure 2-3. Material buckli ng of simple bare homogeneous SFR as a function of TRU enrichment (Hand Calculation) Therefore, for increasing TRU enrichment, th e geometric buckling required for criticality must also increase. Thus, for a given hei ght H the axial buckling can be increased by decreasing the cores critical radius R. The critical radius re quired to equate geometric and material buckling will vary depending on core he ight. The critical radius for different core heights and varying enrichment is given in Figure 2-4. 0 20 40 60 80 100 120 20%30%40%50%60%70%80%90%100%TRU EnrichmentCritical Radius (cm) H=200 cm H=100 cm H=75 cm Figure 2-4. Critical radius re quired to equate geometric buckling with material buckling for increasing TRU enrichment (Hand Calculation)

PAGE 78

78 The converse of the above statement is also tr ue. If the geometric buckling is increased (either by reducing height or radi us) the TRU enrichment that is necessary for criticality must increase. Note that this is the reason for flattening the SFRs height-to-diameter ratio to make a pancaked design. The flattened design increa ses buckling and decreases neutron economy. Hence, in order to achieve criticality, the TRU enrichment must be high. Increasing TRU concentration and decreasi ng uranium concentration equates into a reduction in conversion ratio. For the purpose of this example, consider that the only source of TRU breeding results from neutron capture in U-238. Also assume that any neutron absorption in Pu-239 for this simplified model results in TRU destruction. This is not entirely the case in reality because some neutron ab sorptions in Pu-239 result in Pu-240 and the rest of TRU. Therefore, the conversion ratio as a function of enrichment is given by: 239, 238, 238, 239, 238, 238,1aTRU f aTRU a f a dgeneralizer r ndestructio TRU production TRU CR (2-6) Where : CRgeneralized is the generalized conversion ra tio definition for the two isotopes considered in this sample calculation. This e quation is plotted using th e values from Table 2-2 for varying TRU enrichment in Figure 2-5. Therefore, an enrichment of 30% yields approximately a CR of 0.25 in this bare (unreflected) sample problem. From Figure 2-4, this corresponds to a critical radius of over 120 cm for a core height of one meter. This generalized bare reactor conversion ra tio is significantly less than the actual reference ABR (CR=0.5) due to the absence of reflectors. This draws attention to the importance of reflection in a SFR design. Th e mean-free-path betwee n interactions of HM atoms is significantly high in a SFR (Figure 2-6). Therefore it is probable that a neutron can travel great distances in the reactor core wit hout interacting with another fuel atom which

PAGE 79

79 increases the likelihood of escape by leakage. The axial and ra dial reflectors in SFR designs are necessary to maintain some base level of neutron economy. 0.00 0.20 0.40 0.60 0.80 1.00 1.20 0%10%20%30%40%50%60%70%80%90%100%TRU EnrichmentGeneralized Conversion Ratio Figure 2-5. Generalized convers ion ratio of a simple bare ho mogeneous SFR as a function of TRU enrichment (Hand Calculation) 14.35 14.40 14.45 14.50 14.55 14.60 14.65 14.70 14.75 14.80 14.85 0%10%20%30%40%50%60%70%80%90%100%TRU EnrichmentMean Free Path (cm) Figure 2-6. Mean-free-path of neutron travel between interactions (of any reaction type) between HM atoms (Hand Calculation) The correlation between leakages, TRU enri chment and CR can be observed by plotting the leakage fraction which is given by:

PAGE 80

80 2 2 21 1 1 1 1 1 1ga g leakgenonBD BL PLF (2-7) Where : Pnon-leakage is the probability of non-leakage a nd L is the diffusion length. The leakage fraction is plotted in Figure 2-7. 0.0E+00 5.0E-04 1.0E-03 1.5E-03 2.0E-03 2.5E-03 3.0E-03 3.5E-03 0%10%20%30%40%50%60%70%80%90%100%TRU EnrichmentLeakage Fraction Figure 2-7. Leakage fraction of a simple bare homogeneous SFR as a function of TRU enrichment (Hand Calculation) Conversion Ratio and High TRU Enriched Fuels Assume that a very low CR is desired. But th e prospect of core flattening and a pancake design may be unattractive for other reasons such as the expense of a large core barrel. Therefore, it would be necessary to fix the geometric buckling with a fixed set of core dimensions: H=100 cm, R=113 cm. Because th e geometric buckling has been fixed, the material buckling is also a fixed quantity. As previously stated, in order to achieve a low CR, a very high TRU enrichment is necessary (e.g., rTRU>90%). Because the geometric/material buckling and TRU enrichment is fixed by E quation 2-5, the density of HM atoms (Ntot) in the core must decrease to meet the criticality conditi on. The decrease in HM atom density is caused by the fact that as the TRU enrichment increases the fuel becomes mostly fissile. Therefore,

PAGE 81

81 less fuel (i.e., heavy metal) is needed to be cr itical. Figure 2-8 shows this relationship between the CR, TRU enrichment and the HM core loading. This is the approach taken to achieve decr easing low conversion ratios by a homogeneous ABR design. The ABR height is 101.6 cm and its active core radius is 109 cm. A final decision on the design conversion ratio of the ABR has not been made in the GNEP program. Currently, there are several variations of the ABR design th at range from having a CR of 0.25 to 0.75. 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0.0030.00350.0040.00450.0050.00550.0060.00650.007Heaavy Metal Charge Loading (atom/b*cm)Conversion Ratio and Corresponding TRU Enrichment (%/100) Conversion Ratio TRU Enrichment Figure 2-8. Conversion ratio a nd the corresponding TRU enrichment as a function of HM atom charge density for a bare homogeneous SFR of fixed size (Hand Calculation) The decision on conversion ratio is primarily di ctated by the feasibility of fabricating and irradiating high TRU with low uranium concen tration in the SFR driver fuel. In the homogeneous burner reactors propo sed for the ABR, reductions in CR are achieved by removing from the fuel pin, the vol ume corresponding to uranium [ 7,52,53,54]. This decrease in volume for constant fissile inventory in creases the TRU enrichment which reduces the CR. Since the volume reduction is done such that the fuel pin radi us is decreased, the fa brication feasibility of very small diameter fuel pins becomes a design fact or. There is also a separate feasibility issues

PAGE 82

82 related to reactor kinetics and safety of a low conversion ratio SFR with virtually zero uranium content. Physics of the Axially Heterogene ous Fast Transmutation Reactor The AHFTR design concept seeks to increase the neutron economy of the overall SFR design by using axial targets to recover the axial leakage produ ced by the active core. This enables the active core to maintain some le vel of high leakage which is important for establishing a baseline convers ion ratio. Establishing a reas onably low conversion ratio without the targets is important because, as will be shown, the axial targ et region operates in a converter mode (i.e., MA converted into Pu with little decrease in tota l TRU). This is an important point to make because the design ratio nale of the targets is to convert MAs into plutonium which is a different philosophy from a purely MA burning ta rget or a plutonium breeding blanket. The enhanced neutron economy without significantly sacrific ing conversion ratio allows the AHFTR to have a TRU enrichment that is more akin to past SFR fuel experience than the lower conversion ratio forms of the homogeneous ABR. The more feasible TRU enrichment also solves the problem of fabricating small pin diameters. This is because the higher uranium loading in the fuel also adds volume and hence di ameter to the fuel pin. Also, increasing the uranium content of the driver fuel generally enha nces the resonance feedback attributes of the reactor design. Axial Targets and Axial Leakage Recovery To highlight the potential for neutron econom y improvements, the ability of the axial targets to trap axial leakage is tested. A series of parametric studies on core and pin geometry are performed in the next chapter using the REBUS code. These analyses ultimately culminated in the AHFTR design described in the Introduction. For the purpose of showing the axial target

PAGE 83

83 effectiveness in this chapter, the region hom ogenized atom density data was imported to an MCNP calculation to recreate the BOEC DI F3D calculation from the REBUS fuel cycle simulation. No attempt was made to unfold the actual batch or exact fuel assembly compositions from this region-by-region data The MCNP calculation was performed for both the AHFTR and ABR reference design. Within the MCNP model, neutr on current tallies were made fo r increasing heights in the core geometry. In MCNP, the current is tallie d by simply totaling the number of particles, crossing a given surface within a sp ecified angular range. This range corresponds to the limits of integration of the summation over d about in the equation: ), ,, ( tErndAddtdEJA tE (2-8) Where : J is the net leakage. E is neutron energy, t is time, A is the area of the tally surface, is the solid angle over which the tally is take n. n is the unit normal vector to the tally surface and is the region, energy and time dependent angular flux. Integration of Equation 2-8 gives the net current with a positive sense in the direction of the normal vector. For the following calculation, this normal direction was chosen to be positive upward in the direction towards th e top of the core. For neutrons leaving through the bottom of the core, the absolute value of the current is ta ken so that the current appears positive at all locations. The current and flux distribution for the AHFTR is given in Figure 2-9 and Figure 210. The change in current per change in height is, for all practical purposes, the average axial leakage. It can be observed from Figure 2-9 th at the net number of neutrons moving out of the core is less at the top of the targets (90 cm) than it is at the top of the active core (70 cm). This represents the combined effect of reflection and absorption by the targets. To quantify reflection

PAGE 84

84 at the top of the core, the targets were replaced by : (1) a region of sodium, (2) axial reflector or (3) driver fuel. When compared with a sodium or steel reflector, the targets have less net current at the top of the target region than either of the two reflect or compositions. This indicates the added effect of absorption in the targets. When the axial target region is replaced by fuel, the shape of the axial current profile flattens out as the neutrons enter the gas plenum above the core. Therefore, it is apparent that even though the target region contains some TRU, its neutron production is less than that produced by the active core region. 0.0E+00 2.0E+13 4.0E+13 6.0E+13 8.0E+13 1.0E+14 1.2E+14 1.4E+14 1.6E+14 1.8E+14 2.0E+140102030405060708090100Axial Distance from Bottom of Active Core (cm)Absolute Value of the Net Axial Current (cm^-2 s^-1) AHFTR ABR Reference Target Material Replaced by Sodium Target Material Replaced by Steel Reflector Target Material Replaced by Fuel Figure 2-9. Axial current distribu tion of the AHFTR and ABR (MCNP) Thus, it can be inferred from this comparison that the decrease in current in the axial targets is caused by a combination of axial reflection as well as capture reactions by the MAs. The reduction in the net current from the top of the active core to the t op of the target region equates into a reduction in leakag e by the target region. This reduction in leakage can also be seen by comparing the curvature of the scalar flux in the target region compared to that of the active core (Figure 2-10). Active Core Target Region

PAGE 85

85 0.0E+00 5.0E+14 1.0E+15 1.5E+15 2.0E+15 2.5E+15 3.0E+15 3.5E+15 4.0E+15 4.5E+15 5.0E+15 020406080100120Axial Distance from Core Bottom (cm)Total Flux (cm^-2*s^-1) AHFTR ABR Reference Figure 2-10. Axial flux distributi on of the AHFTR and ABR (DIF3D) Axial Targets and Minor Actinide Conversion The capture effect of the MA s in the moderated target re gion can best be qualified by examining the energy dependence of the capture a nd fission reaction rate of Am-241. To do this, the MCNP model used for the above neutron current calculation was modified to tally the neutron flux in the target region and in the active core region. This flux tally was discretized into 33 equal lethargy sized bins (same as that used by MC2-2) in order to creat e a point-wise neutron energy spectrum for both regions. This binned fl ux tally was weighted with a cross section multiplier corresponding to the isotopes and reacti on rates of interest. Using the Monte Carlo method, this multiplier effectively integrates the reaction rate in each bin from an energy continuous flux and cross section. To calculate a volume averaged flux, MCNP sums the length of all neutron track-lengths crossing the tally region and then divides this summation over the regions volume. The flux multiplier is used in a similar way except the track-length of each

PAGE 86

86 tally is multiplied by the continuous energy micros copic cross section of the desired reaction and isotope type. V EEL EdEE ERi E length track EE E ii i)()( ~ ) ~ () ~ ()(2/1 (2-9) Where : Ltrack-length is the total length of travel of a neutron as it passes through the tally region. V is the volume of the tally region. i(E) is the microscopic cross section for the reaction and isotope of interest for the tally multiplier. Using the 33 energy bins, the binned reac tion rates give an approximately smooth distribution of the reaction rates as a function of energy. The binned capture and fission reaction rate spectra for Am-241, Pu-238 and Pu-239 are pl otted in Figure 2-11 and Figure 2-12. It is important to mention that these plots give the microscopic reaction rate which is normalized per atom and not the macroscopic reaction rate which would be weighted by the atom density of each isotope. 1E+09 1E+10 1E+11 1E+12 1E+13 1E+14 1E+15 1E+16 1E-071E-061E-051E-041E-031E-021E-011E+001E+011E+02Energy (MeV)Fission and Capture Reaction Rates (b*cm^-2*s^-1/lethargy) Capture: Pu-238 Fission: Pu-238 Capture Pu-239 Fission: Pu-239 Capture: Am-241 Fission: Am-241 Figure 2-11. Binned microscopi c reaction rate spectra of sele ct isotopes as a function of incident neutron energy on an atom in the moderated target region (MCNP)

PAGE 87

87 As can be seen in Figure 2-11, the fission re action rate of Am-241 is several orders of magnitude less below one MeV than for any of the plutonium atoms. However, the Am-241 capture reaction rate is of the same order of ma gnitude as any of the pl utonium fission reactions in plutonium. Therefore it is evident that transmut ation of an Am-241 atom into a plutonium (i.e., Pu-238) atom will increase its fissile worth. It is useful to note that the Pu-238 fission rate falls off sharply at energies below 100 eV (Figure 2-13). This is due to the fact that the Pu-238 fission cross section falls off sharply below 100 eV. 1E+09 1E+10 1E+11 1E+12 1E+13 1E+14 1E+15 1E+16 1E-071E-061E-051E-041E-031E-021E-011E+001E+011E+02Energy (MeV)Fission and Capture Reaction Rates (b*cm^-2*s^-1/le thargy/lethargy) Capture: Pu-238 Fission: Pu-238 Capture Pu-239 Fission: Pu-239 Capture: Am-241 Fission: Am-241 Figure 2-12. Binned microscopi c reaction rate spectra of sele ct isotopes as a function of incident neutron energy on an atom in the active core (MCNP) It is also noteworthy to point out that neutron absorptions in Pu-238, that do not result in fast fission transmutes into fissile Pu-239, which has a high fission cross sec tion at all energies. Unlike the purely fast spectrum (Figure 2-12), isot opes in the axial target are equally exposed to neutron fluxes in both energy ranges. Because the practical energy range for fission reactions increases with each successive neutron capture (F igure 2-11), the target is essentially breeding fissile worth, even though fissile atoms are not a direct result of the first neutron capture.

PAGE 88

88 1E-04 1E-03 1E-02 1E-01 1E+00 1E+01 1E+02 1E+03 1E+04 1E+05 1E-081E-061E-041E-021E+001E+02Energy (MeV)Cross Section (Barns) Am-241 Capture Am-241 Fission Pu-238 Fission Figure 2-13. Comparison of capture and fissi on cross sections between Am-241 and Pu-238 (ENDF-VI) If the MA atom concentration in the targets is sufficiently high, the Am-241 capture reactions will energy shield the fission reaction of both Pu-238 and Pu-239. Because neutrons are being absorbed in Am-241 instead of plut onium, this energy shield ing effect is roughly analogous to the spatial resonan ce shielding provided by a burnable poison in an LWR fuel pin. Because of this energy shielding, the fissile worth created via transmutation in the target region can not be truly realized until the transmuted pl utonium atoms are reprocessed and placed in the fast spectrum as driver fuel. The flux in the active core is entirely fast (Figure 2-12). Theref ore, the corresponding reaction rate spectra are also in the fast energy range. Hence, th e energy shielding ability of Am241 over Pu-238 or Pu-239 is not available. Th e capture cross section of Am-241 falls off sharply after one MeV, whereas the fission cross sections of Pu -238 and Pu-239 are much higher in this energy range. Also, the concentration of Am-241 is significantly less in the driver fuel than it was in the targets because of the much smaller Am-241 concentration in the driver fuel

PAGE 89

89 than in the targets. Hence, without the ener gy shielding by Am-241, the fissile worth of Pu-238 and Pu-239 is higher in the active core than in the target region. The change in fissile worth can be seen by observing the differen ces in fission-to-absorption rati o between the target region and the active core region (Table 2-3). Table 2-3. Fission and capture one-group cross sections and fissi on-to-absorption ratios for the AHFTR axial target and ac tive core regions (MCNP) Axial Targets Driver Fuel/Active Core Fission (barns) Capture (barns) Fission per Absorption Fission (barns) Capture (barns) Fission per Absorption U-234 0.32 3.00 0.100.350.46 0.44 U-235 3.93 1.69 0.701.640.42 0.80 U-236 0.12 1.73 0.060.100.33 0.23 U-238 0.04 0.63 0.060.040.22 0.15 Np-237 0.31 4.83 0.060.341.19 0.22 Np-238 10.16 0.36 0.973.520.11 0.97 Pu-236 8.54 1.66 0.843.300.50 0.87 Pu-238 1.36 2.02 0.401.100.56 0.66 Pu-239 3.58 1.91 0.651.670.33 0.83 Pu-240 0.37 2.19 0.150.390.38 0.51 Pu-241 5.74 1.45 0.802.180.33 0.87 Pu-242 0.26 1.78 0.130.270.33 0.46 Am-241 0.28 4.75 0.060.281.29 0.18 Am-242m 11.54 1.71 0.873.310.26 0.93 Am-243 0.20 4.68 0.040.201.15 0.15 Cm-242 0.15 1.57 0.090.160.21 0.44 Cm-243 8.58 1.20 0.882.240.17 0.93 Cm-244 0.45 2.24 0.170.440.66 0.40 Cm-245 6.31 0.83 0.882.000.25 0.89 Cm-246 0.28 0.93 0.230.270.17 0.61 Cm-247 2.59 0.98 0.731.940.25 0.88 Cm-248 0.35 1.96 0.150.310.18 0.64 Bk-249 0.18 5.14 0.030.170.97 0.15 Cf-249 6.03 1.45 0.812.330.56 0.81 Cf-250 0.95 5.78 0.141.190.29 0.81 Cf-251 6.09 1.66 0.792.160.25 0.90 Cf-252 1.94 0.80 0.710.620.23 0.73 Combined 0.05 0.12 0.270.060.06 0.47 Based on fission-to-capture ratio data from Table 2-3, one Pu-238 atom is produced for every 1.38 neutron captures in Am-241 in the target region. This newly formed Pu-238 atom than fissions for every 2.49 neutron absorptions. This is a smaller neutron investment compared to LWR IMF; which is 5.47 absorptions per fissi on (Table 1-2). An additional capture in Pu-238

PAGE 90

90 produces Pu-239 which has a fission-t o-absorption ratio in the targets of 0.65. In the fast active core region, 1.586 neutrons are re quired to convert an Am-241 atom into a Pu-238 atom. This newly formed Pu-238 atom than fissions for every 1.51 neutron absorptions. The much high fission-to-absorption rate of Pu238 in the active core (i.e., 1/1.51=0.66) than the axial targets (i.e., 1/2.49=0.40) shows the ut ility of the spectrum shift from the thermalized to the fast spectrum. An additional capture in Pu-238 pro duces Pu-239 which has a fission-to-absorption ratio in the active core of 0.83. Combining Leakage and Capture Effects As can be seen from Figure 2-14, even though the target region cont ains moderating pins, the moderating effect places most of the flux in the epithermal to fast energy range. In this energy range, most of the resolved resonances ar e very close to each other and not sufficiently wide to create local flux depressi ons in the core or pin geometry. This is different from thermal spectrums where neutron slowing down places most of the neutrons in the thermal energy range where the resonances are much wider and well se parated. Therefore, spatial self-shielding effects between well resolved and well separated resonances have less im portance in the axial target region than for completely thermal spectrums such as for LWRs. Hence, the samespectrum region-wide homogeniza tion assumptions used in cross section and flux calculations by the MC2-2 and DIF3D/REBUS codes ar e adequate for approximating the lattice physics in the active core and target regions. Also, it was demonstrated in this chapter that SFRs inherently have a high TRU enrichment for criticality reasons because of th eir inherent high rate of neutron leakage. Therefore, the relative change in concentrati on of fissile atoms (i.e., TRU enrichment) as a function of burnup can be small, even for high burnups Hence, it should be expected that the neutron spectrum in the core is fairly insensitive to the effect of fissile atom depletion.

PAGE 91

91 Therefore, the need for spectrum updated cross section sets (i.e., micro-depletion ) in analyzing the ABR and AHFTR is not as strong as it is for LWR spectrums. In LWR fuels, the enrichment is small compared to SFRs. Therefore as the LW R fissile atoms are deplet ed, the percent change in the fissile concentration is si gnificantly more than it is for SFRs. The burnout of isotopes with resonances that are well resolved and well sepa rated changes the magnitudes of flux depressions that are caused by those resonances. Therefore, for LWR fuels it is important to create several cross section libraries correspondin g to different stages in depletion. Because, SFRs such as the ABR and AHFTR do not have significant changes in the neutron spectrum with burnup, this spectrum updating or micro-depletion is not necessary. 1E+10 1E+11 1E+12 1E+13 1E+14 1E+15 1E+16 1E-071E-061E-051E-041E-031E-021E-011E+001E+01Energy (MeV)Flux (cm^-1*s^-1/lethargy) Target Region Active Core Region LWR IMF Figure 2-14. Comparison of flux spectrums betw een LWR IMF (thermal), AHFTR target region (epithermal-fast) and AHFTR active core region (fast) (MCNP) In fact, the ability of the AHFTR to breed plutonium through Am-241 transmutation is made possible because the target regions spectr um does not change significantly with burnup. The moderation effect causes the energy shieldin g of plutonium unresolved and poorly resolved resonances by Am-241. This en ergy shielding minimizes the in situ burnup of plutonium isotopes in the target region. Therefore, transmuted plutonium isotopes are allowed to

PAGE 92

92 accumulate with irradiation. The fissile useful ness of this transmuted plutonium is only fully realized when these isotopes are placed in th e active core where the spectrum is much harder. The negligible spectrum change in the target region is mostly due to the fact that the target regions epithermal spectrum is not strongly a ffected by the well resolved and well separated resonances occurring at thermal energies. The epithermal flux can be explained by the fact that the target volume (despite the presence of mode rating rods) is still mostly filled with sodium, steel and HM atoms. In addition, the curvature of the flux (i.e., buckling) in the targets is not completely flat and is similar to the active core. Therefore, th e axial leakage th rough the targets is still fairly significant. This fact suggests that the target region is not a black absorber and that a significant number of neutrons are lost through leakage before they can be moderated and absorbed. Hence, the axial target region behaves mo re as an integral component of the core than a neutron sponge outside of the core. The target composition contains a starting amount of plutonium and uranium that makes the total plutonium content and correspondi ng spectrum roughly constant throughout the irradiation. This constant level of plutonium (regardless of isotopic composition) helps minimize spectrum shifts in the target region. A detailed explanation of the targ et composition of MAs, some plutonium and some uranium is given in the next chapter.

PAGE 93

93 CHAPTER 3 AXIAL TARGET DESIGN ANALYSIS As discussed in the previous ch apter, in order to give a high net TRU destruc tion rate it is necessary to greatly reduce the SFRs conversion ratio. This has the advantage from a waste management point of view of decreasing th e amount of net TRU production from neutron capture in U-238. The first necessary require ment for reducing the amount of U-238 capture reactions is to decrease the amount of U-238 in the SFR. This decision leads to an ABR design with no U-238 radial blankets. Furthermore, th e removal of blanket assemblies reduces the overall reactor size from an S-PRISM design which has nine rows of driver and blankets to the current 1000 MWth ABR design with seven driver rows [ 7,19,55,56]. Waste Management Philosophies a nd Conversion Ratio Definition Traditionally, the fissile CR (fCR) has been defined as the fissile atom production rate divided by the fissile at om destruction rate [ 14]. The fissile production ra te is defined as the sum of neutron capture reactions (inc luding reactions of fissile atoms) that lead to production of a fissile atom. The fissile destruction rate is the sum of neutron capture and fission reactions which remove fissile atoms. Therefore, the fCR is simply the ratio or balance between the mass production rate (source) divided by the rate of mass destruction (sink). This more traditional CR definition is the one calculated and reported by the REBUS code [ 57]. d d f d c d p p c pN N nDestructio Atom Fissile Production Atom Fissile CRf (3-1) Where : (d) is the daughter product resulting fr om the neutron capture and subsequent decay (transmutation) form the (p) pare nt isotope. N is atom density and c and f are the single group neutron capture and fission cr oss sections respectively.

PAGE 94

94 This definition has been altere d slightly by authors in recent years to mean the net TRU produced from uranium divided by the net TRU destroyed by fission giving a transuranic burning CR (tCR) [ 58]. The difference between the fCR and tCR is that the tCR does not give credit to capture reactions in the denominato r because these reactions only transmute existing TRU into other TRU. From a waste management standpoint only fissi on removes transuranic waste from the fuel cycle. Additionally, onl y capture reaction of uranium (specifically U-236 and U-238) that lead to TRU atoms ar e allowed in the denominator of the tCR because these isotopes directly transmute into Np-237 and Pu-239. d TRU f TRU U c U U c UN N N 238 238 236 236 tFission TRU Production TRU CR (3-2) The change in the definition of CR is made to reflect the ABRs principle design objective, which is to destroy TRU. The only real disc repancy between these two definitions is the accounting of MA transmutation. The tCR does not consider fissile plutonium generation from MAs as a source term in the numerator. The fCR gives an indication of the balance between transmutations over total absorption reactions in th e fuel but does not give an indication of the actual TRU burning performance of the core. In both cases, MA transmutation is simply treated as existing TRU isotopes being converted into ot her TRU. For example, if only plutonium and no uranium are supplied to the SFR, than the tCR is zero. This is because the net TRU production in the numerator of the tCR is zero. The fCR would also be near zero because the transmutation parent of Pu-239 is U-238. The tCR and fCR would be in good agreement because the contribution of U-238, which is the domina ting parent isotope in the numerator of the fCR, would be near zero.

PAGE 95

95 However, if neptunium and americium is th e only fertile species and the only fissile species is their transmutation daughter Pu238 (and ultimately other TRU through successive neutron capture), than the tCR is still zero. Yet, the fCR would be near unity. In this hypothetical situation, there is no net production of TRU. However, this hypothetical core would generate its own fissile worth, in the form of transmuted plutonium, in order to meet the excess reactivity and cycle length requirements. Therefor e, the core would be self sufficient from a reactivity standpoint and thus require no external supply of fissil e isotopes. Similar to a breeder reactor, only an external supply of fertile MA isotopes would be required. It is important to remember that transmuted plutonium (Pu-238) is not fissile in the classical sense because in order for it to fissi on, additional kinetic energy must be added to overcome the critical energy for fission. This is why the fission cross section of Pu-238 is so much higher in the fast spectrum than in the th ermal spectrum (Table 1-1). Pu-239 is fissile by definition because fission can occur simply by c ontact with a neutron of virtually zero kinetic energy. The AHFTR design concept uses the fertile na ture of SNF MAs to reduce the amount of externally supplied SNF Pu-239 needed to supplem ent the plutonium bred internally. This has the overall effect of reducing the tCR while increasing the fCR. Meeting the reactivity demands of the fuel cycle using MAs reduces the demand for externally supplied fissile material from SNF. The resulting surplus of SNF plutonium could be better used to fuel other SFRs such as the ABR or perhaps be used to make plutonium based Mixed Oxide (MOX) fuel for LWRs. Odd mass number plutonium isotopes (Pu-239,2 41) are fissile in any spectrum which makes them a viable fuel for thermal reactors. In effect, SNF plutonium in a MOX fuel form can have a tCR between 0.6 and 0.7 when irradiated in a PWR [ 14]. Given that several of the newer

PAGE 96

96 classes of LWRs (Generation III and III+) are desi gned with the capability to be fueled with full MOX cores, these reactors are a possibility for burning the pl utonium component of SNF. However, for the purpose of bounding the scope of this work, only the ABR and AHFTR are considered as the final destination for SNF Pu and MAs in the closed fuel cycle analyses. For the closed fuel cycle analyses to follow, the emphasis of the AHFTR is on MA consumption, whereas the ABR focus is to destroy the plutonium component of SNF TRU. This chapter will show that the AHFTR is well suited for the task of burning the MAs from TRU. Therefore, a combination of ABRs and AHFTRs can be used to burn the TRU produced by LWRs. In this scenario, most of the TR U mass, which is plutoni um, would be burned by ABRs. A fraction of the mixe d-fleet (ABR and AHFTR) would be reserved for AHFTRs which are dedicated to MA burning (Figure 1-9). An economic evaluation that compares this hybrid ABR and AHFTR fuel cycle with the single reactor ABR fuel cycle will be evaluated in Chapter 7. Transmutation Based Reactor Design In this chapter, a parametric design study is conducted on the effects of core flattening, moderation and heavy metal fuel content upon th e transmutation performance of the axial targets. The plutonium generation in the axial targets is also a functi on of reactor and fuel geometry. Therefore, the primary purpose of th e parametric study is to evaluate the tradeoff between core performance factors and the maximum MA destruction rate. Also evaluated is the: cycle length, excess reactivity, and total core D oppler coefficient and so dium void worth. All these factors are related by core geometry and fuel composition. In this parametric study, the targets height and composition are held constant. The targets comprise a 20 cm tall axial blanket placed betw een the plenum and active core. In this parametric study, the lateral core layout is assumed to be iden tical to the ABR case, reported by

PAGE 97

97 Hoffman, with a tCR of 0.5 [ 7]. The lateral core layout for the ABR with the axial target modification is shown in Figure 3-1. Starting with the ABR re ference height of 101.6 cm the active driver core height was varied to 51.6 cm in 10 cm increments. For each core height, the total core power was reduced by the fraction of volume removed from the active core by the height reduction. The reduction in power with height was done in the parametric study to give an equally comparable power density, depletion rate and cycle length for each height. Table 3-1 gives the reactor power of the core shown in Figur e 3-1 for each increment in active core height. Table 3-1. Reactor thermal power and core heights evaluated in parametric analysis Total core height (cm) 121.6111.6101.691.681.6 71.6 Active core height (cm) 101.691.681.671.661.6 51.6 Reactor Thermal Power (MW) 1000900800700600 500 For each one of these active core heights, a range of pin pitch-to-diameter ratios, which correspond to variable CR, is evaluated. For comparison purposes, the ABR reference (Figure 15) was evaluated for each increment in pin pitch-to-diameter ratio. These ABR reference cases provide a benchmark for comparison of reactor parameters for down-selecting to a practical AHFTR core height and pin diameter. The pin pith-to-diameter ratios and the corresponding fuel assembly volume fractions are shown in Table 3-2. These values represent th e fuel pin dimensions, used by Hoffman, to vary the ABR tCR from 1.0 to 0.25. Table 3-2. Reference fuel assembly de sign for varying pitchto-diameter ratio. p/d=1.1 p/d=1.176 p/d=1.293 p/d=1.357 tCR of the reference ABR 1.000.7500.500 0.250 Fuel pin diameter (cm) 0.8080.7550.623 0.464 Fuel pin pitch (cm) 0.8880.8880.806 0.630 Pins per assembly 271271324 540 Fuel Assembly Volume Fraction Fuel 34.27%29.30%22.08% 17.40% Bond 11.42%9.77%7.36% 5.81% Structure 25.73%25.68%26.41% 29.15% Coolant 28.79%35.25%44.15% 47.60%

PAGE 98

98 Figure 3-1. Preliminary AHFTR core design used in parametric analyses Finally, a core height and pin diameter that gives the highest tran smutation efficiency while at the same time ensuring reactivity coeffi cients within the boundar ies of the homogeneous AHFTR Fuel Assembly Configuration ABR Fuel Assembly Configuration Handling Socket Middle Enrichment Zone Driver Ultimate Shutdown Rod Assembly Primary Control Rod Assembly Inner Enrichment Zone Driver Outer Enrichment Zone Driver Radial/Axial Reflectors Shield Assembly Axial Target Portion of Driver Assembly Fuel Rod Gas Plenum Coolant Inlet Nozzle

PAGE 99

99 reference was selected. Using this core height and pin diameter combina tion, additional driver assemblies were added to the AHFTR core radius to bring the thermal power rating back to 1000 MWth as it is for the ABR design. This gives the AHFTR a commer cial scale by roughly equating the overall active core volume and power to that of the ABR. Transmutation Targets and Accompanying Fuel Cycle Target rods are placed adjacent to zirconium hydride (ZrH1.6) dilution rods in an axial blanket configuration above the dr iver fuel. The term dilution was adopted to describe blank fuel pins within the CAPRA core design that were filled with steel instead of fuel. A zirconium metallic fuel alloy is assumed for both driver and target rods. Zirconium hydride was selected as the moderator for its high thermal conductivity and melting temperature. A hydrogen-tozirconium stoichiometric ratio of 1.6 was sele cted for zirconium hydrid es delta phase which retains composition for temperatures up to 1000oC (Figure 3-2) [ 59]. Figure 3-2. Zirconium hydride phase di agram for varying hydrogen content [ 59]. These driv er fuel and axial target rodlets, s lugs, are collocated with in the same fuel pin and share the same plenum space. The term slu g is common in metal alloy fuel literature which indicates the injection casting process used in fabrication. Therefore, the fuel rod contains

PAGE 100

100 a stack of metal slugs as opposed to the ceramic oxide pellets typical of LWRs. The ratio of target to ZrH1.6 containing fuel pins is a pproximately five to one (Fi gure 3-3). Therefore, for a typical hexagonal fuel assembly, 226 of the 271 pins contain target and 45 contain ZrH1.6 slugs. Figure 3-3. Representation of the axial target pi n-lattice showing the orie ntation of targets (red) and zirconium hydride pins (green)1 The accompanying fuel management strategy allo ws the targets and driver fuel to be discharged and recycled together in the same pyroprocessor ba tch. After pyroprocessing, the mixed target and driver mass str eams supplies the fabrication of fr esh driver fuel. Fresh targets are fabricated from the americium and curium co mponent of SNF. This americium and curium stream is provided by the UREX+3 separations process (Table 18) for LWR fuel. The UREX+3 neptunium and plutonium stream provides the external fissile feed to the driver fuel as well as the targets. The neptunium is kept with the plutonium by th e UREX+3 process for the purpose of enhancing proliferation resist ance. Some plutonium is supplied to the targets to minimize the power swing as a result of Pu-238 and Pu-239 gene ration during the irradiation. The reasons and effectiveness of this approach are discussed in Chapter 6. 1 The color pattern indicates the five-to-one relationship between targets and moderating rods. However, the actual number of pins in the picture is 210 targets and 61 zirconium hydride rods which is 77% and 23% of the total 271 pins respectively. ZrH1.6 Moderating Rod Transmutation Target Rod

PAGE 101

101 As depicted in Figure 3-4, the SNF americium an d curium is irradiated only once in targets before it rejoins the neptunium and plutonium in the driver fuel. The fuel cycle scenario in Figure 3-4 (and also in Fi gure 1-8) is used as the model fo r the out-of-core fuel management operations performed by the REBUS code (Figure 21). The combined fuel cycle represented in Figure 1-9 is performed by two separate REBUS calculations: one for the ABR and one for the AHFTR. Figure 3-4. AHFTR fuel cycle scenario s howing connections between partitioning and transmutation technology2 2 Fuel cycle calculations performed for the LW R-through-UREX+ portion of the fuel cycle were performed by TRITON. Equilibrium fuel cycle calculations performed for the boxed portion of the flow chart were performed by REBUS. A separate REBUS calculation is performed to model the equilibrium cycle of the ABR. (See Figure 1-1 and Figure 1-9). Reprocessing Losses ABR Fuel LWR SNF UREX+ Separation and Fabrication Plant UOX SNF AHFTR Spent Fast Reactor Fuel Pyroprocessing, Blending & Driver Fuel Casting Fresh Fast Reactor Fuel Assembly AHFTR spent fuel Light Water Reactor Np+Pu Np+Pu & RGU Make-up Material for Driver Fuel Np+Pu & Am+Cm & RGU Target Material Target Casting and Fuel Assembly Fabrication

PAGE 102

102 Therefore, sufficient americium must be destroyed in the targets such that the unburned americium is less than 5 w/o of the HM in th e driver. The 5 w/o limit for MA driver fuel concentrations is selected as a guideline for en suring acceptable reactivity ki netics features of the fast reactor. This limiting MA concentration also ensures that the expected irradiation performance of the driver fuel does not significan tly deviate from the current experience base on metallic SFR fuels. The decision to use 5 w/o of MA per total HM is based on results obtained by the CAPRA program conducted in Europe to evaluate the EFR [ 16,17]. Sim ilar to the ABR, the AHFTR has no radial bl ankets and uses three enrichment zones to flatten the radial power pr ofile across the core [ 7]. For the purpose of simplification of nom enclature, the term enrichment zone is dr opped and only the terms: inner core, middle core and outer core are used instead. The driv er fuel composition is a TRU/U/10Zr (by weight) alloy. The ratio of middle and outer core enrichments to that of the inner core is: 1.25 and 1.50 respectively. The driver fuel composition de finition is the same as the homogeneous ABR reference with the exception of the external feed being UREX+3 Np+Pu. The external feed for the reference metallic fueled ABR design groups all of TRU elements together using the UREX+1a process at the SNF aque ous reprocessing facility. The target fuel fixed composition is 10N p+Pu/10Am+Cm+Bk+Cf/40U/40Zr by weight. Some of the UREX+3 separated Np+Pu stream (i.e external feed) is diverted from the active driver fuel supply to the target s. None of the pyroprocessed pl utonium is used to refuel the targets. This is mainly because the current pyroprocessing technology does not readily allow for elemental separation. The choice to use pyropr ocessing allows the hi gher mass actinides of curium and californium that were generated during the target i rradiation to be diluted over a larger fuel inventory. This dilution is intended to reduce the in tensity of radiation fields and

PAGE 103

103 hence shielding requirements to fuel fabrication workers dealing with recycled material. The 10 w/o Am+Cm+Bk+Cf concentration in targets was selected to retain the fuel performance characteristics observed in the AFC-1 irradiatio n experiments which were performed at INL. The AFC-1 tests and the reasons for proposing this specific combin ation of Am+Cm+Bk+Cf with Np+Pu, U and Zr is discu ssed in detail in Chapter 6. Transmutation Target Physics As mentioned in the introduction, the reposit ory space benefit stems primarily from the removal of americium from the fuel cycle. This is because the repositorys waste emplacement drift spacing is limited by the maximum rock temperature at the midpoint between them. This rock temperature is principally a function of the decay heat produced by Am-241 in the SNF [ 13]. Because of its long half -life, radiotox icity, high solu bility and low sorption in Yucca Mountain tuffs, Np-237 is the principal environm ental concern to the biosphere, if water does come into contact with the SNF. Because, Np-237 is the alpha decay product of Am-241, americium destruction in transmutation targets also minimizes the Np-237 accumulation in the repository. Am-241 and Np-237 have an even neut ron number and thus the binding energy contribution of an absorbed neutron is not suffi cient to overcome the critical energy required for fission. In fact, the addition of a neutron to an odd-neutron nucleus (fissile isotope) to form an even-neutron compound nucleus gives a binding energy change that is about one MeV greater than for changing an even-neutron nucleus in to an odd-neutron compound nucleus. This explains the fission threshold at one MeV for the long lived MAs, Np-237 and Am-241. A neutron capture in Np-237 generates the fissile Np-238 nucleus. However, Np-238 quickly beta decays into Pu-238 with a short 2.117 day halflife. A neutron capture in Am-241 produces the fissile Am-242,242m isotopes. The yield fraction to th e ground state is estimated to be

PAGE 104

104 approximately 85% in the fast spectrum [ 10]. This Am-242 ground stat e beta decays into Cm 242 with a branching ratio of 83%. The other 17 % of Am-242 electron captures to become Pu242. Cm-242 decays into Pu-238 with a 163 day half-life. Because of the fission threshold, the Am241 fission cross section below one MeV and above the resonance range (where the SFR neutron spectrum is ve ry small) is two orders of magnitude less than for fission above one MeV (Figure 3-5). However, the capture cross section is almost as high as the Pu-239 fission cross section in the same energy range. 1E-04 1E-02 1E+00 1E+02 1E+04 1E+061.E-071.E-051.E-031.E-011.E+01 Energy (MeV)Flux (1E-10 cm^-2^-1/lethargy) X-Section (barn) Am-241 Capture Am-241 Fission Pu-239 Fission Neutron Flux Figure 3-5. ENDF-VI americium and plutonium cross section pl ots versus a metal fuel SFR neutron spectrum Therefore, the AHFTR axial target neutron spectrum is mode rated slightly in order to reduce the neutron energy to just below one MeV. This increases the neutron capture in americium relative to plutonium fission in the targ ets. The effect can be seen by evaluating the ratio of total absorption in americium over total ab sorption in Pu-239. This increase can be seen by observing a capture and a fission based fuel utilization factor for th e driver and the targets. In Table 3-3, a transmutation utilization factor is used to quantify the reaction probability of capture

PAGE 105

105 in a given isotope divided by any absorption in the region of space being studied. The transmutation utilization factor is defined in the equation below. i f ii c ii r ii N N N nutilizatio ion transmutat (3-3) Where : i represents a given isotope in targets or driver fuel and r represents capture or fission Similarly, the fission utilization factor is defi ned as the reaction probability of fission in a given isotope divided by any absorption in the region of the core being studied. Also supplied in Table 3-4 is a single group cross section ratio. The cross section ratio shows the spectral effect on the microscopic cross sections alone w ithout being weighted by number density. Table 3-3. Transmutation utilization factor (MCNP*) Driver Target CAP/ABS FISS/ABS CAP/ABS FISS/ABS Am-241 1.59% 0.34% 15.64% 0.93% Pu-239 6.02% 30.36% 8.24% 15.43% *Single group cross sections for Table 3-3 and Table 3-4 were performed by MCNP benchmarks of REBUS calculations using the final AHFTR core design. Table 3-4. Single group cr oss section ratio (MCNP) Am-241 CAP over Pu-239 FISS Am-241 ABS over Pu-239 ABS Driver 0.77 0.78 Target 1.33 0.91 It is important to note that the ratio of the Am-241 capture to the Pu-239 fission cross section is greater in the target than the driver fuel because of moderation. It is also apparent that the total absorption cross section ratio of Am-241 over Pu-239 is greater in the targets than the driver fuel because of this in creased neutron capture. This effect explains how the capture utilization factor for Am-241 is nearly equivale nt to the fission utiliz ation in Pu-239 in the targets. The slight target moderation results in a relatively epithermal flux compared to that for the active driver regions as shown in Figure 3-6. It is important to note, that though the total flux in the targets is less than in the active driver, much of it is being depresse d by resonance absorption

PAGE 106

106 in the epithermal range. Much of this resonanc e absorption is in Pu-239 and U-238. However, resonance absorption in Pu-239 l eading to fission serves to ge nerate more neutrons. These neutrons have a relatively short mean-free-path because of the soft er spectrum. Therefore, they remain locally within the targets neutron popul ation. Moreover, resonance absorption in U-238 serves to generate more Pu-239. The comb ination of epithermal spectrum and plutonium breeding by both Am-241 and U-238 creates enough pl utonium to replace the initial plutonium loaded in the targets at BOEC. 1E+10 1E+11 1E+12 1E+13 1E+14 1E+15 1E+16 1E-051E-041E-031E-021E-011E+001E+01Energy (MeV)Neutron Flux (cm^-2*s^-1)/lethargy Targets Inner Core Middle Core Outer Core Figure 3-6. Target flux spectrum compared with the inner, middle and ou ter core for a 101.6 cm tall AHFTR with p/d of 1.1 (REBUS) This breeding gives a fCR in the targets close to one but a tCR more near to 0.75. As can be seen by Figure 3-7, the total amount of Pu-239 stays relatively constant. However, the total plutonium content increases slightly, main ly from Pu-238,242 production from Am-241 transmutation. However, the total transuranic mass is being reduced as americium is being transmuted. Parametric Study Because the transmuted plutonium remains in the fuel cycle via the pyroprocessor, the amount of space reserved in the r eactor for targets plays a signifi cant role in the demand for the

PAGE 107

107 UREX+3 supplied Np+Pu. From a geometry st andpoint, the ratio of ta rget volume to core volume is dictated by the active core axial height. From a physics standpoint, the amount of Pu239 breeding in the active core is a function of co re flattening. Since, breeding excess Pu-239 in the active core is not a primary objective of th e AHFTR, it is desirable to minimize the active cores CR. Conversely, since br eeding even-plutonium is an indicator of americium destruction, it is not obligatory to minimize the c onversion ratio in the target region. 0.0E+00 5.0E+00 1.0E+01 1.5E+01 2.0E+01 2.5E+01 3.0E+01 3.5E+01 0.000.921.852.773.694.615.54Effective Full Power YearSingle Batch Mass (Kg) CM245 CM244 CM242 AM243 AM242m AM241 PU242 PU241 PU240 PU239 PU238 NP237 Figure 3-7. Isotope masses as a function irradia tion time within the target region a 101.6 cm tall AHFTR with p/d of 1.1. Each color bar represents an isotopes mass (REBUS) Effects of Pin Diameter and Core Height Because of the vast range of options presented that affect the reactors overall physics and transmutation performance, the following assumpti ons are made for the purpose of confining the parametric analysis of pin and core dimensions. Core design from Figure 3-1. Core height varied from 101.6 cm to 51.6 cm (Table 3-1) Fuel pin diameter varied from 0.808 cm to 0.464 cm (Table 3-2) Number of fuel pins per assembly (both dr iver and target region): 271 (Table 3-2) Peak fuel burnup constrained to 18 at. % (Figure 2-1) Ratio of zirconium hydride to target rods is: 1/5

PAGE 108

108 The rate of americium destruction is a function of the target spectrum, active core axial leakage and target rod volume. Increasing the di ameter of the target rod also increases the volume of americium charged to the AHFTR per cycle. If the reacto r cycle is shortened, then the time rate of americium fed to the core is increa sed because the target ma ss loaded per cycle is constant due to the fact that the target rod co mposition has been fixed. This does not, however, indicate the spectrum eff ect of the zirconium hydride content or flux intensity in the targets. Because the target irradiation time is not held co nstant, simply comparing the ratio of charged to discharged target americium is not an adequate indication of the destruction efficiency. Instead, a half-life is defined for Am-241 transmutation which quantifies the magn itude of the capture reaction rate. Because the rate of Am-241 production rate via beta decay from Pu-241 is relatively slower than its destruction rate via ne utron capture, the transmutation half-life can be approximated using a first order differential equa tion. Therefore, Am-241 destruction can be characterized with a purely exponential behavior just as with radioactive decay. This exponential behavior is given by the Am-241 production and decay rate equation. Table 3-5 shows the irradiation time required to reduce the americium mass by half. capture Am ion transmutat capture Am capture Am decay PuTN dt dN NN dt dN241 2/1 241 241 241)2ln( (3-4) Table 3-5. Transmutation ha lf-lives for a preliminary AHFTR design (Years) (REBUS) Pitch/Diam. p/d=1.1 1.176 1.293 1.357 101.6 cm 2.49 2.31 2.04 1.87 70.6 cm 2.73 2.54 2.24 2.06 50.6 cm 3.23 2.99 2.64 2.44 Notice the half-life decreases for decreasing pi n diameter. This can be explained by the higher sodium fraction in the active core for de creasing fuel pin diameter. Increased leakage invests more neutrons in the targ ets. The decreasing half-lives w ith height are also related to axial leakage. However, the expected result is a decrease in half-life due to an increase in

PAGE 109

109 leakage with the core height reduc tion. This contradiction in the half-life trend with core height can be explained by more evenly distributed neut ron utilization in the active driver core. Because of the enhanced axial leakage, the dr iver fuel radial power profile develops a depressed region in the inner co re with decreasing height. The increase in the americium transmutation half-life with decreasing core heig ht in Table 3-5 is caused by a reduction in the intensity of neutrons leavi ng the inner core region. It is important to note the c onformity of the three power profiles plotted in Figure 3-8 in the area of the outer core. This behavior is indicative of the cosine shape of the radial flux profile as neutrons leave the core. The increas ed axial leakage shown in Figure 3-9 affects the curvature of the radial power profile in the inner core region more so than in the outer region. This is because the axial flux gradient is decrease d more in the inner core than the outer core. The flux in the outer core regions are already suppressed everywhe re by radial leakage. This explains the conformity of the three different ra dial power profiles in Figure 3-8 at the outer edge. Because the power density becomes more evenly distributed axially and radially, the power density on the active core top surface becomes reduced. A flatter radial power profile also reduces the radial power peaki ng. The reduction in power peaking in the inner core makes possible raising the reactor power, which was reduced in the parametric study for decreasing core height Even though the decrease in core height increases axial buckling and hence axial leakage, the larger buckling in the radial direction relative to the axial direction causes flux in the inner core to be reduced which gives the flat radial power profile. This reduction of the inner core flux is also a result of the reduction in power for each reduction in core height which was done for the purpose of the parametric study.

PAGE 110

110 Because of the depressed flux in the inner core the power of the final flattened AHFTR core geometry is increased to give a more realistic power density. 50 100 150 200 250 300 350 400 1234567Fuel Assembly Row NumberPower Density (MW/m^3) 101.6 cm Tall 71.6 cm Tall 51.6 cm Tall Figure 3-8. Active core radial power density profile for a preliminary AHFTR with active core height: 101.6, 71.6 and 51.6 cm (p/d=1.1) (REBUS) 70 120 170 220 270 320 370 420 470 20%30%40%50%60%70%80%90%100%Percent Height from Bottom of CorePower Density (MW/m^3) 101.6 cm Tall 71.6 cm Tall 51.6 cm Tall Figure 3-9. Inner core axial power density pr ofile for a preliminary AHFTR with active core height: 101.6, 71.6 and 51.6 cm (p/d=1.1) (REBUS) After increasing reactor power back to the practical limit, the increased power density in the targets will translate into a higher flux de nsity for transmutation. Therefore, the transmutation half-life for the targets with a mo re realistic thermal pow er rating will be higher

PAGE 111

111 than for the parametric study. A following section describes two possible AHFTR core geometries with two different active co re heights and a 1000 MW power level. Effects of Moderating Pins The affect of moderation on th e transmutation efficiency of the target region was also evaluated. Using a pin pitch-dodiameter ratio of 1.176 from Table 3-2 and a core height of 101.6 cm, the number of zirconium hydride rods per target rods was varied. The ratio of zirconium hydride rods per target rods was varied from: 1/5, 1/6, 1/7, 1/8 and 1/9. All of the other design variables used in the pr evious section were kept the same: Core design from Figure 3-1. Core height of 101.6 cm, Fuel pin di ameter of 0.755 cm (Table 3-2) Number of fuel pins per assembly (both dr iver and target region): 271 (Table 3-2) Peak burnup constrained to 18 at. % (Figure 2-1) Vary the number of zirconium hydride to target rods from 1/5 to 1/10 It was found that the affect of moderator did have an affect on tran smutation efficiency. Table 3-6 gives the transmutati on half-life of Am-241 for each moderator-to-target rod ratio. Table 3-6. Transmutation halflife of Am-241 for varying number of moderating ro ds per target rods in the target region for the preliminary AHFTR design (REBUS) Description* Transmutation Half-life (Years) One ZrH1.6 per Nine Target Rods 3.903 One ZrH1.6 per Eight Target Rods 3.759 One ZrH1.6 per Seven Target Rods 3.605 One ZrH1.6 per Six Target Rods 3.411 One ZrH1.6 per Five Target Rods 3.193 *Total number of pins is held constant by the fuel assembly design It is important to note that the transmutation half-life decreases for increasing moderator. This means that the addition of hydrogen to the targ et region does contribute a spectrum effect to the transmutation efficiency. Figure 3-10 and Figure 3-11 shows the average neutron spectrum in the target region for each moderator-to-target rod ratio.

PAGE 112

112 1.0E+11 5.0E+13 1.0E+14 1.5E+14 2.0E+14 2.5E+14 3.0E+14 3.5E+14 4.0E+14 1E-061E-051E-041E-031E-021E-011E+001E+01Energy (MeV)Flux (cm^-2 s^-1)/lethargy One ZrH1.6 Per Five Targets One ZrH1.6 Per Six Targets One ZrH1.6 Per Seven Targets One ZrH1.6 Per Eight Targets One ZrH1.6 Per Nine Targets Figure 3-10. Average neutron flux in the targ et region of a preliminary AHFTR design for a varying number of moderating rods per ta rget rods (lin.-log. scale) (REBUS) 1E+11 1E+12 1E+13 1E+14 1E+15 1E-061E-051E-041E-031E-021E-011E+001E+01Energy (MeV)Flux (cm^-2 s^-1)/lethargy One ZrH1.6 Per Five Targets One ZrH1.6 Per Six Targets One ZrH1.6 Per Seven Targets One ZrH1.6 Per Eight Targets One ZrH1.6 Per Nine Targets Figure 3-11. Average neutron flux in the targ et region of a preliminary AHFTR design for a varying number of moderating rods per ta rget rods (log.-log. scale) (REBUS) Notice that the fast flux decreases for increasi ng concentration of moderating rods. This is a sign that the fast component of the flux is be ing reduced by the moderation. However, because the y-axis of Figure 3-10 is linear, it is difficult to see from the plot whet her or not there is an

PAGE 113

113 increase in the neutron flux at lower energies. Figure 3-11 shows the same data that was plotted in Figure 3-10 but on a logarithmic scale. Notice that the thermal-epithermal component of the spectrum increases for increasing number of moderator per target rods. The in crease in thermal-epithermal flux for increasing zirconium hydride rods explains the improvement in transmutation half-life. As the spectrum softens, more of the flux is placed in the reso lved-unresolved resonance range below one MeV. Recalling the cross section plot in Figure 3-5, th e Am-241 capture cross section in this range has the same overall magnitude of that of the Pu-239 fission cross section. Because, the Am-241 loading and Pu-239 loading are approximately equa l, the Am-241 capture reaction rate provides a degree of energy shielding effect over the fi ssion rate of Pu-239 (and the other plutonium isotopes). The small amount of energy shielding gained by Am-241 allows neutrons invested in the target region by leakage from the active core to be invested in neutron captures in Am-241 as opposed to fissions in the Pu-239 which enables high transmutation efficiency. The transmutation efficiency gained by the moderation can be visualized by tracking the change in the relative Am-241 content of the fuel normalized to the initia l Am-241 loading of the fresh fuel, which is shown in Figure 3-12. In addition to the destruction of Am-241, the gain in transmutation efficiency by increa sing neutron captures (i.e., transm utation) can be visualized by tracking the change in the rela tive Pu-238 content of the fuel normalize to the initial Pu-238 loading in the fresh fuel, wh ich is shown in Figure 3-13. Because the case with one zirconium hydride r od per five target rods demonstrated the highest transmutation effi ciency over all other cases, this combination was selected for the final down selection of the AHFTR design. A decision was made to limit the concentration of moderator rods to one per five targets due to a concern that too much moderation would cause

PAGE 114

114 power peaking effects in the targ et region and near the interface w ith the driver fuel. Increasing the thermalizing effect decreases the neutron m ean-free-path. The concern over power peaking arises from the possibility that discontinuities in the neutron flux between regions with dissimilar neutron spectrums (due to isotope depletion) could emerge if the neutron mean-free-path becomes too short. These discontinuities woul d not only lead to unacceptable power peaking, but would also invalidate the accura cy of the fast reactor codes, MC2-2 and DIF3D, for the target analysis. These codes assume minimal flux disc ontinuity from region-to-region in order to homogenize the fuel over large regions of the core. 0% 20% 40% 60% 80% 100% 120% 00.511.522.533.544.5Effective Full Power YearPercent Change in Am-241 from BOEC (Mass(t)/Mass(t=0) One ZrH1.6 per Five Targets One ZrH1.6 per Six Targets One ZrH1.6 per Seven Targets One ZrH1.6 per Eight Targets One ZrH1.6 per Nine Targets Figure 3-12. Percent of the in itial Am-241 mass remaining in th e target rod as a function of irradiation time for varying number of moderating rods in a preliminary AHFTR (REBUS) There is an additional feasibil ity limitation on the practical number of pins that can be loaded with zirconium hydride as opposed to targ ets. The addition of mo derator rods reduces the number of target rods available for loading MAs. Therefore, to burn an equal amount of mass in the targets, the MA concentration in the target slug must increase for decreasing number of target pins. As will be discussed in Chapter 6, for metal alloy fuels there is a practical limit on the

PAGE 115

115 concentration of MAs that can feasibly be loaded into the fuel which are related to the volatility of americium in its melted form. 0% 100% 200% 300% 400% 500% 600% 700% 800%00.511.522.533.544.5Effective Full Power YearPercent Increase in Pu-238 from BOEC (Mass(t)/Mass(t=0) One ZrH1.6 per Five Targets One ZrH1.6 per Six Targets One ZrH1.6 per Seven Targets One ZrH1.6 per Eight Targets One ZrH1.6 per Nine Targets Figure 3-13. Percent of the initial Pu-238 mass created in the target rod as a function of irradiation time for varying number of moderating rods in a preliminary AHFTR (REBUS) Tall and Flattened Axial Heterogeneous Core Designs Using the reference ABR, with pin diamet ers from Table 3-2, as a standard for comparison, the AHFTR core height and pin diam eter were varied. The reactor performance traits considered were the excess reactivity, tota l core void worth and Doppler coefficient. The largest diameter fuel pin design (p /d=1.1) from Table 3-2 is selected (for this section) to give a Doppler coefficient with increased negativity than the reference case. This is expected because the volume of fuel in the active core is increased and the percent of that fuel being uranium is also increased. After down-selection, the final active core geometry is 91.6 cm with a 1.1 pin pitch-to-diameter ratio. The 10 cm reduction in active core height was found to give a total core void worth that was slightly less than the homogen eous reference ABR with the 1.1 pin pitch-todiameter ratio (Table 3-2). Though the active driver core height is 10 cm less than the reference ABR, the total core height is 10 cm taller due to the addition of the 20 cm height of the targets.

PAGE 116

116 This core design, as well as a much flatter co re design, is compared to the reference ABR in Table 3-7. It should be noted that these comparisons are made for a pin pitch-to-diameter ratio of 1.1. The flat AHFTR core design, discussed next, is given in Figure 3-14. Table 3-7. Core design summary for the reference ABR with tall and flat AHFTR (REBUS) ABR (Ref.) Tall AHFTR Flat AHFTR Total Core Height 101.6 110.6 91.6 Active Driver Height 101.6 91.6 71.6 Total Core Volume (m3) 3.3 3.6 4.0 Active Core Volume (m3) 3.3 3.0 3.1 Rows of Driver Fuel 7 7 8 Pitch-to-diameter ratio 1.1 1.1 1.1 Inner Core Enrichment 14.92%16.07%18.65% Enrichment Split (IC/MC/OC) 1.0/1. 25/1.501.0/1.25/1.501.0/1.12/1.25 Cycles per Enrich. Zone (IC/MC/OC) 6/6/7 6/6/7 6/6/7 Cycle Length (EFPD) 322.86319.90350.58 For the flattened version of the AHFTR, a 71.6 cm height is evaluated to observe a much larger target-to-driver volume ratio. An additional row of outer core drivers is added to this flat core so that the power density is made roughly comparable to the tall AHFTR (Figure 3-14). Radial and Axial Power Profiles Because of the inherently flatter power distri bution, the gradient of enrichment splitting for the flat AHFTR can be decreased from that used for the ABR. Therefore, the flat cores middle and outer zones are enriched to 1.12 and 1.25 times that for the inner core respectively. This tailored enrichment splitting gives the fl at core a much more evenly radial power distribution than the tall core. The axial leakage works in paralle l with the enrich ment splitting to create an almost completely flat radial power prof ile across the inner and middle enrichment zone. Figure 3-15 shows the radial power distri bution for six axial slices through the tall and flat versions of the AHFTR. In Figure 3-15 and Figure 3-16, the power density of the flatter core geometry is noticeably greater than that of th e tall core geometry. This greater power density is linked to the axial

PAGE 117

117 leakage coming from the active core. As was seen in the transmutation half-life discussion in the parametric study, flattening the core reduces the ra dial curvature, or buckling, of the power (and flux) profile. Figure 3-14. Preliminary Flat AHFTR design with eight rows of fuel instead of seven Ultimate Shutdown Rod Assembly Primary Control Rod Assembly Inner Enrichment Zone Driver Middle Enrichment Zone Driver Outer Enrichment Zone Driver Radial Reflector Assembly or Axial Reflector Region of Driver Assembly Shield Assembly Axial Target Portion of Driver Assembly Fuel Rod Gas Plenum AHFTR Fuel Assembly Configuration Handling Socket Coolant Inlet Nozzle

PAGE 118

118 0.0E+00 5.0E+01 1.0E+02 1.5E+02 2.0E+02 2.5E+02 3.0E+02 3.5E+02 4.0E+02 4.5E+02 5.0E+02 1234567Radial Fuel Assembly Row NumberPower Density (MW/m^3) 20% 40% 60% 80% 100% 120%A 0.0E+00 5.0E+01 1.0E+02 1.5E+02 2.0E+02 2.5E+02 3.0E+02 3.5E+02 4.0E+02 4.5E+02 12345678Radial Fuel Assembly Row NumberPower Density (MW/m^3) 20% 40% 60% 80% 100% 120%B Figure 3-15. Radial power density profile for six axial slices through the core: Axial height is represented as a percentage of the full core he ight. (A = tall, B= flat core designs) (REBUS) As can be seen in Figure 3-15, the curvature of the radial power profile is less for the flat core than it is for the tall core. Remembering ba ck to the discussion in Chapter 2 on the critical buckling of a hypothetical bare cylindrical SFR, th e sum of radial and axial geometric buckling must be equal to the material buckling. Ther efore for approximately equal materials buckling between tall and flat geometries, a decrease in radial buckling requires an increase in axial

PAGE 119

119 buckling. Hence, the axial buckling of the flat core is greater than that of the tall core. It is because of this increase in axial buckling that the flat core has more axial leakage than the tall core. The greater axial leakage gives the flat core a greater po wer density in the target region than in the tall core, which is shown in Figure 3-16. 0.0E+00 5.0E+01 1.0E+02 1.5E+02 2.0E+02 2.5E+02 3.0E+02 3.5E+02 4.0E+02 4.5E+02 5.0E+02 20%40%60%80%100%120%Axial Level (20%=bottom,100%=top,120%=target)Power Density (MW/m^3) ROW 1 ROW 2 ROW 3 ROW 4 ROW 5 ROW 6 ROW 7A 0.0E+00 5.0E+01 1.0E+02 1.5E+02 2.0E+02 2.5E+02 3.0E+02 3.5E+02 4.0E+02 4.5E+02 20%40%60%80%100%120%Axial Level (20%=bottom,100%=top,120%=target)Power Density (MW/m^3) ROW 1 ROW 2 ROW 3 ROW 4 ROW 5 ROW 6 ROW 7 ROW 8B Figure 3-16. Axial power density profile for each ro w of fuel: Axial height is represented as a percentage of the full core height. (A = tall, B= flat core designs) (REBUS)

PAGE 120

120 The increased power density in the target re gion also provides for a reduced transmutation half-life of 2.22 years. This transmutation half -life corresponds to a transmutation efficiency which is comparable to the small pin diameter (p/d=1.357) from Table 3-5. This small pin diameter was used by Hoffman et al to attain a tCR equal to 0.25. Therefore, the flat AHFTR core design achieves the MA destruc tion efficiency of a core with a thin fuel pin but does this using a larger pin diameter. The larger pin diam eter used for the flat version of the AHFTR is more representative of the ABR with a tCR=1.0. This higher efficiency combined with an incr eased volume of the target region (physically more fuel assemblies) gives a higher destruction rate of Am+C m+Bk+Cf when compared with the tall core. This higher Am+Cm+Bk+Cf destru ction rate equates into a higher plutonium transmutation breeding rate. This increases the amount of Am+Cm+Bk+Cf and reduces the amount of Np+Pu drawn from the UREX+3 plant. Hence, the size of the surplus Np+Pu feed (outside the boxed off portion of Figure 3-4) decreases. In fact for both the tall and flat designs, the external Np+Pu feed for the active core is reduced to zero for the equilibrium fuel cycle. Consequently, for both the tall and flat core designs, the UREX+3 plant only needs enough Np+Pu and Am+Cm+Bk+Cf to produce fresh targets. For either case, the radial power profile falls o ff sharply in the outer tw o rows of fuel. This is attributed to the dominating radial leakage eff ect on the flux gradient in the outer core. It is important to note, the volume ratio of these outer two rows to the rest of the fuel is 0.5 and 0.44 for the flat and tall cores respectively. So the volume of the target material located in the low flux of these outer two rows decreases as the core radius is increased. Hence, the flat AHFTR has the smallest fraction of targets located above the low flux, low power driver fuel. This gives

PAGE 121

121 the flat AHFTR the advantage of having the high est achievable transmutation rate over a greater share of the targets than the tall AHFTR. In addition to the geometrical improvement in target exposure in the radial direction, the axial volume ratio of target to driver fuel incr eases with core flattening. This is because the target volume in the numerator of this ratio is fi xed for a given radius by its 20 cm height, but the denominator is decreasing as the core height is decreasing. core active core active core core driver etth cm hR cm R V V_ 2 2 arg20 20 (3-5) Where : Vtarget is the approximate volume of the targets in the core. Vdriver is the approximate volume of driver fuel in the core. Rcore is the core radius. hactive_core is the height of the active core. The height of the ta rgets is held constant at 20 cm. Therefore, the overall MA charge rate per cy cle increases as the active core height is decreased. The combined effect of the incr eased axial leakage from the active core, the increased share of fuel having a higher power an d an increase in the MA charge rate per cycle makes the flat AHFTR design the most attractive transmutation system because of its physical ability to consume MAs. Table 3-8 shows the fuel cycle parameters for the homogeneous reference compared to the tall and flat AHFTR designs. As expected the flat AHFTR has a greater amount of Am+Cm+Bk+Cf being destroyed in the targets. This can be observed by examining the americium transmutation half-life for the flat ve rsus the tall core design. In addition to a decreased transmutation half-life (increased cap ture reaction rate), the overall Am+Cm+Bk+Cf target destruction rate is also increased. The en hanced destruction rate is the combined effect of transmutation half-life and the target assembly charge rate.

PAGE 122

122 Despite having a larger destruction rate in the targets, the flat core has a smaller overall Am+Cm+Bk+Cf consumption rate (Table 3-8). This is because the flat core has a longer cycle length (i.e., refueling interval) than the tall core wh ich is also equivalent to a smaller fresh target charge rate. Though the tall AHFTR has a more efficient target (s horter transmutation half-life), these targets discharge a larger Am+Cm+Bk+Cf vol ume to the pyroprocessor. The tall core has a less efficient target but the shorter cycle le ngth increases the rate that Am+Cm+Bk+Cf is charged to the targets and consequently the rate that un-transmuted MA mass is received by the pyroprocessor. The difference in target destru ction rate and total core consumption rate illuminates the importance of distinguishing between the transmutation rate of a given isotope and the actual charge rate of that isotope. The is otope destruction or reacti on rate is a function of neutron flux and energy in any gi ven region of the core, whereas the charge or consumption rate is more influenced by the logistics of the fuel cycle operating at steady state. Table 3-8. Fuel cycle comparison for the re ference ABR with tall and flat AHFTR (REBUS) ABR (Ref.) Tall AHFTR Flat AHFTR Active Driver Height 101.691.6 71.6 Maximum MA in Driver Fuel OC Am/HM 0.00%1.72% 1.79% OC MA/HM 1.06%3.20% 3.40% TRU Externally Supplied Feed Rate Am+Cm+Bk+Cf Feed* Rate (kg/EFPY) 2.79E+003.63E+01 3.31E+01 Np+Pu Feed Rate* (kg/EFPY) 4.84E+013.60E+01 4.31E+01 HM Feed Rate (kg/EFPY) 3.75E+023.80E+02 3.75E+02 Transmutation Half-Life IC Target Am Half-Life (Yr) --2.49 2.22 *See Figure 3-4 for a pictorial representati on of the source of the external feed. In both cases, the amount of MAs diluted in the driver fuel is comparable to that of the reference ABR and is less than the 5% limit set by the CAPRA recommendation as discussed earlier. Note that the MA concentration in th e outer core is higher than the other two zones because the transuranic enrichment is the highest in that region due to enrichment splitting

PAGE 123

123 (Table 3-8). The Am+Cm+Bk+Cf in the driver fuel is effectively burned in the active core region and does not accumulate in the AHFTR fuel cycle. Reactivity Feedbacks The Doppler coefficient and total core void worth are compared in Table 3-9. These performance parameters are calculated for the cores at BOEC, using the calculation methods discussed in Chapter 2. The Doppler coefficients for these cores are fair ly comparable with the values reported by Hoffman et al [ 7]. It should be noted that th e ABR reference considered in these scoping calculations is not the exact sam e core desi gn proposed by Hoffman et al [ 7]. The core layout, cycle length, num ber of batches, etc. were not cons erved between this analysis and that of Hoffmans. The core layout given in Figure 1-5 is the one proposed by Hoffman for a tCR of 0.5. The core design that Hoffman proposed for the larger fuel pins (p/d=1.1) is slightly different (Table 3-9). The tCR=0.5 core layout was considered a middle-of-the-road design that could be used for comparison purposes in th e scoping calculations in this chapter. The reference ABR design and the flat version of the AHFTR both have a slightly negative Doppler coefficient. However, the tall versi on of the AHFTR has a s lightly positive Doppler coefficient. The positive void coefficient of the tall core is due to the higher concentration of MAs in the driver fuel than the reference ABR and an absence of core flattening. The affect of fuel temperature increase, for the tall core causes re sonance broadening of the U-238, TRU and fission products. This resonan ce absorption can cause the neutron spectrum to harden which causes an increase in multiplication. The positive worth of this multiplication can be greater than the negative feedback of the U-238 resonance abso rption if the core has inadequate leakage to remove this extra reactivity. He nce, a degree of leakage is beneficial for Doppler feedback to ensure that the inner regions of the core are not overly reflected An increase in axial leakage

PAGE 124

124 also decreases the positive void worth, because of an improvement in axial streaming compared to the other two more symmetric cores. As discussed in Chapter 2, the conversion ratio is also sensitive to core flattening. As leakage is enhanced, the concentration of fiss ile material must increase. Hence, the TRU increases. The excess reactivity may be viewed in the same way. As leakage increases, the TRU enrichment required to achieve the maximum fuel burnup (taken to be 18 at. % in these scoping calculations) also increases. This can be seen in Table 3-9. For the homogeneous case, this would normally reduce the cycle length because a reduction in U-238 capture also reduces the production of Pu-239 which is needed to achieve the fuel burnup. However, the flat AHFTR design has the longest cycle length. This is attributed to the reac tivity breeding effect of the axial target region. Since, the overall core size is larg er for the flat versus th e tall or ABR cases, there is actually less neutron escape when the targets are factored into the consideration. Therefore, neutrons are efficiently i nvested in creating fissile material in the targets as the initial plutonium charge in the driver burns out. Also shown in Table 3-9 is the effect of to tal core size on the AHFTR dominance ratio. The dominance ratio gives an indication of the closeness of the first and second lambda-mode eigenvalues and is generate d by the DIF3D calculation [ 42]. A larger dominance ratio (nearer to 1.0) indicates less neu tronic coupling between regi ons of the core. Because the mean-free-path for fast neutrons is much larger than for ther mal neutrons, fuel regions in an SFR are normally well coupled. This is confirmed by the small domi nance ratios in Table 3-9. It is desirable to have a tightly coupled fast reactor from the standpoint of transient response Since, the effective delayed neutron fraction for fast reactors is significantly less than for thermal systems; it becomes advantageous to reduce the complexity of the types of reactiv ity feedbacks that are

PAGE 125

125 expected to occur in the fast reactor. The flat AHFTR has the largest total core volume compared to the other two designs and also has a moderated region (reduc ed mean-free-path). Nevertheless, the increase in dominance ratio from the reference ABR is not significant. Table 3-9. Physics comparison for the refe rence ABR with tall and flat AHFTR (REBUS) Hoffman [ 7] ABR (Ref.)* Tall AHFTR Flat AHFTR Active Driver Height 101.6 101.691.6 71.6 Pin Pitch-to-Diameter Ratio 1.1 1.1 1.1 1.1 Total Core Void Worth ($) 6.29 9.168.82 8.27 Doppler Coefficient (/K) -0.11 -0.130.06 -0.14 Excess Reactivity (%) -0.06 1.311.12 1.70 BOEC k-eff 1.0 1.013251.01312 1.01727 Cycle Length (EFPD) 370 322.86319.90 350.58 Max. Burnup (at.%) 11%18%18% 18% Dominance Ratio** --0.37980.3813 0.3958 Conversion Ratio (tCR) Core Conversion Ratio --0.890.87 0.84 Target Conversion Ratio ---0.91 0.90 Inner Core Conversion Ratio --1.010.97 0.90 Middle Core Conversion Ratio --0.850.82 0.81 Outer Core Conversion Ratio --0.770.75 0.77 *The performance characteristics for the refere nce ABR listed here are not that of the actual point design of the GNEP ABR design which is currently evolving and not final. See Chapter 2 for an explanation of kinetics calculations. **T he dominance ratio is not reported by Hoffman et al in the report, ANL-AFCI-177. Fuel Performance Indicators One issue identified in previous heterogeneous target studies, by Sanda et al, was that the transmutation of MAs into plutonium can create an undesirable power peaking effect [ 28]. This issue was most pronounced for target irradiations in high flux regions in the inner core or for long deep-burn irradiations in the outer core. This work sought to dim inish or eliminate any changes in the target thermal power throughout its life in the core (i.e ., from beginning-of-life (BOL) to end-of-life (EOL)). For axial targets th is is especially important because the targets share the same coolant channel flow orifice as the driver fuel. SFR fuel assemblies often are, by-and-la rge, designed with a metal shroud that encompasses the fuel pins. This shroud prevents cross-flow of sodium to adjacent fuel

PAGE 126

126 assemblies. Therefore, sodium coolant flow in the fuel assembly is controlled on an assembly basis using an orifice at the bottom of the fuel assembly near where it meets the bottom grid plate. The shroud prevents differences in pre ssure losses in each fuel assembly that would otherwise cause sodium coolant to move laterally in order to equalize the lateral pressure gradient. Therefore, despite a large radial power gradient (e ven for the AHFTR) compared to LWRs, the flow orifices are used to make the coolant outlet temp erature at the top of the core fairly constant. If there were to be a wide change in thermal power of the target region, the amount of flow could be sufficient to cool the target at BOL and inadequate fo r EOL. Also, undesirable power peaking in the driver fuel could result if the power sharing between driver and targets shifts significantly throughout the irradi ation. Power peaking and shifting was minimized by choosing a target composition with a small amount of starti ng plutonium and uranium to ensure a slow net destruction of Pu-239 over the cour se of the irradiation. Also the high importance of the Am241 capture cross section in the ep ithermal flux ensured that neutr ons could be equally absorbed in americium, as could be absorbed into Pu239 (Table 3-3 and Figure 2-11). This gives the americium the purpose of a burnable poison in the target region as well as a fertile source for breeding even-plutonium isotopes. The power produced by these even-plutonium isotopes are sufficiently suppressed by the cap ture reactions in Am-241. Table 3-10 evaluates the fuel performance indicators for the AHFTR. The peak Linear Heat Generation Rate (LHGR) in the targets was highest for the flat case. In all cases, the peak LHGR for the entire core occu rred in the driver fuel not the targets. This peak driver LHGR occurs in the innermost row of fuel at the active cores mid-plane. This is also true of the volume and exposure integrated fi ssion density for the target and driver fuel.

PAGE 127

127 Table 3-10. Fuel performance comparison for the reference ABR with tall and flat AHFTR ABR (Ref.) Tall AHFTR Flat AHFTR Active Driver Height 101.691.6 71.6 Linear Heat Generation Rate (kW/m) Peak Driver LHGR (kW/m) 38.136.9 35.5 Peak Target LHGR (kW/m) --20.1 25.3 Peak Fast Fluence: E>0.1 MeV (1E23 cm-2) Inner Core 5.725.66 5.53 Middle Core 5.585.49 5.61 Outer Core 5.054.95 5.88 Max Integral Fission Density: (f/cm3) Driver Fission Density (1E21 f/cm3) 6.967.67 7.09 Target Fission Density (1E21 f/cm3) --3.64 4.62 Average Discharge Burnup (MWD/kgiHM) 124.4127.3 134.1 Historically, the fission density has been used as an indicator of irradiation induced swelling and gas release in metal fuels. It a llows equal comparison of fuel performance when the amount of HM loading is not constant in th e comparison. This is the case for the AHFTR because the metal driver fuel and targets have diffe rent zirconium contents in their alloying. For all cases, the peak fission density is less in the ta rgets than it is in the fu el. Hence, the expected swelling and gas release, includ ing transmutation helium, should be similar to that of past experience with metal driver fuel for equivalent fission density. This assumes that the rate of void formation and interconnected porosity in th e target is roughly equivalent for the lower zirconium content driver alloy. The irradiati on induced interconnected porosity of fission gas bubbles coming together early in the irradiation was one of the significant developments of the Mark II driver fuel used at the EBR-II [ 60,61]. The effect slows the rate of swelling by allowing fission gas to escape to the plenum and increases th e time before cladding interaction. Because of the high LHGR and burnups associated with both metallic and oxide based SFR fuels, typically 60-80% of all fission ga s is released to the plenum [ 62]. This explains a plenum height of about 1.5 to two tim es the height of the co re in Figure 1-7, Figur e 3-1, and Figure 3-14.

PAGE 128

128 The smaller radial power gradient for the flat case gives its driver fuels the smallest peak LHGR. The flatter power profile al so explains the higher peak fast fluence in the outer core for the flat case than the other core designs. A more evenly distributed fast flux also results in a more level fast flux exposure between the inner, middle and outer core re gions. Note that the peak fluence limit for all the core designs mentioned is greater than 423 cm2. This number is the fast fluence limit assumed in the Hoffman report in consideration to the maximum displacements-per-atom (dpa) for fast reactor grade steel. HT-9 is a fast reactor grade martensitic/ferritic steel that was used for claddi ng and in-core structures for EBR-II and the Fast Flux Test Facility (FFTF) [ 63]. For the final down selection disc ussed in the next section, a lim it of 200 dpa is assumed. The design basis and perf ormance criteria for the AHFTR driver fuel and targets are discussed in detail in Chapter 6. Similar to the LHGR distribution, a flatter flue nce distribution allows more of the fuel to be irradiated to a level closer to the 423 cm2 limit. As discussed in Chapter 2, the constraining parameter on the fuel cycle for the parametric study was the peak fuel burnup at the midpoint of the first row of fuel. The lim it on fuel burnup as opposed to maximum cladding damage during the equilibrium cycle convergen ce process is a limitation of the REBUS code (Figure 2-1). The next section di scusses a final down selection that reduces the cycle length such that the technical irradiation damage limits fo r the HT-9 structural components are met. Final Down-Selection: The AHFTR Design A final down selection in the AHFTR core desi gn is made based on the transmutation and reactor physics attributes of the above flattened design case. In order to give a more favorable fluence and dpa, the cycle length from this case is reduced. In order to give a shorter cycle length, the pin diameter of the flat design is reduced to increase the TRU enrichment. The only parameter changed in the fuel assembly design is an increase in the pin pitch-to-diameter ratio

PAGE 129

129 from 1.1 to 1.176. The resulting pin diameter (0.755 cm) is equal to the ABR version with a tCR of 0.75 which was proposed by Hoffman et al (Table 1-6 and Table 3-2) [ 7]. This pin diameter and num ber of pins per assembly also closely matches the S-PRISM driver fuel assembly design (Table 1-3) [ 8]. The slight increase in enrichment also had the benefit of decreasing the AHFTRs CR (both fCR and tCR) from the case analyzed above. The core geometry of the flat case was also m odified slightly. Due to the low neutron flux in that part of the target region residing in the outer core (row 7 and 8), the target region above these assemblies was removed. Without these outer targets, the core design shown in Figure 314 now becomes the core design originally shown in Chapter 1 in Figure 1-7. Additionally, the removal of these outer core targets resulted in a reduction in the un-tran smuted MAs discharged from the target region and sent to the pyroprocesso r. Because of the higher efficiency, a higher concentration of transmuted plutonium and a smaller concentration of un-transmuted MA in the mass flow were sent to pyroproces sing from the target region. Fuel Cycle Performance of the Final AHFTR Design The generation of plutonium isotopes by th e transmutation of MAs can be seen by comparing the mass flow rate of Am-241 entering and the Pu-238 exiting the target region. The resulting mass flow rate of plut onium exiting the target region can then be viewed as a supply rate (via pyroprocessing) of in ternally provided plutonium to th e driver fuel. This internal plutonium supply rate can then be compared with the external plutonium supply rate from the aqueous reprocessor. From Table 3-11, it can be seen that the discharge rate of Am-241 coming out of the target region (0.0202 g/MWD) is 40% that of the Am-241 charge rate entering the targets (0.0580 g/MWD). Therefore, 60% of the Am-241 mass entering the target region is destroyed by transmutation or by fission. Also, it is important to note that the target regions Pu-238 discharge

PAGE 130

130 rate (0.0175 g/MWD) is 6.7 times the charge ra te going into the target s (0.0026 g/MWD). Also, it can bee seen from Table 3-10 that the target regions Pu-238 discharge rate (0.0175 g/MWD) is three times the externally supplied Pu-238 feed rate from the aqueous reprocessing center (0.0058 g/MWD). This means that the target regi on provides more than three times the supply of Pu-238 to the active core via pyroprocessing than the Pu-238 supplied by the make-up transuranic supply from the a queous SNF reprocessing. Table 3-11. Final AHFTR mass flow of each isot ope entering and exiti ng the target region and active core region per MWD Charge and Discharge Rate of Mass Flows Normalized to grams/MWD Entering Target Exiting Target* Entering Core** Exiting Core* Np-237 0.0053 0.0024 0.0264 0.0207 Pu-238 0.0026 0.0175 0.0969 0.0736 Pu-239 0.0480 0.0471 0.9963 0.9201 Pu-240 0.0217 0.0284 0.6255 0.5802 Pu-241 0.0105 0.0086 0.0947 0.0850 Pu-242 0.0062 0.0108 0.1457 0.1312 Am-241 0.0580 0.0202 0.0829 0.0577 Am-243 0.0302 0.0138 0.0761 0.0623 Cm-244 0.0116 0.0188 0.0815 0.0666 Supply Rate from Aqueous Reprocessing to Target Region Supply Rate from Aqueous Reprocessing to Active Core Region Np-237 0.0053 0.0033 Pu-238 0.0026 0.0058 Pu-239 0.0480 0.0291 Pu-240 0.0217 0.0169 Pu-241 0.0105 0.0011 Pu-242 0.0062 0.0037 Am-241 0.0580 0.0050 Am-243 0.0302 0.0000 Cm-244 0.0116 -0.0039 *The sum of these two columns is the total mass stream seen by the pyroprocessor. **This column is the total mass of exte rnally supplied and th e internally supplied pyroprocessed TRU. The balance between Am-241 dest ruction versus Pu-238 generation can also be seen by examining the isotopic inventory of the core at BOEC and EOEC which is given in Table 3-12. Also from Table 3-12, it can be seen that the tota l concentration of MAs per HM is maintained to be much less than the imposed 5% limit imposed on the active core composition.

PAGE 131

131 Table 3-12. Target and active co re region isotopic fuel inventory compared with the ABR at BOEC and EOEC (REBUS) Target Region Active Core Region ABR (CR=0.5) BOEC EOEC BOEC EOEC BOEC EOEC Heavy Metal Loading (kg) 714 69812,58312,3809,368 9,127 TRU Loading (kg) 244 2402,7912,7413,021 2,913 MA Loading (kg) 115 106374360343 332 Pu Loading (kg) 129 1342,4172,3812,678 2,581 Pu-238 Loading 13 16110105100 97 Pu-239 Loading 61 611,2361,2201,140 1,088 Am-241 Loading 50 42928688 83 MA/HM Ratio 16.11% 15.23%2.97%2.91%3.66% 3.64% The buildup and depletion curve for both the target region an d the active core region is given in Figure 3-17 and Figure 3-18, respectively. These curves give the change in fresh driver fuel and target composition as a function of irradi ation time. Notice the small relative change in the overall plutonium content of the fuel with irra diation time. This veri fies the statement made in Chapter 2 regarding the lack of spectrum shifti ng as a result of depletion in the SFR. Because, the relative fuel composition of the core does not change consid erable during the course of the irradiation, the change in resonance and unr esolved resonance shielding between various isotopes is negligible. 1.0E+00 5.1E+01 1.0E+02 1.5E+02 2.0E+02 2.5E+02 3.0E+02 3.5E+02 01234 Effective Full Power YearMass of Isotope (Kg/MTiTRU) Np-237 Pu-238 Pu-239 Pu-240 Pu-241 Pu-242 Am-241 Am-243 Cm-244 Figure 3-17. Buildup and depletion curve for fresh fuel in the target region of the final AHFTR design (REBUS)

PAGE 132

132 The same can be said about the target regi on to a limited extent. Though the change in MA concentrations is significant as intended, the Pu-239 concentration varies slowly with irradiation. This slow Pu239 depletion was intended by design by adding a small amount of uranium to the metal alloy of the target slug in order to buffer changes in spectrum with irradiation. The insensitivity of the target regions neutron spectrum to burnup is verified by calculations performed in Chapter 5. 1E+00 1E+01 1E+02 1E+03 01234 Effective Full Power YearMass of Isotope (Kg/MTiTRU) Np-237 Pu-238 Pu-239 Pu-240 Pu-241 Pu-242 Am-241 Am-243 Cm-244 Figure 3-18. Buildup and depleti on curve for fresh fuel in the active core region of the final AHFTR design (REBUS) Reactor Performance Characterist ics of the Final AHFTR Design A comparison is made between the final AHFTR down-selection and the homogeneous metal ABR (CR=0.5) design proposed by Hoffman et al [ 7]. It was the lateral core layout of this ABR (CR=0.5) design case that was adopted for the ABR refe rence cases which was used for comparative analysis of the earlier parametric study. This is also the lateral core layout shown in Figure 1-5 and Figure 3-1. To achieve the conversion ratio of 0.5, Hoffman used a pin diameter of 0.623 (p/d=0.623) which is less than the pin diameter of 0.808 (p/d=1.1) which was used in the previous section to compare the tall and flat versions of the AHFTR. Also, Hoffman

PAGE 133

133 increased the number of fuel pins from 271 (tCR=1.0 and tCR=0.75 designs) to 324 for a tCR of 0.5. Because of the significant change in fuel pin size and assembly design, the ABR (CR=0.5) design is somewhat more exotic than the fuel de signs used in past react or designs such as EBRII, FFTF and the proposed S-PRISM. This is wh y the AHFTR fuel assembly design is based on the ABR with a tCR=0.75 instead of the tCR=0.5 design. Table 3-13 gives a comparison of various reactors and fuel cycle attributes of both the final AHFTR and two ABR designs. The AHFTRs transuranic enrichment is just inside the current experience database with Pu-U-Zr fuel alloys tested duri ng the IFR program in the 1980s [ 64]. Also because of the sm aller enrichment, the excess reacti vity for the AHFTR is less than the ABR. It is important to note the selection of the pin di ameter of 0.755 cm (p/d=1.176) for the AHFTR was not optimized to make the fluence and dpa limit as close to the 4x1023 cm-2 and 200 dpa limits as possible. Instead, it was statically selected on the basis of most probable a nd feasible fuel composition and assembly design. Extending the cycle length is possible by decrea sing the driver fuel enrichment. However, as discussed in Chapter 2, an increasing convers ion ratio results from decreasing TRU enrichment. The AHFTR core design was made to have approximately the same cycle length as the ABR (CR=0.5) case. However, because of the higher neutron leakages from the AHFTR active core, due to the flattened geometry, a larger uran ium fraction in the core is required to increase internal breeding, which extends the fuel burnup, in order to meet the cycle length requirement. However, per the discussion Chapter 2, the addi tion of uranium (i.e., re ducing the transuranic enrichment) does not necessitate the removal of fissile TRU. It does however indicate that the fuel fraction is increased and the sodium fraction in the core is decreased (i.e., a larger fuel pin

PAGE 134

134 diameter). Therefore, the transuranic enrichment of the AHFTR driver fuel is less than that of the ABR (CR=0.5), but the transuranic loading is comparable and actually slightly larger (Table 3-12). Table 3-13. Initial reactor physics and fuel performance co mparison between the AHFTR and ABR design proposed by Hoffman et al ABR (CR=0.5) ABR (CR=0.75) AHFTR Fuel and Reactor Dimensions* Total Core Height 101.600101.600 91.600 Active Driver Height 101.600101.600 71.600 Pin Pitch-to-Diameter Ratio 1.293 1.176 1.176 Pin Diameter (cm) 0.623 0.755 0.755 Pins per Assembly 324 271 271 Calculated Conversion Ratio Fissile Conversion Ratio (fCR) 0.64 0.84 0.84 Transuranic Conversion Ratio (tCR) 0.53 0.77 0.72 Excess Reactivity and Cycle Data Inner Core Enrichment 26.6 16.1 20.77% Excess Reactivity 2.85% 1.47% 1.23% Cycle Length (EFPD) 219 232 214 Fuel Assembly Residence Times (cycles) Inner Core 6 6 6 Middle Core 6 6 6 Outer Core 7 6.5 6 Enrichment Splitting Factor (multiple of Inner Core) Inner Core 1.00 1.00 1.00 Middle Core 1.25 1.25 1.12 Outer Core 1.50 1.50 1.25 Reactivity Kinetics and Feedbacks Delayed Neutron Fraction** 0.0035 0.0035 0.0033 Total Core Void Worth ($) 9.17 6.82 8.34 Doppler Coefficient (/K) -0.08 -0.10 -0.14 Peak Fast Fluence: E>0.1 MeV (1E23 cm-2) IC: Row 1 Mid-plane 4.00 3.86 3.55 Displacements per Atom in HT9 (dpa) IC: Row 1 Mid-plane 182.7 176.3 160.4 *These dimensions are also give n and discussed in more detail in Table 1-6 and Table 1-7 in Chapter 1. **The delayed neutron fraction was calculated with MCNP using fission source libraries with and without delayed neutron fraction data [ 65]. Because of the higher co ncentra tion of uranium in the AHFTR fuel, it should be noted that the Doppler coefficient of the AHFTR design is slightly more negative than that of the ABR (CR=0.5). Also, because the transuranic, partic ularly MA, concentration in the AHFTR driver

PAGE 135

135 fuel is less than that of the homogeneous core, th e positive void worth of the AHFTR is slightly less than that of the homogene ous reference (Table 3-13). Transmutation Analysis of the Final AHFTR Design Table 3-14 shows the rates of MA and plutonium consumption for the AHFTR versus the ABR designs. From Table 3-14, the AHFTR bur ns less total TRU per fission energy generated than the ABR (CR=0.5) but roughly twice the am ount of MAs. Also the AHFTR burns almost as many MA as it does plutonium. For comparis on, Table 3-14 also shows the amount of TRU produced for a typical PWR design fuel assembly with the composition given in Table 2-1. Table 3-14. Mass production and destructi on rates per installed megawatt per year ABR (CR=0.5) ABR (CR=0.75) AHFTR External Supply Mass Streams Provided by UREX+3 Aqueous Plant* External Np+Pu Supply (kg/MWY) 0.1596 0.0695 0.0549 External Am+Cm+Bk+Cf Supply (kg/MWY) 0.0092 0.0051 0.0370 External Mass Supply Broken Down by Pu and MA Pu Consumption (kg/MWY) 0.1516 0.0654 0.0519 MA Consumption (kg/MWY) 0.0172 0.0092 0.0400 UOX PWR** MOX PWR** Pu Consumption (kg/MWY) -0.0758 0.2033 MA Consumption (kg/MWY) -0.0074 -0.0482 Mass streams calculated from the equilibrium cycle search calculation performed by REBUS on the fuel management scenario in the boxed off region of Figure 3-4. **UOX PWR calculations were performed using TRITON for an initial enrichment of 4.5% and burned for 50 MWD/kg. MOX PWR calculations were performed using TR ITON for an initial Np+Pu concentration of 10% and burned for 50 MWD/kg. Also shown in Table 3-14, is the amount of plutonium consumed and MAs produced in a typical PWR MOX fuel assembly. As a side no te, it is worth mentioning that the MOX-PWR fuel assembly has approximately the same net TRU consumption rate as the ABR. In fact, it can be easily calculated that most PWR MOX fuels have a conversion ratio between approximately 0.6 and 0.8 depending on the TRU enrichment. However, the MOX has a negative net consumption (i.e., production) of MAs. This empha sizes the fact that a re actor design can have a low CR with a high net destruction of TRU but still do very little for reducing the isotopes most

PAGE 136

136 important to HLW management and the reposito ry design (i.e., MAs and long lived fission products) By comparing the mass flows per unit of in stalled thermal reactor capacity, the ABR (CR=0.5) would require approximately one MWth to destroy the transuranics produced by two MWth of UOX PWRs. Coincidently, if the fu el design and core layout of the metal ABR (CR=0.75) is used, the ratio of ABR installed power to PWR installed power is near unity. The AHFTR also destroys TRU at about the sa me rate that it can be produced in a PWR UOX. However, the TRU destroyed by the AHFTR is more than 40% MAs. Hence, the AHFTRs MA consumption is roughly 2.3 times greater than that of the ABR (CR=0.5) design. These comparisons demonstrate that the transuranic consumption of the AHFTR is comparable to the range of ABR options. More importantly, the fissile (i.e., plutonium) requirements of the AHFTR are significantly less than the ABR. The AHFTR requires only a third of the externally supplied plutonium than the ABR (CR=0.5) a nd 70% of the plutonium burned in the ABR (CR=0.75). Note that the use of the term exte rnally supplied can be used synonymously with the consumption rate. The fact that the fissi on cross section for most MAs is much smaller than that of fissile plutonium suggests that the 40% MA concentration in the AHFTR TRU requires transmutation before ulti mately being destroyed by fission. In order to show that the MAs are not producing 40% of the fission reactions in the core, th e contribution to the reactor average fission rate for each actinide is otopes is broken down in Table 3-15. Because 81% of the reactors thermal power is produced by fissions of plutonium isotopes but only 14% of the externally supplied HM is actually plutonium, Table 3-15 shows that the remainder of the plutonium that is undergoing fission is provided by transmutations from the fertile isotopes: U-238 and the MAs. A closer look at the amount of pl utonium supplied to the

PAGE 137

137 core versus actually undergoing fi ssion shows that 92% of the TRU that is undergoing fission at any given time is plutonium. In actuality, only 56% of the TRU being supplied to the core is plutonium. One important fact to observe is that the total rate that HM is actually supplied to the core is 0.00103 kg/MWD (1.03 g/MWD). This is equivalently the amount of HM mass that is destroyed by fission in order to produce one megawatt of power in one day. Am-241 is supplied to the core at a rate of 5.8E-05 kg/MWD. Th e fraction of Am-241 in the HM make-up feed is 5.65%. This is equivalent to saying that 5.65% of all fission reactions in the core are derived from the introduction of Am-241. However, onl y 1.01% of the fission power is being produced by Am-241. Table 3-15. Concentration of isotopes suppl ied to replace the mass destroyed by fission compared to the contribution to fission by each isotope (REBUS) External Heavy Metal Supply (Kg/MWD) Fraction of Each Isotope per HM of External Feed Fraction of Each Isotope per HM of All Fission Reactions Fraction of Each Isotope per TRU of External Feed Fraction of Each Isotope per TRU of All Fission Reactions U-234 7.61E-08 0.0074% 0.16% --U-235 3.96E-06 0.3860% 0.48% --U-236 2.34E-06 0.2278% 0.07% --U-238 7.68E-04 74.8441% 11.57% --Np-237 8.38E-06 0.8170% 0.33%3.33% 0.37% Pu-238 4.15E-06 0.4047% 3.58%1.65% 4.09% Pu-239 7.65E-05 7.4637% 59.36%30.42% 67.66% Pu-240 3.46E-05 3.3718% 8.55%13.74% 9.75% Pu-241 1.68E-05 1.6334% 7.64%6.66% 8.71% Pu-242 9.96E-06 0.9714% 1.43%3.96% 1.62% Am-241 5.80E-05 5.6534% 1.01%23.04% 1.15% Am-242m 1.97E-07 0.0192% 1.17%0.08% 1.33% Am-243 3.02E-05 2.9460% 0.70%12.01% 0.79% Cm-242 1.92E-09 0.0002% 0.03%0.00% 0.03% Cm-243 9.33E-08 0.0091% 0.14%0.04% 0.16% Cm-244 1.16E-05 1.1342% 1.34%4.62% 1.52% Cm-245 1.01E-06 0.0989% 2.21%0.40% 2.52% Cm-246 1.19E-07 0.0116% 0.16%0.05% 0.18% Total 0.00103 100% 99.9%100% 100% As an appropriate comparison, for the same analogy applied to the ABR (CR=0.5), the Am-241 concentration in the HM make-up feed is only 1.5% which is set by the isotopic content

PAGE 138

138 of SNF TRU and the TRU enrichment in the fuel. Also, the fraction of fission power contributed by Am-241 in the ABR (CR=0.5) is only 0.8%. Hence, about half of the Am-241 externally supplied to the core is destroyed by fission (as Am-241) and the remainder is transmuted. In contrast, the AHFTR destroys only 20% of the ex ternally supplied Am-241 through direct fission of Am-241 atoms. The remaining 80% of the Am-241 external mass supply is transmuted. This result demonstrates that the AHF TR use MAs for conversion into fi ssile isotopes. This strategy is different from the ABR which does not us e transmutation targets to precondition the americium into plutonium.

PAGE 139

139 CHAPTER 4 REACTOR REACTIVITY CONTROL STRATEGY The standard control rod poison for most SFR designs has historically been natural boron or boron enriched in the more absorbing isotope: B-10 [ 29]. The preference of boron over other poison m aterials, such as silver or gadolinium, is primarily due to the order of magnitude larger unresolved resonance cross section of B-10 over ot her conventional neutron poisons in the fast spectrum. The necessity to enrich the boron in B10 is determined by the ex cess reactivity of the core. The ABR, like the EBR-II and many othe r SFR designs, has been proposed to have enriched B4C as the control rod material. As a side note, FFTF used a sintered boron powder as the control rod material. Because the AHFTR has a smalle r excess reactivity than the ABR reference case (Table 313), it requires less control rod wort h to suppress this reactivity. Therefore, boron enrichment is not required. As demonstrated by calculations pe rformed by Hoffman et al and Morris et al, the excess reactivity of an actin ide burner SFR (such as the A BR) increases with decreasing conversion ratio [ 7,20]. The increase in ex cess re activity is a result of the requirement to enhance neutron losses by leakage in order to attain the low fCR. As mentioned before, for homogeneous cores, the fCR is closely coupled to the tCR because transuranic breeding is a function of the neutron balance between parasi tic absorption and neutron escape losses. The AHFTR has less neutron losses because the axial ta rgets lessen their probability of escape. Tc-99 versus B-10 as a Control Rod Neutron Poison Though it will be demonstrated in this chapter that B4C would be an acceptable neutron poison for the AHFTR, the fundamental objective of this work is to destroy TRU, and if possible other HLW, separated from SNF such as fission products. As identi fied in the introduction, Tc99 has a non-trivial concentration in SNF (approximately equal to the MA concentration). Also,

PAGE 140

140 because of its radiotoxicity, and long half-life (in a geologic time frame), it requires a permanent disposal solution, such as in a geologic repository Because of this need to remove Tc-99 from the fuel cycle, previous studies have focuse d on ways of transmuting it in a SFR. It is noteworthy that the UREX+ aqueous reprocessi ng technology was developed partly for the purpose of removal of Tc-99 from the repository destined SNF waste stre am. For destroying Tc99 by transmutation, Yang et al pr oposed burning metallic Tc-99 in moderated targets within an accelerator driven SFR system [ 66]. However, Yangs studies c oncluded that the transm utation half-life of Tc-99, even in a moderated target within the SFR, is in the range of decades. This result precludes the use of Tc-99 as an Integral Burnable Absorber (IBA). IBAs are burnable poisons that are an integr al part of the fuel assembly that can not be removed once it has been manufactured [ 3]. An example of an IBA is the Westinghouse Integral Fuel Burnable Absorber (IFBA) fuel rod which has a coating of zirconium diboride on the fuel pellets [ 3]. IBAs are used extensively in m any PWR fuel de signs for the purpose of leveling power peaking and power shifting that can occur between beginn ing-of-cycle (BOC) and end-of-cycle (EOC). Using IBAs allows all the fissile material in an LWR to be consumed at a more level rate which enhances fuel utilization and extends the fuels reactivity limited burnup. In order to control pin and assembly power peaking between BOC and EO C, an IBA must be able to be almost completely depleted by the end of the first cycl e. Otherwise, the pres ence of the IBA would penalize the excess reactiv ity of the core during the next i rradiation cycle. Because a Tc-99 target can not be completely depleted during a singl e irradiation cycle, it is unlikely that it could be made a feasible IBA. Despite the difficulty of burning Tc-99 as an IB A, there is still a cost advantage to using technetium as a neutron poison in control rods. The isotopic enrichment of boron is an

PAGE 141

141 expensive process that could be avoided if Tc99 were used instead. Since, the UREX+ process produces a technetium waste stream that is se parate from the other fission products; it is a material that would be readily on hand in the LW R-to-SFR fuel cycle (Fig ure 1-9). If not used for the production of control rods technetium would require geologi c disposal as a HLW stream. If the AHFTR consumes Tc-99 as a service to the fuel cycle, it would be entitled a credit towards the cost of its fuel purchase as payment for the de struction of Tc-99, which is considered HLW. A detailed explanation of the economic incentiv e to destroy HLW is given in Chapter 6. Tc-99 has a virtually identical unresolved resonance neutron cap ture cross section to that of U-238 in the fast spectrum (Figure 1-6). By comparison, the unresolved resonance capture cross section of Am-241 has a magn itude three times as great (Figur e 4-1) as Tc-99. This fact explains why Tc-99 is more difficult to transmut e than americium in a SFR. Similarly, when compared to B-10, the Tc-99 capture cross section is also less by approxima tely a factor of three (Table 4-1). It is important to remember that it is B-10 s highly absorbing property that makes enriched B4C the primary candidate for reactivity control in the ABR. However, the excess reactivity hold down requirement (i.e., excess reactivity) of the AHFTR is less than that of the ABR (CR=0.5) by roughly a factor of two (Table 3-13). Therefor e, enriching the B4C is not as significant of a prerequisite for the AHFTR. Also it is important to note that metallic Tc-99 has an atomic density that is roughly 3.5 times gr eater than that of the B-10 constituent of 90% enriched B4C (Table 4-1). Therefore, Tc-99 can be considered a candidate for the neutron poison in the movable control rods in the AHFTR Because a control rod can be inserted or withdrawn per the excess reactivity requirements of the core, it is not necessary to completely burn out the Tc-99 by EOC as it would be if were implemented as an IBA.

PAGE 142

142 Table 4-1. Atomic concentra tions of absorber atoms in B4C versus technetium metal (atom/barncm) Atom Concentration Natural B4C Enriched (90%) B4C Metallic Technetium B-10 4.39E-03 1.98E-02 -B-11 1.76E-02 2.20-03 -C 5.49E-03 5.49E-03 -Tc-99 --6.69E-02 1E-07 1E-06 1E-05 1E-04 1E-03 1E-02 1E-01 1E+00 1E+01 1E+02 1E+03 1E+04 1E-071E-061E-051E-041E-031E-021E-011E+001E+011E+02Energy (MeV)Cross Section (barns) B-10 B-11 Am-241 Tc-99 Figure 4-1. ENDF-VI total neutron absorption cr oss section plots for select AHFTR absorber materials Because Tc-99 is not as strongly absorbing as B-10 or Am-241, it is more suited as a gray absorber material. As mentioned in the Introduct ion, Messaoudi et al proposed that Tc-99 could be used as a neutron absorber in dilution pins within the driver fuel assembly for the purpose of resonance feedback [ 32]. To minimize the affect of ener gy shielding with U-238, which could happen during a sodium void, Kim et al has shown that the poison rods shoul d be concentrated in their own standalone poison assemblie s as opposed to evenly distributed in the driver fuel. In the case of the study performed by Kim et al, B4C was proposed as an IBA [ 31]. However, it should be expected that the sam e princi ple applied by Kim et al also app lies to control rods. Therefore, distributing the control rods in evenly distributed clusters thr oughout the fuel, as is done in LWRs, is not practical in SFRs. In fact, in most SFR designs cont rol rods are commonly

PAGE 143

143 bunched together into a movable subasse mbly that travels inside a fixed control assembly that takes its own discrete location within the core. Therefore, the technetium control rods will be located in dedicated control assemb lies as shown in Figure 1-7. Control Assembly Design The AHFTR reactivity control strategy consists of three different types of control rod mechanisms: (1) a reactivity shim control rod bank, (2) a safety shutdown rod bank and (3) an ultimate shutdown rod bank for achieving shut down margin. Technetium is selected to be the absorber material for the shim and safety rod banks. The shim and safety rod clusters operate independently of each other, but are together co-located in the primary control assembly locations (Figure 1-7). The ultimate shutdown rods compose a separate co ntrol assembly (Figure 1-7). The dimensions of each control system, wh ich will be discussed in the following sections, are given in Table 4-2. Table 4-2. Control assemb ly types and dimensions Control Assembly Type Ultimate Shutdown Primary Control Gas Expansion Module Control Assembly Pitch (cm) 16.142 16.142 16.142 Assembly Duct Thickness (cm) 0.394 0.394 0.394 Inter Assembly Gap (cm) 0.432 0.432 0.432 Subassembly Duct Thickness (cm) 0.394 --Intra Assembly Gap (cm) 1.000 --Rod Type Shutdown Shim Safety Reflector Poison Type B4CTc-99 Tc-99 Void CRGT Outer Diameter (cm) --0.755 0.755 0.755 CRGT Inner Diameter (cm) --0.643 0.643 0.643 Control Rod Cladding Outer Diameter (cm) 0.7550.455 0.455 -Poison Slug Outer Diameter (cm) 0.6430.343 0.343 -Length of Poison Section (cm) 91.10056.900 91.100 91.100 Number of Rods 271133 138 271 Traditional SFR Control Assembly Design: B4C Ultimate Shutdown Assembly It is important to indicate the current state of typical SFR control assembly design such as that proposed for the ABR before discussing the de tails of the technetium primary control. To

PAGE 144

144 illuminate this point, the design of the ultimate shutdown control assembly is selected to be based on that of the ABR. Therefore, the ul timate shutdown control a ssembly draws its design from more conventional SFR contro l designs. Similar to the ABR, a small number of the control rod assemblies are designated for attaining the ultimate shutdown margin of the core. This margin is considered necessary to bring the co re reactivity down to the point of the cold shutdown condition [ 55]. Most of the length of the ultimate shutdown asse mbly consists of just the outer hexagonal HT-9 shroud (Figure 4-2). Sodium coolant is allowed to flow through this empty assembly structure. The ultimate shutdown rods make up a 271 pin bundle array that is located inside a secondary hexagonal HT-9 shroud which is concentr ic with the outer assembly shroud. This secondary shroud and pin bundle comprise a subassembly which is allowed to move freely inside of the outer assembly shroud. This contro l subassembly is connected to a drive shaft that extends to control motors out side of the reactor vessel. Just as with traditional SFR core designs, the neutron poison material for the ultimate shutdown system is B4C. For the AHFTR, natural boron is used instead of enriched boron. For the AHFTR primary control rods, technetium metal is adopted for the poison material for its high atomic density of Tc-99 atoms. Tc-99 metal wa s also proposed for the technetium transmutation targets by Yang et al. However, unlike the Tc-9 9 targets proposed by Yang et al, the primary control assemblies proposed in th is dissertation are not implemen ted with a moderating material in their design. Technetium Based Primary Control Assembly Design Originally, an internal subassembly design, ex actly the same as that proposed for the ABR, was considered for the AHFTR primary control sy stem. In the ABR, the downward motion of a primary control hexagonal subassembly into the top of the active core is responsible for

PAGE 145

145 reactivity shim. These rods used for reactivity sh im also have a dual role as the ABRs safety rods. Thus, the ABR uses the same primary control subassembly to provide enough negative reactivity to shut the core down to a safe conditio n in case of an accident situation. Therefore, the ABR uses this subassembly design to provide reactivity shim as well as part of the shutdown margin. Figure 4-2. Ultimate shutdown control assembly configuration When a top loaded primary control subassembly is inserted into the top of the AHFTR, the resulting absorption of the neutron flux in th e upper half of the core also decreases the transmutation efficiency of the axial targets. Therefore, the mechanics of the primary control Ultimate Shutdown Control Assembly Configuration Handling Socket Coolant Inlet Nozzle Sodium in the Control Assembly Duct Concentric Hexagonal HT-9 Ducts Control Assembly Wall B4C Control Rod natural boron Rod Bundle Sub-Assembly Wall Sub-Assembly Fully Inserted Sub-Assembly Fully Withdrawn Region of the Active Core Sodium outside the rods Direction of Ultimate Shutdown Subassembly Movement Drive Shaft

PAGE 146

146 assembly design diverges slightly from the ABR. This decision was ma de to allocate a portion of the primary control assembly rods, to be used for shim control, to be inserted through the bottom of the core so that they would not significantly perturb the flux in the axial target region. Instead of a moveable subassembly filled with control rods, two separate clusters of control rods are created. Rods from both a shim rod cluster and a separate safety rod cluster are evenly distributed throughout the primary contro l assembly. Rods of the same cluster are attached to one another through a spider-like tie rod subassembly si milar to that used in PWRs. Therefore, there is a spider-lik e tie rod subassembly for the shim rods and a separate one for safety rods within each primary control assembly Because the shim rods are inserted into the core from the bottom, the shim rod cluster is attached to long tie-rods that are equal to the height of the core. To insert shim rods into the core, its spider subassembly is pulled upwards by the action of the rod drive. When the shim rods are inserted to their desired locati on, a mechanical clamping mechanism can be actuated to prevent them from falling out of the core. Alternatively (or additionally), the shim rod drive mechanism can be designed with sufficient internal friction in the gearing system that uninten tional rod drop out of the core is highly improbable. It is important to note that the incorporation of bot h top and bottom loaded control rods does not increase the physical size of th e reactor and associated core ex ternals because neither shim or safety rods are withdrawn beyond the extent of the gas plenum (top) or the axial reflector (bottom). Figure 4-3 shows the mechanical deta ils of the primary control assembly design. The AHFTR primary control assembly consists of a tube bundle of c ontrol rod guide tubes (CRGT) that are wrapped in a hexagonal asse mbly shroud. These CRGTs are capped at the bottom to prevent sodium coolant from entering the tube. Half of these CRGTs are allocated for

PAGE 147

147 the shim control rods whereas the other half are allocated for the safety control rods. The shim and safety rod clusters move independently of each other in the CRGTs through the locomotion of their respective connecting spider subassemblie s. These spider-like connectors are then each connected to two concentric drive shafts that ex tend to the control motors outside of the reactor vessel. Figure 4-3. Primary contro l assembly configuration It is envisioned that a rack-and-pinion style control rod drive motor would be used for the inner safety rod shaft. A worm-screw type dr ive mechanism would be used to move the outer Handling Socket Coolant Inlet Nozzle CRGT Bundle within the Primary Control Assembly Control Assembly Wall Tc-99 Safety Rod Concentric Drive Shafts Tie Rod Connecting to Shim Rod Tc-99 Shim Rod Empty CRGT Empty CRGT Tie Rod Connecting to Shim Rod Direction of Safety Rod Movement Direction of Shim Rod Movement Region of the Active Core Sodium outside the CRGTs Spider-Like Tie Rod Subassembly Primary Control Assembly Configuration Region of the Axial Reflector Region of the Gas Plenum

PAGE 148

148 annular shim rod shaft. The safety system woul d operate on an electric servo with a circuit interrupter such that interruption of power to the servo would cause the safe ty rod clusters to fall into the core. The shim system would only move the shim rod cluster when power was applied to the worm gear servo. Hence, if power is in terrupted to the system, the shim rods would not fall out of the core, but rather be held in place by the friction in the worm-gear system. It is envisioned that the CRGT tube would be filled with heliu m. The lack of sodium flow through the tube prevents the ac tion of hydraulic frictional drag on the control rod in the tube that could hinder its motion. This is an importa nt design feature because of the possibility of a stuck control rod cluster due to a control assemb ly being axially bowed or warped. The lateral and axial temperature gradients, which are typica l in a SFR core, frequen tly create expansion and bending forces that cause assembly bowing. In fact, assembly bowing is sometimes relied upon as a negative reactivity feedback mechanism relate d to leakage. This feedback mechanism will be discussed later in this chapter. An additional design consideration is attrib uted to gamma ray heating of structural components, which is the result of gamma ra ys that are produced by fission and neutron activation, imparting kinetic energy into the atoms of structural materials. The number and dimensions of the CRGT tubes in the primary cont rol assembly are exactly the same as that of the fuel cladding used in the fuel assemblies. Because the CRGTs receive the same level of sodium flow that fuel pins receive, it is exp ected that gamma ray heating of control assembly structures will be sufficiently dealt w ith by the coolant outside the tubes. Gas Expansion Module A measure of passive safety control is in cluded by incorporating six gas expansion modules (GEM) in each six corn ers of the core periphery (Fi gure 1-7). GEMs are special assemblies placed at the perimeter of a SFR whic h are designed to enhance neutron leakage in

PAGE 149

149 the event of the loss of primary coolant flow. GEMs were originally designed for reactors concepts of the IFR program [ 67]. A typical GEM design is shown in Figure 4-4. Figure 4-4. Gas expansion m odule assem bly configuration The lower end of the GEM assembly consists of a lower rod bundle of HT-9 pins that is an integral component of the cores lower axial reflec tor. A separate bundle of tubes fills the space above the axial reflector. These tubes are open at their bottom e nd to allow sodium coolant to enter the tube. A helium gas bubble fills the upper end of the tube, which is capped at the top near the handling socket to prev ent the sodium from passing through the tube. In the event that the bulk coolant circulation pumps fail, the pr essure differential betw een the non-circulating sodium and the helium gas bubble forces sodi um out of the tube causing the gas bubble to Gas Expansion Module Assembly Configuration Handling Socket Coolant Inlet Nozzle Control Assembly Wall Helium Bubble HT-9 Reflector Rods Sodium Filled into Tube Region of the Active Core Sodium outside the rods Sodium direction during loss of coolant flow

PAGE 150

150 expand. The displacement of the sodium and ex pansion of the helium creates a voided space in the radial reflector that extends down into the active core region. The removal of reflection by the sodium at the core periphery enhances neutron streaming from the active core. Control Rod Worth The shim rods are inserted through the bottom of the active core by the withdrawal of the shim cluster spider subassembly. The purpose of the shim rods is to absorb the excess reactivity produced by the addition of fresh fuel at BOC. As can be seen in Figure 4-5, the reactor cores BOEC excess reactivity can be completely suppressed when the rods are inserted 55 cm. This distance is 75% of the height of the active core. 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 0102030405060Rod Tip Position from Bottom of Active Core (cm)Core Reactivity at BOEC ($) Figure 4-5. Shim control rod bank reactivity worth at BOEC As the fuel is irradiated, the excess reactiv ity is consumed by the burnup of the fuel. Therefore, the shim rods are pushed out of th e bottom of the core, by the downward motion of the spider subassembly, until they are fully withdr awn by the end of the irradiation cycle. The reactivity worth of the shim rod bank at EOEC is shown in Figure 4-6. Figure 4-5 and Figure 46 show the excess reactivity worth of the core at BOC and EOC as a fu nction of the length of

PAGE 151

151 control rod inserted into the core. Notice, that the relative reactivity worth of the shim rod does not change considerable between BOC and EOC. -4.00 -3.50 -3.00 -2.50 -2.00 -1.50 -1.00 -0.50 0.00 0102030405060Rod Tip Position from Bottom of Active Core (cm)Core Reactivity at EOEC ($) Figure 4-6. Shim control rod bank reactivity worth at EOEC (DIF3D) The safety control rods are inserted through th e top of the core above the axial targets. These rods do not move throughout the irradiation. Instead they are raised to the top of the core during startup where they remain during normal opera tion. In the event that rapid shutdown is required, the reactor monitoring and safety systems can trigger the drop of these rods into the core by interrupting power to the control rod drive servo. The sa fety rod bank reactivity worth at BOEC and EOEC is shown in Figur e 4-7 and Figure 4-9 respectively. There is a slight positive reactiv ity insertion as the rod tip pass es through the axial targets. This positive reactivity feedback is caused by th e spectrum hardening of the flux in the targets due to the epithermal absorbing worth of techne tium. Figure 4-8 shows the flux spectrum in the targets when the safety rod tip is inserted into the axial target region. However, it is important to note that the preferential absorp tion of neutrons in Tc-99, as the control rods pass through the targets, can only be observed in the epithermal ener gy range. This is beca use, the capture cross section of Tc-99 falls off sharpl y at neutron energies above one MeV (Figure 4-1). It is also

PAGE 152

152 important to note that the spect rums plotted in Figure 4-8 are virtually indistinguishable. Therefore, the spectrum hardening from Tc-99 en tering the target region can be considered negligible. Hence, the initial positive increase in reactivity in Figure 4-7 has a total value of only ten cents. It is likely that this ten cents can be compensated by allowing for passive safety feedbacks such as the leakage control afforded by the GEMs. -1.50 -1.00 -0.50 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 0102030405060708090Rod Tip Position from Top of Axial Target (cm)Core Reactivity at BOEC ($) Figure 4-7. Safety control rod ba nk reactivity worth at BOEC (DIF3D) 1.0E+12 1.0E+13 1.0E+14 1.0E+15 1E-051E-041E-031E-021E-011E+001E+01Energy (MeV)Flux (cm^-2*s^-1)/lethargy Safety Rods All Oout Tip Half Inserted into Target Tip Fully Inserted Through Axial Target Tip Fully Inserted Past Axial Target Figure 4-8. Neutron spectrum in the targets as a function of safety rod insertion (DIF3D)

PAGE 153

153 It is important to note that the positive in sertion exists only at the BOEC. As the americium is depleted in the targets, the loca l spectrum dependence on neutron capture is also lessened. This is because the reactivity suppr ession provided by the combination of moderation and americium capture is lessened. Therefore, the insertion of the rod and resulting absorption of neutrons in the epithermal range has less imp act on reactivity at EOEC. This is reflected in the safety rod reactivity worth curve at EOEC in Figure 4-9. As can be seen from Figure 4-7 and Figur e 4-10, the safety rod bank has sufficient reactivity worth by itself to completely shut the r eactor down to at least one dollar below critical. Therefore, even if for any reason, the entire sh im rod bank were to be absent from the core during an emergency shutdown, the shim rod bank could completely take its place and still have one dollar of shutdown margin remaining. -6.00 -5.00 -4.00 -3.00 -2.00 -1.00 0.00 0102030405060708090Rod Tip Position from Top of Axial Target (cm)Core Reactivity at EOEC ($) Figure 4-9. Safety control rod ba nk reactivity worth at EOEC (DIF3D) Additional shutdown margin is provided by the ultimate shutdown subassembly bank. Like the safety rods, the ultimate shutdown s ubassemblies are inserted through the top of the core. These subassemblies are withdrawn through the top of the core during startup where they remain during normal operation. At the end of the irradiation cycle, the safety and ultimate

PAGE 154

154 shutdown subassemblies are inserted to provide for the full shutdow n margin of the core. Figure 4-10 and Figure 4-11 show the reactivity worth of the ultimate shutdown system for BOEC and EOEC respectively. 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 0102030405060708090Rod Tip Position from Top of Axial Target (cm)Core Reactivity at BOEC ($) Figure 4-10. Ultimate shutdown system worth at BOEC (DIF3D) -2.50 -2.00 -1.50 -1.00 -0.50 0.00 0102030405060708090Rod Tip Position from Top of Axial Target (cm)Core Reactivity at EOEC ($) Figure 4-11. Ultimate shutdown system worth at EOEC (DIF3D) It is important to note that the overall reac tivity worth of the ultimate shutdown system is $1.75 despite being composed of natural boron in B4C. The three ultimate shutdown assemblies are located in the inner most region of the core where the flux is strongest. Therefore, despite

PAGE 155

155 being composed of natural boron, the ultimate shutdown assemblies have the largest average reactivity worth per assembly. There are thr ee ultimate shutdown assemblies with a combined worth of $1.75 which gives an average reactivity worth of 58 per assembly. Conversely, the safety rod system and shim rod systems both onl y have an average reac tivity worth of 22 and 28 per primary control assembly, respectively. By inspection of Figure 4-7 and 4-10, it is apparent that the total shutdown worth of the reactor at BOEC is $6.25. The complete s hutdown at BOEC accounts for zero net reactivity when the shim rods are inserted and also $4.50 and $1.75 of total negativ e reactivity insertion by the safety and ultimate shutdown systems, respectiv ely. By inspection of Figure 4-6, Figure 4-9 and Figure 4-11, it is apparent that the total shutdown worth of the reactor is $9.75. The complete shutdown at EOEC accounts for the total negative reactivity insert ed by the safety and ultimate shutdown systems as well as an additional $3.50 by fully inserting the shim rod system. Thus, if an emergency shutdown is required at EOEC with all shim rods removed, the safety and ultimate shutdown system can provide $6.25 of negative reactivity. Reactivity Worth of Boron versus Technetium It is important to compare the reactivity worth of metallic Tc-99 with the more conventional SFR reactor poison: B4C. As noted earlier, the atom concentration of metallic Tc99 is approximately 3.5 times that of B-10 in enriched B4C. However, the unresolved resonance absorption cross section of B-10 is roughly three time s greater than that of Tc-99 for most of the fast spectrum. Figure 4-12 contrasts the reactivit y worth of the shim rod bank if natural boron or enriched boron is used in the form of B4C as opposed to metallic Tc-99. It is apparent from Fi gure 4-12 that enriched B4C has a greater reactivity worth in the AHFTR than Tc-99. However, the use of B4C would only reduce the length of control needed to

PAGE 156

156 shut the core down by approximately 10 cm. Ther efore, the technetium rod is considered as a viable alternative to enriched B4C as a control material in the AHFTR. 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 0102030405060Rod Tip Position from Bottom of Active Core (cm)Core Reactivity at BOEC ($) Tc-99 metal B4C (natural) B4C (90% enriched) Figure 4-12. Shim control rod bank reactivity wo rth at BOEC for metallic Tc-99 compared to natural and enriched boron in the form of B4C (DIF3D) Top versus Bottom Inserted Shim Rods One of the key design features of the primary control assembly was to have the shim rods inserted through the bottom of the core ensure that they did not steal neutrons away from the axial targets. For the sake of comparison, th e volume average radial flux profile of the active core and targets has been plotted for the case wher e all shim rods are inse rted through the bottom of the core versus all shim rods being inserted through the top of the core. These flux plots are given in Figure 4-13. Notice that the flux in the targets is slightly le ss for the top inserted shim rods than it is for the bottom inserted shim rods. The reduction of flux indicates that the tr ansmutation efficiency of the axial targets would be reduced if the shim rods were inserted through the top of the core. Unlike the safety or ultimate shutdown rods, the shim rods require being in the core at all times. Hence, the neutron absorbing effect of a top inse rted control rod would ha ve a constant impact

PAGE 157

157 on the targets regions transmutation efficiency, regardless of its gradual removal from the active core. Thus it is more feasible to have a botto m inserted shim rod cluster design as was discussed in the primary control asse mbly proposed for the AHFTR. 0.0E+00 5.0E+14 1.0E+15 1.5E+15 2.0E+15 2.5E+15 3.0E+15 3.5E+15 4.0E+15 4.5E+15 020406080100120140Radial Distance from Core Center (cm)Total Flux (cm^-1*s^-1) Target-Bottom Loaded Shim Rod Active Core-Bottom Loaded Shim Rod Target-Top Loaded Shim Rod Active Core-Top Loaded Shim Rod Figure 4-13. BOEC radial flux distributions of the active core (volume averaged over core height) and axial target fluxe s for shim rods being inserted through: the top or the bottom of the core (DIF3D) Axial Power Tilt and Shim Rod Insertion Figure 4-14 shows the axial power distribution for various shim rod positions in the AHFTR core. As discussed previously, the larg e mean-free-path of neutrons in the core combined with the relatively gray neutron absorb ing worth of technetium reduces the affect of localized flux depressions on the overall core flux distribution. Therefore, the axial flux distribution is virtually identical regardless of the length of control rod inserted into the active core region. Other Reactivity Feedbacks As indicated previously, the AHFTR, like the ABR and many other SFR designs has a positive void coefficient. The positive reactivity feedback is a direct consequence of neutron

PAGE 158

158 energy spectrum hardening resulting from the lo ss of the slight moderation provided by the sodium coolant. When the spectrum hardens, th e lack of down scatteri ng causes the number of neutrons above the fission threshold of fertile isotopes (U-238 and MAs) to increase. An increase in above-threshold fissions causes an increase in the neutron multiplication contribution of these fertile isotopes. Figur e 4-15 gives the neutr on spectrum of the AHFTR inner core region during normal steady state operation and a scenario where all of the sodi um coolant is voided from the core. 40% 50% 60% 70% 80% 90% 100% 110% 120% 130% 0102030405060708090100Axial Distance from Bottom of Active Core (cm)Relative Power Density to Inner Core Average (%) 0 6.7125 13.425 20.1375 26.85 33.5625 40.275 46.9875 53.7 Figure 4-14. Axial power distribution for increasi ng shim rod length into the active core (inner enrichment zone) region (DIF3D) This spectrum hardening can be unfavorable if there is no other competing feedback mechanism that can negate the void induced pos itive reactivity insertion. For SFRs, some negative reactivity feedback comes in the form of Doppler re sonance broadening of capture cross sections as the fuel temperature increases. Additionally, because SFRs typically exhibit a high degree of leakage, they can rely on feedback mechanisms that increase leakage as the fuel and structural materials increase in temperature. Th e negative reactivity feedback caused by this fuel expansion is most pronounced in metallic fuel s and was a key control aspect of EBR-II [ 35].

PAGE 159

159 0.0E+00 2.0E+14 4.0E+14 6.0E+14 8.0E+14 1.0E+15 1.2E+15 1E-051E-041E-031E-021E-011E+001E+01Energy (MeV)Flux (cm^-2*s^-1)/lethargy Normal Void Figure 4-15. Neutron spectrum at BOEC of the inner core region for steady state operation versus a complete loss of sodium coolant (DIF3D) The void and Doppler reactivity worth for the AHF TR are given in Table 4-3. Also given is the negative reactivity provided by leakage fee dback when fuel expansion changes the radius and height of the active core. Leakage feedback is also provided by the GEM. The void effect of the helium bubble expansion in the GE M is also given in the table. As can be seen from Table 4-3, if the sodi um coolant is completely lost from the active core (coolant remains in the gas plenum, axial and radial reflectors), a positive reactivity insertion of about $8.60 occurs. However, if thermal expansion e ffects cause both the height and diameter of the core to increase by only 2.5%, th e core will lose approximately six dollars worth of reactivity. If both the height and diameter increase by 5.0%, the expansion would remove approximately $13.30, which is sufficient reactivity to completely negate the positive reactivity insertion by the void. Also, an additional dollar of negative re activity can be supplied by the combined contribution of the GEM and Doppler f eedbacks. Assuming a loss of primary coolant flow, the GEMs provides can provide $0.52 of ne gative reactivity. The Doppler feedback can provide an additional $0.0014 per degree increase in fuel temperature.

PAGE 160

160 Table 4-3. Core reactivity worth for indepe ndently separate reactiv ity feedback effects BOEC Difference from Normal EOEC Difference from Normal Normal Steady State (all rods out) $3.79 --$0.0-All Coolant Voided (all rods out) $12.13 -$8.34$8.60-$8.60 Doppler (+100K) (all rods out) $3.65 $0.14-$0.15$0.15 Normal Steady State (shim rods in) $0.0 -n/a n/a All Coolant Voided (shim rods in) $8.62 -8.62 n/a n/a Doppler (+100K) (shim rods in) -$0.14 $0.14 n/a n/a Values Below This Point Taken with All Rods Out GEM Voided $3.27 $0.52 -$0.50 $0.50 Radial Expansion (2.5%) -$0.35 $4.14 -$4.17 $4.17 Radial Expansion (5.0%) -$5.01 $8.80 -$8.88 $8.88 Axial Expansion (2.5%) $1.82 $1.97 -$2.01 $2.01 Axial Expansion (5.0%) -$0.71 $4.50 -$4.58 $4.58 Axial Fuel Expansion During the IFR program, axial fuel expansion of Pu-U-Zr (and U-Zr) EBR-II fuels were tested at the Transient Reactor Test (TREAT) facility. The resu lts of these tests are summarized by Rhodes et al [ 68]. For these tests, fuel pins were fi rs t irradiated to va rying burnups in the EBR-II. These irradiated fuel pins were then su bjected to transient overpower (TOP) at TREAT. These power levels were in excess of four tim es their nominal power level in EBR-II. All overpower transients caused extensive melting in th e test fuel which amounted to one-half of the original fuel inventory. The prefailure data from these tests indicated that the metallic fuel slugs expanded axially in excess of the 1% attributed to purely thermal expansion. Rhodes et al attributed this additional expa nsion to swelling caused by diss olved fission product gasses being liberated by the partial melting of the metallic fuel. Post test examination of the fuel pins that did not exhibit cladding rupture sh owed large bubbles in the fuel that formed from fission gasses which coalesced into voids as the fuel partially me lted and then expanded. It is this expansion of

PAGE 161

161 coalesced gas bubbles that caused the fuel to expand during the TOP. The level of axial expansion was later quantif ied as a function of: The amount of molten fuel cr eated during the overpower The concentration of fission gas made available when the fuel melted The initial bubble size due to surface tension effects The pressure of gas in the fuel pin gas plenum resisting the expansion The results obtained by Rhodes et al showed th at one of the test pins achieved an axial elongation of 3.7%. Also, the axial expansion co uld be well predicted using a simplified model where only gas trapped in the bubbles was available for expansion. The results of these tests are promising from the standpoint of reactor control. If these models can be developed to a higher fidelity, they might be used to optimize the axial expansion to meet the TOP scenarios for a given SFR desi gn with a safety margin to cladding breach. Thermal expansion and fuel assembly bowing was used to validate the safety case for future ALMRs during the IFR program. This passi ve feedback capability was demonstrated in the Shutdown and Heat Removal Te st (SHRT) conducted at EBR-II [ 69,70]. The SHRT tests dem onstrated passive reactor shut down using fuel expansion a nd natural circulation during two LOCA scenarios: an Unprotected Loss of for ced circulation Flow (ULOF) as well as an Unprotected Loss of Heat Sink (LOHS) [ 71,72]. Radial Fuel Bowing Radial expansion of the core through fuel assembly bowing is a phenomenon that is highly coupled between the core physics of the TOP re activity insertion (either by LOCA or other accident initiators) and the thermal-hydraulic an d thermo-mechanical behavior of the sodium coolant and fuel assemblies. The bowing of fuel assemblies in the radial direction is caused by lateral temperature gradients across the crosssectional area of the fuel assembly. The temperature gradient across the fuel assembly is related to the power gradient in the reactor

PAGE 162

162 during the transient. The temperature gradient ca uses differences in the axial thermal expansion of the assemblys HT-9 duct wall from one side of the fuel assembly to the other. It is this difference that induces a bending moment on the lengt h of the fuel assembly. Assuming the core temperature gradient during the transient is pos itive for decreasing radius (core hotter at the center), the bending moment pushes the fuel assemb lies outwards. At a fundamental level, the amount of bowing created by these temperature gradients is roughly ap proximated by treating the assembly as a tubular beam with an appl ied bending moment. Varying complexities of calculations of this type have been implemented in computer codes since the 1970s to try to accurately predict fuel bowing. However, thes e codes to date have all assumed steady state conditions and were typically benchmarked against cr itical pile simulations (also at steady-state) using perturbation theory [ 73,74,75, 76]. Fuel bowing was used to enhance the passive safety of the EBR-II. The EBR-II fuel assembly was located in the core by two closely placed lower grid plates. To minimize radial expansion below the mid-plane of the core, a me tal button was incorporated onto the surface of each six of the hexagonal sides approximately ha lf-way up the length of the fuel assembly. These buttons held the fuel assemblies tightly in place below the mid-plane. EBR-II did not have an upper grid plate. The combination of the buttons and the two lower grid plates effectively cantilevered the fuel assembly below the core mid-plane. If a strong bending moment were to be induced on the fuel assembly, the lack of an upper grid plate would allow the fuel to blossom (analogous to a flower) at the top of the core. This bl ossoming effect was an integral feature of the passive safety at tributes of the EBR-II. The E BR-II passive reactivity feedback by expansion forces was demonstrated by the SHRT te sts. Similar negative reactivity feedbacks for

PAGE 163

163 fuel assembly bowing have also been verified in critical pile tests performed at steady state conditions by various groups [ 73]. Despite the success of these tests, the transien ts initiated were dem onstrated on fairly long time periods compared to the neutron lifetime in a fast reactor. Due to the small delayed neutron fraction in fast reactors compared to thermal reac tors, the neutron lifetime is much shorter and is in the range of tenths of micro-seconds. If a void were instantaneously introduced into the core for some hypothetical situation, the reactivity insertion and re sulting TOP could possibly occur on a timescale much faster than fuel bowing could occur. However, it is important to realize that the total core void worth cal culations in Table 4-3 were performed assuming that sodium was onl y voided in the core and not the surrounding reflectors. The likelihood of such a hypothetical reactivity insertion from void formation in the fueled region alone may not be probable. If this is the case, this hypothe tical void insertion may be considered in the realm of core disruptiv e accidents (CDA) and may or may not require a robust reactor design for compensating for this remote possibility [ 77]. If such a robust design is requir ed, then it is outside the realm of this dissertation work due to the fact that the ABR designs currently under c onsideration also have large positiv e void induced reac tivity insertions for this accident event (Table 3-13). It should be briefly mentioned that the c oolant outlet temperatur e of the current ABR design is approximately 780 K. The boiling point of sodium (at atmospheric pressure) is 1156 K. If a pool type reactor with the bulk sodium at or just above atmospheric pressure is assumed, the temperature margin before sodium boiling occurs is roughly 370K. Sodium has a large thermal conductivity (60 W/m-K) and fairly large heat cap acity (1,250 J/kg-K). Therefore, a pool type design could offer a large thermal momentum whic h minimizes the rate of increase in the core

PAGE 164

164 coolant temperature. If a l oop type design is used, the pressurization of the sodium could increase the margin to boiling. However, a loop design would possibly have to sacrifice the large thermal momentum offered by a large sodi um pool. The decision for pool or loop type SFR reactor cooling strategies is currently the subject of much debate in the SFR community. Both methods have their own pros and cons. Th e choice for a pool or loop type reactor does not play a role in the transmutation performan ce of the reactor and doe s not enter into the calculations of this dissertation. For the purpose of discussion in the fuel performance discussion in Chapter 6, a pool type design is assumed. Technetium Transmutation Rate When Tc-99 absorbs a neutron, it transmutes into Tc-100 which decays with a half-life of 15.86 seconds by beta particle emission into a stab le atom of Ru-100, ruthenium. Because, Ru100 is followed by two more stable isotopes of ruthenium on the chart of the nuclides, it is unlikely that successive neutron capture from the original Tc-99 transmutation will produce a significant mass of radioactive material. Therefor e, it can be assumed that Tc-99 transmutation by neutron capture will remove the radiotoxicity associated with Tc-99 from the fuel cycle. Hence, the technetium shim rod, though not a true transmutation target such as that proposed by Yang et al, performs a secondary purpose as a Tc-99 burner. Given that some Tc-99 material is inserted into the reactor to some extent in the form of the shim rods, it is expected that some Tc-99 will be destroyed as a function of the amount of rod that is inserted into the core. To evaluate the rate at which th e shim rod Tc-99 is destroyed, a REBUS calculation was performed for a single cy cle (non-equilibrium depletion mode) with the shim rods fully inserted into th e core throughout the enti re irradiation. Using this simulation, the overall depletion rate of Tc-99 was computed.

PAGE 165

165 This depletion rate was then di vided by the entire length of th e shim rod to find the average Tc-99 depletion rate per length of rod. Thus, the average Tc-99 depletio n rate per length of control rod is: 1.4E-6 kg/MWD per cm or shim rod. If it can be assumed that the amount of excess reactivity is linear as a function of burnup, than it can also be assumed that the rate that the shim rod is removed is also a linear function of irradiation time. For a shim rod length of 55 cm and a cycle length of 214 days, the shim rod re moval rate is: 0.26 cm/EFPD. Integrating the Tc-99 depletion rate per unit length of rod over th e entire cycle length then gives the total Tc-99 consumption rate of the core accounting fo r the shim rod movement. Thus, the Tc-99 consumption rate of the shim rods is: 0.0143/MWY. A summary of results on Tc-99 consumption by the AHFTR is given in Table 4-4. Table 4-4. Summary of Tc -99 consumption by the AHFTR Tc-99 Depletion Rate per Unit Length of Shim Rods (kg/MWY/cm) 1.4E-6 Shim Rod Removal Rate (cm/EFPD) 0.26 Cycle Length (EFPD) 214 Tc-99 Consumption Rate by Shim Rods (kg/MWY) 0.0143 Tc-99 Production Rate by Fissions in the Fuel (kg/MWY) 0.0053 Tc-99 Net Consumption Rate by the AHFTR (kg/MWY) 0.0089 Tc-99 Production Rate of reference PWR fuel (kg/MWY) 0.0074 Using the TRITON calculation of the referenc e UOX PWR fuel assembly, which was used to generate the isotopic vector of the ABR and AHFTR external feed, the rate of Tc-99 produced by the LWR fleet was found to be: 0.0074 kg/MWY. Therefore, the AHFTR shim rods can consume approximately 1.9 times more Tc-99 th an that produced by a PWR for equal amounts of energy produced by each reactor type. In othe r words, the shim rods associated with one megawatt of installed AHFTR capacity can cons ume the Tc-99 produced by 1.9 megawatts of installed PWR capacity. Despite the attractiveness of burning Tc-99 in the shim rod system of the AHFTR, these calculations up till now have not reflect the Tc -99 produced by the ABR or the AHFTR in the

PAGE 166

166 fast reactor fleet. This is partly because, unlike the UREX + aqueous reprocessing technology, pyroprocessing does not currently offer a separation strategy for Tc-99. However, if Tc-99 separation were possible with pyroprocessing the net Tc-99 consumption of the entire AHFTR core can be found. The AHFTR fuel produces approximately 0.0053 kg/MWY of its own Tc-99 through fissions of fu el atoms. Therefore, the net destruction of Tc-99 by the AHFTR is only: 0.0 143 kg/MWY 0.0053 kg/MWY = 0.0089 kg/MWY. Hence, if it were possible to re cover the Tc-99 produced by fission, the support ratio of PWRs to AHFTRs would be reduced to 1.2.

PAGE 167

167 CHAPTER 5 DIFFUSION VERSUS TRANSPORT BENCHMARKS A benchm arking effort is conducted to eval uate the validity of the diffusion theory calculation used for the core physics and transpor t analysis. The Variat ional Anisotropic Nodal Transport (VARIANT) code is an add on mo dule that comes with the current publicly available version of the DIF3D code [ 78]. The VARIANT spherical harm onics treatment can be applied to the hexagonal-z nodal discretizati on which describes the AHFTR geometry. VARIANT can be invoked with minimal cha nges to the DIF3D input. Therefore, the equilibrium cycle calculation performed by REBUS is re-evalu ated using the VARIANT option of the DIF3D code. Due to the inherent archaic memory allocati on structures availabl e in the DIF3D/REBUS system, only three moments of the angular flux (P 3) were calculated. Also, because of these memory restrictions, the number of energy groups was reduced from 33 to eight. Therefore, the amount of energy information is reduced in lieu of increased angular information. Because of these limitations, the MCNP code was used to evaluate the axial and radial spatial flux profiles as well as the driver and target energy sp ectrums. A short FORTRAN processing code was written for copying the batched homogenized isotopi c number densities from each core region in DIF3D/REBUS model into the MCNP model. The corresponding hexag onal-z nodal geometry scheme used in the DIF3D model was al so preserved in the MCNP model. Also tested in this chapter is the spatial and energy shielding treatment used in the MC2-2 cross section collapsing algorithm. The collapsed cross sections (group constants) used for the deterministic calculations are generated by comp letely homogenizing the pin and fuel assembly geometry detail into a representa tive zero-dimensional infinitely dilute mixture. One of these mixtures is defined for every region in the co re sharing a similar is otopic composition and,

PAGE 168

168 hence, neutron spectrum. These regions include th e: inner, middle, outer enrichment zones and the axial target region above the active core. Also, a mixture is created for the shield and reflector compositions. Using a cr itical buckling search calculation, the group constant library is generated for each zero-dimensional mixture in separate calculat ions performed by MC2-2. Therefore, the spatial shielding and the region-t o-region neutron shadowing effects in the reactor core are not accounted for during the cross section gene ration. This task would generally be performed in a lattice calculation at the pin or assembly level for most thermal spectrum applications before using the group cons tant set in a core simulator. For fast reactors, the mean-free-path is significantly longer than in thermal spectrums allowing for the homogenization over the fuel assemb ly. However, the axial targets are slightly moderated to an epithermal or hard-epithermal spectrum. Thus, it is important to test whether or not spatial shielding occurs within the target fuel rod placed adjacent to a moderator rod. A unit cell calculation of one zirconium hydride rod surrounded by six targ ets with reflective boundary conditions is modeled using the MCNP code. The fl ux for one of these six targets is evaluated as a function of radius going into th e fuel slug. This radial dist ribution is determined by dividing the fuel slug into five equal volume zones a nd tallying the neutron flux entering each of these zones. The region-to-region shadowing effect is also evaluated using the MCNP code. The six pins from the cell calculation are homogenized into an annular region of equal volume surrounding the zirconium hydride pin. Shadowing is determined by tallying the neutron flux as a function of radial distance away from the center of the annulus. DIF3D, VARIANT and MCNP Methods VARIANT solves the multigroup steady-state ne utron diffusion and transport equations in two and three dimensional Cartesian and hexagonal geometries using variational nodal methods.

PAGE 169

169 Anisotropic scattering is treated in these calcul ations. However, for the coupling of VARIANT with REBUS, it was found out that the REBUS code is ill-equipped to accept the higher order scattering matrices in the microscopic cross s ection data file format (ISOTXS) provided by the MC2-2 code. ISOTXS is a binary data file fo r transferring microscopic cross section data between codes by different authors which uses the standardized format adopted by the Committee on Computer Code Coordination (CCCC). The difficulty of REBUS to process the higher order scattering data in ISOTXS is caused by the method in which REBUS homogenizes fuel isotopic number densities over each region of the core (i.e., Inner, Middle, Outer Core and Targets). Recalling Figure 2-1, REBUS homogenizes fuel compositi ons of all fuel batches independent of depletion stage within a user defined same spectrum region. This new region averaged number density is then used in the physics calculation to generate the flux in that region. This flux is then multiplied by the original un-homogenized number densities at each region to get the reaction rate at that location in the core. These reaction ra tes are then applied to the original isotopic compositions of each fuel batch at each stage within that region to perform the depletion process. Therefore, the exact position of each individu al fuel assembly within the core is not tracked. It is only necessary to track the mass of each batch as it moves from region to region as a function of cycle or stage number of its depletio n. For many fast reactor core designs (ABR, AHFTR, etc.), the fuel is not shuffled. Instead, a fuel assembly typically resides in the same location in which it was originally loaded from BOL through EOL. Because the fuel is not shuffled, the core can be divided into se parate enrichment zones throughout the core.

PAGE 170

170 For this reason, even though the MC2-2 code provides the high er order scattering data, REBUS does not have the data stor age structures to receive it. The creators of VARIANT made substantial modifications in DIF3D to handle the input of anisotropic cro ss sections. However, the VARIANT module expects this data to be input in macroscopic data file format (COMPXS). Very little formatting changes were made fo r the absorption and fission principal cross sections, which can be in ISOTXS or COMPXS form. Nevertheless, REBUS can only accept the ISOTXS format due to the homogenization routin e. Logistically, it is impossible to input COMPXS formatted data into REBUS because the nature of the equilibrium search manipulates the atom densities of the fresh fuel compositi on (i.e., TRU enrichment) in order to meet the constraints of equilibrium cycle length and bur nup. Because of these limitations, the higher order approximation of the scattering cross se ction had to be dropped for the coupling between VARIANT with REBUS. Flux Spectrum Analysis A P3 approximation of the flux and current are performed by VARIANT but with zero order scattering in the PN equations. To this end, MCNP is used to compare the BOEC radial and axial flux distributions as well as neut ron energy spectrum with the DIF3D/REBUS and VARIANT/REBUS results. Figure 5-1 shows the flux spectrum fo r the AHFTR mid-plane at the inner most row of fuel. It is important to note the close agreement between the MCNP and the DIF3D spectrum. The major observable difference between the two curves is the general magnitude of the flux which can be seen to be centered on 0.25 MeV. The flux depression at 0.25 MeV is a result of several well resolved but overlap ping sodium resonances in th at energy range. Note the flux depression at 30 keV which corresponds to the well resolved iron capture resonance at that energy. The VARIANT spectrum does not capture the true shape of the energy distribution due

PAGE 171

171 to the eight energy group resoluti on. Figure 5-2 shows the flux sp ectrum for the target zone for the innermost row of fuel. 0.0E+00 2.0E+14 4.0E+14 6.0E+14 8.0E+14 1.0E+15 1.2E+15 1.4E+15 1E-051E-041E-031E-021E-011E+001E+01Energy (MeV)Flux (cm^-2*s^-1)/lethargy Dif3D VARIANT MCNP Figure 5-1. Neutron energy spectrum for the AHFT R at the core mid-plane for the inner most row of driver fuel (taken at BOEC) 0.0E+00 1.0E+14 2.0E+14 3.0E+14 4.0E+14 5.0E+14 6.0E+14 1E-051E-041E-031E-021E-011E+001E+01Energy (MeV)Flux (cm^-2*s^-1)/lethargy Dif3D VARIANT MCNP Figure 5-2. Neutron energy spectrum for the AHFTR at the target region located directly above the inner most row of driver fuel (taken at BOEC) Na-23 Fe-56 Na-23

PAGE 172

172 Here again, the DIF3D spectrum shows good agr eement with the MCNP spectrum. In fact, the difference in overall magnitude is less than in the driver fuel. Note that because of the epithermal spectrum, the sodium resonance at 3 keV is much more pronounced than in Figure 51. However, much of the flux is still in the epithermal-to-fast energy range such that heavy metal resonances dominating from 0.1 eV through 100 eV have very little importance. For a later discussion, it appears that the resonance shie lding of the lighter metal elements (comprising the structure and coolant), with resolved resonances between 1 keV and 1 MeV, have the most affect on the flux. The difference in total flux magnitude betw een the three calculation methods may be attributed to an increasing accuracy of th e angular treatment provided by VARIANT and REBUS. VARIANT provides a better approximati on of the angular distribution of the flux and transport cross section. MCNP provides full anisotropic treatment of scattering as well as the flux gradient as opposed to the diffusion approximation made by DIF3D or the isotropic scattering assumption made with VARIANT. Spatial Flux Analysis The diffusion approximation under predicts the curvature of the axial and radial flux gradient [ 79]. This is observed in the axial and radi al f lux profile for the AHFTR. The axial flux profile for the inner most row of fuel is given in Figure 5-3. The erro r bars for two standard deviations of the flux in the MCNP calculation are shown in the plot but are of the same size as the plotted data point. Notice the peak flux o ccurring at the mid-plane is less for the DIF3D calculation than the other two curves. The total flux calculated by VARIANT is actually much closer to the MCNP result than the DIF3D result. This indicates that the fewer energy groups used in the VARIANT calculation was a valid simplification even though the fine spectral

PAGE 173

173 resolution of the flux energy spectrum is lost. Fi gure 5-4 shows the radial distribution of the flux at the core mid-plane. 2.0E+15 2.5E+15 3.0E+15 3.5E+15 4.0E+15 4.5E+15 5.0E+15 0102030405060708090Axial Distance from Core Bottom (cm)Total Flux (cm^-2*s^-1) DIF3D VARIANT MCNP Figure 5-3. Comparison of the total BOEC flux axial profile fo r the inner most row of fuel 1.5E+15 2.0E+15 2.5E+15 3.0E+15 3.5E+15 4.0E+15 4.5E+15 5.0E+15 020406080100120140Radial Distance from Core Center (cm)Total Flux (cm^-2*s^-1) DIF3D VARIANT MCNP Figure 5-4. Comparison of the to tal BOEC flux radial profile cal culated at the core mid-plane The differences in the mid-plane radial fl ux distribution between calculation methods are negligible for the inner regions of the core. However, for increasing radial and axial distance from the center, some differences in the flux gr adient become observable. The target region radial distribution (Figure 5-5) shows notably more difference for regions nearest the center of

PAGE 174

174 the core. These differences may be explained by the differences in angular treatment between the three methods. It is interes ting to note that the ta rget flux calculated by VARIANT is greater than the MCNP calculation for the inner rows of targ ets. This may be attributable to the lack of anisotropic expansion of neutron scattering used for this calculat ion. The anisotropic scattering off of target hydrogen could cause more neutr ons to pass through the target without being absorbed. Not having this anisotropic effect in the VARIANT calculation causes some of the neutrons to be reflected back into the target and active core regions. Radial Flux Distribution in the Targets1.5E+15 1.6E+15 1.7E+15 1.8E+15 1.9E+15 2.0E+15 2.1E+15 2.2E+15 2.3E+15 2.4E+15 2.5E+15 02 04 06 08 01 0 0Radial Distance from Core Center (cm)Total Flux (cm^-2*s^-1) DIF3D VARIANT MCNP Figure 5-5. Comparison of the total BOEC flux radial profile calculated for the axial target region The DIF3D curve under pred icts the flux given by both MCNP and VARIANT. Therefore, it is expected that th e MA destruction rate should be on the conservative side of the full transport flux evaluation. This result can be seen in Table 51. As expected, the rate of transmutation is higher for VARIANT than it is for the DIF3D calculation. However, all of the values in the comparisons made by Table 5-1 are in close agreement with each other. Therefore, the use of the diffusion theory is deemed as an acceptable solution to the transport equation for fast reactor core and epithermal target physics an alysis. It is important to make the connection

PAGE 175

175 that there is very little impact on the fuel cycle performance evaluation when using diffusion or transport theory to calculate the AHFTRs reactor and transmutation performance. Table 5-1. Select reactor parameters for the AHFTR given by DIF3D and VARIANT. DIF3D VARIANT BOEC k-eff 1.012377 1.013683 tCR 0.7180880.715677 Cycle Length (EFPD) 213.957 212.078 Inner Core Enrichment 20.77% 20.70% Am-241 Transmutation Half-Life (EFPY) 2.31 2.26 Am-241 Consumption Rate (kg/EFPY) 2.12E+01 2.14E+01 MA Consumption Rate (kg/EFPY) 4.00E+01 4.04E+01 Differences between BOEC and EOEC Fluxes As discussed in the Chapter 2, the relative or percent change in the initial fissile concentration in the fuel as a function of burnup is small. This is the fissile requirement of the core is high due to the high leakage (i.e., geomet ric buckling). Therefore, it should be expected that the neutron spectrum in the core is fairly in sensitive to the fissile at om depletion. Figure 56, shows the neutron spectrum at the mid-plane and target level in the first row of fuel in the AHFTR. 0.0E+00 2.0E+14 4.0E+14 6.0E+14 8.0E+14 1.0E+15 1.2E+15 1.4E+15 1E-051E-041E-031E-021E-011E+001E+01Energy (MeV)Flux (cm^-2*s^-1/lethargy) BOEC Active Core Region EOEC Active Core Region BOEC Target Region EOEC Target Region Figure 5-6. Comparison between the BOEC and EOEC neutron spectrums for the inner most row of fuel (DIF3D)

PAGE 176

176 Notice that there is a slight change in the magnitude of the flux between BOEC and EOEC. This difference is attributed to the slight change in the average fissile concentration in the core between BOEC and EOEC from the fuel burnup. It should be expected that the average flux in the core should generally increase as the fissile material is being decr eased by depletion due to the fact that generally number density and fl ux are inversely proportional to each other. To illustrate this point, the radial flux distribution at the core mid-plane is plotted for both BOEC and EOEC in Figure 5-7. The axial flux profile is plotted for the innermost row of fuel for both BOEC and EOEC in Figure 5-8. 1.5E+15 2.0E+15 2.5E+15 3.0E+15 3.5E+15 4.0E+15 4.5E+15 5.0E+15 020406080100120140Radial Distance from Core Center (cm)Total Flux (cm^-2*s^-1) BOEC Active Core Mid-plane EOEC Active Core Mid-plane BOEC Target Region EOEC Target Region Figure 5-7. Comparison between the BOEC and EOEC radial fl ux profiles (DIF3D) Depletion Test To check the affect of the difference in flux values on the results of the REBUS depletion calculation, a benchmark with MONTEBURNS was performed to compare the k-eff as a function of irradiation time. In equilibrium-mode, the REBUS code perfor ms a minimum of two DIF3D calculations: BOEC and at EOEC. The BOEC calculation is perf ormed to determine the flux values needed to deplete the fuel. The EOEC calculation is performed to determine whether or not the

PAGE 177

177 uncontrolled k-eff has converged to one, indicating that the code has re ached the end of the irradiation (Figure 2-1). A middle-of-equilibri um-cycle (MOEC) calculation is performed to update the flux values during the depletion to ensure that any spectrum changes, though small, will be reflected in the depletion. Due to the small change in flux spectrum and intensity throughout the irradiation (Figure 5-6, Figure 5-7 and Figure 5-8) this three-point check is usually acceptable. The limitations of the RE BUS code in equilibrium-mode, only allows a maximum of two DIF3D calculations to be perf ormed between BOEC and EOEC. For the many core design calculations performed in this dissertation, using the e quilibrium-mode, only one MOEC calculation was performed between BOEC and EOEC. This decision was made in order to minimize computational time. 2.0E+15 2.5E+15 3.0E+15 3.5E+15 4.0E+15 4.5E+15 5.0E+15 0102030405060708090Radial Distance from Core Center (cm)Total Flux (cm^-2*s^-1) BOEC First Row of Fuel EOEC First Row fo Fuel Figure 5-8. Comparison between BOEC and EOEC axial flux profiles (taken at the inner-most row of fuel) (DIF3D) To check the validity of this three-point check, a non-equilibrium-mode REBUS calculation was performed using the geometry and atom density data extracted from the equilibrium-mode output. In non-equilibrium mode the REBUS code allows an arbitrary number of DIF3D calculations to be performed between BOC and EOC. Therefore, the BOEC atom

PAGE 178

178 density and geometry used in the previous st andalone DIF3D calculati ons (Figure 5-1 through Figure 5-8) of this chapter was also used for BOC values in a representative non-equilibrium coupled DIF3D/REBUS calculation. In a similar manner, the atom density and geometry data used in the previous standalone MCNP calculati ons of this chapter wa s applied to a coupled MCNP/MONTEBURNS calculation. Figure 59 shows the k-eff results from the nonequilibrium DIF3D/REBUS calculation and the MCNP/MONTEBURNS calculation. 1 1.005 1.01 1.015 1.02 1.025 1.03 1.035 1.04 1.045 1.05 0 50 100 150 200 250Effective Full Power Dayk-eff REBUS MONTEBURNS Figure 5-9. Reactivity curve compar ison between REBUS and MONTEBURNS There is a bias of approximately 30 millik-eff between the REBUS and MONTEBURNS calculations. Part of this difference can be attributed to the difference in the diffusion approximation made by DIF3D and the transport calculation performed by MCNP. However, much of the difference exists due to a discre pancy between fission product masses at BOC in the DIF3D and MCNP models. This discrepancy is caused by the MC2-2 cross section library (used by DIF3D and REBUS) having more fission product isotopes than the standard ENDF libraries available to MCNP and MONTEBURNS. MC2-2 also uses ENDF libraries to produce the group Note: Error Bars have two sigma confidence

PAGE 179

179 constants used in DIF3D/REBUS. However, the reason MC2-2 has more ENDF isotopes than MCNP is because almost all of the fission produc ts in the comprehensive ENDF database are important to fast reacto r calculations. In general, all fiss ion products have approximately the same cross section value in the unresolved reso nance range. However, not all of these fission products have appreciable value in the thermal ra nge. Therefore, for thermal reactor calculations these extra fission product isotopes are not needed and do not come standard with the publicly available release of MCNP. There is also a slight difference in the sl ope of the reactivity curve between REBUS and MONTEBURNS. The smaller MONTEBURNS slope is explained by the difference in fission product isotopes, which are generated by the ORIGEN part of MONTEBURNS, which do not exist in the MCNP cross secti on library. When these isotopes are generated by the ORIGEN code but can not be recognized by the MCNP code, MONTEBURNS simply decides not to continue tracking these isotopes in the depletion calculation. Th erefore, the neutron absorption of these dropped fission products are not reflected in the MONTEBURNS reactivity curve causing the shallower slope. Before REBUS was adopted for fuel cycle anal ysis in this work, a benchmark calculation was performed between REBUS a nd MONTEBURNS using the Adva nced Burner Test Reactor (ABTR) design as a reference for fuel composition and geometry [ 80]. The ABTR is a smaller prototypic version of the larger ABR design, proposed by ANL as a proof of concept reactor with a prim ary function for materials testing. The ABTR benchmark started with a fresh fuel composition and not one obtained from an equilibrium cycle. The results of this fresh core benchmark calculation are given in Figure 5-10.

PAGE 180

180 Reactivity curve 1.25500 1.26000 1.26500 1.27000 1.27500 1.28000 01836547290108126144162180 Cycle Lenght (days)k-eff REBUS-BOC REBUS-MOC REBUS-EOC Monteburns Linear (Monteburns ) note: Error bars have two sigma confidence Figure 5-10. Advanced Burner Test Reactor benchmark using fresh core fuel composition showing the gradual departure of MONTEBURNS from REBUS1 This benchmark was conducted i nternally by the members of the INL fuel cycle analysis team (which the author is a member of) to id entify the practicality of using REBUS (an ANL code) for fast reactor calculations. Th e non-equilibrium REBUS and MONTEBURNS calculations in the ABTR benchmark showed ve ry good agreement with each other at BOC. However, as fission products were dropped from the MONTEBURNS calculation, the EOC result by MONTEBURNS diverged from the REBUS calculation. Spatial Self-Shielding Test The method by which the group constants are generated in MC2-2 is consistent with zero dimensional slowing down theory techniques th at were in common use at the time the MC2-2 code was developed. The use of such methods is generally considered an acceptable practice for fast reactor core simulation due to the fast neutron mean-free-path being generally larger than the 1 The non-equilibrium REBUS calculations were performed using BOC, MOC Middle-ofCycle and EOC isotopics to generate the MC2-2 cross section library.

PAGE 181

181 pin-cell. There is little change in the neutron spectrum over the local space domain due to this long mean-free-path. Therefore, the use of corre ctions of the flux at node, cell or the boundaries between different fuel compositions such as discontinuity factors, is not applied in the DIF3D or VARIANT codes. However, inco rporation of moderati ng pin-cells within th e target geometry raises the question of the applicabil ity of such generic assumptions. An MCNP sub-lattice model of one zirconium hydride pin su rrounded by six target pins was created to represent an equivalent lattice calculation for this arrangement (Figure 5-11). The neutron flux was tallied in five equal volume zone s within the zirconium hydride pin. Similarly, the fuel slug of one of the six targets was sub-divided into five equal volume tally zones. The 33 energy bin group structure used in the MC2-2 calculation was used to tally these fluxes as a function of energy. The height of this MCNP model is 20 cm. Reflective boundary conditions were used on all sides to simulate a repeating lattice of this geometry. It should be noted that the geometry shown is Figure 5-11 is not a true lattice calculation. This is because the geometry can not be simp ly folded over at the problem boundary condition into a hexagonal lattice. A B Figure 5-11. MCNP sub-lattice representation (A) of a repeat ing pin-cell arrangement (B) Five Equal Volume Tally Zones

PAGE 182

182 The picture in the right-handside of Figure 5-11 shows how the seven pin arrangement would actually look within a hexagon al lattice. Nevertheless, it will be seen that the geometry shown can be used as a valuable analysis tool for evaluating the neutron spectrum variations within the repeating lattice. Figure 5-12 and Figure 5-13 shows the 33 group energy spectrum as a function of the five equal volume tally zones for both the zirconium hydride and target slugs respectively. These plots show a similar spectr um shape and magnitude to that shown in Figure 5-2. Indeed both of these plots are very simila r to each other. The small change in neutron spectrum as a function of penetr ating depth into the moderator is similar to that found by Konashi et al [ 81]. Therefore, there is virtually zero sp atial shielding as a function of fuel slug radius. Because of this uniform irradiation, the target slug should not experience a non-uniform radial burnup distribution or rim-effect. The uniform irradiation across the fuel slug ra dius can be attributed to the epithermal mean-free-path in this target spectrum to be grea ter than that of the pin diameter in the seven pin model of Figure 5-11. The hexago nal flat-to-flat dimension of th e model is 2.31 cm, whereas the neutron mean-free-path (for all reaction types in cluding scattering) is 2. 97 cm. Note the fuel slug outer diameter is 0.557 cm. As a comparis on, the mean-free-path calculated for a generic PWR IMF fuel assembly (Table 1-2) is 2.34 cm, whereas the lattice pitch and pellet diameter is 1.26 and 0.82 cm respectively. Spatial Shadowing Test Now that the irradiation distribution in the pelle t has been analyzed, th ere is still one final check in the applicability of the infini tely homogeneous approximation used by MC2-2. The region-to-region neutron shadowing effect is evaluated by modifying the seven pin sub-lattice model shown in Figure 5-11. A new MCNP unitcell model is created by homogenizing the six target pins into an annulus that surrounds the zirconium hydride pin (Figure 5-14).

PAGE 183

183 2. 0 7E-0 7 1 8 2E -05 2.26E0 4 2.69E 03 3.28 E-0 2 4. 0 0E -01 4.87E + 00 0.062 0.150 0.196 0.232 0.264 0E+00 1E+14 2E+14 3E+14 4E+14 5E+14 6E+14Flux(cm^-2*s^-1)/lethargyEnergy(MeV) Radius (cm) Figure 5-12. Neutron spectrum as a function of energy and the zirconium hydride slug radius (MCNP) 2.07E-07 1.82E-05 2. 2 6E04 2. 6 9E -03 3.28E0 2 4.00E-01 4. 8 7E +00 0.062 0.150 0.196 0.232 0.264 0E+00 1E+14 2E+14 3E+14 4E+14 5E+14 6E+14Flux(cm^-2*s^-1)/lethargyEnergy(MeV)Radius (cm) Figure 5-13. Neutron spectrum as a function of energy and the target slug radius (MCNP) The inner and outer bounds of this annulus ar e the nearest and furthest points of the original six targets from the zirconium hydride pin. The HT-9 cladding and sodium bond and coolant are also homogenized with the fuel slug into the annuluss composition. A white albedo boundary condition was used at the an nulus outer perimeter. The white boundary condition replaces the incident angular dependence with an isotropic distribution. This is

PAGE 184

184 different from a reflective boundary condition which reflects neutron flux with a reflection of the angular distribution that is incident upon it as if it were a perfect mirror. Figure 5-14. MCNP unit-cell model with hom ogenized fuel annulus: zirconium hydride (turquoise), sodium bonded gap (olive), HT-9 cladding (yellow), sodium coolant (red), homogenized fuel annulus (blue) The white boundary condition was used for this problem to give the annulus a representative flux at the boundary to that of a he xagon despite using the circular shape. Similar to the sub-lattice geometry from Figure 5-11, the annulus is divide d into five equal volume zones for tallying the neutron flux. Figure 5-15 shows the 33 group neutron spectrum as a function of radius for these flux tally zones. As can be seen from Figure 5-15, there is ve ry little difference in the neutron spectrum flux magnitude as a function of radial distance from the zirconium hydride pin. This is expected because the epithermal mean-free-path discussed in the previous section is on the dimensional level of the annulus which is filled with mostly sodium. The annulus outer diameter is 2.53 cm compared to a mean-free-path of 2.97 cm. This is compared to the LWR IMF case which has a mean-free-path of 2.34 cm which is roughly twic e the rectangular pitch of 1.26 cm. Note, the 0.727 cm 1.265 cm 0.279 cm

PAGE 185

185 volume fraction of water in a PWR pin-cell is approximately 65 v/ o, whereas the volume fraction of ZrH1.6 in the annular pin cell is approximately 5 v/o. 2 .07E 07 1. 8 2 E -0 5 2 26 E -0 4 2 .6 9 E0 3 3 .2 8 E0 2 4.00E-01 4. 8 7E+00 0.727 0.893 1.032 1.155 1.265 0E+00 1E+14 2E+14 3E+14 4E+14 5E+14 6E+14Flux(cm^-2*s^-1)/lethargyEnergy(MeV) Radius (cm) Figure 5-15. Neutron spectrum within the homogenized annulus as a function of radial distance from the unit-cell origin (MCNP) It is important to distinguish between the m ean-free-path between total interactions versus the mean-free-path between fission events. Tabl e 5-2 indicates the mean-free-path for total interactions versus that for only scat ter, capture or fission reactions. Table 5-2. Comparison of mean-free-paths of different reaction types for various irradiation regions having different spectrums (MCNP) Total Interaction (cm) Scatter (cm) Capture (cm) Fission (cm) ABR (CR=0.5) 4.154.44 386.42 457.24 AHFTR Active Core 3.754.04 339.86 390.53 AHFTR Target Region 2.973.11 166.07 451.81 LWR IMF 2.342.97 21.16 27.89 The loss of energy per collision in sodium is significantly less than that of hydrogen in water. Therefore, without slowing down to ther mal energies, neutrons are more likely to fission at fast energies, where the probability (i.e., cross section) is small compared to thermal energies. Therefore, the average length of travel between fission (or capture) reac tions for a SFR neutron

PAGE 186

186 spectrum is much greater than for an LWR spectrum. This is also true of the slightly moderated target region of the AHFTR as shown in Table 5-2. Calculation Validation Remarks Despite the lack of higher order scatteri ng in the VARIANT/REBUS calculation, close agreement in the neutron spectrum and spatia l flux distribution was found between DIF3D, VARIANT and MCNP. The DIF3D/REBUS an d VARIANT/REBUS BOEC k-effective was 1.012377 and 1.013683 respectively. The MCNP calculation with BOEC isotopic number densities based on the DIF3D/REBUS calculatio n was 1.04071 with a standard deviation of 0.00024. As is seen in the benchmark analysis given in the VARIANT users manual, the full transport evaluation gives a k-eff that is higher than for diffusion by roughly 25 milli-k-eff [ 78]. Naturally, th is bias is not consta nt for any core geometry. It is expected that for increasing core size, the difference between the core average fl uxes of the full transport evaluation and the diffusion approximation would decrease as a resu lt of spatial heterogeneities becoming less significant compared to the overa ll governing core physics. The much higher k-eff obtained by MCNP is a result of the lack of fission product isotopes that were available to the MCNP code. Despite the lack of these isotopes, the MCNP calculations showed similar trends with regards to flux values and a loosely repres entative reactivity curve in the MONTEBURNS calculation. Due to the close agreement in neutron spectrum and spatial distribution between all three calculation methods, little difference was observed in the MA transmutation rate. Therefore, the diffusion approximation is considered an acceptable method for evaluating the fuel cycle and reactor performance analysis of the AHFTR core design. It was also found that the arrangement of moderating and target rods in the AHFTR demonstrate negligible resolved resonance spa tial self-shielding and sh adowing effects. Though neutrons are being slowed down by scatters in the zirconium hydr ide rod, the neutron energies

PAGE 187

187 are still to fast to be considered a true therma l spectrum. Most of the neutrons in the target regions neutron spectrum have energies above 1 00 eV which is at the upper end (or above) the resolved resonance range for e ssentially all heavy metals. At these energies, the resolved resonances of heavy metals are poorly resolved. The flux depre ssions observed in the neutron spectrum of the AHFTR are, in fa ct, a result of the lighter sodium and iron atoms with resolved resonances in the fast range. The nearly fa st epithermal spectrum has a mean-free-path sufficiently greater than the dimensions of th e fuel slug; allowing ne utrons to pass through it without being absorbed non-uniformly in the slug periphery which would give a rim-effect. This uniform irradiation makes it possible to assume infinite dilution in the target region for the group constant calculation performed by MC2-2. Also because the epithermal neutron meanfree-path is on the dimensional le vel of the repeating target-and-moderating rod arrangement, the neutron shadowing effect is negligible.

PAGE 188

188 CHAPTER 6 THE AHFTR FUEL DESIGN The AHFTR design em ploys flattened core ge ometry, epithermal upper axial targets and metallic ternary alloy fuel. The combination of axial targets, combined with the copyroprocessing approach draws from the Integral Fuel Cycle (IFC) strategy demonstrated by the EBR-I and later by EBR-II. In th e IFR program, it was envisioned that both driver blanket fuel would be used to breed fissile plutonium [ 40,82, 83]. Once discharged from the core, driver and fuel assem blies (axial blankets were included as part of the dr iver fuel assembly) would be chopped into small segments. Next the sodium bond occupying the fuel-to-clad gap would be extracted and the remaining chopped cladding hulls and fuel slugs placed in perforated steel baskets. These baskets would be taken to th e electrorefiner for pro cessing. All of these processes were performed in a system of interco nnected hot-cells at a facility adjacent to the EBR-II primary containment building. A more deta iled synopsis of metal fuel electrorefining will be provided in a following section. In the AHFTR design, the moderation effect in the targets suppresses the fission of plutonium isotopes while increasing the capture rate in Am-241. Similar to the Pu-239 conversion from uranium in the IFR blankets, plutonium isotopes are generated by this Am-241 transmutation. The moderating effect enhanc es this transmutation conversion process as discussed in Chapter 3. The combined effect of americium transmutation into Pu-238, and conversion of Pu-239 from the U-238 in the targ et, provides a pluton ium source within the AHFTR fuel cycle analogous to the IFR bla nket fuel. The co-pyroprocessing approach, envisioned by the IFR scenario, is also applicable to the AHFTR because of this plutonium creation. The feasibility of this scenario is validated by the fact that the MA concentration accumulating in the fresh driver fuel can be kept to below 5 w/o (MA/HM) as a guideline

PAGE 189

189 adopted from the CAPRA program. The 5 w/o limit was imposed for the foregoing AHFTR parametric design study in Chapter 3, for the purpose of preserving the fuel irradiation performance established within the IFR experience database for U-Pu-Zr metal alloy fuels. The 5 w/o limit was also considered important to maintain a low MA driver fuel concentration from the standpoint of reactor kinetics. As di scussed in the introduction and Chapter 4, MAs can create an unacceptably high void coefficient due to the resulting spectrum hardening. Fuel Pin Design The neutron trap effect afforded by the modera ted target region allows the neutron leakage leaving through the top of the reactor to be captured in the ta rgets as opposed to being lost. In the absence of the targets, axially leaked neutro ns would leak out of the top of the active core into the gas plenum region which is mostly voi ded space. Placing the axial targets below, in addition to above the driver fuel, is possible and was considered in the co nceptualization of this design. However, including a second target region of equal size to that of the upper region would double the amount of unburned americium discharged to the pyroproce ssor by the targets. Hence, the MA content in the driver fuel w ould roughly double. Remember, from a reactor kinetics and fuel reliability st andpoint, it is desirabl e to maintain the americium content below the 5 w/o limit adopted for this work. Also lower axial targets would act as a neutron trap below the core thus robbing the AHFTR of the reflection necessary for optimizing the cores reactivity requirements. Therefore, a bottom axial reflect or comprised of stainless steel, which is a traditionally common design aspect for SFRs, was adopted. For the AHFTR, S-PRISM and ABR designs, this axial reflector constitutes an HT-9 plug approximately a meter long which comprises the bottom end of the fuel pin. This endplug could either be cast as an integral part of

PAGE 190

190 the fuel cladding or inserted as a separate co mponent and diffusion bonded into the pin. Figure 6-1 shows the general design config uration of the AHFTR fuel pin. Figure 6-1. Conceptual AHFTR fuel pin design In Figure 6-1, the axial target slug is a zirconium metal alloy similar to the driver fuel. The elemental constituents of this alloy are 0. 5Np/9Pu/9Am/1.5Cm/40U/ 40Zr by weight. These weight percents are derived from the ble nding of the Np+Pu, Am+Cm+Bk+Cf and U mass streams produced by the aqueous separations plan t shown in Figure 3-4. As mentioned in Chapter 3, these streams are blended with th e ratios: 10Np+Pu/10Am+Cm+Bk+Cf/40U/40Zr. The SNF TRU isotopic vector was assumed to be that for a typical 17x17 PWR fuel assembly having a discharge burnup of 50 MW D/kg and cooled for five years before being transported to the aqueous separation plant. An additional de cay time of two years was assumed for the time after separation which includes: re processing, fuel fabrication a nd transportation to the AHFTR. The SNF TRU isotopic data was generated by an infinitely repeating fuel assembly lattice calculation using the coupled NE WT/TRITON depletion module of the SCALE5.1 code system. The isotopic vector was given previously in Table 2-1. Target Alloy Selection and Design Considerations The elemental composition of the target was selected, based on the irradiation experience gained by the AFC-1B and AFC-1F tests performed at the Advanced Test Reactor (ATR) at INL [ 84]. The AFC-1B tests consisted of four uran iu m-free metal alloy fuel samples with varying MA, plutonium and zirconium concentrations. Gas Plenum Axial Reflector Driver Slu g Axial Target or ZrH1.6 Slug Sodium Level Bottom Fittin g

PAGE 191

191 A1B1 and A1B4: 48Pu/12Am/40Zr A1B2: 40Pu/10Am/10Np/40Zr A1B3: 60Pu/40Zr A1B5: 40Pu/60Zr The AFC-1F test consisted of four uranium b earing metal alloy fuel samples with varying MA, uranium and zirconium concentration. A1F1: 28Pu/4Am/2Np/66U/30Zr A1F2: 27Pu/3Am/2Np/28/40Zr A1F3: 34Pu/4Am/2Np/60U/20Zr A1F4: 29Pu/7Am/64U The AHFTR target composition is a hybrid compilation of these test compositions: combining a modest amount of uranium (~40 w/o in some cases) and zirconium content (~40 w/o in some cases) and high MA (~10 w/o in most cases) content. The choice to mix fertile and fissile materials in this composition was made with the intent to decrease the expected fission density in the axial target fuel alloy. Metallic fuels experience a swelling incubation period (low rate of swelling) followed by a transition period (high rate of sw elling) as a function of burnup [ 60,85]. The AFC-1 test Post Irrad iation Examination (PIE) perf ormed by Hilton et al. revealed that the time of incubation is proportional to the amount of fission damage impart ed to the overall fuel matrix. Because the zirconium content was varied in these tests, th e total fission energy re leased per unit HM mass (MWD/kgiHM) (iHM stands for initial heavy metal of fuel) or equivalently atom % heavy metal destruction) could not be used as a metric to measure cumulative fission damage. As an alternative, the time integrated cumulative fissions per unit volume of fuel alloy was used instead. As is common in metal fuel irradiation behavior, many of the metal fuel samples experienced fuel-to-clad contact due to swelling. This behavior was also common to metal fuels

PAGE 192

192 tested at EBR-II. The samples that experienced th is behavior were the ones placed in the highest flux and also had the highest plutonium content. However, also similar to EBR-II experience, there was no cladding breach despite this c