ORNL-4119.txt

3 4u5sk p515889 7

ORNL-4119
UC-80 — Reactor Technology

MOLTEN-SALT REACTOR PROGRAM
SEMIANNUAL PROGRESS REPORT
FOR PERIOD ENDING FEBRUARY 28, 1967

LIBRARY LOAN COPV
DO NOT TRANSFER TO ANOTHER PERSON
1f you wish someone else to see this
document, send in name with document
and the library will arrange a loan.

OAK RIDGE NATIONAL LABORATORY

operated by
UNION CARBIDE CORPORATION
for the
U.S. ATOMIC ENERGY COMMISSION

A ORNL-4119
UC-80 — Reactor Technology

Controct No. W-7405-eng-26

MOLTEN-SALT REACTOR PROGRAM
SEMIANNUAL PROGRESS REPORT
For Period Ending February 28, 1967

M. W. Rosenthal, Program Director

R. B.-Briggs, Associate Director
P. R. Kasten, Associate Director

JULY 1967

OAK RIDGE NATIONAL LABORATORY
Qak Ridge, Tennessee
operoted by
UNION CARBIDE CORPORATION
for the
U.S. ATOMIC ENERGY COMMISSION

!
This report is one of a series of periodic reports in which we describe briefly the progress of
the program. Other reports issued in this series are listed below. ORNL-3708 is an especi'o”y
useful-report, because it gives a thorough review of the design and construction and supporting
development work for the MSRE. It also describes much of the geheral technology for molten-salt

reactor systems.

Period Ending January 31, 1958

ORNL-2474
ORNL-2626 Period Ending October 31, 1958
ORNL-2684 - Period Ending January 31, 1959
ORNL-2723 Period Ending April 30, 1959
ORNL-2799 Period Ending July 31, 1959
ORNL-2890 Period Ending October 31, 1959
ORNL-2973 Periods Ending Januory 31 and Aprii 30, 1960
ORNL-3014 Period Ending July 31, 1960
ORNL-3122 Period Ending February 28, 1961
ORNL-3215 Period Ending August 31, 1961
ORNL-3282 Period Ending February 28, 1962
ORNL-3369 Period Ending August 31, 1962
ORNL-3419 Period Ending January 31, 1963
 ORNL-3529 Period Ending July 31, 1963
ORNL.-3626 Period Ending January 31, 1964
ORNL-3708 Period Ending July 31, 1964
ORNL-3812 Period Ending February 28, 1965
ORNL-3872 Period Ending August 31, 1965
ORNL-3936 Period Ending February 28, 1966
ORNL-4037 Period Ending August 31, 1966
SUMMARY oottt et e e e b e e b b e et e e et e et n e e et e e s e e e e et ae e et b et e 1
INTRODUCTION ....ooovecerrnrneesmnesensseseressesssesmsoess oo OSSOSO 9
PART 1. MOLTEN-SALT REACTOR EXPERIMENT
1. MSRE OPERATIONS ........ccoiiii e et PRV 11
1.1 Chronological Account of Operations and Maintenance ...................... e 11
1.2 Reactivity BAIANCE .........cccooiieieiiieieiieeie ettt ettt st e e ea e e et 14
OB 4 0 1=3 8 1= 1 Lo =T OO P OO PO SRS TSP P PP TP PP PP DRSO ‘14
Circulating Bubbles ... ...ooo.vi v e e e e 17
1.3 Thermal Effects of Operation........c..ccccvveiiiiiiiiiiiiiiicieee e 18
" Radiation Heating .......cocooieoiieee et e e 18
Thermal Cycle HiStOIY ..ottt bbbt e e - 19
1.4 Reactor DYNAMICS ....cocovieiiiieieeeiie e eeteieae sttt eoiis e ebesses sear e e era s e e s e s et s e e s e s e e st bt n e s ens e 20
1.5 Equipment PerformancCe............cccoouvriviiienviciccien e et e et e aa e 21
Heat Transfer .......cccocciiviiiiiiiii e e e FOTURUUP 21
Main BIOWELS ..ooooiiiiiiiiiiieeeieeeeeeeee e e e e e reraeen e 25
| S Ts BT Tae) gl D8 Vel L 1-3 1 b L= OO UUS PSP RRPPP PR 26
Off-GaS SYSLEIMS ....vrv vt everoseee e ee e eeeeeessse e es b s s 12 et ems s es et ettt een e s aa e 27
Cooling-Water SyStems ........c.ccocooervinccrincnnn et e eee et e aaeeba e ere et b e a et ee e 31
Component-Cooling SYSOM . ... oot et e bbb 32
Salt-Pump Oil SYSEEMS ... e et e 34
N =Yoo s (ot D RS AT (=Y 1 FOUU U U O ORI UTU PSP PPSPP IO PR PPPTP P 34
| 3 (R 1 c< ST U OO OO PP PP PPTT PP PISTOPPPPPPS 35
Control Rods and DIIveS ..........ooooiiiiiiiiii i e 36
Samplers ..........ccocoeeennns TSROSO PR P PP PO VPPPPPPPP R 36
(OFe T N R et 11 = « LAUUURUUTUUOT U PPN 37
2. COMPONENT DEVELOPMENT .........oooeoorooooeseeesooecereoessssssesooees o sosmoes oo sciers o S S 39
2.1 Sampler-Enricher et et e v 39
COMEAIMATIA IO oo e e e e e e e e et e et e e e et et e e ne e e nre e e 39
CaAPSULE SUPPOTt WATE. ..o oot iiitietitsieie et etetes et ceteet et es et et etses s na s s ekt - 40
Seal Leakage ..o SO U OO PTOOP PO 40
| R Te =y 111 Lo - B UUUTUTUUUTUTU U OO OO VOO U OO OO OO PRSI - 40
2.2 COOLANE SAMPIET ..o eeee et eeeeerae e seses s s b s bbb 41
2.3 Fuel Processing Sampler ..............ccccunne. s e et 41
2.4 Off-Gas Sampler............. OO e e re e e et eeeeaeaeaertreaeareee e e et e eeennnte s 41
2.5 Off-GAS FIIter — MK Il oot eeie oot e et e e ettt e e e e e s eb s e se e s emta e e e e e e e e et e eat srn e e e 42

Confe nts

iii
2.6

2.7

2.8
2.9

Feltmetal Capacity TOS Lottt eee e
ASSUMPLIONS ..ot et e TR U

Experimental Procedure ... e oo _

RESUIES ..o et e ettt a e e e e e e e e naeeas evetsieenies

Examination of the Mk I Off-Gas Filter ............... s e
Pressure DIOP TeStS ...ttt et e e e s e e e e ars e e s aen e nrasre e e nteeraare
Disassembly ... e R
ObserVations .......................... ettt eeetheeete e e o e e et ee—eae e tet e s eh et e e e s e tesho et e et t e et esene et e e e e ereenne

Migration of Short-Lived Gaseous Products into the Graph1te .....................................................

Remote MaintenanCe ............oooiiiiiiiiiie e e e e
Summary of Remote Maintenance Tasks Performed ...........ovoeiiiiomeee e e,
Radiation Levels ..........ccooooiiiiiiiiice e et e e,
Contamination..............cccccevvvvvennn.n. OO TTOTO L TR
Conclusion ............... ettt e Eeeeeieteeieeeeeeeettsineeeneontsntehhteeaeeeethateee s beteteeetaaeibeeeteeee ne e aenn e eeteesaes

3. PUMP DEVELOPMENT ... e ettt ettt e et e

3.1

3.2

MSRE PUMPS . ....o .o oo et e e e s et e e s e e e ee s e oo oo oo s e
Molten-Salt Pump Operation in the Prototype Pump Test Fac1l1ty ..............................................
MEK-2 FUEL PUMP....oioiii ettt e et et et e et e et e et ete et b st e e s
Stress Tests of Pump Tank Discharge-Nozzle Attachment ...
Spare Rotary Elements for MSRE Fuel and Coolant Salt Pumps............coovvviiveeeeennn. SUSUTU
Lubrication R 253 (= 1 TP USSR PSSR URUROR SRR

Other Molten—Salt Pumps ..........................................................
Fuel-Pump High-Temperature Endurance Test Fac1l1ty .......... JEUTRUORR e e

4. INSTRUMENT DESIGN AND DEVELOPMENT ..ot e e e

4,1

4.2

4.3
4.4

Instrumentation and Controls DeSIEN .......ccovviiiviiiiiiie e ettt
Off-Gas Sampler Design ........ et e T TSRS
Control System DeSign ... e e et o

MSRE Operating EXRpPerienCe ... i e e
Control System Relays ..., vt e
Temperature Scanner SYSEEM ........c.ccooviiiiii it
VaBLVES Lo e e e e

TR IMOCOUPLES oo e e e e e
Coolant-Pump Radiation Momtor ......................................................................................................
Safety SYSLEM ..ot e e

Data System ...... et eeeeteeeeteeeetaeetntaeeaieeern——entaeaeas e et e

Instrument Development ..................... P PPRUB e et eae et s
Performance — GENeral............ccccooiiiiiii ettt sttt
TEMPEIAUIE SCAMMET ........ooviiiiviiiieieeeeee oo ettt ee e et eee et ese e st re et e setereees
High-Temperature NaK-Filled Differential Pressure 'I‘ransmltter ................................... e
Ultrasonic Level Probe . ... e e e e,
MSRE Fuel Distillation System LeVel PrOBE ..oooooooeoeeeeee e
Bell-Float-Type Level Indicator for the Mark II Pump........ OO OSSO OIO
' 5. MSRE REACTOR ANALYSIS ............ ettt USRS RURUTRURTUSPUN 79
5.1 Neutron Reaction Rates in the MSRE Spectrum...............c.cooiin e e 79

5.2 Isotopic Changes and Associated Long-Term Reactivity Effects During
Reactor Operation ..............cceoeviiiiiiiiiiiereiereecni e s e e e 83
5.3 Analysis of Transient 135Xe POiSOMIME.....c.ccovioieiiieiiet et ettt n et e 86

" PART 2. MATERIALS STUDIES

6. MOLTEN-SALT REACTOR PROGRAM MATERIALS............ccccoiiirincann e 95
6.1 MSRE Surveillance Program — Hastelloy N ... e 95
6.2 Mechanical Properties of Hastelloy N ... SRRV RPUOPPUURRRIN 103
6.3 Precursors of MSBR Graphite.......ccccccooviiviviiinniiii e e e 108
6.4 Graphite Irradiations ............ccccooviviininiciinn et e e, e 110
6.5 Brazing of Graphite ... FO TP ORPR SR OUUUUURURT e 111

Large Graphite-to-Hastelloy-N Assemblies .............ccccoviiiiiiiiiiiii 111
6.6 Corrosion Resistance of Graphite-to-Metal Brazed Joints........cccocciviiiiiiiiiiiiiiini e 111
6.7 . Thermal Convection Loops ........ccooovvivieiiiieiiiic e e e 115
6.8 Evaluation of MSRE Radiator Tubing Contaminated with Aluminum ............ e, 116

A O 7§ 0 0 K3 1 5 SO PO U 118

7.1 Chemistry of the MSRE ......oocooivioueieeereeeeeesoeresse e, et 118
Fuel Salt Composition and Purity................... e J SO ST PR ORRUPPTTPRRY 118
MSRE Fuel Circuit COorroSion ....c.ccocoiiiiiiiiiimiiinii e N 120
Extent of UF, Reduction During MSRE Fuel Preparation ..., e 121
Adjustment of the UF, Concentration in the MSRE Fuel Salt..............cooioiii 123

7.2 Fission Product Behavior in the MSRE ... e 124

' Long-Term Surveillance SPeCIMEens ........ccoiiviiiiiiiiiiiiii e e e 125
Uranium Analyses of Graphite Specimens ................ccccooeeiiniiiienn, e et e et es 128
FUEL SALt SAMPIES ...ooimioii ittt e e et ettt b et et et e 130
Effect of Operating Conditions ... OO PSPV TO VPP oS 131
Effect of Beryllium Additions .........ccooeoiiimiiiii oot e 131
Pump Bowl Volatilization and Plating TeStS ..o 134
Uranium on Pump Bowl Metal Specimens ........cccocciiiiiiiniviiiicii e 138
Freeze Valve Capsule EXPEriments .........ccccovvoricriiiiiiiiiiiiiiciin st 138
Special Pump BoW]l TeSES ...ttt sttt 141
General Discussion of Fission Product Behavior.............cooooviiiiiiiiin et 142

7.3 Physical Chemistry. of Fluoride MeItS ............ccocoviiririiiiioriiieicess e et e s s 144
. The Oxide Chemistry of LiF-BeF ,-Z1F , Mixtures ... e 144
Solubilities-of SmF ; and NdF, in Molten L1F BeP, (66-34 mole %) ..o 144
Possible MSBR Blanket-Salt M1xtures .............................................................................................. 146

7.4 Separations in Molten FIUOrides ........ococcoooiiiiiiniiiiiiec s 149
Removal of Rare Earths from Molten Fluorides by Precipitation on Sol1d UF, e, 149
Extraction of Protactinium from Molten Fluorides into Molten Metals............... SSUUUTRTRUUR 150
Extraction of Rare Earths from Molten Fluorides into Molten Metals .............cccccoiiiinninnnns 152 7
Protactinium Studies in the High-Alpha Molten-Salt Laboratory .................. e e 153

Preliminary Study of the System LiF-ThF ,-PaF,............. TR UPRSUPTOUROORO 155
vi

7.5 Development and Evaluation of Analytical Methods for Molten-Salt Reactors ......................... 156
Determinations of Oxide in MSRE Salts ...........c.cccccoevrivennnen. e e, 156
Determination of U3*/U%* Ratios in Radioactive Fuel by a Hydrogen Reduction

MEEHOM . ... e e e et et b et e e bttt te e eneaas tevrernrrrenes 158
EMF Measurements on the Nickel-Nickel(Il) Couple in Molten Fluorides...........ccoooveinnnn. 162
Studies of the Anodic Uranium Wave in Molten LiF-BeF -ZrF ... 163
Spectrophotometric Studies of Molten-Salt Reactor Fuels .. ..o 163

7.6 Analytical Chemistry Analyses of Radioactive MSRE Fuels ... 164
Sample Analyses ..........c........ et s et 165
Quality Control Program .........c.ccooooviiiiiiiiiiniinen et e e n e e e e aaes TR 165

8. MOLTEN-SALT CONVECTION LOOPS IN THE ORR ........ccccocoiiiiiiiii 167

8.1 Objectives and DESCIIPLION ......ccoooiiioiiiieeieceee ettt ettt st st saesr s srerenesneresaseeee 167

8.2 First Loop EXPEIIMENt . ..ottt e ee e e e e et st e e ste e et eeneb e e enbe et e anennes 168
In-Pile Irradiation ASSEMDBLY ........oooiomiiioteeeete sttt st sttt et es e st ess et etes s sresst et es st eennees 168
Operations....... e te e areeeee iiiteesiiieeeeeseerterteieionererieeatht i bt sbeeeane aeeee e aaaneeaeseenannrees e 168
Chemical Analysis of Salt ... ..., e, 168
[000] ¢ o131 + WU USSR " b e bttt ieraaaanaaaaaraaaaaaaaas 169
Fission Products ...........c.ccooveveinninnn... N s eeerarereeteresrere e b nbeto tteeasaen i bebes e ratba b et aeanan e e srnretene 169
Nuclear Heat and Neutron FLUX ...t e st e ee s sra e s e sree e 169
Hot-Cell Examination of COMPONENTS ..........ooeiiiiiioreiiie i s eeie e et et ee e s et e enerne e 169

8.3 Evaluation of System PerformancCe .............cccocvvriiiiiiiiiniiiei e 171
HeaterS ....cciiiiiiiiieie et e e e b e breeaeeesaasnsreeaaneaeeesraraaenae e 171
0o Yo =3 <=3 U OO O ST UUUUPRIOPPPR PR 171
Temperature Control ... e e n e .1
Sampling and Addition ............cccooveviiiiiiiniii OO U UD PSPPI 171
SAIt CIrCULALION ..oooiiieiiiii e et ettt et et e e et e e s e etrtn e e e e s r e e e e 172

8.4 Second In-Pile Irradiation ASSEmDbIy.......c..cciiiiiiiiiiiir e e 172
OPEIAtioN ...oeeeeeeeeieeeciiiee et e s s e 173

. PART 3. BREE}ER REACTOR DESIGN STUDIES

9. MOLTEN-SALT BREEDER REACTOR DESIGN STUDIES ........coooiomvvioreeeeeeeeeeeeee e e 174
0.1 GENEIAL ..o e e et ee e et e e s ettt et e s snaseennenas 174
9.2 Flowsheet ........cccoeeeviinnes e R s ettt 177
9.3 Reactor Cell Component Arrangement............occeerenees R s eeteererteeereetreeaeeesetreareeeearasanen 179
9.4 Component Design ................. SRRSO T veterennreernreenransaareseeansannraeeaarnreeeeernns 182

REACEOr VESSEL.. .. ittt ettt e ettt et e e e et e e e n bt e et anen et nmr e 182
"Fuel Heat EXChanger...........cccccooiviiiiiiei et sree e e erreuresratesaesmee i aaeseanreans £ sneeneeent£ens 186
Blanket Heat EXChanger ... 190

0.5 ReACtOr PhYSICS .ottt ettt et 193
9.6 MSBR Gas Handling SYStEmM .......oooueereeereoreeeeeeeeeseeer e e et ettt 199
Xenon Removal...........cccooiiiiiiii S SRS ORUORPO 199
Mechanical DeSiEn. ..ot ittt et et et e e e 200

10.

wvii

Gas Injector Systém ........................................................................... e et e 200
BUbble Separator SYSLEML.. .. c..ooiuieiieitiieeieneeteet et esee e ete sttt e eae e iee e e ree et e e e .. 202
Volume Holdup SYSEEM . ..oooiiiiiiiii et e e s te e e snie e e e s e 202
NONCritiCal COMPONENES ..iivviiieiiieiieieei e et e e ettt e e e et ee e e e e eaeeeaee e aeesamseeeeeseeaneeeen e sneinbeeeeean 203
MOLTEN-SALT REACTOR PROCESSING STUDIES .. ... ettt e 204
10.1 Continuous Fluorination of a Molten Salt.............. et e, 204
" NONPIOLECEA SYSEEIM ..ottt ettt et et e ee et e e e e 204 -
Protected SYStOM .. ....occoiiiiiiiii et e e e e 205
10.2 Molten-Salt Distillation Studies.............c.cccceeiinne, PO OEO DR PTOORPUPUROPPSORRY. s 206
Relative Volatility Measurement.................... bt eht b e er e eae e et en e aee s .. 206 -
Vaporization Rate StUAIiES .........cccciiiiiiiiiioiiii et e et e ch et etnen e ae e 206
Buildup of Nonvolatiles at a Vaporizing Surface ... 208
10.3 Vacuum Distillation Experiment with MSRE Fuel Salt................ccoovi 211
Summary

PART 1. MOLTEN-SALT REACTOR EXPERIMENT

| 1. MSRE Operations

The maintenance and other shutdown operations started during the last period were completed,
and power operation of the MSRE was resumed in October with only one of the main radiator
blowers in serviée. After 17 days’ operation (run 8) at the maximum power attainable with one
blbwer, the second blower was installed, and a 12-day run (run 9) ‘was made with both blowers in
service. _Thé nextv run (run 10) began in December and was continued at full power for 30 days
without interruption. A fourth power run (run 11) was in progress at the end of the report period
with 31 days accumulated. '

Further refinements were made in the reactivity balance, and on-line calculations were used
as a guide during operation. Application of these refinements showed that the unaccounted-for
reactivity change was only about +0.05% 0k/k through the end of run 10 (16,450 Mwhr). Addi-
tional dynamics tests were performed at the start of run 11, which showed that the dynamic char-
acteristics of the system were unchanged.

The performance of equipment in the priméry and auxiliary systems was ge»nerélly satisfactory.
Detailed study has revealed no change with time in the heat-transfer coefficients of the main heat
exchanger or the radiator. Salt plugs in the reactor off-gas line, from an accidental overfill of the
fuel purfip, caused some difficulty, but the line was finally cleared before run 10, and this problem
has not recurred. Partial plugging of the particle trap and charcoal beds in the off-gas system
was also encountered, but this did not restrict power operation. A new particle trap, of different
design, was installed, and no further plugging was noted at that location. Only routine mainte-

nance problems were encountered with other equipment.

2. Component Development

The sampler-enricher was used to isolate thirty-six 10-g samples and eleven 50-g samples on
a routine basis. In addition, eight special samples for use in cover-gas analysis and five cap-
sules for beryllium addition to the fuel were handled. In-an effort to rc?'duce the contamination
‘ llevel in the sampler and adjacent areas, the inside of the sampler was»cleaned with s'p'onges, and

an additional ventilation duct was installed near the transport cask position. The design of the
sample capsules was changed to provide 5;1 nickel-plated magnetic steel top instead of the orig-
inal copper top, so that a magn.elt can be used for retrieval in fh“e event a capsule is dropped. A
mechanical method of assuring that the maintenance valve is. closed is being substituted for the
existing pneumatic system, which has given difficulty because I}of the gradual increase in leakage
of buffer gas through the upper seat. The valve itself will not be replaced at this time. Several
minor maintenance tasks were performed, including replacementli of the manipulator boots, inspec-
tion of the vacuum pumps, énd replacement of a small hinge pin and cotter key which had worked
loose inside the sampler. : | ,

Eight 10g coolant salt samples were isolated. The valve seats of the removal valve in the
coolant sampler were replaced. |

Installation of the fuel process sampler is complete except for the shielding, and the opera-
tional checks have been cornpléted. _

The off-gas sampler was altered to include a hydrocar‘bon analyzer section in pléce of the
chromatographic cell, which was not ready. The intemal piping was also rearranged to permit
sampling upstream of the 522 line filter. |

A redesigned MK II off-gas filter was in-stalled in place of the old particle trap and charcoal
filter assembly. The filter was enlarged from 4 to 6 in. ID and was arranged so that the Yorkmesh
entrance section can be heated above the temperature of the rest of the filter. There were other
changes in the arrangement of the intemals, which were made to correct deficiencies found in the
first filter. Measurements showed that the pressure drop was less than 1 in. H,O at three times
_rated flow and that both filters had efficiencies greater than 99.9%. Tests of the Feltmetal sec-
tions of the filter indicated that the expected life might be as little as three weeks or as great
as 70 years, depending on the character of the particles in the rleactor off-gas system.

The old patticle trap was taken to the hot cells for testing and examination. Tests indicated
an increase in the pressure drop by a factor of 20 over the cleanw filter and that the Fiberfrax sec-
tion was essentially clean. We concluded that most of the pressure drop was in the Yorkmesh
entrance section, where material had filled in the space betweér:{ the wires and plugged the open-
ing of the inlet pipe. Since the inlet pipe expanded longitudinally due to fission product decay
heat, it is believed that operation at power caused the inlet pipe; to push into the plugging ma-
terial, thereby increasing the resistance to flow in the manner of a thermal valve. Metallographic
examination of the deposit area of the Yorkmesh showed heavy carburization of the wire, and
there were indications that the wire had been heated to at least }200°F. The deposit itself con-
tained much carbonaceous material, as well as high Ba and Sr fractions. There was very little.
Be or Zr, indicating that there was essentially no salt carried to this point, and most of the fis-
‘sion products found were daughter products of Xe and Kr. The Fiberfrax section was very clean
except for a small deposit of oil in the first layer. :}

A model was developed which evaluates the concentration of “‘very short lived’’ noble gases
in the graphite while the reactor is at power. Reasonable agreement with measured concentration

distributions was obtained for 14°Ba, 1“Ce, and °'y.
[\

Among the maintenance tasks completed were: (1) replacement of all of the air line quick
disconnects in the reactor cell with metal compression fittings after the elastomer in the orig-
inal disconnects became embrittled from radiation, (2) replacement of the particle trap in the off-
gas line, (3) removal of frozen salt obstructions in several of the lines coming from the pump bowl,

(4) replacement of the core sample array, and (5) replacement of a control rod and drive.

3. Pump Development

The restriction in the annulus between the pump shaft and shield plug in the MK-1 prototype
pump, which was discussed in the previous semiannual report, was found to have been caused
by the salt level being raised accidentally into the annulus during a fill of the system.

The MK-1 fuel pump tank was removed from the prototype pump facility and installed on a

_test stand. This stand was constructed for room-temperature measurement of the stresses pro-
duced in the weld attachment of the discharge nozzle by forces and moments imposed by the
pump-tank discharge piping. During initial testing, a crack was found in the heat-affected zone
of the weld attachment. The crack was repaired by welding, and further exploratory tests are
being made to measure the stresses.

The spare rotary element for the MSRE fuel pump was prepared for reactor service and is
being held in standby. The shaft-seal oil leakage problem on the spare rotary element for the
MSRE coolant pump was resolved, and the assembly is being completed for standby service. The ‘
lubrication pump endurance test was continued. Shaft defllection' and crit\ical speed tests were

completed on the MK-2 fuel pump rotary element, and the MK-2 pump tank is about 70% fabricated.

- 4. Instrument Design and Development

_The design of instrumentation and controls for the off-gas sampler is essentially complete.
Changes in the design of the sampler system requitred changes in the instrumentation and controls
design.

Performance of developmental instrumentation has continued to be generally satisfactory. No
problems were encountered that required redesign or initiation of new development.

Results of investigation indicate that none of the commercially available solid-state multi-

plexers are suitable for use as a direct replacernerit for the mercury switches used in the MSRE

temperature scanner. No further progress has been made in determining the cause of failure of
an NaK-filled differential-pressure transmitter in MSRE service.

The effectiveness of modifications to ultrasonic level probe circuitry has not been determined
because no salt has been transferred to the fuel storage tank.

A new conductivity level probe was developed for use in a fuel distillation system. Design
of this probe is similar to that of probes used in the MSRE drain tank but differs in that it is

smaller in size and will provide a usable indication of level changes.
5. MSRE Reactor Analysis
I

Neutron energy spectra in the MSRE were calculated and used to estimate important isotopic
changes and associated long-term reactrvrty effects during power operation. The changes con-
sidered include depletion of ?3*U, 235U, and ?*3y, production of ?3°U and ?3°Pu, burnout of
initial °Li in the salt and %8 in the graphite, and production of tritium and '®0 as products
of n,a reactions. With the exception of 23%U, the principal contributors to long-term reactivity

changes were found to be the ®Li burnout and 23Pu productron
Further studies were made in the correlation of the observed time behavior of the '3%Xe poi-

soning in the MSRE with calculations from a theoretical model.il Graphic comparisons are given

of calculated buildup and removal of '?*Xe reactivity, following changes in power level, with

some of the experimental reactivity transients observed from op“eration to date. The results, in N
good accord with previously reported evidence, point to the cofi‘clusion that a small amount of

undissolved helium gas is in circulation with the salt, which enhances the mass transfer and .
removal of xenon from the reactor. In addition, the transient arralysis supports the assumption

of a fairly high efficiency of removal of xenon directly from the}gas bubbles by the external .

stripping apparatus. Approximate rarrges of the circulating bubble volume fraction and bubble-

stripping efficiency obtained from the analysis were 0.10 to 0.1;5 vol % and 50 to 100% respec-

tively.

’

PART 2. MATERIALS STUDIES
6. Molten-Salt Reactor Program Materials

‘There was no microscopically visible corrosion or coatings ion the Hastelloy N reactor-vessel-
wall surveillance specimens exposed to molten fldoride fuel in the core during a 7800-Mwhr opera-
tion in which the specimens accumulated a thermal neutron dose of approximately 1.3 x 10°° nvt,
~ However, a carbide deposit about 0.001 in. thick was found on specrmens in' contact with the
graphite. '

A loss in ductility of the irradiated specimens wao found at elevated temperatures, as ex-

" pected. In addition, however, there was an unexpected 20% redrrction_in ductility at low tempera-
ture, which is thought to be related to extensive grain-boundary lcarbide precipitation. The doses
received by the metal specimens are higher than the reactor vessel is anticipated to receive over
its lifetime, and the test results give reassurance that the mechanical properties of Hastelloy N
are more than adequate for the service planned. | |

A family of curves was obtained from tests of Hastelloy N at various strain rates and tempera- -
tures. - These curves will allow one to predict the strain rate sensitivity of the ductility of Hastel-
loy N at any given temperature. The strain rate sensitivity changes markedly wit}r temperature.

We examined seven experimental grades of isotropic graphiteli. None satisfied all the require-
ments for molten-salt breeder reactors, but one had a good pore spectrum, and four appeared to

have potential for MSBR use.

1}
1]

5

Experiments have been designed and are being fabricated which will pemit irradiation of

_graphite to the high exposures that will be incurred in an MSBR. Irradiations in HFIR, DFR,

EBR-II, and ORR are planned.

The search is continuing for corrosion-resistant alloys that are suitable for brazing graphite to
Hastelloy N. A furnace for brazing large graphite-to-Hastelloy-N assemblies is being constructed.
The lofig-term thermal convection loops of Hastelloy N and type 304L stainless steel have
continued to circulate fused salts, acquiring 43,024 and 31,749 hr respectively. Weight losses
from specimens inserted in the stainless steel loop are less than what was measured on earlier

samples.

7. Chemistry

‘ o ! _
Chemical analyses of the uranium concentration in the fuel salt show a measurable decrease.

This results from dilution of fuel salt by the remnants of flush salt that remain in the reactor after
flushing and from the transfer of about 7 kg of uranium from the fuel to the flush salt in each drain-
flush-fill sequence. _ |

The chromium conéentration has remained s‘teady at about 60 ppm, indicating the absence of
corrosion in the reactor fuel circuit. | | |

At termination of MSRE run 7, 1.66 gram-atorris of uranium had been Burned,v and, as a conse-
quence, about 1.66 equivalents of oxidizing species had been produced in the fuel. To neutralize
this oxidizing effect, and to make the fuel more reducing in character, the fuel was treated with
beryllium metal to reduce a small amount of UF, to UF,. To date, 27.94 g of beryllium has been
introduced; this has converted 0.65% of the UF, to UF ,.

Most fission products behaved as expected with the exception of rather noble metals, which
continued to show an apparent tendency to volatilize and to plate on metal surfaces. Attempts to
decrease the volatilization by chemical reduction of the fuel with elemental beryllium were unsuc-
cessful. Detectable volatilization apparently continued for long periods after shutdown; a three-
day shutdown reduced the volatiiization of molybdenum by a factor of only 5.

Further studies of the soiubility of oxide in fuel—flush-salt mixtures have been carried out.

A minimum solubility occurs when the mole fraction of Z1F , reaches 0.01.

In connection with a study of methods of reprocessing MSBR fuel, the solubilities of SmF
and NdF , in fuel solvent have been measuréd as a function of temperature. \ Salt compositions
for possible use as a blanket for the MSBR have been reviewed.

The feasibility of removing rare earths from fuel by precipitation on solid UF , has been
studied, and the results are moderately favorable. Activify coefficients associated with the

reductive extraction of rare earths from fuel into bismuth amalgams have been measured. The

. ptocess appears quite attractive.

The use of reducing agents for protactinium removal from blanket melts was investigated

further. Electrolytic reduction gave disappointing results, but the use of thorium as a reducing
agent gave good results, especially when there was a large surface area of iron metal available
to receive the protactinium. |

In addition to the regular salt samples, several special samples were analyzed. These in-
cluded capsules used to make beryllium additions to the fuel éalt, MSRE off-gas samples, and
highly purified LiF » BeF2 samples. The absolute standard délviation for oxide determined in
ten radioactive fuel samples taken from the MSRE over an eight-month period was 8 ppm.

A transpiration method has been developed for the determination of U3 /U ** ratios in radio-
active fuel samples. The method is based on the measurement of the HF produced by the reduc-
tion of oxidized species when the molten fuel is sparged with hydrogen. Increases in the udt/utt
“ratio from about 0.0005 to 0.005 were observed when metallic geryllium was added to the fuel in
the reactor. |

An experimental reference electrode, consisting of an Ni/Ni 2* half-cell electrically connected
to the fluoride melt through a wetted boron nitride ‘““‘membrane,’ exhibited satisfactory Nernstian
reversibility. On the basis of limited stability tests, this elecfrode appears to be suitable as a
reference for electrochemical measurements in molten fluoridesl. An anodic oxidation wave result-
ing from the voltametric oxidation of U** at +1.4 v was studilzd and found to have properties
most consistent with oxidation of U** to U5, followed by catalytic disproportionation of the US*

The _spectrophotometr'ic determination of U3" in molten fluoride salts was investigated by a
new technique in which U3* is generated voltametrically in th!e optical path of a captive-liquid
cell. The feasibility of determining 50 ppm of U3" in the presénce of 2% U*" was demonstrated.
Measurements of the absorption spectra of Er3+, Sm 3+, and Ho 3+ in LiF-BeF2 indicated that

these ions would not interfere with the determination of U3,

8. Molten-Sali Convection Loops in The ORR

Irradiation of the first molten-salt thermal convection loop experiment in the Oak Ridge Re-
search Reactor was terminated August 8, 1966, because of a leak through a broken transfer line.
A power density of 105 w/cm® was achieved in the fuel channeis 6f the graphite core before fail-
ure of the loop. A second loop, modified to eliminate causes of failure encountered in the first,
began long-term irradiation in ]-anuary 1967. An average core pnower density of 160 w per cubic

centimeter of fuel salt was attained and maintained in the first ORR irradiation cycie.

PART 3. BREEDER REACTOR DESIGN STUDIES
9. Molten-Salt Breeder Reactor Design Studies

|
Breeder reactor design studies have been concerned primarily with making a choice of the
basic reactor on which design effort will be concentrated. The'modular concept has been chosen,

and the power for which the module is to be used is set for the moment at 556 Mw (thermal). The .
»;

avérage core powet density, and therefore the flux, has been arbitrarily cut from the 80 kw/liter
used in previous studies to 39 kw/liter to give greater core life expectancy. X

Further optimization studies have been made on reactor parameters. A durable core config-
uration has been established. The core is 10 ft high and contains 336 fuel cells. The volume
is 503 ft 3, of which 16.5% is fuel salt, 6% fertile salt, and 77.5% graphite. A blanket 11/4' ft thick
axially and 11/2 ft thick radially surrounds the core. A 6-in. graphite reflector is placed between
the blanket region and the container vessel. The fuel cells are joined to the dished head plenum
by pipe thread connections.

The fuel heat exchanger and the blanket heat exchanger are flanged into place, reducing the
number of pipes to be remotely cut and welded if replacement of these items is necessary. A con-
centric coolant line connects the primary and blanket coolant circuits. Flowsheets and design
criteria are being developed for the gas sparging system and the off-gas system.

The layout of the reactor cell has been revised to eliminate some stress préblems that were
found in the original layout. A first attempt at a better mounting for reactor cell components has
been made and is being analyzed.

Some of the more basic MSBR nuclear calculations have been started, and from the first re-
sults some changes have been made in the unit cell dimensions of the core. The reactor as now
contemplated has a yield of e;pproximately 6% per year, a breeding ratio of 1.07, and a fuel cycle
cost of 0.43 mill/kwhr on an 80% plant factor.

Work on reactot physics included (1) a series of cell calculations performed to examine the
the sensitivity of the MSBR cross sections and reactivity to various changes in cell structure °
and composition and (2) several two-dimensional calculations of the entire reactor. The refer-
ence cell contained ~0.2 mole % 233U in the fuel salt and 27 mole % 23271 in the fertile salt.
The fuel volume fraction was 16.48% and the fertile volume fraction 5.85%. The results of the
cell calculations indicated a reactivity advantage dssociated with increasing cell diameter, and
a nominal diameter (flat to flat) of 5 in. was selected. Detailed radial and axial flux distributions
were obtained from the two-dimensional calculations. The radial and axial peak-to-average flux
ratios calculated from these distributions were 1.58 and 1.51, respectively, giving a total peak-to-
average ratio of 2.39. | |

The central cell of the reactor was examined for reactivity control purposes. If a completely
empty graphite tube of 5 in. OD and4 in. ID is filled with fertile salt, the change in reactivity
is 8k/k = —0.018%. If the empty tube is filled with graphite the reactivity change is 8k/k =
+0.0012%. Thus there appears to be a substantial amount of reactivity control available by vary-

ing the height of the fertile column in the tube.

10. Molten-Salt Reactor Processing Studies

The concept of an integral processing plant based on a fluorination and distillation flowsheet

has matured in the last year. Studies on continuous fluorination techniques have ascertained that
high recoveries and good fluorine utilization are feasible, and the measurements of relative vola-
tilities for the distillation stép have been highly encouraging.- Further analysis of the operations
has revealed no new problems which could thwart this approach.

Continuvous Fluorination of a Molten Salt. — The recovery of uranium from the fuel salt of an
MSBR by continuous fluorination embraces two §ignificant problems: (1) the establishment of an
adeciuate concentration gradient in the tower to effect both higfi recovery and reasonable fluorine
utilization and (2) the operation of the system with a frozen layer of salt on all surfaces to pro-
tect them from oxidation by fluorine. Studies with nonprotected? systems using l-in.-diam towers
Etanium with fluorine utilization

of 15%. Studies on column protection involve the construction of a 5-in.-diam nickel tower with

have demonstrated steady-state recoveries up to 99.9% of the u

provision to generate heat fluxes to create a frozen wall of salf.

Molten-Salt Distillation Studies. — Relative volatilities measured at 1000°C and 0.5 mm Hg
pressure for CeF,, LaF,, NdF ;, and SmF , with respect to LiF were 3 x 1073, 3x10"% 6 x
10=% and 2 x10~* respec-tively. The consistency of the results assures that these relative vol-
atilities are accurate. Data have been acquired on rate of vaporization as a function of system
pressure which show that the processing rates necessary in an MSBR system can be achieved
in stills of reasonable size. However, analysis of the buildup of nonvolatile salts at the vapor-
izing surface indicates that some method of salt circulation is mandatory.

Vacuum Distillation Experiment with MSRE Fuel Salt. — An experiment is planned in which
about 48 liters of MSRE fuel salt will be processed by vacuum distillation after the 235U has
been removed by fluorination. The equipmen't, which has been t!:lesigned and is being fabricated,
wi‘ll be used in an experimental program with nonradioactive sait to study still perfformance before

it is installed at the reactor site for use with irradiated salt.
/
[

Introduction

The objective of the Molten-Salt Reactof Program is the development of nuclear reactors
which use fluid fuels that are solutions of fissile and fertile materials in suitable carrier salts.
The program is an outgrowth of the effort begun 17 years ago in the ANP program to make a
molten-salt reactor power plant for aircraft. A molten-salt reactor — the Aircraft Reéctor Ex-
periment — was operated at ORNL in 1954 as part of the ANP program.

Our major goal now is to achieve a thermal breeder reactor that will produce power at low
cost while simultaneously conserving and extending the nation’s fuel resources. Fuel for this
type of reactor would be 23“Q‘UF4 or 235UF4 dissolved in a salt of composition near 2LiF-BeF ,.
The blanket would be ThF , dissolved in a carrier of similar composition. The technology being
developed for the breeder is also applicable to advanced converter reactors.

- Our major effort at present is being applied to the operation and testing of the Molten-Salt
Reactor Experiment. This reactor was built to test the types of fuels and materials that would
be used in thermal breeder and converter reactots and to provide several years of experience
with the operation and maintenance of a small molten-salt reactor. ‘ The experiment is demon-
strating on a small scale the attractive features and the technical feasibility of these systems
for large civilian power reactors. The MSRE operates at 1200°F and at atmospheric presdsure
and produces about 7.5 Mw of heat. Initially, the fuel contains 0.9 mole % UF ,, 5 mole % ZrF ,,
29.1 mole % BeF , and 65 mole % LiF, and the uranium is about 30% 235y. The melting point
is 840°F. In later operation we expect to use highly enriched uranium in the lower concentration
typical of the fuel for a breeder. The composition of the solvent can be adjusted in each case to
retain about the same liquidus temperature.

The fuel circulates through a reactor vessel and an external pump and heat-exchange system.
All this equipment is constructed of Hastelloy N, ! a nickel-molybdenum-chromium alloy with ex-
ceptional resistance to corrosion by molten fluorides and with high strength at high temperature.

The reactor core contains an assembly of graphite moderator bars that are in direct contact with

- the fuel. The graphite is a new materialv2 of high density and small pore size. The fuel salt does

not wet the graphite and therefore does not enter the pores, even at pressures well above the
operating pressure. .

Heat produced in the reactor is transferred to a coolant salt in the heat exchanger, and the
coolant salt is pumped through a radiator to dissipate the heat to the atmosphere. - A small facil-
ity installed in the MSRE building will be used for processing the fuel by treatment with gaseous
HF and F,,.

Design of the MSRE was begun early in the summer of 1960. Orders for special materials
were placed in the spring of 1961. Major modifications to Building 7503 at ORNL, in which the
reactor is installed, were started in the fall of 1961 and were completed by January 1963. .

1Also sold commercially as Inco No. 806.
2Grade CGB, produced by Carbon Products Division of Union Carbide Corp.

9
10

Fabrication of the reactor equipment was begun early in .1962. Some difficulties were experi-
enced in obtaining materials and in making and installing the equ.ip.ment, but the essential instal-
lations were completed so that prenuclear testing could begin in August of 1964. The prenuclear
testing was completed with only minor difficulties in March of 1965. Some modifications were
made before beginning the critical experiments in May, and the reactor was first critical on ]une_ 1,
1965. The zero-power experiments were completed early in July. Additional modifications, main-
tenance, and sealing of the containment were required before the reactor began to o'pegaté at ap-
preciable.power. This work was completed in December.

Operation at a power of 1 Mw was begun in January 1966. At that power level, trouble was
experienced with plugging of small ports in the control valves in the off-gas system by heavy
liquid and varnish-like organic materials. These materials are believed to be produced from a
very small amount of oil that leaks through a gasketed seal and into the salt in the tank of the
fuel circulating pump. The oil vaporizes and accompanies the gaseous fission products and
helium cover gas purge into the off-gas system. There the intense beta radiation from the krypton
and xenon polymerizes some of the hydrocarbons, and the products plug small openings. This dif-
ficulty was largely overcome by installing a specially designed filter in the off-gas line.

Full power — about.7.5 Mw under design conditions — was reached in May. The power is lim-
ited by the heat-removal capability of the salt-to-air radiator heat-dump system. The plant was
operated until the middle of July to the equivalent of about one month at full power when one of
the radiator-cooling blowers — which were left over from the ANP program — broke up from me-
chanical stress. While new blowers were being procured, an array of graphite and metal surveil-
‘lance specimens was taken from the core and examined. o

Power operation was resumed in October with one blower; then in November the second blower
was installed, and full pbwer was again attained. After a shutddwn to remove salt that had acci-
dentally gotten into an off-gas.line, the MSRE was operated in D%—:-cember and January at full power

for 30 days without interruption. A fourth power run was begun l‘iate in January and was still in
' i
i

In most respects the reactor has performed very well: the fujel has been completely stable,

progress after 31 days at the end of this report period.

the fuel and coolant salts have not corroded the Hastelloy N cont:ainer material, and there has
been no detectable reaction between the fuel salt and the graphit!e in the core of the reactor. Me-
chanical difficulties with equipment have been largely confined tfo peripheral systems and auxil-
iaries. Except for the small leakage of oil into the pump bowl, tl:le salt pumps have run flawlessly
for over 10,000 hr. | \
Because the MSRE is of a new and advanced type, substantial research and development ef-
fort is provided in support of the operation. . Included are engineering development and testing of
réactor components and systems, metallurgical development of materials, and studies of the chem-
istry of the salts and their compatibility with graphite and metals both in-pile and out-of-pile.
Conceptual design studies and evaluations are béing made of large power breeder reactors
that use the molten-salt technology. ‘Some research and development is being directed specif-
ically to the requirements of two-region breeders, irfcluding work on materials, on the chemistry

of fuel and blanket salts, and on processing methods.
i»

1Y

Part 1. Molten-Salt Reactor Experiment

1. MSRE Operations

P. N. Haubenreich

1.1 CHRONOLOGICAL ACCOUNT OF OPERATIONS AND MAINTENANCE

. Guymon H. C. Roller

R. H

J. L. Crowley R. C. Steffy

T. L. Hudson V. D. Holt

P. H. Harley A. 1. Krakoviak
H. R. Payne B. H. Webster
-R. Blumberg C. K. McGlothlan

The reactor shutdown that started in July! continued through September. The first of the two
specially construc_:téd replacement blowers was delivered on September 28, ten weeks after the
reactor was shut down. Meanwhile, the time was fully occupied with a host of other jobs that were
completed about the time the replacement blower was received. These included removing and re-
placing core samples, work on control-rod drives, replacement of the special fuel off-gas filter,
modification of the radiator door seals, and repairs and modifications in the cooling-water system.

The first step in the reactor startup in run 8 was seven days of flush-salt circulation (see Fig.

1.1). During this time the salt that had frozen in the sampler line at the pump bowl was thawed.

After the temporary heaters for this job were removed, the reactor cell was sealed, and the leakage

~ was shown to be acceptable by a test at 10 psig. By this time the first blower was ready to run,

but delivery of the second replacement unit was not expected for several weeks. Therefore,
nuclear operation was resumed early in October with only one blower.

The reactor was operated for 17 days at the maximum power attainable with one blower: 5.8
Mw. During this time the pressure drop across the new off-gas particle trap increased to several
psi.. The inlets to the main charcoal beds also became restricted and had to be relieved by back-
blowing with helium. Two days after the start of power operation, the fuel off-gas line became

plugged near the pump bowl, causing the off-gas to be diverted through the overflow tank. This

1MSR Progr&frl Semiann. Progr. Rept. Aug. 31, 1966, ORNL-4037, p. 9.

11
12 ! .
ORNL-DWG 66-12792R

INVESTIGATE GO TO OPERATE  REMOVE CORE SPECIMENS REMOVE

OFF-GAS PLUGGING FULL POWER AT'POWER  REPLACE MAIN BLOWER SALT PLUG.
CHANGE REPAIR THAW FROZEN LINES ’ CHECK
FILTERS AND VALVES SAMPLER TEST CONTAINMENT CONTAINMENT
LOW-P —
DYNAMICS CHECK \ OPERATE
TESTS CONTAINMENT : AT POWER
8
36
2 NUCLEAR POWER Tl 1
; 4
o)
o 2 !
0 III | 3
[ ] T F
SALT CIRCULATING | FYEL @
(o] ! !nlm:ml[ ¢ e —n

J F M A M J J A

Fig. 1.1. MSRE Activities in 1966.

’

complicated the routine recovery of salt that gradually 'accumulétes in the tank, and during some
recovery operations, activity was forced into the line that drains oil leakage away from the pump.

The reactor was shut down to install the second blower unita, which had just been delivered,
and to attempt to relieve the plugged line at the purfip bowl. The system was flushed, the reactor
cell was opened, and heat was applied to the line. While the line was hot, pressure was applied
and the line opened up. Tests showed that, within the accuracy of the available instrumentation,
the pressure drop was then normal. The heaters were removed, ‘%the cell was sealed, and nuclear
operation was resumed seven days after the fuel was drained for the shutdown. Also during this
time, a gas flow element in the vent from the oil catch tank was'li removed and an alternative flow
measurement provided. The flow element had become plugged \x;hen, as the last step in the fuel
drain, the overflow tank was emptied and gas from the pump bowl again vented out through the
oil drain line. |

When the power was raised on November 7 for run 9, the temi)erature of the off-gas line showed
that it was already plugged and that the gas was again bypassing through the overflow tank. To
prevent the transfer of activity into the oil drain line when salt was recovered from the overflow
tank, the nuclear power was reduced several hours before eaéh t:fransfer. Limitations on the amount
of salt that can be tolerated in the overflow tank and the heel th“at remains after a recovery re-
quired that the power be reduced for a transfer every two days after the initial heel had accumu-
lated. Nuclear operation was continued in this fashion for 12 days while heaters, tools, and
procedures for more positively clearing the off-gas line were devised. Then the reactor was shut

down to work on the line and also to check what appeared to be a high inleakage of air into the

reactor cell.
13

Search for the restriction in the off-gas line revealed thin plugs in the flanges at both ends of
a removable section near the pump bowl. The plugs were easily poked out, and the line was then
shown to be clear by viewing, probing, and pressure-drop measurements. The plugs were attributed
to flush salt almost completely blocking the line as a result of the overfill in July. An incon-
sequential amount of this salt was also seen in the 4-in. holdup pipe.

The high inleakage into the reactor cell proved to be from valve-operator pneumatic lines.
Since these lines are protected by automatic block valves, the leaks did not violate containment.
Therefore,‘flowmeters were installed so their input could be taken into account in the routine
monitoring of cell leak rate during operation.

During this three-week shutdown, we also made some repairs and modifications to the cofilpo—
nent cooling blowers and removed and repaired an air valve in the reactor cell.

"~ Run 10 began on December 14 and continued for 30 days at full power.

During the first two weeks of the run, the pressure drop across the particle trap in the fuel
off-gas line repeatedly built up and had to be relieved by forward- or back-blowing with helium.
Examination of th_e first partiéle trap, removed in August, had shown that heating the central
inlet tube would tend to jam it into the first-stage filtering medium (see p. 47). To test the effect
of reducing the heating, for the last two weeks of run 10, the off-gas was delayed on its way to
the particle trap by routing it through the two empty drain tanks. When this route was followed,
the particle trap pressure drop came down and stayed down. '

The inlets to the main charcoal beds had to be back-blown during the first week of run 10 but

not afterward. . _
During this run the UF ; concentration in the fuel salt was increased by the addition of 16 g

of beryllium metal through the sampler-enricher. One purpose of inbreasing the reducing power
of the salt was to investigate the effects on volatile fission product compounds (see p. 123).
Another was to alleviate concern over possible corrosion. Practically no corrosion had been
seen (<0.1 mil of generalized corrosion in 20 months), but the absence of cotrosion depends on
maintaining a reducing environment, and a larger margin was desired.

Operation at full power was to be interrupted after 30 days to permit inspection of the new
blower hubs and blades, which had by then been run over 1000 hr. But toward the end of the run
two conditions developed which caused us to drain the reactor and extend the shutdown. The heat
exchanger between the treated-water and tower-water systems began to leak at an increasing rate,
and the leakage from the air lines in the reactor cell became so large that the measurement errors
clouded the determination of the cell leak rate.

When the reactor was shut down, inspection showed that the blowers were in excellent condi-
tion. The leaks in the air lines were traced to deterioration of neoprene seals in some quick-
disconnects. All the disconnects in the reactor cell were replaced with metal-compression fit-
tings, and the leakage was stopped. The heat exchanger leaking water was replaced. The filter
assembly and the pressure control valve in the fuel off-gas line were removed, and a new filter
assembly was installed. This consisted of two filters in parallel, each with much larger frontal ‘

area than the old filter (see p. 42).
14

Nuclear operation was resumed in run 11 on January 28 anqi continued without interruption
(except for 2 hr to investigate a false temperature alarm) through the end of the report period,
February 28. No difficulties of any consequence were encount‘_lered, and the program of adjusting
the fuel UF , concentration and observing the effect on volatile fission products continued.

Details of operations and maintenance during this report per1od are given in the sections
which follow. Although the emphas1s tends to be on the troubles the reactor was in operation
most of the time, and the operatmn was in most respects quite/satisfactory. Table 1.1 summarizes
some of the history. Salt was circulated in the fuel and coolar%.t loops for 60 and 82%, respectively,

of the time in this report period. The reactor was critical 53% of the time.

3

Table 1.1, Summary of Some MSRE Operating Data

Aug. 31, 1966 Feb. 28, 1967 . Increase
Time critical, hr : 1775 4092 ' 2317
Integrated power, Mwhr 7823 _' 21,514 13,691
Salt circulation, hr ) ‘
Fuel system 4691 7337 2646

Coolant system 5360 1 8946 3586

1.2 REACTIVITY BALANCE
J. R.'Engel

The purpose of the reactivity-balance calculation dufing po%ner operation of the reactor is to
provide current information about the nuclear condition of the sj}stem. During this report period,
improvements were made in the calculations, making it possible to detect vety small anomalies
in reactivity behavior; none was observed. Calculations made éluring previous ’periodsz of opera-
tion did not include the !33Xe poisoning term because the mathematical model, with the coef-
ficients then available, did not adequately reflect the xenon behavior in the reactor. When power
operation was resumed in October 1966 (run 8), we included a calculation‘of the xenon effect to
provide complete reactivity balances. Subsequently, the overall calculation was improved by
modifying some xenon stripping parameters to improve the descr”1pt1on of the xenon transients,

and by including long-term isotopic change effects that had beep previously neglected.

Experience

Figure 1.2 shows the results of some .of the on-line reactivify balances calculated in runs

8, 9, and 10. During runs 8 and 9 the parameters used to calculate the xenon poisoning were

2Ibid., pp. 10—13.
15

ORNL—DWG 66—41944R

+o4 REACTIVIT o
] Y j
4008 e A "&L* sy { | 005 BT — 4 i
z 0 3 ‘M\h ,4 v, \"“'J"E‘ o % & sl ..-“ii . Lg. 4
3 > ] : D i PRIV R
@ _ 4 . o ¢ =
o0 : s » o0 REACTIVITY
04 s -0 | | |
-045 : *L. -0.45 = :
8 POWER : 8 ;
i | [ ] ] z
6 gl 6 —H—t I a—
x 4 c <
3 =% 4 —
2 i 2 PO‘|NE_R
o . - 0
8 10 12 14 16 18 20 22 24 26 28 30 7 9 1 13 15 17 19
) OCTOBER, 1966 . . , NOVEMBER, 1966
+040 | l ‘ ‘ T T 7 (—— |
+0.05 |—= 5 REACTIVITY L. w __| 41* l_ s,
ey
0 WBhe $upa, P ! PRI L m;.r A i
3 L’ O AR T NPT &
Z_ 005 ‘“‘l Ford !‘7* —'|F' il : :
~ g ) i .
o F
2 -0.10 +
-045 —t -
COMPUTER OUT * i
-020—}  OF SERVICE :
ot I I
B :
. I E
| { g ]
x4 POWER B
=
2
0 : ‘
14 18 18 20 22 24 26 28 30 1 3 5 7 9 T 13 15

DECEMBER, 1966 JANUARY, 1967

Fig. 1.2. Residual Reactivity During Power Operation in Runs 8, 9, and 10.

found to describe the steady;state condition reasonably well, but the transients were not well
described. This difference between the calculated and actual xenon transients produced the
cyclic variation in the residual reactivity in November when the nuclear power was cycled between
0 and 7.2 Mw. Detailed analysis of the observed xenon transients led to changes in some param-
eters which produced much better agreement between the calculated and observed transients in run
10. However, there is still a small difference (~0.03% 8k/k) between the calculated and observed
steady-state poisoning at 7.2 Mw. This causes the apparent change in residual reactivity between
zero power and the condition with steady-state xenon poisoning.

The larger negative reactivity transients in Fig. 1.2 are all due to off-normal operating condi-
tions not accounted for in the reactivity balance. In all except one of these transients, the nega-
tive reactivity resulted from an increase in the circulating void fraction while the salt level-in the
fuel-pump tank was abnormally low. The low level, in turn, was caused by low system temperatures
which followed instrument-initiated power reductions. In all these cases the excess voids were
stripped out and the reactivity recovered when the normal pump level was restored by increasing
the system temperature. The negative reactivity excursion on October 23, 1966, was caused by an
abnormally high salt level in the pump tank which reduced the efficiency of xenon removal by the
pump spray ring. The negative reactivity was produced by the transport of additional 135Xe into
the graphite moderator. Since the removal of xenon from the graphite is a slower process than

bubble stripping, the recovery in this transient was slower than in the other cases.
16

Operation at power was resumed on January 28, 1966, and has continued without major inter-
ruption through the end of this report period. After the initial blilildl.lp of 135Xe poisoning, the
residual reactivity has been between 0 and + 0.04% &k/k except.fo'r one incident in which excess
circulating voids were introduced by abnormal operation (low salt level in the pump tank).

The long-term drift in the reactivity balance is summarized in Fig. 1.3, which shows the
residual reactivity as a function of integrated power from the start of power operation through run
10 (ending January 15, 1967). The reference condition for this figure is the system condition at
the start of run 4 (December 20, 1965). Because of the remaining small uncertainty in the !35Xe
term, the represeritative results shown here are taken at zero polwer with no xenon present. There
appears to have been a positive shift of about 0.05% 8k/k during the first 1000 Mwhr of operation
that has remained relatively constant since that time. No speci‘fic cause has yet been established
for this shift. However, the change is nearly as small as the efitimated confidence limits of the
calculation (£0.04% &k/k) and is much smaller than the operating limit on the reactivity anomaly,
which is %0.5% 8k/k. .

Previous reports of the reactivity behavior? suggested the possibility of a significantly larger
positive reactivity anomaly. However, those tentative conclu»sio‘ns were based on data from early
results which have since been corrected. The earlier balances did not include the reactivity
effects of isotopic changes (other than ?35U) or of flush-salt dilutions. Correction for these
effects made a net reduction in the magnitude of the apparent an‘omaly. (The calculation of the
- isotopic-change effects is described on p. 83 of this report.) In .addition, preliminary analysis of
some pressure-release experiments indicated a circulating void f!r_action of 1 to 2% by vélume in

the fuel salt.? If such a void fraction had been present at steady state, the negative effect of the

31bid., pp. 22-24.

ORNL-DWG 67-1076A

0.08
® [ ]
= .
\ .
® 004 s 3 PR *
(3 °
- [ ]
E 3
S :
5 0 "
<
wi
ax
.}
<L
=2
9 _0.04
(723
w
x
-0.08 : ‘ - -
0 2000 4000 6000 8000 10,000 12,000 14,000 16,000 18,000

INTEGRATED POWER (Mw-hr)

Fig. 1.3. Long-Term Residual Reactivity in the MSRE.
17

;

voids would have required an anomalous positive reactivity effect of 0.1 to 0.2% &k/k in order to’
produce the net results that were observed. However, as explained in the following section, later

. experiments indicate that the bubble fraction is much less.

Circulating Bubbles

We have concluded, after intensive analysis of the data, that the pressure-release experiments
do not permit explicit evaluation of the steady-state circulating void _fraction. In the preliminary
analysis of these experiments, the level rise in the pump tank and the reactivity loss associéted
with a pressure release were attributed entirely to circulatifig voids, and a void fraction of 1 to 2
vol % was required to account for the observed effects. However, the excess voids in circulation
immediately after the pressure release disappeared from the system very rapidly (indiéated bubble
stripping efficiencies for the spray ring in the pump tank were 50 to 100%), and this was incon-
sistent with a large steady-state void fraction. In addition, the reactivity loss due to 1 or 2 vol
% voids (0.2 to 0.4% ok/k) that should have been observed upon starting circulation was apparently
absent.

The existence of circulating voids immediately after a rapid decrease in pressure does not
require the presence of a comparable circulating void fraction prior to the decrease; it requires
only a source of gas inside the body of the liquid. This source could be fixed void that is inac-
cessible to the salt so that its volume is essentially independent of the system overpressure.
(One pdssibility might be the pores in the graphite and the spaces between the graphite stringers
in the core.) The inventory of gas in such void would increase with increasing overpressure, and,
when the pressure is released, this excess inventory would go into circulation with the salt. The
pressure-release experiments give only an upper limit of 1 to 2% for the circulating void fraction,
if circulation of all the voids is assumed. Therefore, we have used other measurements to esti-
mate the circulating void fraction.

At the start of run 8, in October 1966, a test was performed to evaluate the reactivity loss
associated with the buildup of the steady-state circulating void fraction. Prior to this startup
the fuel salt had been in a drain tank for 11 weeks and should have been free of bubbles. When
the circulating loop was filled with this salt, there was no tendency for the pump-tank level to
decrease with increasing overpressure. This incompressibility of the fluid Supporfed the assump-
tion that no voids were present initially: The reactivity change from stationary fuel salt to steady-
state circulation was —0.23 to —0.25% 8k/k. An early measurement of this effect, when there was
no evidence for circulating voids, gave a reactivity change of —0.21%, in good agreement with the
calculated effect due to the loss of delayed neutrons. Thus the reactivity loss due to circulating
voids could be only about 0.02 to 0.04% &k/k. This corresponds to a circulating void fraction of
0.1 to 0.2% by volume. A void fraction in this range is consistent with the results obtained from.
the detailed analysis of the !33Xe transients and is also within the range of values that would
explain the observations in the pressure-release experi’m‘ents. Therefore we conclude that the cir-

culating bubble fraction is in fact only 0.1 to 0.2%, under normal conditions, and the presence of
18

the bubbles does not seriously affect the conclusions regarding|the long-term changes in residual
reactivity, |

1.3 THERMAL EFFECTS OF OPERATION
|

‘Radiation Heating

C. H. Gabbard H. B. Piper

The evaluation of radiation. heating effects onlthe fuel pump and on the reactor vessel con-
tinued. “

Fuel-Pump Tank. — The temperature distribution on the u'ppzler surface of the pump tank as a
function of reactor power was presented in the previous semianx-;ual report“_ for the condition of
30-cfm cooling-air flow. Since the startup for run 8, there has bl‘een an unexplained shift downward
in these temperatures. A comparison of the present temperature‘is with those observed previously is
shown in Fig. 1.4. The two lines show the variation in temperautures with reactor power that ex-
isted during runs 6 and 7 with 30-cfm cooling air. The solid points are the temperatures observed
during runs 8 and 9 with 30-cfm cooling air. The open points are those observed during runs 9, 10,

and 11 after the cooling air was turned off. The pomts at 5.5 Mw were taken when the reactor

outlet temperature was at 1225°F rather than the normal 1210°F! This downward shift in tempera-

*Ibid., pp. 27—28.

ORNL-DWG 67-4761

1200 —= I I 7
A
400 i | &7 B
i}A A
1000 /—— o
— [— O
.|5I_-. | H COOL‘M—#"W/ A
w a | 07 DATA W__\_L___-——' A
% B-RUN 5 AN . .
g 900 | : AR -
< H S
i ATA WITH =
Z : .
[ ol |’
800 ‘ ‘
‘|
L ® RUN 8 AND 9 DATA WITH COOLING POSITION "a"
700 A RUN B8 AND 9 DATA WITH COOLING POSITION “B"
© RUN 9,10 AND {1 DATA WITHOUT COOLING POSITION “A"
4 RUN 9,10 AND {1 DATA WITHOUT COOLING POSITION "B"
600 :
0 i 2 3 q 5 ‘ 6 7 8

POWER (Mw)

Fig. 1.4. Comparison of Fuel-Pump-Tank Temperatures.
19

’ture was first believed to be the result of a salt cake that remained.on the metal surfaces after
the overfill incident at the end of run 7. However, a review of the pump-tank temperatures during
this and other filling operations showed that the temperatures were normal before the salt reached
the pump and before the cooling air was tumed on. The calibration of the air-flow meter was
checked and was found to be correct. '

The cooling air to the pump tank was turned off during the attempts to melt out the salt plug
in the _522 line, and although the temperatures on the pump-tank surface were higher than with the
cooling air, the temperature gradient was less. Since this temperature distribution is satisfactory
and may actually be better than with the air cooling, the use of air cooling was discontinued
during run 9. Fuel-pump cooling air had been used during all power operation of the reactor prior
to run 8, and no comparison with previous data is possible. The temperature distribution for
full-power operation has remained essentially constant since .the use of air cooling was dis-
continued.

The thermocouple response when the air was turned off was checked to éee if there was any
indication that the thermocouple attachments had loosened. The transient data indicated that
the largest temperature error, caused by air flow over the thermocouple, was of the-order of 20
to 40°F. Although a direct comparison is not possible, these results appear to be consistent
with previous data. The most probable explanation for the decrease in temperatures is either
that the cooling shroud was accidentally shifted durin.g the maintenance operations between runs
7 and 8 or that some shift has occurred in the air-flow measurement even though the calibration |
of the differential pressure cell was correct.

Reactor Yessel. — dertain temperatures on the reactor véssel are monitored continuously to
determine the effects of radiation heating and to determine if there is any indication of a sedimen-
tation buildup in the vessel, The temperature differences between the reactor inlet and the lower
head and between the inlet and the core-support lugs are monitored by the computer. Previous
data® indicated temperature differences of 1.5 and 2,0°F/Mw for the lower head and for the: core-
support lugs respectively. The temperature differences have now increased to 1.64 and 2.44°F /Mw,
but this increase is not believed to be serious or indicative of a sedimentation buildup, because
this represents a maximum increase of only 3°F at 7.2 Mw, which could easily be the result of
temperature measurement errors. The past and future data will be examined more closely to

determine if a trend in the temperature differences does exist.

Thermal Cycle History
C. H. Gabbard
One of the factors that limits the life of the reactor system is the degree of thermal cycling

on certain critical components, particularly the freeze flanges. The current thermal-cycle history

of all the components thought to be sensitive to thermal-cycle damage is shown in Table 1.2, Of

SIbid., p. 28.
20

Table 1.2. MSRE Cumvulative Cycle History Through February 1967

| Thaw
Component Heat/Cool Fill /Drain Power On /Off Thaw and
) ’ ! ' Transfer

Fuel system - 7 32 40
Coolant system 5 10 36
Fuel pump 8 28 40 409
Coolant pump 6 11 36 103
Freeze flanges 100, 101, 102 7 28 40

- Freeze flanges 200, 201 6 10 36
Penetrations 200, 201 6 10 36
Freeze valve 103 5 ! 29 30
Freeze valve 104 11 8 22 -
Freeze valve 105 15 17 42
Freeze valve 106 17 25 38
Freeze valve 107 10 11 18 -
Freeze valve 108 9 17 14
Freeze valve 109 -9 20 18
Freeze valve 110 2 .2 3
Freeze valve 111 5 4 4
Freeze valve 112 2 2
Freeze valve 204 6 : 15 22
Freeze valve 206 6 13 20

the cycles shown, the ‘‘Fill/Drain,”’ ‘“Heat/Cool,”’ and ‘“Power’’ cycles are the most important,
y

in that order.

The “F111/Dram” cycle consists in filling the System at 1200°F, starting circula-

_ txon and then draining the system. The ‘‘Heat/Cool’’ cycle cons1sts in heating the empty system
14 Y g

from room temperature to 1200°F and then cooling the empty syStem back to room temperature.

The power cycle consists in raising the reactor power from zero to full power and then returning to

Zero power.

. i o
Some of the reactor operations do not fall clearly into single cycles of the types shown; partial

power cycles and overall system temperature changes are examples. These effects are accumu- ' "

lated and charged as equivalent power cycles, which is why there are more power cycles in the

fuel system than in the coolant system. Approx1mate1y 54% of the design cycles have been used

to date.

S. J. Ball

1.4 REACTOR DYNAMICS

Dynamics tests made during the initial approach to full power5 had indicated that the inherent

stability characteristics of the MSRE were quite satisfactory and in good agreement with the

51bid., pp. 29—34.
21

| Ifiredictéd behavior.” Subsequent frequency response tests were made at the start of run 11 with
the reactor at three different power levels in order to detecf any cfianges in reactor dynamic
behavior resulting from a year of power operation (16,750 Mwhr). Pseudorandom binary reactivity
insertioné were used in tests at power levels of 1, 5, and 7 Mw.

The results of these tests showed no detectable changes 'in the MSRE dynamic characteristics
due to aging, thus indicating continuing satisfactory behavior. 7

Two problems encountered in the analysis of the first series of tests were offsets, or biases,
in the magnitude ratio curves for tests run at a given power level, and some unexpected low-
frequency periodic inputs that showed up as spikes in the autocorrelation functions of the rod
motion input signal.® Neither of these problems recurred in the latest tests. The appearance

and disappéarance of these anomalies are as yet unexplained.

1.5 EQUIPMENT PERFORMANCE

Heat Transfer

C. H. Gabbard H. B. Piper

| Primary Heat Exchanger. — The evaluation of the effects of prolonged operation on the heat
tfansfer in the salt-to-salt heat exchanger was continued. Heét transfer data at six different time
periods wére evaluated, using the derivative method,® with the results shown in Fig. 1.5. Although
the calculated coefficients show a slight downward trend with time, we believe the apparent trend
in heat transfer is within the uncertainty band of the calculations. A simple but sensitive indi-
cator of any change in heat transfer is obtained by dividing the reactor power by the overall tem-
perature difference between the fuel leaving the core and the coolant leaving the radiator. A
lower value of this ratio indicates poorer heat transfer, since the fuel and coolant flow rates are
not variable, Data from full-power operation in runs 6—11 are shown in this form in Fig. 1.5.
Taking into consideration both the calculated coefficients and the simple indexe‘s, we conclude
there has been little or no change in the heat transfer performance of the heat exchanger.

Radiator. — The reactor was operated at various power levels at the beginning of run 11, and
data were taken to evaluate the heat removal at different radiator settings. The heat-removal
cépability of the radiator as a function of radiator setting is shown in Fig. 1.6. - .

The same set of data was used to calculate the overall heat transfer coefficient of the radiator
as a function of the air pressure drop across the radiator. The results of these calculations are
shown in Fig. 1.7. The increase in the heat transfer coefficient when the second blower was
energized is probably caused by direct impingement of the air from the blower discharge against

the radiator tubes. The various-air pressure drops were obtained with the radiator doors fully

7S. J. Ball and T. W. Kerlin, Stability Analysis of the Molten-Salt Reactor Experiment, ORNL-TM-1070
(December 1865). - '

8T. W. Kerlin and S. J. Ball, Experimental Dynamic Analysis of the Molten-Salt Reactor Expetiment,
ORNL-TM-1647 (October 1966).

"MSR Program Semiann. Progr. Rept. Aug. 31, 1966, ORNL-4037, pp. 35—37.
22

CRNL DWG 67-4762

700

"y

° 600 S
o e | . -
. A i‘ L L
[ oy ; [+]
£ 500 005 )
~ =
2 s
o >
| S—— . w
. 400 : : 0.04 2
5 erezre 22 =z ’ 7727077777, Z7772 -
s o
] L
re %
W 300 [ S 003 2
(@] ) <L
o x
x ‘: [
w I ) =
$ 200 1‘ ]
2 ® HEAT TRANSFER COEFFICIENT ooz Y
i ZZHEAT TRANSFER INDEX (HTI)

> . HTI = HEAT BALANCE POWER .

g 100 REACTOR CUTLET TEMPERATURE — RADIATOR OUTLET TEMPERATURE

I.:.;:J 0.04

APR MAY JUN JUL AUG SEP ocT NOV DEC JAN FEB MAR
1966 1967

Fig. 1.5. Observed Performance of MSRE Main Heat Exchanger.

open by changing the bypass-damper position. A direct measure of the radiator-air outlet tem-
perature can be made only when the bypass is fully closed, and even then the air from the an-
nulus blowers would cause an error in the measured temperature rise across the radiator. The
outlet air temperatures for the other conditions were calculated from air and salt heat balances
around the radiator. Corrections were applied for the effects of the annulus blowers. However,
the measured temperature rise of the air across the radiator tubes was about 17% higher than the
calculated rise in the two cases where the bypass damper was‘ closed and a measurement was
possible. This discrepancy indicates that either the air-flow Jneasurernent, the temperature meas-
urement, or the reactor heat balance is incorrect. The heat balance and the temperature measure-
ments were assumed to be correct, and the calculated outlet air temperatures for the other radiator
conditions were ;:orrected to be consistent with the values measured when the bypass damper was
closed.

The coefficients for the two conditions when the bypass damper was closed were 28.5 and

38.5 Btu hr—! ft—2 (°F)~!, These are the same values that w“ere calculated from June 1966 data

for similar operating conditions. ‘
The apparent error mentioned.above is about the same as the 15 to 16% discrepancy between

‘the air heat balance and the salt heat balance reported previously!® for similar conditions. This

19pid., pp. 37—38.
23

seems to indicate that the discrepancies are the result of some specific error in measurement
rather than random variations in the data. The radiator air instrumentation was not intended to be
accuratej enough to permit a precise heat balance on the air. The stack is not;sufficiently long to
ensure either a uniform velocity distribution or a uniform temperature distribution across the
stack cross sectién. Since the velocity and tempetrature are measured only at a single point in the
stack, any flow ot temperature asymmetry could cause a significant error. There is also the pos-
sibility that the initial calibration of the air flow instrument was incorrect.

Methods of Improving the Heat Transfer. — A stu?i\y was completed to determine whether the
maximum power capability of the reactor could be raised by some convenient method.

For normal operating conditions, an upper limit of 1210°F has been placed on the reactor

outlet temperature. This temperature was selected on the basis of thermal stress cycling and

© ORNL-DWG 67-4763

100 /3 . a s
—_ i ' /
s & . | - !
7] A
o
| R
3 / /
(]
E:E 60 i — ‘ :
&o
. Eao A
25 | |/ /
&340 / :
& / A
a
2 | /
[ ]
(]
20 / /
2 ' TWO MAIN
= ‘ l BLOWERS
H RUNNING
77 0] & - A »
g /
<f
S / BOTH MAIN BLOWERS OFF
80 / 7
—~— N ‘
z / ;
a
5 /
82E e0
z&
Q4 / . ONE MAIN BLOWER RUNNING
Ew
wa
o= / o .
Qo 40
O ! A,
o / /
o
[
20 / 7 INLET AIR TEMPERATURE =40°F
/| T
O .
o 1 2 3 4 5 6 7 8

" HEAT BALANCE POWER (Mw)

Fig. 1.6. MSRE Radiator Performance.
24
|
| * »

|
ORN L‘- DWG 67-2155

100
'—
=z
w
o
(T8
w
< |
O . 50 -
O T 1
m ° 1} n ‘
§NC MB—4 AND MB-3 P—N“r“,_—o-'
2 J./”d/ .
é £ /.'.—
= ™~ 1
3 -t
|<—I o — MB—" ||0Nu
w = 20 :
T
4
3
a
o
wt
>
© 10
{ 2 5 iC ~ 20
AR PRESSURE DROP (in. HZO) . .
i
I
Fig. 1.7. Effect of Air Pressure Drop on Radiator .

Heat Transfer Coefficient,

|
stress rupture life of the reéctor system. The minimum coolant éalt temperature in normal opera-
tion has been se‘t at 1000°F, in order to reduce the probability of freezing the radiator in case
salt flow is interrupted. These two temperature limitations and the heat transfer capability of
the main heat exchanger limit the reactor power to about 7.4 MW.!i In the summer months this
power limitation about coincides with the capacity of the coolant radiator. However, during
colder weather the radiator heat transfer capacity increases, and a portion of the cooling air
must be bypassed to avoid overcooling the coolant salt.

\_ The reactor power can be increased only by improving the hqiat transfer performance of both
the radiator and the main heat exchanger. Other than replacement with a larger unit, the heat-
exchanger capability can be increased only by increasing the flow rates of the fuel and coolant
systems. Increasing the temperature difference between the fuel and coolant system is undesirable
because of adverse effects on the thermal-cycle and stress-rupture life of the reactor system.

that could be obtained and to

A study was completed to determine the maximum flow rates
determine the increase in heat transfer performance that would result from these flow increases.
The pumps are capable of accepting larger-diameter impellers and of operating at a higher horse-
power rating so that increased flow is possible. The flow can also be increased by using a
,higher-freqfiency power supply to increase the rotational Speed.olf the pumps. Slightly higher flow
rates can be obtained with the higher rotating speed, because the horsepower rating of the drive
motors can be increased at the higher speed. Calculations indicate that the maximum possible

flow rates for the fuel and coolant systemé would be 1530 and 1644 gpm respectively, assuming
that the pumps themselves are the limiting factor. Using these maximum flow rates and operating
between the temperature limits that now permit 7.2 Mw, the heat fransfer capability of the heat

exchanger would be increased to 8.1 Mw. Even this modest increase might not be practical, how-
25

ever, because of undesirable effect_s of increasing the flow. For example, development tests
suggest that increasing the flow causes more gas bubbles to be introduced into the circulating
salt by the stripper jets in the pump bowl. This would be undesirable, because it would introduce
more uncertainty into the reactivity balance,

In the radiator, increasing the salt flow has very little effect, since over 95% of the heat
transfer resistance is on the air side of the tubes. There is no way to improve the radiator per-
formance without major expense. Additional surface area could be provided by adding more tubes
or by adding some type of fins to the tubes, but either would be difficult and time consuming.
Additional air capacity could be provided, but this too would be a major undertaking. The radiator
air flow would have to be incrgased by a factor of about 1.8 to remove 10 Mw. This would in-
crease the air pressure drop to ~35 in. H,O and the power requirements' of the blowers to about
2900 hp, as compared with 500 hp for the present blowers. Neither the present blowers nor-the
building electrical system is capable of meeting these requirements. Oné additional blower could
possibly increase the radiator heat removal to the 8.1-Mw level, which would be consistent with
the maximum possible power of the heat exchanger, However, the blowers would be operating very
close to their surge limit, new drive motors might be required to avoid an overload condition, and
the existing building electrical system would be unable to supply the third blower.

In conclusion, the difficulties.in raising the power capability of the reactor far outweigh any
advantages that could be gained from the relatively small power increase that can be reasonably
achieved. Since the objectives of the MSRE can be met with the present heat removal system, no
attempts to increase the power capability are planned. ‘

"

Main Blowers

C. H. Gabbard.

The main blowers, MB-1 and MB-3, which had failed at the end of run 7,!! were rebuilt by the
manufacturer and returned to normal service. By the end of the report period, MB-1 and MB-3 had
operated without difficulty for 2100 and 1675 hr, respectively, since they were repaired. Main
blower 1 has been inspected twice since it was repaired, and main blower 3 was inspected once,
the first inspection of MB-1 coming after 350 hr of operation -while MB-3 was being installed. Bot-h
blowers were inspected at the end of run 10 — after 1350 and 930 hr of operation. The inspections
included dye-penetrant inspections of the blading and hubs, a visual inspection of the coupling,
an alignment check, and a retorquing of all the bolts. With the exception of a few bolts that were
retightened, both blowers were completely satisfactory.

We attempted to find the cause of the failures at the end of run 7, but the specific cause is
uncertain. The best explanation appears to be that one of the blades failed first along one of

several existing cracks and that the resulting unbalance, or impact with the blade fragments,

rbid., p. 40.
26

caused the remaining damage. The origin of the existing cracks is also unknown, but they
probably occurred either during fabrication or during the overspéed_ test.

The rebuilt units have reinforced hubs and have magnesium alloy blades that are 35% lighter
than the original aluminum ones. The new units were also giveP a 30% overspeed test, with a
dye-penetrant inspection of the hubs and blading before and after the test. The manufacturer had
difficulty in casting the magnesium alloy blades free of heavy 'surface porosity and cracks.
Rotor assemblies were rejected on three occasions because of ¢racking in the blades after the
overspeed test. In each of these cases, cracking had been preshent prior to the test, but it had
been removed by surface grinding. The units that were accepted, including a spare unit, were
free of objectional defects. “

The rebuilt rotor assemblies were installed in MB-1 and MB:3 under the supervision of the
manufacturer’s service engineer. The rotor and drive-motor sha%ts were carefully aligned, and
the rotors were dynamically balanced in place. Instrumentationiiwa's provided to monitor the
vibrations and the bearing temperatures of the two blowers whil(“-:‘ they are in operation. Vibra-
tion normally runs below 1 mil, although greater vibrations deve'iope(‘i on two occasions, once
when ice built up on the blades of MB-1 and once when its bearings became very cold (below
15°F).

Radiator Enclosure
M. Richardson |
' ‘i

The coolant-salt radiator operated continuously with salt cifculating at temperatures between
1000 and 1200°F for the last five months of the report period.

The modifications of the door seals'? proved effective in reducing air leakage and consequent
heat losses to a satisfactory minimum. Although close examina%ion was not possible after the
radiator went into operation, external examination on January 171; showed little deterioration in
more than 3400 hr at high temperature. The inlet door was in ex!cellent condition, with good con-
tact at top and bottom between the linked hard seals and the sofjjt seal on the face of the enclosure.
On the outlet door the hard seals appeared to be in good condition. There was a tighi seal across
the top but a gap of 1/8 to 3/8 in. across the bottom of the door. Ah 1-ft section of the soft seal that
had become detached was found in the outlet duct. |

Heat leakage through penetrations in the top and sides was no problem. The thermocouples
and power wiring showed no evidence of overheating, v
As a result of the blower failure in July 1966, some antimissile proteétion was provided

‘steel aircraft cables, shock

between the blowers and the radiator. A grid of 1/4-in. stainless
mounted, on 2-in. centers was installed just downstream of the blowers. Heavy wire screens with

0.4-in. mesh were installed just ahead of the radiator,

121bid., pp. 67—70.
27

The performance of the magnetic brakes on the door-lifting mechanism became marginal toward
the end of run 10, and some slipping occurred when door positions were adjusted, Preparations
were made for replacement of the worn and broken brake shoes at the next shutdown or if the brakes

become inoperable.

Of#f-Gas Systems

P. N. Haubenreich

The fuel off-gas system continued to présent some problems throughout the period. One prob-
lem area was associated with the overfill of the fuel pump that got flush salt in some gas lines
near the pump. The other was a continuation of the difficulty that appeared when the reactor first
operated ‘at power: the accumulation of polymerized cfil residues in small passages. The former
interfered with operations and required considerable work in the réaétor cell to remedy. The latter
was a nuisance through run 10, but after the installation of a new particle trap caused no more‘ |
trouble. |

Plugging near Fuel Pump. _ After the accidental overfill of the fuel pump, ! 3 the bubbler
reference line was cleared by remotely applied, external heaters. Salt in the sampler line was
melted the same way, but ran down and froze at the junction with the pump bowl. The obstruction
that formed here cleared out when the pump was heated up with salt in the bowl. No further trouble

was encountered with these lines. The off-gas line, shown in Fig. 1.8, was a different matter.

131bid., pp. 24~25.

ORNL-DWG 67 — 4764

——{— . E— e m— }
41_5 JUMPER OFFGAS HOLDUP
P VOLUME , 4in. DIAM

;|

TO PARTICLE TRAP ‘ J
AND CHARCOAL BEDS —~—}§ j

PUMP BOWL—»—| |-

OVERFLOW PIPE —=]

OVERFLOW TANK

Fig. 1.8. Off-Gas Piping Near Fuel Pump and Overflow Tank.
28

Although some salt entered the off-gas line, as indicated by a I‘temporary rise in temperature at
TE-522-2, pressure drop measurements showed no significant difference from the clean condition.
This was attributed to the blast of compressed helium from the drain tank that was released back-
ward through the off-gas line, before the salt had time to freez? completely, when the overfill
triggered an automatic drain. Therefore the only action taken before the startup in September was

to replace the short, flexible ¢

‘jumper’’ section of the line, where the convolutions would certainly
hold somé salt.

The off-gas flowed freely, with no unusual pre‘s.sure drop through 26 days of circulating helium,
flush salt, and fuel salt at low power. Then, two days after poflwer operation was resumed at 5.8
Mw, a plug developed in line 522 somewhere between the pump bowl and the junction of the over-
flow tank vefit with the 4-in. holdup line. The first indication was a decrease over a few hours
from 225°F to 160°F at TE-522-2, as the plug caused the off-ga:ls to bypass through the holdup
tank. The presence of the plug was confirmed when HCV-523 was closed to build up pressure
to return salt from the overflow tank to the pump bowl: pressure in the pump bowl also built up.
Efforts to remove the restriction by applying a 10-psi differential either forward or backward were
unsuccessful. 4
| The bypassing of the off-gas through the overflow tank did not hi_nder operation, except for one
specific job: recovery of salt from the overflow tank. Salt slowly but continuously accumulates
in the tank during operation, and it is therefore essential to return salt to the pump bowl two or
three times a week in order to maintain proper levels. With the plug in the line from the pump
bowl, it was necessary to greatly reduce helium flows into the pump so that the overflow tank
pressure could be increased faster than that in the pump bowl to make the salt transfer. Through
the remainder of run 8, salt was returned from the overflow tank six times, and on at least four of
these occasions some fission product activity was blown or diffused up the pump shaft annulus
into the oil collection space. This was a consequence of the reduced helium purge down the shaft
annulus and the unavoidable, sudden pressfirization of the pum;;- bowl that occurred at times in the
procedure. The charcoal trap in the vent from the oil catch tank prevented any serious release of
activity to the stack. The activity level in the oil tank and the‘; line increased, however, and the
last two releases from the pump bowl caused the flow element in the vent to plug partially and
then completely. |

After run 8 was terminated, stéps were taken to clear the plug from the off-gas line so that the
normal salt recovery procedure could be used. Frozen salt was suspected of causing the plug, so
the fuel loop was flushed to reduce radiation levels, the reactor cell was opened, and specially
built electrical heaters were applied to the line between the purwnp' bowl and the first flange.
Heating alone did not clear the plug,‘ but when, with the line hot, 10 psi was applied backward _
across the plug, it blew through. The pressure drop came downjas more helium was blown through
until it became indistinguishable from the normal drop in a clea;l pipe. The temporary heating

apparatus was then removed. ,
29

-

In run 9 the power was raised only 8 hr after fuel circulation had commenced, but TE-522-2
came up to only 150°F, indicating that the line was already plugged again. Plans were immedi-
ately set in motion to do a more thorough job of clearing. While tools and procedures were being
devised, the reactor was kept in operation, but great care was taken to avoid getting fission
products or salt spray up the pump shaft annulus again. This entailed lowering the power to 10
kw, 24 hr before the overflow tank was to be emptied, then stopping the pump 4 hr beforehand to
let the salt mist settle. Then the fuel p:lmp was vented through the sampler and the auxiliary
charcoal bed during the transfer. After three cycles of this, the reactor was drained and flushed
again in preparation for working on the off-gas line.

This time heat was applied to the short section of line between the second flange and the top
of the 4-in. decay pipe. When heating to about 1100°F did not open the line, the flexible jumper
was disconnected to permit clearing the obstruction mechanically. In the flange above the 4-in.
line, the 1/z-irl. bore was completely blocked, but the weight of a chisel tool broke through what
appeared to be only a thin crust of salt. Borescope inspection showed that the rest of the vertical
line was practically clean, and there was only a thin layer of salt in the bottom of the horizontal
run of the 4-in. pipe. Helium was blown through the line at five times the normal flow, and the
pressure drop indicated no restriction, Turming then toward the pump bowl, we saw a similar plug
in that flange. This too was thin and wlas easily broken out. A 1/4-ir1. flexible tool was then
inserted all the way into the pump bowl to prove that a good-sized passage existed. A new jumper
line was installed and operation was resumed. No further difficulty was encountered with this
section of the off-gas system. ’_ ‘

Because obtaining samples remotely without spreading contamination would have been most
difficult, no analyses were made of the material in the flanges. But it appeared that salt had
frozen in the line, almost completely blocking it, during the overfill. Material in the off-gas
stream during operation then plugged the small passages. The heaters melted the salt out of
the pipe, but the flangés were not as hot, and a thin bridge remained.

Particle Trap. — The first particle trap was designed on the basis of a rather brief period of
investigation and development after plugging in the off-gas system halted the planned approach to
full power. It served its purpose in that it protected the pressure control valvg and to some extent
"the main charcoal beds, permitting the experimental program to proceed. When the pressure drop
across the trap began to build up, an identical replacement was prepared, so that the first could be
removed and examined as an aid in designing a better trap while operation continued. This re-
placement was deferred until after rufi 7, because the pressure drop across the trap never became
prohibitive, | | .

The examination of the first particle trap and the design of the new model are described on p.
47. The second unit served through runs 8, 9, and 10. This unit behaved in runs 8 and 9 much as
had the first particle trap. The pressure drop occasionally built up to 5 to 10 psi, beginning two
days after power operation startéd in run 8, but backblowing with helium was effective in reducing

the pressure to 2 to 4 psi. ‘In run 10, however, after the first week of power operation, backblowing

~
30

effected only temporary relief at best. Various tactics were used to get gas to the particle trap
with as little delay as possible, to see if increasing the fissi?n product heating would drive off
organic material from the place where it was causing plugging% After these efforts proved inef-
fectual, the opposite approach was used: the off-gas was delayed as long as possible. This was
done in recognition that heating caused the central inlet tube to expand farther into the Yorkmesh
filtering material. The delay was obtained by routing the gas through an equalizing line to the
empty drain tank, through the tank and the salt fill lines to the other tank where it bubbled up
through several inches of salt heel, then out through the drain-tank vent line to the particle trap.
‘This gave the gas about 8 hr of delay and also bubbled it throtilgh salt in the freeze valves and
drain tank before it reached the particle trap. The pressure dr-fop across the trap was 16 psi when
the gas was rerouted, but within a few hours it was below 2 ps:i. For the last two weeks of run
10, the pressure drop across the particle trap remained below 2 psi when the delayed route was
followed, but began to build up almost immediately when the original route was used..

The charcoal filter, just after the particle trap, held up fission products, as shown by the
heating during power operation, but the temperature profile showed no significant change to indi-
cate poisoning by organics. The pressure control valve was operated full open, with the pump-
bowl pressure determined by the drop in other parts of the sys{jern. Therefore it appeared that
nothing would be lost by removing these items to make room fc;r new, dual-particle traps.

The new (third) particle trap had been in service for 5100 Mwhr of reactor operation by the end
of the report period. The measured pressure drop showed no increase at all, remaining below 0.1
psi. At 7.4 Mw the temperatufe measured near the Yorkmesh was around 275°F, the thermocouple
nearest the Feltmetal (but shielded by a pipe and a bellows) average 108°F, and the lowest couple
indicated 61°F, only 10°F above the temperature of the water §urrounding the trap (see p: 42 for
a description of the new traps). |

. Main Charcoal Beds. — The main charcoal beds continued to perform their function of holding
up the fission products: there was no breakthrough of activity other than the normal 10-year 85Kr.
Some difficulty was encountered at times, however, when the pressure drop at the inlet built up to . .
an inconvenient level. . '

Bed sections 1A and 1B had been used almost exclusively during earlier runs, and they were
on line when run 8 started. As had happened before, the pressfure drop at the inlet of these sec-
tions began to increase a few hours after the power was raised!. When the pressure drop reached
7 psi, sections 2A and 2B were also brought on line. The preésure drop through the four sections
in parallel remained below 1 psi through the end of the l;un, but the pressure drop through 1A and
1B remained high, and backblowing did not bring their pressure drop below 4 psi.

A situation similar to that in 1A and 1B also existed in the auxiliary charcoal bed, where the
pressure drop was abnormally high and did not respond to backblowing. Tests showed that the
restriction was near the inlet end, and it was strongly susp‘ect(;;ed that it was due to organic mate-
rial clogging the steel wool at the opening of the inlet pipe into the bed. Therefore, a remotely

placeable assembly of electric heaters was designed for this section, and in the brief shutdown
31

betweenA‘runs 8 and 9 it was tried. Some improvement was observed when the heater temperatures
reached 720°F, and after these temperatures were raised to 1235°F and then cooled, the -pressure
drop was down by a factor of 5, to a satisfactory level.

In light of this success an attempt was made to clear up the restriction in section 1B by heat-
ing the inlet with a torch. (Electric heaters to fit this bed were not then available.) Although the
temperature in the bed reached 870°F, there was ho improvement in pressure drop. In run 9 all
four sections of the main bed were operated in parallel, with no further effort to clear 1A and 1B
and no indication of any change in pressure drop. After this run, electric heaters were .installed
on the inlets of 1A>ar1d 1B. Heating to 750°F for 8 hr brought the pressure drop back to the normal
range for clean beds. Although the heaters could not be used in normal operation because the
cooling-water level is above the inlet sections, they were left in place with leads brought out
through the shield. |

Run 10 began with 1A ;and 1B on line, but after one day of power operation the pressure drop
began to build up, and the flow was switched to 2A and 2B. Three days later, for the first time,
the pressure d‘rop across these beds began to increase, reaching 5 psi before they were backblown
to reduce the drop below 2 psig. The flow was returned to 1A and 1B, after their pressure drop
had been reduced below 3 psi by backblowing.i For the last three weeks of run 10, the pressure
drop did not again build up.

In run 11, through February, only 1A and 1B were used. Thel pressure drop remained about
2.5 psi for two weeks and then began slowly to increase. After four weeks of operation it had
reached 5 psi. |

4 and the re-

The early finding of organic material in the original charcoal-bed inlet valves!
sponse of the pressure drop to heating are gobd evidence that most of the trouble with pressure
drop was caused by organics. Why the restriction built up so slowly in run 11 is not known.

Designs were completed during run 11 for more positive remedies for the high pressure drop.
These include a particle trap just ahead of the bed inlet manifbld and piping to bypass the inlet

section of each bed where the plug is believed to be.

Cooling-Water Systems

R. B. Lindauer

Treated-Water Cooler. — Befc;re run 8, 17 leaking tubes in the treated-water cooler were
plugged. ! In runs 8 and 9 and through part of run 10 the leakage from the treated-water system
averaged 2 gal/day or less. Then early\in January the leakage increased to about 6 gal/day and
a few days before the end of run 10 began to increase again, reaching about 50 gal/day by the
shutdown. Leakage in the cooler to the tower;water system was confirmed by the appearance of

lithium in the tower water. Before run 11 the cooler was replaced by an available surplus heat

14M’SR Program Semiann. Progr. Rept. Feb. 28, 1966, ORNL~-3936, pp. 124-—-29.
- 15SR Program Semiann. Progr. Rept. Aug. 31, 1966, ORNL-4037, pp. 47—48.
32

exchanger. The new exchanger has only two-thirds the tube area of the old cooler, but the heat
transfer coefficient is higher, resulting in about the same performance. Through the end of
February there was no evidence of leakage in the new cooler.

Space Coolers. — Leaks had occurred in both reactor-cell space coolers at brazed joints on
the brass tubing headers, and repairs had proved to be difficult because repair of one joint tended
to open adjacent joints. Prior to run 8, both coolers (and all otfier in-cell components) were
proved to be leak-tight, but when operation was resumed, condeqsed moisture began to appear in
the system recirculating the cell atmosphere. Although leakage; from the space coolers was
suspected, they could not be isolated for leak testing because of the high temperatures in the cell.
In view of the discouraging experience with repairs, two replacément coolers were procured in
Which the headers and nipples wete made of copper and the troublesome joints were welded. (The
Heliarc welding of copper to copper was done at ORNL.) |

Radiolytic Gas. — As reported previously, '* when the reactor is at full power about 3 scfh of
radiolytic hydrogen and oxygen is evolved in the treated-water slystem. Steady-state accumulation
in the system was reduced from 8 ft3 to 6 ft3 by supplying the tl‘lermal shield slides with \;rater
" pumped from a small vented tank. Before run 8, a 350-gal degassing tank was installed in the
return line from the main thermal shield and slides. With this tank in the system, the gas accumu-
lation at full power was reduced to 3 ft3. The holdup was the same whether or not the slide supply
tank was used in conjunction with the main degassing tank. The radiolytic gas being stripped
from the water is diluted with air to below the explosive limit ifi the degassing tank and is then
vented outside the reactor building. , _

Chemical Treatment. — In the course of testing and repair of leaks in the treated-water cooler,
cooling-tower water leaked into the treated-water system. Before the reactor was started up in
September, the resultant sodium contamination was reduced by dilution with condensate, after
which the required corrosion inhibitor concentration was restored by adding more lithium nitrite.
Power 6peration was resumed with 1.6 ppm sodium in the treated water. Activation of the sodium
raised the radiation levels around the water system, but did not' interfere with operation. After
September, no more additions of corrosion inhibitor were requirgd until January, when some makeup
was needed because of leaks. | ‘ '

In the cooling-tower system, corrosion inhibition was by daiily additions of balls of Nalco 360
(sodium chromate and phosphate). To eliminate troublesome delposition of calcium phosphate on
flowmeter tubes, the inhibitor was changed to a mixture of potassium dichrofnate and zinc sulfate.
Continuous-feed addition equipment was also installed at the cgoling tower.

J
Component-Cooling System

P. H. Harley

During thié report period the component-cooling pumps, CCI‘D-I and CCP-2, operated 1490 and

2159 hr respectively. Although some difficulties were encountered with the system, the output was

adequate and there was no interference with reactor operation,
|
33

~ Previous operatién of the pumps had indicated insufficient capacity, and the speed of both
units was ihcreased during the shutdown after run 7, Operation was satisfactory in run 8, but in
run 9 it became evident that there was a malfunction either in the output-pressure transmitter or
in the pressure-control valve, PdCV-960. During the shutdown after run 9, inspection of PACV-960
showed that the valve stem was stuck with the valve partly open. Deterioration in the operation
of this valve probably also accounted for at least part of the apparent loss of capacity of the
pumps. The valve oberated satisfactorily after the valve was repacked and the stem polished and
lightly lubricated., A drain was installed at the discharge of the gas cooler which should reduce
moisture collection in the valve bonnet, and PACV-960 is cycled periodically to ensure free stem
operation. The pressure transmitter was also checked and found to be operating properly. How-
ever, the piping was rerouted to eliminate sections where condensate could collect.

During this same maintenance period, the butterfly check valve from CCP-1 was found to be
inoperative and was repaired. The failure occurred at a silicone rubber hinge which supports the
two wings of the check valve. This check valve would prevent CCP-2 from developing pressure,
because the gas could short-circuit back through CCP-1 to the blower suction. This was the
second reported failure of a check valve; it had been in service between 1548 and 3074 hr. A
similar check valve had failed on CCP-2 in January 1966, after 1640 hr of operation, and had been
replaced by one made at ORNL as there was no sparé on hand. The locally fabricated check valve
was replaced in September 1966, although it had not failed after ~.1800 hr.

Pump CCP-1 had to be stoppéd during run 10 because of low oil pressure, and CCP.-Z was used
for the remainder of the run. After run 10, a cracked copper fitting in the oil system was repaired,
~and 2 gal of oil was added to bring the oil level back to normal.

The drive belts on the component-cooling pumps caused intermittent difficulties through run 7.-
Prior to run 8 the motor mounts were strengthened to reduce flexing, and deflectors were installed
to try cooling of the belts with the incoming gas. Even with the deflectors, a thermocouple near
the belt of CCP-1 indicated an ambient temperature of 150 to 160°F. Although this is above the
130°F for which the belts are rated, inspection in December, after 1400 hr of operation, showed
thét the belts were in good condition and only moderate tightening was required. There has been
no evidence of belt slippage' since the motor mounts were reinforced.

The rupture disk installed to eliminate the pressur!e relief valve leakage would not withstand
the shock of starting the blowers and has been removed.

No inspection has been made on the strainer installed in the gas piping in August. However,
there has been no increase in system pressure drop to indicate collection of material in the
strainer.

Pump CCP-3, which cools out-of -cell freeze valves, operated the entire period without trouble.
34

- Salt-Pump Oil Systems

J. L. Crowley

The systems supplying oil for lubrication and cooling in the salt pumps operated.continuously
for the last 51/2 months of this report .period. Operation was practically free of trouble.

Oil leaking past the lower shaft seal in a salt pu.mp collects“in an external tank. Oil from the
fuel pump, which had accumulated in earlier months at rates up to 20 cc/day, !® accumulated in

October and November at about 5 cc/day. There was no accumuiation in December, but in January

| the oil began to collect again at about 5 cc/day. Seal leakage fi-om the coolant pump, which had
accumulated previously at aBout 2 cc/day, increased to about 17lcc/day. These rates are far
below the 1000 cc/day that was originally set as acceptable, I;

Leakage of oil into the salt pump bowls is, in principle, detectable by changes in oil in-
ventories. The inventories in the two systems both indicated a ;mall, unaccounted loss over
the 51/2 months of operation, as shown in Table 1.3, In four mon%hs of earlier operation,’® the -
calculated loss from the fuel system was —0.3 liter (an apparent gain), and from the coolant sys-
tem there was an apparent loss of 0.3 liter. The significance ofjthese changes is somewhat ques-
tionable because uncertainties in the inventories are relatively farge (around 1 liter or more). But

during the last months of operation, the decrease in reservoir levels was rather steady, and there

was no evidence of more oil being trapped in the motor housings.

Table 1.3. Oil Systems Inventory Changes

i
September 1966—February 1967

Changé (liters)

Item Fuel Coolant -
System System
Removed in samples 6.49 ” 5.64
“Accumulation in catch tank 0.57 “ 1.35 _ -
Total accounted for - 7.06 . 6.99
Decrease in reservoir 8.34 | 7.56
Apparent loss 1.28 0.57 . -

Electrical System

T. Mullinix T. L. Hudson
R. H. Guymon

During the shutdown prior to run 8, the entire electrical system at the reactor site was over- -

hauled and put in good operating condition. Most of the major electrical breakers were tested and

16/bid., pp. 50—51.
'35

calibrated to assure proper operation. One of the emergency-power diesel generators (DG-3), which
had a cracked engine block., was replaced with a unit of the same size obtained from another instal-
lation. In additioh, the 250-hp.motors for the main radiator blowers and the two 48-v dc generators
were inspected. and reconditioned, | ‘

In general, the electrical system performed satisfactorily during this period. The emergency-
power generators are operated under load routinely for test purposes, and they have always oper-
ated when required. On September 30 a major power failure in the X-10 area caused an automatic
transfer to the alternate feeder for the MSRE. During this outage, two of the MSRE diesel gen-
erators were operated in parallel with TVA to help supply the X-10 electrical load. This opera-
tion was entirely satisfac‘tory.' The only difficulty encoun'tered with the electrical system was a
fuse failure in the high-reliability ac power supply (static dc-ac inverter) which.resulted in an un-

scheduled control-rod scram.

Heaters

T. L. Hudson

During the shutdown (in September) a resistance check was made on all in-cell heaters at
junction boxes located outside the reactor and drain-tank cell. Other than heater HX-1, which had
partly failed before the shutdown, only one heater, H102-1S, was found that had failed. This was
one of three installed spare heaters on'the vertical section of pipe under the heat exchanger. No
repairs were required, since there is adequate capacity in the other heaters on this section of
pipe.

~ The resistances of the coolant-system heaters located outside of the reactor cell were also
checked. Three heaters were found that had failed. One was on the coolant system 5-in. piping,
one on the coolant-pump furnace, and the other one, a spare heater, was ofi the fill-line piping. "It
was necessary to replace only the heater on the 5-in. piping, since there was adequate heater
capacify at the other locations to preheat the coolant system.

Aside from heater-element failures, a remote disconnect for one of the fuel-pump heaters
(heater FP-1) was damaged during remote operations associated with the tha\;ving of the salt plug
in the fuel-pump off-gas line. The heater was plu'gged into a spare disconnect to restore it to )
service. _ ’

Some additional heater-element failuréé héve occurred since the operation of the reactor sys-
tem was resumed. As mentioned above; part of the elements in heater HX-1 had failed earlier,

On October 12, additional elements failed in this heater, and on October 28 the last of the ele-
ments failed, Using the other heaters, it has been possible to keep this section of the heat ex-
changer adequately heated. This was aided by the coolant system, sihce coolant-salt circulation
was maintained during the period. ‘When the fuel and coolant systems are drained, tests will be
made to determine whether this heater must be replaced. '

One of the elements on heater H-106-4 failed on January 19, 1967. However it has been pos-

sible to maintain adequate temperatures using the other elements and adjacent heaters.
36

Control Rods and Drives

R. H. Guymon M. Richardson

The performance of .the control rods and drives has been within the operating limits throughout
the operation. None of the rods has ever failed to scram on request, but some difficulty was en-
countered in withdrawing one of the rods.

After run 7 we found that the drop time for control rod 3 had increased slightly.!? While we
were investigating this problem during the shutdown before run 8, the drop time increased still
further and approached the limit of 1.3 sec. In addition, occasional hanging was observed when
the rod was raised 2 in. from its fully inserted position. We removed the rod-drive assembly and
found that the long drop times were caused by a bent cooling-air tube. The air tube presumably
was bent when the drive unit was last reinstalled (February 1966). The hanging on rod with-
drawal was attributed to interference between sharp edges at the bottom of the rod and a protru-
sion of ‘a warped guide in the thimble. We replaced both the rod drive and the control rod with
spare units after first rounding the sharp corners at the lower end of the spare rod. This relieved
the hanging problem and gave an initial average rod-drop time (for 35 drops) of 0.92 sec with a
maximum drop time of (.95 sec. Subsequent measurements of the drop for rod 3 gave 0.85 sec.
(The drop time normally decreases because of increasing flexibility of the rod with continued use.)

Shortly after the new assembly for control rod 3 was put in service, we observed a shift of
about 0.4 in. in the reference zero position as indicated by a special single-point indicator near
the bottom of the rod thimble. There has been no subsequent change in the reference position,
and the cause of the shift is not definitely known. However, when a similar shift occurred pre-
viously,!® it was attributed to slippage of a drive chain on a sprocket gear.

The lower-limit switch on rod 3 sometimes sticks after a scram, and the rod must be fully
withdrawn to dislodge it. This may be a recurrence of a similar, previously reported difficulty,!®
but it presents no difficulty in normal operation. |

On February 13, 1967, the fine synchro on rod 2 failed. The coarse synchro is f‘uvnctioning

properly and provides sufficient information for continued operation.

Samplers

R. H. Guymon R. B. Gallaher

The fuel and coolant salt samplers met the requirements of the experimental program without
delays or serious difficulties. As reported previously, flush salt that froze in the fuel sampler
line during an accidental overfill was melted with external heaters. However, some obstruction

femained at the top of the pump tank until the system was heated when there was flush salt in

171bid., p. 53. .
'8MSR Program Semiann. Progr. Rept. Feb, 28, 1966, ORNL-3936, p. 54.
37

the pump.. Thereafter there was no interference with sampling, nor any mechanical‘ trouble of any
consequence throughou:lt the report period. Contamination, although adequately controlled, did
become more of a problem because of the higher activity of the salt and the nature of some of the
special devices that were lowered into the fuel pump. |

During this report period 66 samples were removed from the fuel system and 8 were removed
from the coolant system. Of the fuel-system samples, 43 were routine 10-g salt samples, 6 were
large salt samples (50 g) for oxide analyses, and 17 were special-purpose samples related to

studies of the oxidation state of the uranium and fission product behavior.

Containment
P. H. Harley H. B. Piper

Reactor Cell Leakage. — The secondary containment, consisting of the reactor and drain-tank

cells and the closure devices in their penetrations, was shown to have acceptably low leakage at
“the beginning of the report period. It remained so throughout the period, but the routine meas-
urement of cell inleakage during operation gave erroneously high results in run 9 and was some-
what uncertain in run 10 because of leaking pneumatic lines, o

In a comprehensive test of penetration block valves and check valves before run 8, only 8 of
160 valves showed leakage above the specified. allowable rate. Metal filings and bits of Teflon
pipe tape were found in some of the valves, and cleaning and replacément of some soft seats cor-
rected all of the leaks, o

The reactor and drain-tank cells were tested at 10 psig three times during the period.' Before
run 8 the measured leak rate was 65 scfd; before run 10, 43 scfd. These rates are factors of 3.0
and 4.5 below the acceptable rate at 10 psig. The 10-psig test before run 11 was only long
enough to show that the leak rate was acceptably low; no exact figure was obtained.

For normal operation the cell is held at —2 psig, and the permissible inleakage has been set
conservatively at 85 scfd. (This was based on an extrapolation from accident conditions which
assumed orifice flow through all leaks.) In run 8 the measured rate was 65 scfd, within the
prescribed limit, but by a much smaller factor than was obtained in the 10-psig test just before the
run. In run'9 the calculated inleakage increased to more than 300 scfd, and one reason for termi-
nating the run was to investigate this problem. Several valve-operator pneumatic lines were found
to be leaking in the cell, but when the run 9 measurements were corrected for their contributions,
the net rate was only 62 scfd. In run 10 the pneumatic-line leaks were measured continuously and
deducted frorfi the total inleakage, giving a fiet inleakage rate around 50 scfd. In run 11, after the
air-line leaks were stopped, the measured inleakage. was only about 10 scfd. |

Leaks in the pneumatic lines do not represent possible routes for activity release in the event
of a spill in the cel‘l, because block valves just outside the cell automatically shut off the lines if
the cell goes above atmospheric pressure. Of course theseleaks must be taken into account in the
calculations of the inleakage through other routes, and, during run 10, flowmeters were used in ‘the

air lines to obtain the necesséry information. The leakage from eight lines became quite large,
38

however (up to 3500 scfd), causing considerable uncertainty in the cell inleakage measurement.
Nitrogen was substituted for air in the leaking lines to keep the oxygen concentration in the cell
below 4% (to rule out fire in the event of an oil leak).

After run 10 the pneumatic-line leaks were traced to quick-disconnects in which the neoprene
seals had become embrittled. There were 18 disconnects with elastomer seals; eight disconnects,
all near the center of the reactor cell, were leaking. Seventeen of these disconnects were replaced
with special adaptors sealed at one end by an aluminum gasket and at the other by a standard
metal-tubing compression fitting. (One disconnect was left because it was on a line that is always
at cell pressure.) After the replacements, all 17 lines were proved to be leak-tight.

Activity Releases. — Operation of the reactor was the direct cause of only two minor releases
of activity to the ventilation stack. These occurred while capsules containing gas from the pump
bowl were being removed through the sampler-enricher and involved 1 and 2 mc of iodine. There
was also minor contamination of the work area at the sampler-enricher, which was easily cleaned
up.

Continuous monitoring of the ventilation stack showed that a total of 206 mc of iodine and 9
mc of particulate activity were released during the six-month report period. Most of this (192 mc)
was released in December while the fuel off-gas line at the pump bowl was being probed and
inspected.

Replacement of the fuel off-gas particle traps in September and again in January reéulted in
considerable contamination of the interior of the vent house. The area was cleaned to below con-
tamination-zone levels before operation was resumed, but after the January incident, this cleaning
necessitated replacing the floor grating and the top layer of stacked concrete block under the
grating.

Stack Fans and Filters. — Stack flow was maintained continuously at a safe level throughout
the period. The east stack fan was improved by the installation of new bearings, sheaves, and
belts as had been done earlier on the west fan, Before fun 8, after other maintenance was

finished, the roughing filters were replaced to reduce pressure drop and increase stack flow,
2. Component Development

Dunlap Scott

The effort of the development group consisted in work on the fission product off-gas removal
system, on the sampler-enricher, and evaluation of the radioactive maintenance of the components,

Several problems with the evaluation of their probable causes and solutions are given below.

2.1 SAMPLER-ENRICHER

R. B. Gallaher

Forty-seven routine samples were taken during reactor runs 8, 9, 10, and part of 11. Of these,
11 were 50-g samples for oxide and U3 */U** analysis. In addition to routine sampling, the
sampler-enricher was used to perform a number of special sampling and addition operations as

follows:

1. three samples containing absorptive wire for investigation of cover gas constituents,
2. two CuO capsules for determination of hydrocatbons in the cover gas,
3. three gas bombs for general cover gas analysis,

4, five capsules for addition of beryllium to the fuel.

Refer to Chap. 7 of this report for a detailed discussion of the special samples. Cumulative
totals to date for the sampler-enricher are 189 samples (including special samples) and 87 enrich-
ments. Operation of the sampler was for the most part satisfactory. A brief summary of nonroutine

operating and maintenance experience is given below.

Contamination

There has been a gradual increase in the radiation level inside the samplef. This is atfributed
to small particles 6f salt which cling to the outside of the capsule and are dislodged during han-
dling in the 1C and 3A areas. Some difficulty has been encountered with the spread of contami- -
nation in the area a;ljacent_ to the sampler, caused apparently by the release of small quantities of
gas or solid material during transfer of the sample into the transport cask. Two steps were taken
to reduce this problem: (1) the interior areas of the sampler were decontaminated by wiping with
adhesive tape and damp sponges, and (2) a ventilation duct, connected to the main building venti- .

lator system, was installed near the tra.nsport cask position.

39
40

Capsule Support Wire

As an empty capsule assembly was being attached to the drive unit latch, the wire which con-
nects the capsule to the key pulled free from the key. Fortunately, the operator retained his grip
on the capsule so that it was not drop})ed into the sample withdrawal pipe. It was found that the
_knot on the end of the wire was too small and had pulled through the key. Since then all existing
capsule assemblies have been reinspected. Also, new capsules have been fabricated using nickel-
plated mild-steel tops in place of the original copper top-s, so that a magnet can be used for re-

trieval in the event a capsule is dropped into the withdrawal pipe.

Seal Leakage

There has been a gradual increase in leakage of buffer gas through the upper seat of the main-
tenance valve. This condition is appérently caused by an accumulation of salt particles on the
seating surface. While the leak rate is relatively small (approximately 25 cm?3/min), it is suffi-
cient to cause difficulty with proper opération of the interlock which indicates that the valve is
closed. Since cleanup of the seating surfaces would be difficult, and since replacement of the
valve is not warranted by the leak rate alone, a mechanical method of assuring that the valve is
closed is being substituted for the existing pneumatic system. The désign of this section of the
sampler-enricher does permit ready replacement of the maintenance valve should this become

necessary.

Maintenance

The sample withdrawal pipe (line 999) was plugged with flush salt on July 24, 1966 (see Chap.
1, this report). The plug, which was located about 2 ft above the pump bowl, was melted by means
of electrical heaters clamped to the outside of the pipe. The melted salt drained down the pipe
and formed a new plug immediately above the pump. This second.pl‘ug was finally melted out by a
combination of external heaters and heat from the pump bowl after the reactor system was brought
up to normal operating temperature.’ ‘ -

Manipulator Buffer System. — At some time during the maintenance period, a small leak de-
veloped in the manipulator buffer system, and both boots were replaced. A small ‘‘pinhole’’ was
found in the atmosphere—side boot. The cause of the hole was not determined. This set of boots
had been used for 35 sample cycles ~

Access Door Mechanism. — Durmg one sampling cycle, a cotter key was noticed on the floor
of area 3A. Close examination of the access port door mechanism showed that the key had come
out of the top of the lower hinge pin.. The hinge pin had then worked down out of the top of the

hinge. Using only the one hand manipulatof, the pin was repositioned properly in the hinge, a new

IR. B. Gallaher, Removal of Flush Salt from Line 999, MSR-66-33 {Oct, 31, 1966).
41

cotter key was inserted into the 0.100-in.-diam hole in the pin, and the key was then spread to
lock it in place. In doing this the lower key was knocked out and had to be replaced also. This
repair work did not interfere with reactor operation or the sampling schedule.

Yacuum P‘umps. — The radiation level in the vacuum-pump enclosure was checked after a two-
month cooling-off period (following shutdown on July 17, 1966). It was found that the radiation
level was less than 100 mr/hr, which is low enough to permit direct maintenance work in this area.
The oil level was checked on both pumps, and oil was added to one of thé pumps.

Spare Parts. — In an attempt to minimize the time required for haintenance, arrangements have
been made to stock certain critical subassemblies. These include a manipulator arm, an area 1C

subassembly, and an operational valve assembly.

2.2 COOLANT SAMPLER
R. B. Gallaher

Eight 10-g samples were isolated by the coolant sampler. A total of 53 samples, including two
50-g samples, have been taken with this equipment. ‘

The leak rate of the removal valve-seat buffer increased after one sample. This was corrected
by replacing the valve seats. Also, an extra washer was added to the valve to permit more pres-

sure to be applied to the seals.

2.3 FUEL PROCESSING SAMPLER

R. B. Gallaher

Installation of the fuel processing sampler is complete except for the shielding. Leak testing
and operational checks have been completed. The system was found to be satisfactory except for
a defective pai't in the containment buffer system.

The éhielding has been designed and fabricated and is ready for installation.

2.4 OFF-GAS SAMPLER
A. N. Smith R. B. Gallahe_r

A system is being fabricated for installation in the reactor off-gas stream to pemit analysis
for fission produét and other gases. Originally? it was planned that the system would include a
thermal conductivity cell, a chromatographic cell, and a refrigerated molecular sieve trap. Three
changes have been made to the sampler system since it was last reported:

1. Several developfiént problems associated with the chromatographic cell could not be re-
solved by the scheduled startup date for the sampler. The chromatograph has therefore been

omitted from the system.

2MSR Program Semiann. Progr. Rept. Feb, 28, 1966, ORNL-3936, p. 67.
42

2. A hydrocarbon oxidizer, a C02-H20 absorber, and a second thermal conductivity cell have
been added to the system. These components, acting in concert, will provide a measure of
the level of hydrocarbon contamination in the off-gas stream. In addition, the oxidizer unit, by
virtue of its position upstream of the molecular sieve trap, will serve to protect the latter from
fouling by organic material.

3. The original design provided for sampling the off-gas stream at a point immediately down-
stream of the particle filter in the reactor off-gas line, specifically line 522. The piping has now been
rearranged to provide an additional sample point immediately upstream of the 522 filter. Except
for the approximate 40-min delay time, samples from the upstream point should be identical with
pump bowl exit gas. Also, by using the hydrocarbon detector (see item 2 above), results from the
two sample points may be used to judge the relative effectiveness of the off-gas filter. Gas will
be routed to the sampler by w;'zly of existing lines (533, 561, and 538), so that the changes asso-
ciated with the new sample point will be restricted to piping internal to the sampler.

A diagram of the revised sampler system is shown in Fig. 2.1. Structural and piping work on

the sampler proper is complete. Instrument and electrical work is approximately 90% complete.

2.5 OFF-GAS FILTER - MK I
A. N. Smith

The off-gas filter installed in line 522 in March 1966 consisted of a particle trap, or prefilter,
and a charcoal filter (see ORNL-4037). Experience during operations plus hot-cell examination of
the particle trap after removal from the system (see elsewhere, this section) indicated that the
particle trap was too small and that the charcoal filter was unnecessary. Accordingly, the par-
ticle trap and the charcoal filter assembly were replaced with two particle traps in parallel. The

new particle traps, shown in Fig, 2.2, are similar to the old unit except for the following revisions:

1. The trap housing was increased from 4 to 6 in. ID, resulting in an increase in cross-sectional

area of 225% in both the Yorkmesh and Fiberfrax sections.

2. The unit was in effect turned upside down so that the Yorkmesh section is at the top of the ,
unit and the Fiberfrax section is at the bottom. This change was made to permit heating of the
Yorkmesh section (using beta decay heat and lowering the water level) while still maintaining
cooling on the other two sections, Pipe connections were adjusted to compensate for this change,

so that the flow path is now downward through the filter.

3. The disposition of the Yorkmesh was modified to provide increased frontal area in the

direction normal to the flow.

4, Only the coarse Feltmetal was used. The original particle trap used a coarse section
(removes 98% of particles <1.4 ) and a fine section (98% <0.1 f) in series. Since the first trap
had shown very little loading in the final section, it was not believed that the fine Feltmetal was

)

needed.
43

SAMPLER CONTAINMENT BOX
/

REFERENCE GAS
(HELIUM)\ VENT
« f————1 THERMAL- 4—’
@ CONDUCTIVITY
S Dt CELL o
&
<1
THERMAL -
CONDUCTIVITY <t —{<}
CELL
| MOLECUL AR '
SIEVE
w '
[m]
=
[}
x SAMPLE
& TRANSFER
] BOTTLE
Q

B RADIATION SAFETY |

\ RECIRCULATING

AR SYSTEM
FAN
VOLUME OFF-GAS @ VOLUME
HOLDUP FILTER > HOLDUP '—E:::
FUEL prd
PUMP MAIN CHARCOAL BED
HCV533

VALVES |
@_T:

v

ORNL-DWG 67-4765

100 cc/min

3 psig

\ DRAIN TANK VENT HEADER

Fig. 2.1, Schematic Diagram — MSRE Off-Gas Sampler.

CHARCOAL BED

AUXILIARY

£

5. The total area of the filter elements was increased to 288 in.2. The two new traps, how-

ever, differ in that the No. 1 unit has the support screen on the upstream side of the Feltmetal,

resulting in an effective filter area of 30% of the total area. The original particle trap had an ef-

fective filter area of 22 in.2, so that the relative increase is 4% for unit No. 1 and 13 x for unit

No. 2.
ORNL-DWG 67-4766

YORKMESH
T.E. FELTMETAL [T‘ E . FIBERFRAX TE.
. N NWIF 7 7 F T 7 7 77 7 7 F 77 7 7 7 F 7 fF  f 7
SN ) f I, = =
Y=t & [ v v/
- - —— — ' - — e e NN N
: \ \ | y /4 1IN
— ; N
H V1
—_— ‘:'l —— .t\\\}\?-jr—'___ = ‘{7 \\'5\\1“ — \_‘\\\u\.l\\\\\\\\‘|
I = e ——————— =
X M ; 8 7.
7 /’ ™~ ™~ ™~ A
\ i ¥ ] V i N
) | | | I '
= —
S\ VA A AW 4 £ L S S L 7 yd
NICKEL BAFFLES
{ o 1t 2 3

INGHES

Fig. 2.2. Line 522 Particle Trap Mark Il, Particle Trap Subassembly.

144
oy

45

6. The depth of the Fiberfrax section was reduced by about 50% (from 6%, to 3% in.), because
examination of the first trap did not show much oil, even in the first Fiberfrax section.

Specific design data are as follows:

Yorkmesh Section

Material . ) ,304 S5

Overall size 6 in. in diam X 4 in. deep
Total weight 0.62 1b

Bulk density 0.16 g/cm3

Wire diam - 0.011 in,

Calculated surface area _ L 894 in.?

Feltmetal Section

Material ' Huyck No, FM 225

Removal ratihg — gas service 98% <1.4

Size of inner element 3 in. in diam x 101/2 in, long
Size of ocuter element 51/2 in. in diam x 101/2 in. long

Effective surface area
Unit No. 1 (effective area is reduced 85 in. 2
by perforated support plate) ‘
Unit No. 2 281 in.?

Fiberfrax Section

Material ‘ Carborundum Co. Long Staple Fiberfrax
Mean fiber diam Spu

Packing density _ . . 81/2 to 9 1b/ft>

Total weight ' 213.5 ¢

Geometry ‘ Six annular compartments; two each at

.1/4 in., 1/2 in., and 1 in. deep

Efficiency and pressure drop tests were made on the new particle traps prior to installation.

The results were as follows: .

Unit No. DOP* Efficiency Pressure Drop at 15 liters/min
1 _ 99,588 <1 in. H,O
S 2 . 99,960 <1 in. H20

*Dioctylphthalate-polydisperse mean particle size, 0.8 [Lt.

2.6 FELTMETAL CAPACITY TEST
A. N. Smith

The particle trap used in the MSRE off-gas system® (line 522) contains three filter elements

in series: a loose-knit bundle of stainless wire (Yorkmesh), a porous stainless steel cylinder

IMSR Program Semiann. Progr. Rept. A‘ug. 31, 1966, ORNL-4037, pp. 7477,
46

i

(Huyck Feltmetal No. FM 225), and a bed of inorganic fibers (Fiberfrax long staple). Sizing cal-
culations for the unit have been hampered in part by a lack of information as to the loading capacity
of the Feltmetal. Accordingly, a bench test was performed to develop data which might be used to

estimate the life expectancy of the Feltmetal under reactor conditions.
rd

Assumptions _.

In designing the experiment the following assumptions were made regardirig the reactor off-gas
stream: _

1. The character of the suspended particles is variable, ranging from crystalline tnoble—g:sls
daughter product) to tarlike (oil decomposition products), with various intermediate mixtures.

2. The particle size is on the order of 1 p. .

3. The concentration of particleé in the off-gas stream is such that the rate of accumulation
on the filter is on the order of 1 g/day. This estimate is jbased on the assumption that salt en-
trainment is negligible, the daughter product deposition rate is ~0.1 g/day (approximately 1% of
7.5 g of 235U per day), and the pump oil leakage averages 4 g/day, of which 25% is converted to

solids which are deposited on the particle trap.

Experimental Procedure

Particles suspended in a helium carrier stream were pas.sed through samples of the Feltmetal
(see Fig. 2.3). Ducting upstream and downstream of the filter was glass, permitting direct obser-
vation of the flowing stream. The geometry was such that particles on the order of 5 y and larger
could be expected to settle out before reaching the filter element.

The particles were generated by evaporation of hexane from very fine droplets of hexane solu-
tion containing the solid under study, which was aspirated into the flowing gas stream. The con-
centration of these solutions and the flow conditions at the aspirator were held constant, such that
the particle size could be expected to be the same for the three solids studied. Dave Moulton of
Reactor Chemistry was responsible for conception and development of the particle generator. The
test was divided into two phases: '

1. The change in load capacity as a function of particle character was checked by using three
different solids: naphthalene, stearic acid, and paraffin.

2. Using stearic acid, the effect of support screen position on load capacity was measured.
The support screen, a thin perforated metal sheet, is normally placed on the downstream side of
the Feltmetal to provide mechanical support in the event of excessive pressure drop across the
filter. In the MSRE, however, routine backblowing operations had made excessive pressure drop
in the reverse-flow direction more likely. It was thus desirable to know the effect of providing
mechanical support on the upstream side of the filter. ‘

Helium flow during all tests was maintained at 5.3 liters/min. - Collection of solids was con-,
tinfied until the pressure drop across the Feltmetal rose to abofit 3.5 psi. The quantity of solids

collected was determined by weighing. t
47
ORNL—DWG 67 — 4767

: VENT
HELIUM HEADER J

20in.
FILTER
. SAMPLE
|1 Il II
]

SECONDARY
FILTER

|
| il
. b T_ 1
] * . 2-in. DIAM GLASS PIPE
FEED . \
SOLUTION —=| |~ e
=+ PARTICLE N
GENERATOR —
. — |=— MERCURY MANOMETER
7 2 =
N
N
GAS FEED /0.0|3in,
_\ HOLE DETAIL OF PARTICLE
4 § 2in " GENERATOR p
2 RN
Z N
N
AT TSI

—u—— SOLUTION FEED LINE

N
N\
N

) N
N

Fig. 2.3. Capacity Test — Feltmetal (Huyck No. FM 225).

. Results

The test results are summarized in Table 2.1, The relative capacity of the Feltmetal for
naphthalene/stearic acid/paraffin was found to be about 1000/5/1. If these figures are applied
to the reactor system at a collection rate of 1 ‘g/day,‘the life expectancy of the new MSRE particle
trap ( see Sect. 2.5) might be as little as 3 weeks or as much as 70 years, depending on the char-
acter of the material collected. This estimate is based on a Feltmetal area of 280 in.2 and a
permissible pressure drop of 5 psi.

The capacity of the Feltmetal with the support plate on the upstream side was found to be
about 25% of the capacity obtained with the support plate on the downstream side. This result is

in reasonable agreement with the calculated value of 30% free area for the support screen.

2.7 EXAMINATION OF THE MK | OFF-GAS FILTER

;

Dunlap Scott - A. N. Smith

The Mark I lparticle trap,* which had been installed in the fuel system off-gas line in April
1966, was replaced in September 1966 with a new trap of identical design. In October 1966 the old.

‘MSR Program Semiann. Progr. Rept. Aug. 31, 1966, ORNL-4037, p. 75,
48

Table 2.1. Loading Capacity of Huyck Feltmetal No. FM 2257

Weight Collected Final AP Unit AP

2

i . : e
Type of Solid (g/m.z) (psi) : (psi g  in.”)

Capacity for Different Types of Solids

Naphthalene 0.96 0.058 0.06
Stearic acid 0.31 3.6 . ‘ 12
Paraffin - 0.06 3.9 65
/
Support Plate Final AP (psi) Weight Collected (g)

Capacity vs Support-Plate Position

Upstream 3.6 0.186

Downstream 3.5 0.73

dcarrier gas was helium at a flow of 5.3 liters/min,

trap was taken to the hot cells for testing and examination, Pressure drop tests were performed,
after which disassembly, examination, and sampling were done; the samples were analyzed by the

Analytical Chemistry Division. A description of this work is given below.

Pressure Drop Tests

The setup for the pressure drop tests is shown in Fig. 2.4. Helium was metered to the particle
trap through a calibrated flowmeter at a metering pressure of 10 psig. Flow was controlled at V-1,
and the pressur;a drop was .assumed to be the PI readin‘g, since the gas was vented to the hot cell
(cell pressure controlled at — 1.5 in. HZO) . After the total drop through the unit was determined at
various flows, three 14-'in.-diarn holes were drilled through\the wall of the trap to permit the flow tq
short-circuit to the cell without passing through the Fiberfrax section, and the pressure drop meas-
urements were repeated. The test results are plotted in Fig. 2.5. Also shown in Fig. 2.5 are the
results of the pressure drop measurements made on the clean filter before installation in the MSRE
off-gas system, |

The pressure drop data before and after reactor service indicate the following:

1. The pressure drop after service was about 20 times the ‘‘clean’’ pressure drop. Assuming
orifice-type flow, this would represent a reduction in flow area of about 80%.

2. The bulk of the pressure drop occurred in the Yorkmesh and Feltmetal sections of the
filter. Thus the Fiberfrax section was essentially clean.

The particle trap was in service in the reactor off-gas line dufing the operating period from
April 1966 through July 1966. During this time the pressure drop across the particle trap exhibited

wide fluctuations., Increases were usually, but not necessarily, associated with periods of power

\
49

ORNL-DWG 67-4768
(@ YORKMESH 304 STAINLESS STEEL |
, ® COARSE FELTMETAL
~ ® FINE FELTMETAL

. F
@ FIBERFRAX
2

QIPI \' E7
p_9

¢ ®
% ’ 'kilvfél——h %_%}l ) 1-—HOT CELL
km~d_77 - = 7] BLDG 3026

T

————

HELIUM
CYLINDER

T

HOLES DRILLED THROUGH
(/| PIPE WALL AT THIS POINT

TO OBTAIN PRESSURE DROP = (&L
EXCLUSIVE OF FIBERFRAX

Fig. 2.4. Pressure Drop Test — MSRE Particle Trap No. 1.

CRNL-DWG 67-4769

10
77
1/
5 - (® OATA TAKEN 10-24-66 74
L SZ
‘ ' 4
ol - ,// >(3) DATA TAKEN
/' 10-25-66
/
1 S
—— (1) CLEAN FILTER BEFCRE INSTALLING
B —— (@) TOTAL DROP ACROSS FILTER AFTER REMOVAL
= — FROM MSRE OFF-GAS SYSTEM
a 05— /
Q (® SAME As CURVE (2) EXCEPT PRESSURE )4
o - DROP THROUGH FIBERFRAX IS EXCLUDED /
o _
i 4 '
=2
7 /
o o2 :
1
a . /
NORMAL MSRE e
ot OFF-GAS FLOW——
) /!
4
V4
0.05 VAN
)4 (1) DATA TAKEN 3-24-66
7 .
0.02 METER NO. A-2784 USED FOR ALL TESTS METERING
: PRESSURE CURVE NO.{ AMBIENT; METERING PRESSURE
CURVE NO. 2 AND 3 10 psig | R
0.01 IR o L

0.2 05 A 2 5 10 20 50 100
HELIUM FLOW (liters [STP) /min)

Fig. 2.5. Pressure Drop Results — MSRE Trap No. 1.
50

operation. Decreases were sometimes unexplained and at other times were associated with delib- -
erate attempts to clear the trap, such as reverse-blowing with helium. Table 2.2 shows steady-
state values for the particle-trap pressure drop at ;larious times during the April to July 1966
period. The fluctuations in pressure drop seem to indicate that material was alternately deposited
and then removed by increased pressure drop or breakdown (such as by thermal or radiation ef-

fects). The maximum pressure drop of 9.7 psi would indicate a flow area reduction of about 95%.

Disassembly

The particle trap was disassembled to permit examination of the three main sections of the
trap. The cuts were made with a band saw which is part of the hot-cell equipment in the HRLEL.
All operations o‘fhandling, cutting, shifting, disassembling, sampling, weighing, photographing,
and preparing samples for shipment were done by the experienced operators of the HRLEL using
manipulators with direct and periscopic viewing, and the work suffered very little from the remote
handling requirement, .

The first cut was made at section A-A, 5 in. above the lower weld (Fig. 2.6). This point was
chosen to pemmit removal of the bellows section of the inlet pipe as well as the upper section of
the coarse and fine filter sections. '

A second saw cut was made through the upper filter at section B-B to get a closer look at a

deposit on the lower surface of the upper end flange.

Observations

Yorkmesh Section. — The area of the Yorkmesh which had been immediately below the inlet
pipe was covered with a blue-gray to black mat which had completely filled the space between

the wires of the mesh (Fig. 2.7). The shape of the mat corresponded to the bottom of the inlet

ORNL-DWG 66-11444R

FINE METALLIC FILTER

FIBERFRAX
COARSE METALLIC FILTER . (LONG FIBER) THERMOCOQUPLE (TYP)

LOWER WELD SAMPLE 97 "' w——

()
I!”W///////////' ///[/f/////' 7 /////////////////////////////,/,l
Er7: 2

II I "z?’(,
i ' 7 fi S | R | BT dfz}.yfi#

B } ;
= (1
il — [TIETS ATIAAVTETRRNIVIARL N 4 2z Iz rrr

== SU———— CYTTTTTTTrTOTYY T m fwcfil M[fl 7. }/

= - 7 _ _ L T %’
| 4 | %/,g ’ ;‘(
it ' ) ]
G l ,m,.mmm”/////// ///
SAMPLES ‘A LsampLe 3A L —-"g"
58 AND 6
STAINLESS STEEL MESH NICKEL BAFFLE (8)-

Fig. 2.6. Location of Sofiaple Points in Particle Trap.
51

R32824

Fig. 2.7. Inlet to Yorkmesh Section MSRE Particle Trap Mark I.
52

Table 2.2. Pressure Drop Ex'perience — Particle Trap Mark |

Particle Trap Pressure Drlop : Reactor Power at
Date and Time at 4.2 liters of Helium per min Time of Reading
(psi) . ' (Mw)
Before installation ‘ 0.06
4-4-66 - k <0.1 0
4-8-66 | 1.1 0
4-9-66 ) | 0.9 0
4-10-66 . 2.8 0
5-20-66 9.7 6
5-31-66 2.2 0
6-8-66 0.8 0
7-16-66 ‘ 8.7 _ 7.1
7-+16-66 2.8 ‘ 0
9-16-66 1.3
(2 months after. shutdown) !
Hot-cell test : 1.2

pipe, and it is likely that this material was the major restriction to gas flow during the operation
in the reactor.

Since the inlet pipe temperature probably increased several hundred degrees during power
operation of the reactor, it is believed that this restriction behaved much like a thermal valve.
This could account for the unexpected increase in pressure drop while at power and the decrease -
in pressure drop when the power was reduced. In laboratory tests® using a sample of the porc;us
metal to filter an oil mist, a rise in temperature caused a decrease in the pressure drop of a
loaded filter. This predictable-result was attributed to boil-off of the oil. .

A radiation survey made around the outside surface of this lower section of the trap gave
readings ranging from 6900 to 8000 t/hr. The probe read 11,400 r/hr when inserted into the
position formerly occupied by the inlet pipe. The radiation level dropped off to 4100 r/hr at the
bottom of the trap. The bottom of the inlet pipe had a deposit of blackish material which corre-
sponded to that in the Yorkmesh. The inside of the inlet pipe at the upper end of the bellows ap-
peared clean and free of deposit. The exterior of the bellows had some of the light-yellow powdery

material on a background of dark brown.

SB. F. Hitch et al., Tests of Various Particle Filters for Removal of Oil Mists and Hydrocarbon Vapor,
ORNL-TM-1623 ( Sept. 7, 1966). ' '

’

-
53

The outer shell of the lower section was slit at the welti, and the Yorkmesh was removed. It
was found that the surface of the outer wires of the mesh bundle was covered with a thin layer of
amber-colored organic material.. Much of this material evaporated from the heat of the floodlamp
used to make the photographs.

As the bundle was unrolled, the color of the film on the wire changed from amber to brown to
black near the center. The black material was thicker than the wire by a factor of 2 or 3. This
material was brittle, as was the wire, and much of it came loose as the wire was flexed. Samples
of this material, designated Nos. 5B and 6, were taken for examination and chemical analysis.

A piece of the wire covered with the black material was removed and mounted for metal-
lografihic examination, and it was found to be heavily carburized with a continuous network of
carbide in the grain boundaries. There was no evidence of melting of the wire; however, the grain
growth and other changes indicated operating temperatures of at least 1200°F. The nonmetallic
deposit observed on the wire mesh was apparently of a carbonaceous nature and appeared to have
been deposited in layers. These ‘‘growth rings’’ were probably the result of off-gas température
and flowrate changes.

Porous Metal Section. — The perforatedrplate of the coarse filter section and the lower flange
of the filter assembly were covered with a stratified scale (view A-A, Fig. 2.6). The color varied
from a very light yellow to orange. One stratum in the lower flange area appeared gray, almost
black. A sample (No. 9) was taken of the light-colored material, including some of the black.

The perforated plate of the fine filter section was covered with a thin, dark-brown coating, which
seemed to be evenly distributed over the surface of the plate. The inner surface of the outer wall
of the trap was covered with a light-amber coétin'g, which was also evenly distributed. It is be-
lieved that these coatings were deposited by condefisation and sputtering of the oil from the adja-
cent filter and that the dark-brown color indicates that the porous metal screen had operated at a
much higher temperature than had the outer wall. The lower surface of the upper flange (view B-B,
Fig. 2.6) contained a deposit which had the appearance of organic residue. The deposit was amber
colored, and the fractured edge (Fig. 2.8) gave the impression that the material was brittle. There
is as yet no explanation for formation of this deposit or how it came to be formed in this particular
location,

The radiation level on the outside of this section of the filter read from 200 to 360 r/hr. A
piece of the inner (coarse) Feltmetal and perforated plate (sample 3A) was removed and examined
in the hot cell. ' '

Fiberfrax Section. — The upper section of the filter containing the Fiberfrax was slit open, and
the layers were examined. The material at the entrance showed an oillike discoloration, but there
was no evidence of any significant accumuzlation of material. Corfiparison of the weights of the. |
different layers with the weights of the material originally loaded indicated changes’'of less than
0.2 & |

An interesting observation relates to the very low radioactivity level in this material, even in
the entrance section which is separated only by the Feltmetal filter from an area containing ma-

terial with activity levels of thousands of r/hr. The only detectable activity above the examination
CYN 1015

FRACTURED EDGE

Fig. 2.8. Upper End of Porous Metal Section — MSRE Particle Trap Mark I.

55

cell background (4.2 r/hr) was at the discharge end of the Fiberfrax section. It is probable that
~

this activity resulted from pressure transients which could have carried gaseous.decay products

from the charcoal trap back upstream into the particle trap: Even so, the activity level was only

1.8 t/hr above background.

Analytical Results

A total of four samples from three different locat1ons (F1g 2. 6) were subJected to a variety of

analytical tests. The samples were identified as follows:

Sample No. Taken from
3A Coarse section of porous metal filter
.9 Scale on lower flange of porous metal section
5B Mat at inlet to Yorkmesh section
6 Mat at inlet to Yorkmesh section

For sections of sample 3A it was found that about the same weight loss (0.2% of sample
weight) resulted from heating to 600°C in helium as from dipping in a trichlorethylene bath. The
material removed by heating was cold trapped and found to be effectively decontaminated; however,
the trichlorethylene wash was contaminated with fission products. ‘

Samples 5B and 9 were compared for low-temperature volatiles; at 150°C No. 9 lost 5% and
No. 5B lost none. When raised to 600°C, the weight losses were about the same, 35% for sample
5B and 32% for sample 9. Analysis of the carbon content gave none for sample 9 and 9% for ‘
sample 5B. This indicates that sample 9 had not reached as high a temperature as had sample 5B.

The mass spectrographic analysis of sample 6 indicated that there was a very high fraction of
fission products. These are estimated to be 20 wt % Ba, 15 wt % Sr, and 0.2 wt % Y. In the same
analysis the salt constituents Be and Zr were est1mated to be 0. 01 and 0.05 wt % respectively.,

In addition the material in samples 5B and 6 contained small quant1t1es of-Cr, Fe, and Ni,
while sample 9 did not. The reliability of these values is compromised by difficulties caused by
the presence of organics and small sections of wire in the sample,

The gamma-ray spectrographic work indicated the ptesence of the following isotopes: 137Cs,
89g; 103Ry or 196Ry, 119mAg 95N, and ! *°La. ‘

All three samples were chemically analyzed for Be and the level was below the detectable
limit of 0.1%. Attempts to analyze for Zr were compl1cated by the presence of large quant1t1es of
Sr. '

The fraction of soluble hydrocarbons was determmed using CS " and the values were 5B, 60%;
6, 73%; and 9, 80%. The extract solutions from samples 5B and 6 were allowed to evaporate and
a few milligrams of the residue was mounted between salt crystals for infrared analyS1S The sam-

ples were identical and were characteristic of long-chain hydrocarbons.__There was no evidence of
56

any functional groups other than those involving carbon and hydrogen. Nor was there any evidence
suggesting double or triple bonds. There was an indication of a possible mild cross-linkage. It:is
likely that there is more cross-linkage ‘of the organic in the gas stream than appeared in these sam-
ples, and the low indication could be due to the insolubility of the cross-linked organic and the
high operating temperatures of the wire mesh, which would cause breakdown of the organic into
elemental carbon and volatiles. |

The following conclusions can be drawn from the results of the examination:

1. Since the spectrographic analysis indicated that the concentrations of Be and Zr were very
small compared with those of the fission products Ba and Sr, the amount of entrained salt mist /
carried to the filter was hegligible_.

2. The high activity and the large amount of barium and strontium in the inlet. section of the
particle trap indicate that a large fraction of the solid daughters of Kr and Xe which decay in the
line is carried down the line with the off-gas stream.

3. The distribution of activity indicates that entrance areas "'of the Yorkmesh were unexpectedly
effective in trapping the solid fission products. Decay heat in this area resulted in temperatures
above 1200°F. |

4. The collection of hydrocarbon mist on the Yorkmesh had probably enhanced the collecting

efficiency of the mesh for solid particles.

2.8 MIGRATION OF SHORT-LIVED GASEOUS PRODUCTS INTO THE GRAPHITE
R. J. Kedl |

The measured concentrations of various fission products in the MSRE graphite samples were
listed as a function of depth in the samples in the previous progress report.® Three of these ma-
terials are daughters of very short-lived noble gases, '*°Ba from (16-sec) 4%Xe, 141Ce from
(1.7-sec) '41Xe, and °'Y from (10-sec) °!Kr. A model has been developed which evaluates the
concentration of “‘very short-lived’’ noble gases in the graphite while the reactor is at power. It
can be shown that with the reactor at power,

—xy\/ EA _ .
vD €/D. e " /e -

e = V5 ¢/
where -

C s = noble-gas concentration in graphite (atoms per volume of graphite),

Q = noble-gas generation rate in fuel salt (atoms/time, volume of salt),
A = noble-gas decay constant,

D = noble-gas diffusion coefficient in molten salt,

s

Dg = noble-gas diffusion coefficient in graphite,

"€ = graphite void fraction available to gas,

SMSR Program Semiann. Progr. Rept. Aug. 31, 1966, ORNL-4037, pp. 181—83.
R T

57

x = distance in graphite.
The very short-lived noble-gas concentration distribution in gréphite is approached very shortly

after the reactor is brought to power. We can therefore compute the longer-lived daughter product

concentration distribution as a function of time with the reactor at power. Also, the daughter

product concentration can be taken thr0ugh the period when the reactor is shut down by consider-
ing decay only. This model assumes that once a noble gas decays, its daulghter is immediately
adsorbed by graphite and does not migrate. ‘

Concentration distributions were computed for 14°Ba, 141Ce, and °'Y in the MSRE core graph-
ite samples using values of Dg obtained from the slope of the distribution curve. The computed
values of the concentrations at the surface of the sample are shown in Figs. 2.9, 2.10, and 2.11,
together with the measured distributions. For sake of clarity, the computed concentration is shown
only at the surface, but the computed internal distribution would be roughly parallel to the measured

internal distribution.

2 ORNL-OWG 67-4770

+ SAMPLE FROM TOP OF CORE—WIDE FACE

® SAMPLE FROM MIDDLE OF CORE—WIDE FACE

X SAMPLE FROM BOTTOM OF CORE—-W|DE FACE
CIRCLED POINTS AT DEPTH=0C INDICATE COMPUTED

©
<
-
o ® VALUES
52 &
Q
Q
© \
e 10! LR
TN
it \\. N
<5 N MIDDLE
3 o~ N\ TOP\
E \X \
’g BOTTOM
T 2 | + :
1 AN
S ‘+ X \ @
z \
= X
{0 X
8 0 \\ \\‘\ ’
= N N\ @ (]
— +
E N\ AN \‘X
S s AN =
Q \\ N
(&)
n \ \
m
¥ \ N\
2 .
\+\
10?

0 0.01 0.02 0.03 0.04 0.05 0.06
DEPTH IN GRAPHITE (in.)

Fig. 2.9. 140p . Distribution in MSRE Graphite Samples.
{4 co CONCENTRATION IN GRAPHITE (dpm /g-graphite at 4100 on 7-17-66)

58

CRNL-DWG 67-47T1

10
@
5 \ -
\. + SAMPLES FROM TOP OF CORE—WIDE FACE
2 ® SAMPLES FROM MIDDLE OF CORE—WIDE FACE  —
\ X SAMPLES FROM BOTTOM OF CORE—WIDE FACE
N CIRCLED POINTS AT DEPTH=0 INDICATE COMPUTED
5)6 VALUES '
\[#]
10 MIDDLE
+
\ v \
\ N\®,
TN
> \\TOP \
'\ BOT TOM
9 ,
10 \ AY AN
\\ \\ AN
_ e
\ N
° \\ \\ \.\
\ N
\ \ \\
2 + ‘\ X
8 M
10 \
AY
\
5
+
2
10
0 0.0 0.02 0.03 0.04 0.05 0.06

DEPTH IN GRAPHITE (in.}

Fig. 2.10. 141¢e Distribution in MSRE Graphite Samples.
59

ORNL-DWG 67-4772

s [\ DATA NOT AVAILABLE FOR TOP AND BOTTOM SAMPLES __|
\ CIRCLED POINT AT DEPTH=0 INDICATES COMPUTED |

\ . VALUE

vd

oL
(=]
/

\
—X
\\
5
AV
2 B
10° .
N\
\,
\

5 \

\MIDDLE

L

91y CONCENTRATION IN GRAPHITE (dpm/g-grophne at 4400 on 7-17-66)

o7 \
10a \
N
X
) \
5 AN
®
» ,
- \

107 :

0 0.01 0.02 0,03 0.04 7 005 0,06

DEPTH IN GRAPHITE {in.)

Fig. 2.11. 1Y Distribution in MSRE Graphite Samples.

2.9 REMOTE MAINTENANCE

R. Blumberg

The reactor shutdown which began in July 1966 extended. through October 30. " After that the
reactor cell membrane was cut three separate times to carry out maintenance operations in the
cell. During this period, power production increased from 7822 Mwhr to 16,277 Mwhr at the

beginning of the last maintenance operation.
60

Summary of Remote Maintenance Tasks Performed

August 26, 7822 Mwhr. — A removable section of off-gas line 522 was replaced. This spool
piece was replaced because it was suspected of having a plug of frozen salt. This was the hottest
piece removed from the reactor, approximately 2000 t/hr. It was removed with blind flanges ofi
both ends for containment.

September 7, 7822 Mwhr. — Both a control rod and a rod drive from position 3 were replaced.
This was routine from a handling standpoint. The lower end of the rod measured 250 t/hr on
contact. |

September 9. — The west space cooler that had been removed the previous period for
repair of a water leak was replaced. The assembly had some wipable contamination which was. re-
moved at the decontamination facility. There were some local spots of induced radiation of up to
400 mr/hr at 6 in. The leak, which existed in a brazed joint, was rebrazed directly, under Health
Physics supervision. r

September 10, October 30. — The sampler withdrawal pipe (line 999, see Sect. 2.1) was
thawed. By using a series of remotely manipulated external heaters, we managed to move the
frozen salt down the pipe to within 3 in. of the end. The final thawing of this line required using
the heat from the circulating flush salt plus that from several small ceramic heaters which had to
be remotely positioned in the complex area around the pump tank afid line 999. This was a totally
unanticipated job, and much of the difficulty was due to the crowding of various equipment into |

one area.

September 15. — The line 522 particle trap was removed and replaced. The trap which was re-
moved was stored in a dry well at the burial ground using blind flanges for containment. This unit
was later retrieved and examined at the HRLEL. :

» September 16, 7822 Mwhr. — We completed installation of the new graphite and Hastelloy N
surveillance samples, This went very smoothly and quickly. (The above operation completed the
shutdown which started in July.)

November 5, 10,067 Mwhr. — We installed three external heaters on line 522 at the pump in an
attempt to thaw a salt plug (see Chap. 1). This proved to be ineffective. The radiation level at
open tool penetration in the portable maintenance shield was 70 r/hr. -

November 23, 11,236 Mwhr. — The plug in line 522 was removed by direct means. External
heaters, chisel-type rods, a flexible shaft snake to clear a pipe with two 90° bends, and a
borescope were used on this job. The flexible pipe spool piece was replaced in order to make
sure that this line was open; 70 r/hr at tool penetrations was a definite factor in slowing down-
the work. |

December 5, 11,236 Mwhr. — We removed, repaired, and reinstalled valve 960, which controls
the component cooling air. Gamma radiation was significantly lower here than for the job on line
522. Induced radiation in the valve (less than 20 mr/hr) was low enough to allow contact repair

of the valve stem.
wites] |

~J

61

Jonuary 16, 16,277 Mwhr. — We replaced 17 “sfiap-tite” quick disconnects in lines which
supply air to operate valves. These disconnects had developed leaks due to radiation embrittle- |
ment of an elastomer seal. The replacement fitting does not have ény elastomers. The work was
started with only three days of decay, and the radiation at open tool holes in the maintenance
shield was 60 r/hr. The difficult part of this job, from a mechanical handling standpoint, con-
sisted in making up a 3/8-ir1. tubing joint on horizontal tubing at a distance of 20 ft.

While the air lines were being refitted in the reactor cell, a particle trap of improved design
was installed in the off-gas system. In this operation all of the existing parts were removed from
the area of the old trap, leaving only the inlet and outlet flanges. The new assembly was fabri-
cated in jigs to assure that it would fit the existing flangeé. -During the course of this job the

working area became severely contaminated and had to be cleaned several times.

Radiation Levels

Radiation levels increased si gnificantly in almost every situation. These are described in
Table 2.3, and a plot of in-cell radiation levels vs time is given in Fig. 2.12,

There are several things of interest that one may note from these curves:

1. Both the reactor cell and drain cell radiatioh levels are increasing as more power is.
produced. ‘ . }

2. The radiation level in the reactor cell, with the fuel salt drained, decays at a faster rate
than the radiation level in the drain cell with the fuel in the drain tank. .

3. The dramatic increase in radiation level in the drain tanks during the period that the off-gas
was vented through the drain tanks (December 23—January 14) gives some indication of how radio-
actively hot the off-gas is. |

4. On December 11 after some 23 days of decay, we find that the radiation level at the north
wall of the reactor cell is down to 250 r/hr from a level of 2000 r/hr on November 18. This is an
indication of the residual radiation effects that-may be of consequence in a shutdown involving

handling of major components.

Table 2.3. Description of Radiation Levels That Affect Maintenance

General background just above portable maintenance shield (PMS) when in place.

Usually less than 5 mr/hr; however, over the off-gas line it was 10 to 20 mr/hr.

At’an open tool penetration. Up to 100 r/hr recorded. Generally 70 r/hr over the
off-gas line. 60 r/hr on PCV 919 air disconnect work.

High bay levels when two roof blocks are off. The handling of these blocks and
movement of the PMS slide must now be done routinely from the remote maintenance
control room. During this operation, there were readings of 20 mr/hr in the reactor

control room and 40 to 60 mr/hr in the hot change house.

Hot spots on top of reactor cell before removal of lower shield blocks but after removal
of the upper blocks. We have run into local hot spots as high as 400 mr/hr before, but
during the most recent shutdown, there was an area reading 2 r/hr located alor_:g the

south edge of block R above the heat exchangér.

ORNL DWG 67-4773

-—— iympW 20G'6H

62

[N

- JyMW 002°9) =K

141

o

- q3and Hltm 1714

GNY T7132°38 4IA0 NHOM LHYVLS
JIONVYNILNIVAN NI938 17130 N3dO

3SNOH AN3A NI

A

-4—— 13Nn4 NIvd0

B

— Jymy 0029+ ¥2ILINDANS

N NN Z
AN R MR ¢
RN // NN AN ///WWW 7/ N

TIME-SCALE ———

Ne—e S

|

i \/\"“'\/ N

1I%Ss HSN1d 3LvIN2HID aNY 714

. 225 3NN
-H— ONI99NTdNN LYVLS

N A

\«<-PT NO. 6 BETWEEN THE TWO DRAIN TANKS (DRAIN CELL)

bW

e
lv

-4—— SX3078 dO0L 3IAOW3Y

<PT NO. { ON NORTH WALL NEAR PUMP (RE. CELL)

Bl

re—— 13Nd4 NIVYHO

-4—— TyItLIHIBNS
- JyMN 9EZH

NiT SV -4d40 Q399074 HLIM

NOILVYH3dO ¥IMOJ n_n_D.._w

AWW.N

10°

(44/4) T3A3T NOILVIQVY

FEB

JAN

1967

» JAN

In-Cell Radiation Levevls at the MSRE.

—

DEC
Fig. 2. 12,

NOV «———— = DEC

. 1966

63 4

Contamination

Control of the spread of contamination is becoming more difficult and requires a greater poftion
of our time and effort. In the effort to remove the plug of frozen salt in the off-gas line 522 (see-
entry for November 23), a number of very contaminated long-handled tools were accumulated.
Rather than risk the spread of contamination in the high bay, these toéls were stored in the re-
actor cell in a sheath of 4-in. pipe. During the replacement of the particle trap in the off-gas sys-
tem (see entry for January 16), airborne contamination required the use of gas masks and many

!
man-~hours of cleanup work.

Conclusion

The work that was accomplished involved the off-gas ‘system and auxiliary valves. Much of

this work was unanticipated, requiring new tools and procedures. In all cases there were tight
’ I4

‘schedule demands which were met. One may conclude that the-increase in skill and experience

has been at least equalled by an increase in difficulty arising from the radiation and contamination.
3. Pump Development

3.1 MSRE PUMPS
Molten-Salt Pump Operation in the Prototype Pump Test Facility

During the 2631-hr test of the MK-1 prototype pump reported previously,' there was evidence
of a restriction in the annulus between the pump shaft and the shield plug at the lower end. Post-
test examination revealed the restriction to be salt of a composition similar to that circulated by
the pump; so it could not have been produced by condensation of salt vapor. We believe that this

“salt entered the annulus during a previous l'oop filling operation. The test loop is filled by pres-
surizing the drain tank with helium to displace salt into the loop through a fill and drain line.
Because there was insufficient salt in the drain tank, the liquid level in the tank dropped to the
fill-line entrance before the loop was filled to the desired level. Gas then passed through the
line into the loop and raised the level in the pump tank high enofigh'to carry salt into the an-
‘nulus. Some of the salt froze there and remained when the level subsided as the gas entered
the pump tank.

All the loop tests for the MK-1 pump and the testing of spare rotary elements for the MSRE
pumps were completed; so the prototype pump tank was removed from the test facility to make

way for the MK-2 pump tank.
MK-2 Fuel Pump

The MK-2 pump was designed to replace the MSRE fuel pump if the pump tank in the reactor
system were to fail. The MK-2 pump tank has the same diameter as the MK-1 tank but was made
longer to provlide the desired expansion volumelfor salt without an overflow tan-i( below the pump.
The longer pump tank made it necessary to increase the unsupported length of the pump shaft.
Also, some changes were made in the mechanics of cortacting salt and cover gas to remove
gaseous fission products from the salt. These and other features of the new pump will be tested
before it is considered suitable for use in the MSRE. '

Sh_aft deflection and critical speed tests were performed on the rotary element for the MK-2
pump. The measured critical frequency was 1860 cpm, which compares favorably with the pre-

dicted value of 1800 cpm. Fabrication of the pump tank? is about 70% complete.

IMSR Program Semiann. Progr. Rept. Aug. 31, 1966, ORNL-4037, p. 81.
21bid., p. 82. '

64
4

65 =

Stress Tests of Pump Tank Discharge-Nozzle Attachment

The discharge nozzles of the MK-1 and MK-2 pumps penetrate the pump tank in the region of

the junction of the bottom head with the cylindrical side. This unusual configuration greatly

increases the uncertainty in calculation of stresses in the nozzle and the tank wall. Hence we
decided to use the MK-1 pump tank to obtain a sin}ple comparison between calculated and meas-
ured stresses produced by internal pressure and by.fprces and moments imposed by the discharge
pipe. | | ' -

A fixture was designed and fabricated for holding the tank and for applying forces .and moments
to the discharge nozzle. Stress-Coat tests were initiated, and a crack in the tank was found in
the heat-affected zone of the weld joining the nozzle to the tank. The Hastelloy N from which the
tank and weldments were macie came from early experimental heats whicl'; are now known to be
crack sensitive. We believe that microfissures were present in the welds when the pump was new,
and that théy enlarged to form a visible crack either under the stresses present during 15,000 hr of
operation at temperatures from 1000 to 1500°F or during the initial Stress-Coat test. The crack was

repaired by welding, and exploratory strain measurements are being made to determine whether we

can obtain valid data.

Spare Rotary Elements for MSRE Fuel and Coolant Salt Pumps

We reported previously on the testing of the spare rotary elements for the fuel and coolant
pumps of the MSRE.? The spare element for the coolant pump ran satisfactorily, but the one for
the fuel pump leaked oil excessively through the lower shaft seal. The rotary elements are
identical except for the impellers and the.arrangement of service piping, which is i.nstalled at
the reactor site. For a time it appeared that replacement of the rotary element of the fuel pump
in fhe reactor might be necessary to overcome the problems of the off-gas system; so we con-

verted the spare rotary element for the coolant pump into a spare for the fuel pump.” It is being

~ held in standby for the reactor.

~ The cause for the excessive seal leakage in the other i'otary element was investigated. The
stationary part of the seal has a Graphitar rinhg supported on a stainless steel bellows. The
Graphitar ring bears on a carbon steel ring on the pump shaft. The bellows has a weld down one
side which increases the stiffness of that side. Thishlack of uniformity and possibly the way in

which the seal was assembled caused the seal surfaces to be not quite parallel or concentric, and

'the seal leaked excessively. The bellows was adjusted to improve the parallelism and concen-

tricity of the seal rings, the pump was reassembled, and the testing was completed. The rate of
oil leakage through the lower seal-was only 21 cm?/day at reactor operating conditions.
A fixture for holding the stationary seal element to attain the required machining precision was

designed and fabricated and will be used for making seal assemblies in the future.
4 66

Lubrication System
The lubrication pump endurance test? was continued, and the pump has now run for 31,350 -
hr, circulating oil at 160°F and 70 gpm. ‘
3.2 OTHER MOLTEN-SALT PUMPS
Fuel-Pump High-Temperature Endurance Test Facility

A new drive motor rated for 200 hp at 1800 rpm is being purchased for the salt pump.? With the

delivery of the motor, the facility will be ready for operation with molten salt at high temperature.
%

4. Instrument Design and Development

4 _ \ 4.1 INSTRUMENTATION AND CONTROLS DESIGN

R. L. Moore J. R. Tallackson

Off-Gas Sampler Design
. | | D. G. Davis

The déSign of instrumentation and controls for the off-gas sampler! was continued. This de-
sign work was temporarily stopped in April 1966 because of uncertainties concerning tlre chro-
matograph and the possible effects of particulate matter found to be present in the off-gas sys--
tem.?'3 Work on the design of a revised off-gas sampler system was resumed in August and is
now essentially complete.

Changes in the original design which affected instrumentation and control were as follows:

The chromatograph system included in the original design was removed.

2. A copper oxide scrubber was added for removal of hydrocarbons from the sample gas.

3. A second themmal conductivity cell was added for checking efficiency of the copper oxide
scrubber.

Revisions to instrumentation and controls included:

new heater controls for both the copper oxide scrubber and the thermal conductivity cell,

2. replacing the single-pen Brown recorder with a two-pen recorder for recording the outputs of
two thermal conductivity cells,

- ~ 3. additional themmocouples,

4. changes in power and control wiring and in panel layout.

Except for the above changes, the instrumentation and control for the off-gas sampler remains
as previously reported. ! |
Panel fabrication of the new design is approximately 95% complete, and assembly work inside

the containment box is approximately 75% complete.

MSR Program Semiann. Progr. Rept. Feb. 28, 1966, ORNL.- 3936 pp. 6769,
[bzd pp. 23-27.
Ibrd., pp. 65—67.

~

67
" 68

Control System Design

P. G. Herndon

Control Instrumentation Additions and Modifications. — Further additions to and modifications
of the instrumentation and controls systems wete made to provide additional protection, improve
petformance, and provide more information for the operators. Fifty-nine requests for changes in
the instrumentation and controls system (or in systems affecting instrumentation and controls)
were received and reviewed during the past report period. Of these, 29 requests resulted in
changes in instrumentation and controls, 1 was (:_anceled; 9 did not require changes in the instru-
mentation or controls, and 7 are active requests for which design revisions are either in progress
or p.ending. The remaining 13 requests required changes to process switch operating setpoints.
Some examples of these changes follow. '

Master Control Circuits. — Two new jumpers and associated circuitry were installed to provide
a bypass around the frozen permissive contacts of freeze valves 105 and 106 in the drain tank
helium supply valve control circuit. This makes it possible for the operator to transfer fuel salt
between drain tanks FD1 and FD2 through the reactor fill lines 105 and 106. To prevent inad- |
vertent transfer of fuel salt to the reactor core, the circuits were originally designed to prevent
transfers by this route. Transfers were permitted through the transfer lines only. Operating ex-
perience, howevef, has proved that direct transfers between tanks FD1 and FD2 are quicker and
less complicated operations under most circumstances than going through the transfer line. 4

The design of the jumper board graphic panel was revised to incorporate recent control circuit
changes and a newly devised system of identification for all jumper plug receptacles and lamps.
Procurement of four new metal-photoengraved graphic panels, required for this revision, is under
way. |

Fuel and Coolant Pumps — The design and installation of circuitry that permits the operators
to override all protective interlocks in the coolant-salt c1rcu1at1ng—pump control circuit were com-
pleted. New operating criteria stipulated that preventing salt from freezing in the radiator is more |
important than protecting the pump from possible damage resulting from the bypassed interlocks. * |
The design included a manual switch connected directly in the power circuit breaker ‘‘close’’ and
“‘trip’” coil circuits. The power distribution system was not altered. _

As part of the system revisions made to prevent plugging of the off-gas system, the fuel pump
bowl pressure control valve was removed from the system. The change was made because the de-
sired degree of control can be effected by other means, and the space was needed for installation
of a particle trap.

Fuel Off-Gas Systems. — To help overcome the operating difficulties caused by intermittent

plugging in the main reactor off-gas line, a helium gas supply system was installed in the vent

*Memo from P. N. Haubenreich to J. R. Tallackson, August 3, 1966.
SMSR Program Semiann. Progr. Rept. Aug. 31, 1966, ORNL-4037, pp. 61~62.
%

69
I\,l
house. The instrumentation and controls for this system, which include dofible-stage pressure

regulation, high-pressqre alarms, variable area flowmeters, and several pressure indicators, were
designed to protect the reactor fuel system from the possible application of overpressure and to
maintain the integrity of primary containment. _

A capillary flow element, FE-524B, was designed, fabricated, and installed in the fuel-pump
upper-gas-letdown line 524. This element replaces the matrix type of flow element originally in-
stalled in this line. Efforts to remove foreign matter that was plugging this element were unsuc- -
cessful.

A bubbler type of level measuring system was installed in the vent house to indicate the water
level in the charcoal filter pit. By use of valves, this system can also be used to check the water
level in the 522 line filter can. A second 12-point temperature recorder was also installed in the
vent house to record charcoal filter bed temperatures. These themocouples were originally con-
nected to a single-point indicator through a manual selector switch.

To provide additional information, three thermocouples were installed on the electric heater
used to heat the top of the charcoal-bed filters. | ’

Load Control. — An instrument system that monitored vibration in the coolant-salt-radiator
cooling-air blowers was designed and installed. The system consists of four self-generating
elements connected to an electronic vibration signal ar.lalyzer. One element is mounted on each
bearing of the two large blowers, and each element produces electrical signals proportional to
the velocity of the mechanical vibrations. The vibration signal analyzer indicates the average
velocity and the peak-to-peak displacement of the vibrations. Prior to installation, the system,
including the interconnecting cables, was tested on a calibr'ated’lshake table in the ORNL engi-
neering mechanics laboratory in the Y-12 area. All the components, valued at $2500, were ob-
tained from Reactor Division stores and were surplus items obtained from CANEL. . In addition
to the vibration instruments, a thermocouple was installed in a well drained into each of the four
bearing housings and was connected to the data logger through the thermocouple patch panel.

The load setback interlocks were removed from the load control circuits. This action, de-
signed to lower the radiator doors, under certain conditions, to the intermediate or 1-Mw output
position, was ineffective because of the difference between predicted and actual loading charac-
teristics of the radiator system. Operating-experience has shown that this action reduces the
load much less than expected, mainly because the radiator is so well ventilated with the doors
closed that the 1-Mw output position is actually at the lower travel limit of the doors.

Auxiliary Systems: — Additional instrumentation was provided for the treated cooling water |
system, which was modified to eliminate radiolytic gas. The newly installed degassing tank was
equipped with an autematic level control system consisting of a differential pressure transmitter,
pneumatic recording controller, and-a 4-in. valve in the tank discharge line. A new air purge
system was also installed to serve both the degassing and surge tenks. Secondary containment

was maintained by installing two solenoid block valves, one in the purge supply and one in the
70

vent line. A low—air-flow alarm switch was also installed in the vent line to give warning that
an explosive H, mixture may exist in the degassing tank.

A pressure-actuated switch was mstalled in line 917 at a point upstream from a newly in-
stalled filter to annunciate high component-coolant-pump discharge pressure.

A number of instruments in one safety circuit channel are supplied from the 48-v dc bus
through a 1-kw inverter; a loss of power to these instruments could cause the reactor to drain
unnecessarily. A circuit was designed and installed that will automatically transfer the load
from the 1-kw inverter to the reliable power distribution panel IPP3 if the inverter should fail.

Four additional thermocouples were added to the sampler surveillance rig;

Secondary Containment. — Just before the January shutdown of the reactor it was becoming
increasingly difficult to maintain the proper atmosphere in the reactor cell because of leaks from
the lines supplying air to the in-cell control valves. Several temporary corrective measures such
as installing flowmeters to locate the leaks and connecting the operators to a nitrogen supply
were taken so that operations could continue until the t‘i'me for a scheduled shutdown. During the
shutdown it was determined that the neoprene O-ring seals in the remote disconnects had failed
on some valves. These disconnects were replaced with standard ?’/s—in. compression-type fittings.
The system is now operating satisfactorily. '

A fixed bypass restriction is installed between the operator supply and vent lines to provide
a path for pressures on the valve operator to equalize so that the valve can move to the fail-safe
position \;vhen the reactor cell containment block valves close. During normal operations the flow
through these restrictors was large enough to cause a pressure buildup in the vent header which
produced undesirable effects on other valves connected to the same header. To correct this prob--
lem, the vent lines from the reference side of the operators on in-cell control valves PdCV-960A,
HCV-919A, FCV-903, and HCV-919B were disconnected from the common vent header and con-
nected to a containment ventilation duct through separate 1/“-in.-OD tube lines.

During a routine test, one of the two high-level switches in the vapor-suppression tank was
found to be defective. Loss of this switch caused the loss of one channel of information needed
to establish that the water level in the tank was correct. Since a second switch failure might
require that the reactor be shut down until the switches could be repaired, and since the removal
of the switches from the tank is a difficult operation, a bubbler-type level measuring system,
which includes a containment block valve and associated safety circuits in the purge supply, was
designed and was installed in the reactor-cell vapor suppression tank. This installation utilized
a dip tube which was included in the original design in anticipation of such need. This new level
system will also enable the operator to check the water level in the tanks as a routine procedure.

Sampling and Enriching Systems. — Except for adjustments to operating switch setpoints and
annunciator revisions, the instrumentation and controls on the fuel sampler-enricher have not
been changed. Increasing leakage, however, from the fuel sampler-enricher operational valve
buffer zone into the sample removal area 1C has made operations difficult with the existing safety

interlock arrangement. Alternate circuit arrangements are now being considered.
71

The annfinciator circuits for high radiation in the containment areas of the fuel sampler-en-
richer were revised for the convenience of the operator. The existing circuits annunciate high
radiation directly each time it occurs, and frequent annunciations are normal during certain stages
of sampling. The revised circuit operates to block out the annunciator after the first alarm is
silenced until the safety interlock circuits are manually reset.

The installation of the fuel processing sampler was completed.. Several instrument and control .
circuit changes which have improved the operation of the fuel and coolant salt samplers were in-

corporated in this facility.

4.2 MSRE OPERATING EXPERIENCE

C. E. Mathews J. L. Redford R. W. Tucker .
Confrbl-Sysfem Relays

The coils in four of the 120-v ac relays failed during this period. All the relays we*e in
freeze-valve control circuits and had been enetgized continuously for long periods of time.

Replacements were received for all 139 relays that operate from the 48-v dc system. These
relays were overheating and destroying the Bakelite frame due to faulty relay coil design. Twenty

of the relays were replaced, and the rest will be replaced at the next convenient shutdown.

Temperature Scanner System

The oscilloscopes on the scanner system that used vaéuum tubes were reblaced with transis-
torized units. This has resulted in better operation and reduced maintenance of the scanner sys-
~ tem. ‘

A calibrating system was added to the écarmers to permit calibration without making external

connections to instruments.

Yalves

Control valve PCV-960 for the component-coolant pumps stopped operating properly but could
not be repaired until it could be removed from the reactor cell. Repair-s were effected by cleaniqg
and lubricating the valve stem and packing. | ! .

Equalizing valves were installed on FT-903, which senses cooling-air flow to the pump shroud,

so that this instrument loop can be checked with the reactor operating.

Nuclear Instrumentation

A modified drive tube unit, which provides for controlled cable bends, was installed in wide-
rangé counting channel 1 after a short occurred in the cable to the preamplifier.
A capacitor was installed in the servo system to prevent noise from a Foxboro converter from

causing the servo to oscillate.
72

The fast-trip comparators used in the nuclear system were found to be inoperative if a suf-
ficiently large signal was applied to the input. A diode added to the FTC circuit removed this
possibility and was installed on MSRE FTC moduies.

Operation of the relay matrix in the nuclear system generated considerable noise, making it .
difficult to reset the safety-system channels. The existing resistor-diode combination for damp-
ing the voltage induced by relay operation was replaced by a Zener diode and a diode combination

that was more satisfactory.

Electronic Switches

A capacitor in an Electra-systems module failed, causing a reactor shutdown. Allvcapacitors.
‘in these modules will be tested during the next shutdown. Multiple setpoints were detected on
~ several modules, These were fixed by changing resistors to compensate for transistor aging.
The zero bias windings in the magnetic amplifiers in four modules were found to be open-circuited.
The magnetic amplifiers were replaced.
/ . /

Thermocouples

Thermocouple performance has continued to be excellent.

Coolant-Pump Radiation Monitor

Radiation monitor RM 6010 has its output recorded by the data logger. However, the range of
this instrument can be changed by a switch on the front panel. Since the range was not recorded
by the data logger, the actual radiation level recorded was not definite. Additional contacts were

added to the selector switch to provide the data logger with the necessary range information.

Safety System

Five spurious rod scrams were experienced during the reporting period. None were caused by
~malfunction or failure in the reliable rod scram instrumentation. Three of these scrams were the
result of human errors, and the remaining two were caused by equipment failures. One of these
equipment failures was in the radiation monitoring instrumentation, RM-528, on the off-gas line
from the coolant circulating purr;p. This is a two-channel system, and a failure in either channel
will stop the fuel circulating pump and thereby reduce the rod scram setpoint to 15 kw. At the
time this failure occurred, the reactor was developing over 7 Mw. The other equipment failure
was in an overvoltage relay in the 60-kva power supply. The resulting momentary loss of power

produced a scram.
73

4.3 DATA SYSTEM

G. H. Burger C. D. Martin

The reactivity-balance calculation was refined to the point where it now produces meaningful
numbers. In order to maintain system continuity in the event of a shutdown and subseqfient re-
start, a program was added that writes information necessary for initialization of the system on
the drum memory once each minute and on magnetic tape once each hour. Certain reactivity-
balance parameters are among these data. In event of a computer malfunction, which causes an
automatic restart of the system, the initialization information last written on the magnetic drum
memory is used to restore the core memory. For restart after shutdowns involving malfunctions
which may have destroyed the drum data, the last information written on magnetic tape is used
to initialize the system. The availability and use of these drum and magnetic-tape data have
made the éyétern easier to use and have given an extra margin of continuity assurance against
system faults.

Several new operator request functions were added to the system, -and a number of others were
modified to eliminate some possible conflicts between functions. Functions added included three
demand log-type requests that list those scan points on which alarm functions have been modified
by previous operator requests. Also added was a request function to facilitate updating (when fuel
is a&ded) the numerical quantity of fuel used in the reactivity-balance calculation. Though no
new analog signals were added to the scan lists, four contact inputs were added to indicate the
range selector switch setting of the electrometer amplifier for the high-level gamma chamber in
the coolant cell (RE-6010). The engineering units conversion equation for this signal was mod-
ified to read the selector switch position and convért the analog input signal accordingly. This
change was made to provide the best possible resolution of the instrument reading under all op-
erating conditions, as this radiation level varies from a few milliroentgens per hour to several
hundred roentgens per hour, depending upon the status of the reactor.

The fully implemented Reactor Operator’s Computer Manual was completed and was made
available in one volume. . The manual contains all of the information required for maximum effec-
tive utilization of the system in conjunction with reactor operations. .

During the reporting period the system was used extensively for routine daily operation and
analysis of the reactor. Its utility was very strongly impressed upon us in December when the
failure of a bearing in the drum memory unit rendered/ the system totally useless for several days.
With the system out of service, those functions normally performed automatically by the computer
had to be done manually — and sometimes painfully. Replacement of the failed bearing was
prompt, and the drum was quickly back in service. For some time after the system was restarted,
sporadic failures of short duration were experienced. When the drum failed, all system power was
shut off. These sporadic failures are thought to have originated with slow changes in the elec-

tronics as the system reestablished thermal and electrical equilibrium during the restart.
TIME (%)

100

mW Wb s NN ® @D Y W
oo U O ¢ 0O OGO UL O g O Wu

15
10

CRNL-DWG 6€7-4774

[] "oN* TIME SCHEDULED SHUTDOWNS SCHEDULED SHUTDOWNS
FOR MAINTENANCE AND FOR MAINTENANCE OF
MODIFICATIONS PERIPHERAL EQUIPMENT

MSRE COMPUTER TIME AVAILABLE

Q- """-- N I D O O e e = I I e 1 ] N Y R =
L \F - I D R T DU R DS P o oy o DN | ot o T T e T ] ~96.42 %
dRRsE ]
5 . —

P

.

|

. -

] . -
1 84522 29I 512 1926'3 10 47 24347 14 2128'4 1118 25’4 1" 4825’ 1 B 1522296 132027| 3 401724' 18 152229| 5142 *926'2 9462330718 2 2814 144821 2 91623306 1320 27|310|724
ocT NOV DEC JAN FEB MAR APRIL MAY JUNE JULY AUG SEPT ocT NOV DEC JAN FEB

1965 1966 1967
' WEEK BEGINNING '

Fig. 4.1. MSRE Data System Service Record from Date of Acceptance Oct. 1, 1965, to Mar. 1, 1967.

vL
v 75

The failures cited above reduced the ovérall availability of the system (Fig. 4.1) from the date
of acceptance to just over 96%, as compared with 97.31% at the end of-the last reporting period.
Although the system is operating satisfactorily, there are still several minor problems, apparently

in the hardware; they act as follows:

1. occasionally cause the system to block out requests from the operator’s console,
2. cause an automatic restart, ’ -

3. cause erroneous, unreadable data to be written on magnetic tape.

These firoblems are not ve.ry serious, and we have devised methods of circumventing them. They

do constitute minor annoyances which we hope to eliminate in the future,

4.4 INSTRUMENT DEVELOPMENT

R. L. Moore G. H. Burger J. W. Krewson

Performance — General

Performance of developmental instrumentation installed on the MSRE continued to be gener-
ally satisfactory. Although some isolated failures and some deviations from optimum petformance
occurred, they were minor in nature and were correctable by routine repair and adjustment. No
~problems were encountered which required the redesign of any of the developmental ifistrumenté-
tion presently installed on the MSRE or the initiation of development of new instrumentation.

Since operation of MSRE developmental instrumentation has become routine, the details of
performance are reported in Sect. 4.2. Comments on performance in the following paragraphs are
limited to those associated with the initiation or termination of development efforts. Where no
mention is made of the performance of a developmental instrument, it may be assumed that per-

formance has been satisfactory and that no modifications or unusual attention has been required.

Temperature Scanner

Investigation of the feasibility of réplacing the mercury switches used in the temperature
scanner system with solid-state multiplexers was continued. ® Results of these investigations
showed that none of the commercially available devices considered was suitable for use as a
direct replacement for the MSRE switches, and that the cost of all was prohibitive for the in-
tended service. Several of the devices, however, showed prémise for use in more. conventional
temperature-scanning systems in their present form or, with further development, for use in future
MSRE-type systems. We have located sufficient ,meréury-switch spare parts to satisfy foresee-
able MSRE requirements, and further investigation of s’olid-gtate multiplexers has been indefi-

nitely postponed. !

®Ibid., p. 84.
76

High-Temperature NaK-Filled Differential Pressure Transmitter

No further progress was made in determining the cause of zero and span shifts in the coolant
salt system flow transmitter that failed in service at the MSRE or in the repair and reconditioning
of this instrument,’

Negotiations with the manufacturer for repair or replacement of the transmitter ordered for use

as an MSRE spare are continuing:

Ultrasonic Level ProBe

Since no salt has been transferred to or from the fuel storage tank during this report period,
there has been no opportunity to check the effectiveness of modifications of probe circuitry previ-
ously reported.® They were made to eliminate the excessive frequency drift experienced in the

excitation oscillator associated with this probe.

MSRE Fuel Distillation System Level Probe

A conductivity-type level indicator has been developed for use in a fuel distillation system
which will be installed at the MSRE and has been successfully tested in the MSRP level test
facility. This level indicator is similar in operation and construction to the conductivity-type
system used to obtain single-point indication 6f level in the MSRE drain tanks, ? but differs in
that the probe is smaller in size, and the new system will provide a usable indication of level
changes. The range of this instrument is limited to approximately 30% of the length of the signal
generating section, because the signal becomes increasingly nonlinear with changes in molten-
salt level outside that range (see Fig. 4.2). With the modification described below, it is con-
sidered adequate for use in the fuel distillation system.

A way of reducing the effect of this nonlinearity in the molten-salt level rangeé required for
this experiment is to increase the length of the signal generating section so that the bottom 30%
of the probe covers the desired range. This was accomplished, without exceeding the limitations
on overall length of the instrument, by folding the excitation section as shown in Fig. 4.3. The
limitations on overall length were due to the limited space around the still. Preliminary drawings
for this system have been completed and are now being circulated for approval. This level indi-

cator will be constructed of INOR-8.

Ball-Float-Type Level Indicator for the Mark 1l Pump

The design of a float level indicator for the Mark Il pump was completed, and fabrication is

under way. This instrument differs from the level indicator now installed on the coolant salt

“Ibid., p. 85.
81bid., pp. 85—86.
SMSR Program Semiann. Progr. Rept. Jan. 31, 1963, ORNL-3419, pp. 40—42.
77 _

ORNL-DWG 67-4775

80
°
/b./
padis
70 -
/.o
o
60 ,/
_ : ®
y
‘, : EXCITATION-1.3 A AT 1 kHz MOLTEN
50 _ SALT TEMP-1300°F PROBE °

) ® "LENGTH-6 in.

40 — l.
/
iy

I SIGNAL LEVEL CHANGE WHEN MOLTEN

SIGNAL (mV ac)

SALT TOUCHED LOWER END OF PROBE

20 T

BACKGROUND NOISE LEVEL
(PROTOTYPE TEST)

0] : i 2 3 4 5
MOLTEN SALT LEVEL CHANGE (in.}

Fig. 4.2. Conducti-vify-Type Level Indicator for Fuel Distillation System — Signal Output vs Level.

pump at the MSRE in that the float is inside the pump bowl, and the range is 7 in. instead of
5in. Also, as previously reported, 10 the method of transformer mounting will not permit removal

for repair or replécement. Once installed, it must remain on the pump bowl until the pump is re-.

moved.

100/SR Program Semiann. Progr. Rept. Aug. 31, 1965, ORNL-3872, p. 71.
78

ORNL-DWG 67-4776

TOP VIEW

SIGNAL AMPLIFIER AND
LEVEL INDICATOR

EXCITATION
SOURCE

SIGNAL LEADS

AR
2 B 1 | 1 HEAD COVER
5 ‘N
| A '
1 | 4 FOLDED EXCITATION
’ SECTION
’
‘s o
o 1
/
‘BN’ CONTAINMENT VESSEL
11N
(1
AN NN N R / OONSOMOIOOONOSANND
5

/SIGNAL' GENERATING SECTION
CUTAWAY SIDE VIEW

AN I RRR AR

Fig. 4.3. Conductivity-Type Level Indicator for Fuel Distillation System — Simplified Schematic.
5. MSRE Reactor Analysis

5.1 NEUTRON REACTION RATES IN THE MSRE SPECTRUM

B. E. Prince

Computer programs presently used for reactor physics analysis permit a faitly detailed rep-
resentation of the neutron energy spectrum in the central, or most important, region of the re-
actor core. In the MSRE a small fraction of the total neutron reactions do occur in the regions
which surround the central graphite-moderated region (i.e., the peripheral downcomer and the
lower and fipper séit plenums). In these locations the neutron spectrum can differ somewhat
from that in the central region. However, it can be shown that this fraction of the total re-
actions is quite small, and that reaction cross sections averaged over the neutron spectrum
in the graphite-moderated region are accurate enough for estimating important isotopic changes

| and their associated reactivity effects during MSRE operation.

Programs which have been used extensively for MSRE physics analysis are the GAM-II!
and the THERMOS2 codes. The former calculates the epithermal energy spectrum of the neu-
tron flux (the flux ab.ove energies where thermal motion of the moderator nuclei exerts a
strong influence on the neutron spectrum). The latter calculates the spectrum in the thermal
energy region, together with its dependence on position within the individual cells comprising

" the lattice of fuel salt and graphite. In each of these programs the actual energy spectrum is
approximated by a finite number of energy groups (99 groups in GAM-II spanning the interval
15 Mev to 0.414 ev, and 30 groups in THERMOS spanning the interval 0 to 0.876 ev). The
programs intrinsically utilize library tapes of neutron cross sections for each nuclide and
energy group, which represent the best current evaluation of experimental cross-section data.
The GAM-II program was acquired subsequent to many of the early studies of the nuclear
characterisfics of the MSRE, most of which are summarized in ref. 3. Results of later anal-
ysis using these two programs, in conjunction with few-group diffusion caléulations, have

been found to be in good agreement with experimental observations in the MSRE. *

lG. p. Joanou and J. S. Dudek,r GAM-IT — A 83 Code for the Calculation of Slowing Down Spectrum and
Associated Multigroup Constants, GA-4265 (1963).

2y, C. Honeck, THERMOS — A Thermnalization Trafisport Theory Code for Reactor Lattice Calcula-
tions, BNL-5826 (September 1961).

3P, N. Haubenreich et al., MSRE Design and Operations Report, Part III: Nuclear Analysis, ORNL-
TM-730 {Feb. 3, 1964). : :

4P. N. Haubenreich et al., MSRE Zero Power Physics Experiments, ORNL-TM (in preparation).

79 '
80

The energy dependence of the epithermal and the thermal neutron spectra in the MSRE core
ét 1200°F, obtained from calculations with these programs, is shown in Figs. 5.1 and 5.2. In
these calculations the core was represented as a right circular cylinder (R x H = 29 x 78 in.).
This choice ensures that the epithermal neutron leakage calculated in the GAM-II idealized
model closely approximates the actual gross epithermal leakage from the graphité-moderated
region of the MSRE core. For the THERMOS calculations the fuel salt channel and graphite
lattice cell was represented as an idealized one-dimensional cell in cylindrical geometry.

The control rods were assumed to be withdrawn, and localized perturbations in the spectra
caused by the control-rod thimbles were neglected.
- In Fig. 5.1 the epithermal neutron flux spectrum is plotted vs lethargy, a dimensionless

variable proportional to the logarithm of the neutron energy, namely,
lethargy = —1n (energy/10 Mev) .

In this energy region, where most of the important resonance effects in the neutron reactions
occur, the lethargy variable is used for theoretical and computational convenience. For ease
of interpretation of Fig. 5.1, we have included a logarithmic energy scale. The calculated
fluxes per unit lethargy for each group in GAM-II are shown as solid points, with the points
connected by straight line segments. The flux spectrum in Fig. 5.1 is normalized to one

sourcé neutron produced from fission. In the GAM-II model this flux spectrum corresponds

ORNL-DWG 674777
ENERGY (ev) ,
4 5 6 7

10° 10 10
'8 I I | | | l I

FLUX PER UNIT LETHARGY (naormalized to one fission source neutron}
™
I

i7 i6 ~ 15 14 13 12 11 10 9 8 7 6 S 4 3 2 f 0
LETHARGY '

Fig. 5.1. Calculated Epithermal Neutron Spectrum in the MSRE.

e
81

ORNL—-DWG 67-4778

0.20

A6 /
" | /AN
0.12 | /9){/ \1\
- CENL.E;;C CiERAIPHIITIIE |SITRINGER-.._17 . \

FLUX PER UNIT ENERGY {normalized)

0.08 |— . :
° [elE)dE=1.0186 ‘//,4 \
: — >CENTER OF FUEL CHANNEL
0.04 ‘ ' /‘e/ 0.876ev ’ ’ l | ||
: A [ E)dE=0.9760
A L
—6/—"" . el
0! :
0001 0.002 0005 001 0.02 005 04 0.2 0.5 1

ENERGY (ev)

Fig. 5.2. Calculated Thermal Neutron Spectrum in the MSRE.

to that in a single-region reactor, consisting of a'homogeneofis mixture of graphite and fuel
salt, in which corrections have been introduced for the fuel channel geometric self-shielding
of the absorption resonances in 238U. The physical neutron spectrum corresponding most
closely to this theoretically calculated flux should be the spectrum within the graphite
stringers. From Fig. 5.1 it may be seen that the flux spectrum above about 0.1 Mev exhibits
the general characteristics of the fission spectrum. At lower energies (increased lethargies),
there is a transition to the spectrum of scattered and moderated neutrons. The irregular de-
pendence on energy of some of the group fluxes in the transition region can be identified with |
the effects of neutron resonance scattering in the light elements in the fuel salt (’Li, °Be,
and 1°F). | |

Figure 5.2 éhows the flux spectrum in the energy region below about 1 ev (0.876 ev has
been chosen as the effective upper cutoff energy for the thermal group in the MSRE calcula-
tions). In this energy region the cross sections of the important nuclides in the MSRE do
not exhibit strong resonance effects, and therefore a smooth curve has been drawn through
the calculated group fluxes. Comparison of the curves in Fig. 5.2 shows that, within the fuel
salt channel, there is a slight depression or flux disadvantage, together with an energy ‘‘hard-
ening,’’ or preferential removal of low-energy neutrons, relative to the graphite. This spatial
flux depression is relatively small from the standpoint of estimating reaction rates (maximum
depression in the energy-integrated flux is less than 5%, volume averages over the graphite
and fuel regions less than 3%). The local flux spectra shown in Fig, 5.2 are normalized such
that the energy-integrated thermal flux, averaged over the total volume of fuel channel and
associated graphite, is unity.

Cross sections for the important nuclide constituents in the MSRE, averaged ovér the

spectra shown in Figs. 5.1 and 5.2, are listed in Table 5.1. To obtain the effective cross
82

Tabie 5.1. Average Cross Sections for Thermal and Epithermal Neutron Reactions in the MSRE Spectrum

Cross Section Averaged Cross Section Averaged

Effective Cross Secticn

Over Thermal Spectrum,  Over Epithermal Spectrum,

Nuclide Energy <0.876 ev Energy > 0.876 ev in Thermal Flux
(barns) (barns) (barns)
6pi® 417.5 17.4 457.6
Boron 330.4 13.8 362.4
149gm 3.57 x 10* 90.8 3.6 x 10°
151gm 2.63 x 103 126.7 2.9 x 103
135xe 1.18 x 10° 84.6 1.18 x 10°
Nonsaturatingé 20.0 (barns/fission) 10.0 (barns /fission) 43.1 (bams/fission)

fission products
234U

39.7 35.4 121.4
233y (abs) 271.9 24.8 329.1
235y (v x fission) 555.4 37.4 641.8
236y 2.6 17.7 43.5
238y 1.2 9.4 22.9
239py (abs) 1404.0 20.5 1451.3
239py (v x fission)  2411.9 36.7 2496.7
58ped 0.53 0.036 0.61
59¢cod 16.3 2.6 22.25

4Cross section for the reaction 6Li(n, O'.)‘?'H.
bNatural-enrichment boron (19.8% 10B).

®Estimated from L. L. Bennett, Recommended Fission Product Chains for Use in Reactor Evaluation
Studies, ORNL-TM-1658 (Sept. 26, 1966). '

d . . -
Used in flux wire monitoring.

sections given in the final column of this table, we have used reactor theory to calculate

the average ratio of integrated epithermal to themmal flux over the moderated region of the

core. These effective cross sections, when multiplied by the magnitude of the thermal .

flux, give the total reaction rates per atom for neutrons of all energies in the MSRE spec-

trum. Unless otherwise specified, the cross sections listed in Table 5.1 are for (n,y) re-

actions. Those nuclide reactions which are of importance in the interpretation of reactivity

changes during operation, plus two nuclide reactions which have been used for thermal-flux

wire monitoring, are listed in Table 5.1.

Certain high-energy reactions, also of interest in interpreting MSRE operations, are listed

in Table 5.2, These include two reactions which have been used for high-energy flux wire

monitoring, and those reactions which produce isotopic constituents of interest from the stand-

point of fuel-salt chemistry. In general, the reactivity effects of isotopic changes’due to re-
83

Table 5.2. Average Cross Sections and Fluxes for High-Energy Reactions in the MSRE

Cross Section, Averaged

o T Over Energy Spectrum Total Neutron Flux Above EO,
Nuclide Reaction Energy Cutoff Above Eo _ Averaged Over Core Volume
11 -2 1 —1
- (Mev) (bams) (10" " neutrons c¢cm™ “ sec” ~ Mw )
%Be n,2n 1.83 0.234 ' 0.750
°Be n, @ 1,00 _0.0408 1.563
g n, p 4.50 ~0.0250 0.0974
PBg n, a 3.01 0.0988 0.271
54pe® n, p 2.02 0.1644 0.629
58N 2 n, p 1.22 0.1262 \ /  1.288
238y - n, 2n 6.06 0.2490 ©0.0201

a l . . .
Used for flux wire monitoring.

actions in this group are quite small com;;ared with those associated with the reactions listed
in" Table 5.1.

As indicated in Table 5.2, the cross sections are averaged ovér that portion of the high-
energy spectrum of Fig. 5.1 for which the magnitude of the cross section is significant. The
lower cutoff energy listed is that value chosen from the GAM-II group structure which corre-
sponds most closely with the physical cutoff energy for the particular reaction considered.
The upper cutoff energy is 15l Mev for all réactions. The calculated maghitudes of the neutron
fluxes in these énergy intervals, averaged over the volume of the core, are also given in
Table 5.2, -

The effects of isotopic changes due to the reactions considered in Tables 5.1 and 5.2 are

described in the following section.

5.2 1SOTOPIC CHANGES AND ASSOCIATED LONG-TERM REACTIVITY
EFFECTS DURING REACTOR OPERATION ‘

In the MSRE the rate of depletion of the 23°U fuel charge is sufficiently low that some con-
venient first-order approximations can be used in calculating the changes in isotopic composi-
tion of the fuel salt. We have assumed that the magnitude and energy spectrum of the neutron
flux remain constant during operation ét a given power level, and we have used the calculated
effective cross sections listed in Tables 5.1 and 5.2. For all isotopic changes occurring in
the fuel salt, account has to be taken of the ‘‘flux dilution’’ effect of the time the fuel spends
in that part of the circulating system extemal to the core, out of the neutron flux. Thus the
2

calculated volume-average thermal flux for the entire fuel loop is 6.7 x 10'! neutrons cm™

sec™! Mw—!, whereas the average thermal flux over the core region is 2.0 x 10!2 neutrons
84

cm~? sec™! Mw~!. Of the isotopic changes being considered here, only the burnout of the
residual !°B in the graphite is determined by the latter flux level.

Rates of production or depletion of important isotopic constituents of the fuel salt, cal-
culated on the basis outlined above, are listed in Table 5.3. The light-element reactions

explicitly considered for this tabulation are

i \ _
19F(n,p)1°0 ——— ('H production) ,
29 sec -
®Li(n, a)3H "(®Li depletion and 3H production) ,
°Be(n, 0)’He ——— > SLj (°Li production) ,
4 sec
19F(n,a)1N——> 160 (180 production) .
7.4 sec

To a good approximation, all the rates given in Table 5.3 may be assumed constant during a
large fraction of the exposure life of the current fuel charge. For example, after one year’s
operation of the MSRE at 7.5 Mw, the maximum corrections due to saturation effects would
reduce the concentrations of 23°Pu and tritium, calculated according to the linear approxi-

mation, by about 11 and 4% respectively.

Table 5.3. !sotopic Changes in the Fuel Salt

During Reactor Operation

- Rate of Production (+) or Depletion (~)
Nuclide . in Fuel-Salt Circulating System®

(g/Mwd)
lq +1.3x 10°
3H . +4,1%x 1074
51

N Depletion ,—/8.3 x 1074
Producéion - +0.4x 10”4
Net depletion —-7.9x10™*
160 +2.2x 1074,
234y -52x%x10"3
235y —1.30
236y +0.26
238y _ —0.19
23%py +0.19

“Calculations based on fuel salt volume of 70.5 ft°.
85

In addition to the bumup of 235U during power operation, the changes in the concentrations
of 5Li, 234U, 236y, 238y, 239Py, and the nonsaturating fission products, all in the salt, and
the residual !°B in the graphite induce reactivity changes which are important in the analys1s
of the nuclearbehavior of the core. We shall exclude 235U from consideration here, since
separate detailed account is taken of its inventory-reactivity changes in the analysis of op-
erations. Most of the remaining effects given above manifest themselves as very slowly de-
veloping positive reactivity changes, dependent on the time integral of the power or energy
generated, The exceptions are 235U (for which there is a slight increase in concentration
resulting from radiative capture in 235U) and the buildup of nonsaturating fission product
poisons.v . ‘

As a typical example, reactivity effects at 10,000 Mwhr (the approxima.te exposure at the
end of MSRE run 8), corresponding to these isotopic changes, were estimated from reactor
theory. The results are given in Table 5.4. In this gfoup the terms of largest magnitude
were found to arise from the production of 23°Pu and the bumout of the ®Lj initially present
in the fuel salt. In addition to the results listed in Table 5.4, an algebralc formula was de-
veloped for the irradiation dependence of all reactivity effects included in this group. This

is being used in the on-line calculations of reactivity changes during MSRE operation. >

SJ. R. Engel and B. E. Prince, The Reactivity Balance in the MSRE, ORNL-TM (in preparation).

Tu.ble-5.4.. Reactivity Effects Due to Isotopic
Changes in the MSRE"

Approximate Reactivity

Nuclide . Effect at 10,000 Mwhr
‘ (% Ak/k)
6L  0.017
10ge 0.007
234yb _0.001
236y —0.003
- 238y . 0.004
239p, . . 0.051
Nonsaturating -0.005

fission products

Total 0.072

“The boron concentration present initially in the MSRE
graphite was estimated from MSR Program Semiann.
Progr. Rept. July 31, 1964, ORNL-3708, p. 376. The
calculation of the reactivity effect neglects a correction
factor accounting for the spatial dependence of the boron
burmnout in the graphite.

Includes react1v1ty effects of both depletmn of 234y

and associated production of 235U,
86

5.3 ANALYSIS OF TRANSIENT '35Xe POISONING

Studies were continued with the purpose of correlating the time behavior of the 135Xe
poisoning observed in the MSRE with calculations from a theoretical model. The mathematical
model us;ed for this purpose has been previously described in refs. 6 and 7. In the present
section we will give only a qualitative description of the main aspects and assumptions in
the theoretical model. We will then compare graphically the calculated buildup and removal
of 135Xe reacti\}ity, following changes in power level, with some of the experimental re-
activity transients which have been observed from operation to date. Finally, we discuss
some tentative conclusions which can be drawn from the currently available evidence.

The theoretical model is an extension of the steady-state model described in ref. 8, to
include the transient behavior of the 135Xe reactivity following a step change in reactor
power level. In the model chosen, we have assumed that all the 13‘51 produced from fission
remains in circulation with the salt. After decay to '3Xe, the xenon migrates to the acces-
sible pores of the graphite at the boundaries of the fuel channels and also to minute helium
bubbles distributed through the circulating salt stream. An effective mass transfer coefficient
was used to describe the transfer of xenon from solution in the circulating salt to the interface
between the liquid and the graphite pores at the channel boundaries. Equilibrium Henry’s law
coefficients were used for the mass transfer of xenon between the liquid phase at the interface
and the gas phase in the graphite pores. The numerical value used for the mass transfer co-

, efficient between the circulating salt and the graphite was based on krypton injection experi-
ments with flush salt circulating in the fuel loop, performed prior to nuclear operation of the
MSRE. ® |

Similar considerations were assumed to apply to the mass transfer of xenon from liquid
solution to the gas bubbles. The coefficient of mass transfer from the liquid to a small gas
bubble, of the order of 0.010 in. in diameter, moving through the main part of a fuel channel,
was estimated from theoretical mass transfer correlations. The equilibrium !3%Xe poisoning
was found to be relatively insensitive to the bubble diameter and mass transfer coefficient,
over a reasonable range of uncértainty for these parameters.’

The computational model provides for different efficiencies of removal by the extemal
stripping apparatus of xenon dissolved in the salt and that contained in the gas bubbles.
The efficiency of removal (fraction of xenon removed per unit circulatedvthrough the spray -
ring) of xenon dissolved in the salt was estimated to be between 10 and 15%, based on some

early mockup experiments to evaluate the performance of the xenon removal apparatus.

®MSR Program Semiann. Progr. Rept. Feb. 28, 1966, ORNL-3936, pp. 82—87.
"MSR Program Semiann. Progr. Rept. Aug. 31, 1966, ORNL-4037, pp. 13-21.

5Rr. J. Kedl and H. Houtzeel, Development of a Model for Computing 1335xe Migration in the MSRE,
ORNL-4069 (in preparation).
87

Calculations of the steady-state '3°Xe poisoning, reported in ref. 7, indicated that the
low apparent poisoning as a function of power level could be described by a variety of com-
binations of circulating bubble volume fractions and bubble stripping efficiency. One of the
purposes of the transient calculations, therefore, was to attempt to separate those parameter
effects which could not be separated in the steady-state calculations.

A number of transient reactivity curves were calculated, with the aid of an IBM 7090 pro-
gram, based on the theoretical model described above. These calculationis were compared with
experimental data logged by Ithe reactor’s BR 340 on-line computer. The apparent transient
135Xe poisoning was determined by subtracting all other known power-dependent reactivity
effects from the reactivity change represented by movement of the regulating rod following a

step change in the power level.

In Figs. 5.3-5.13 we have compared some of the transient reactivity curves obtained from
this analysis with some experimental transients, in the chronological order in which they were
obtained. In each of these figures the solid curves represent the calculated behavior, and the
plotted points show the observed experimental reactivity effect. At this date, only a few rela-
~ tively clean experimental transients corresponding to step changes in power level (for which
the 7090 program was devised) have been obtained. However, several characteristics of the
135Xe behavior are indicated from these curves. These will be discussed by considering the
figures in order.

Figure 5.3 shows the calculated and observed xenon transients for a step increase in re-
actor power from 0 to 7.2 Mw. The calculations (solid curves) were made for a variety of -
circulating void fractions (ab) to show the effect of this parameter on the xenon poisoning,

A single bubble stripping efficiency (¢,) of 10% was used for this figure. This relatively
low efficiency is approximately equal to the efficiency estimated for the stripping of xenon
dissolved in the salt; it was considered to be a lower limit and a reasonable first approximé—
tion for this parameter, in the absence of strohg evidence for assuming a higher value for the
bubbles. The effectiveness of the circulating gas in reducing the poison level is due to the
combined effects of the large overall gas-liquid surface area for mass transfer to the bubbles
and of the large xenon storage capability of the bubbles (because of the extreme insolubility
of xenon'in molten salt). Thus the bubbles compete effectively with the graphite for removal
of xenon from the liquid, and xenon in the circulating fluid is a less effective poison than that
in the graphite because about two-thirds of the fluid is outside the core at any instant. The
plotted points represent the observed 13%Xe reactivity transient at the beginning of run 7
(July 1, 1966). The data indicate that the low apparent xenon poisoning can be explained by
a large void fraction (between 0.5 and 1.0 \;01 %) and a low bubble stripping efficiency. Note,

however, that the transient buildup is not closely fitted by these parameter values.

In Fig. 5.4 the curves indicate the calculated effect of increasing the bubble stripping
efficiency for a fixed, relatively small (0.1 vol %) circulating void fraction. The plotted

points are for the same reactor xenon transient shown in Fig. 5.3. A comparison of Figs.
88

5.3 and 5.4 'shows that the steady-state xenon poisoning is desc.ribed as well by a low void
fraction with a high bubble stripping efficiency as it is by a high void fraction with a low.
stripping efficiency. However, the shape of the experimental transient poison buildup is de-
scribed more closely by the parameter values in Fig. 5.4. .

Figures 5.5 and 5.6 show the calculated and observed transient buildup of '*5Xe poisoning
after a step increase in power from 0 to 5.7 Mw in run 8 (October 1966). The ranges of values
of a, and €, used in these calculations are the same as those used in Figs. 5.3 and 5.4. Again,
one finds that the shape of the observed transient is matched more closely by the calculations
which assume a low void fraction and a high bubble stripping efficiency.

Figure 5.7 shows the calculated and observed !3°Xe realctivity transients for a power re-
duction from 5.7 Mw to 0, with the !35Xe initially at equilibrium. Only the experimental data
for the early part of the transient are shown in this figure, since the reactokr was made sub-
critical before the complete xenon transient could be recorded. However, the calculated
curves reveal an important characteristic of the transient xenon behavior, which is due to
variations in the overall xenon distribution that result from the choice of values for a, and
€,. If the circulating void fraction is low, most of the poisoning effect is due to xenon in the
graphite, and only a small amount of xenon is in the circulating fluid. Xenon that is produced
in the fluid from iodine decay continues to migrate to the graphite for a period of time after the
power (and, hence, the burnout rate in the neutron flux) has been reduced. This produces a shut-
down peak in the xenon poisoning. Eventually, the stripping process reduces the xenon con-
centration in the fluid so that some of the xenon in the graphite can escape and be stripped out.
- This results in a more rapid decrease in xenon poisoning than simple radioactive decay.' As
the circulating void fraction is increased, a larger fraction of the xenon inventory (or poisoning)
is associated with the bubbles, and there is less xenon migration to the graphite. In this case
-the shutdown peak tends to disappear. This characteristic makes the shutdown transients some-
what more sensitive to the values of the bubble parameters and, potentially, more useful in the

analysis of the' xenon behavior. ‘
| For the same experimental decay transient as that plotted in Fig. 5.7, Figs. 5.8 and 5.9 show
the effect of increasing the bubble stripping efficiency, with the circulating void fraction held
fixed at two representative low values (0.10 and 0.15 vol % tespectively). Although the section
of experimental data for this transient is too limited to allow a valid comparison to be made, it
is again seen that the high bubble stripping efficiencies and low circulating void fractions also
provide reasonable representation of the observed data. The lack of any apparent shutdown peak
in the experimental data, however, seems to suggest that a larger amount of gas may have been
in circulation at the beginning of the shutdown (termination of run 8) than was apparent at the
beginning of the run (Fig. 5.6).

A second '33Xe stripping out decay transient, somewhat longer than the preceding, was ob-
served during run 9 (November 1966), following reduction in the powér level from 7.4 Mw to 0.
This transient is shown in Figs. 5.10 and 5.11, where the calculated curves are again based on

the assumption of relatively high bubble stripping efficiencies and low circulating void fractions.
89

In this case a slight rise in the apparent poisoning following shutdown was indeed observed.
From comparisons of the results given in these two figures, it appears that, under nomal op-
-erating conditions, o, and €, might be bracketed between 0;1 and 0.15 vol %, and 50 to 100%
~ respectively. ( ‘

Finally, in Figs. 5.12 and 5.13, we show the most recent shutdown transient obtained at the
termination of run 10 (Jan. 14, 1967). In this case the apparent !3°Xe reactivity transient was
recorded for more than 40 hr after the reduction in power level. The results are also in good
accord with the conclusions indicated above.

Although substantial progress has been made in interpreting the xenon behavior in the MSRE,
_the experimental data which have thus far been accumulated for the transient behavior of the
135¥e poisoning are as yet insufficient for any final conclusions to be drawn concerning the
‘“‘best’’ values of the circulating void fraction and bubble stripping efficiency. As one ex-
ample, it should be noted that, if gas bubbles are continuously being ingested into the main
circulating stream as the evidence indicates, the volume of gas in circulation is probably not
constant, but rather is a élowly varying quantity depending on the level of the liquid in the
fuel-pump tank and th’e transfer rate of salt to the overflow tank. This dependence is as yet
not well understood, and future operation is expected to shed further light in this area.

Other refinements of the model for the xenon behavior may b_e required as reactor operating
data are accumulated. Theserefinements are not expected to strongly affect the conclusions

indicated in the preceding description.

CRNL-DWG 67-1077

4 ’ ap, VOLUME
PERCENT
o e EXPERIMENTAL DATA, OBTAINED CIRCULAEING
' DURING MSRE R™™N NO.7
/ ©
1.0
0.8 C.05

iy

g
7 /’
/

C.4

REACTIVITY MAGNITUDE (% 5k/4)

0.2

4 28 32 36 40
TIME AFTER INCREASE IN REACTOR POWER LEVEL (hr)

Fig. 5.3. Effect of Yolume of Circulating Gas on Transient Buildup of 135%¢
Reactivity. Step increase in power level from 0 to 7.2 Mw; bubble stripping ef-

ficiency, 10%.
REACTIVITY MAGNITUDE (% 34/k)

REACTIVITY MAGNITUDE (%Sk/k)

90

ORNL-DWG 67-1078

0.7 - T T
I l l I I &, , BUBBLE STRIPPING
® EXPERIMENTAL DATA, OBTAINED EFFICIENCY (%)
0.6 R .7
DURING MSRE RUN NO P
0.5 —
I
/ /_____é-é--
0.4 // /,__/
/] / 50
0.3 S
/ ——
| "]
X ' L] ®s0 (0 ®
0.2 ) e . & ———T"00
o _—
/ A
/ /.‘./
0.1 74 /-/
/ 28
o
o L&&° .
0 4 8 12 16 20 24 28 32 36 40

TIME AFTER INCREASE IN REACTOR POWER LEVEL (hr)

Fig. 5.4, Effect of Bubble Stripping Efficiency on Transient Buildup of
135y, Reactivity, Step increase in power level from 0 to 7.2 Mw; vol % circu-

lating bubbles, 0.10.

ORNL-DWG 67—1079

0.7
0.6 [ & EXPERIMENTAL DATA, OBTAINED
: DURING MSRE RUN NO.8 ap, . VOLUME PERCENT
. CIRCULATING BUBBLES | oot
-—-‘——_-
0.5 = 040"

0.4 /

o3 / L oosT

’ / /,

/ — | ~0.50

(0] 2 / _‘.—-—'—'___- ‘s o [ ] d Py .

. / P /’ o ] '{OO

///__-o-.-'l—’-.-.-‘ ——
/i .
(oK - / //T-“ P
N °*
e * ¢
0
0 4q 8 12 16 20 24 28 32 36 40

TIME AFTER INCREASE IN REACTOR POWER LEVEL (hr)

Fig. 5.5. Effect of Yolume of Circulating Gas on Transient Buildup of 135x¢

Reactivity. Step increase in power level from 0 to 5.7 Mw; bubble stripping effi-

ciency, 10%.
REACTIVITY MAGN&TUDE (T 84/k)
( o

REACTIVITY MAGNITUDE (% S4/4)

91

ORNL —-DWG 67—-1080

07 1
0.6 ¢,» BUBBLE STRIPPING ———
EFFICIENCY (%)
e EXPERIMENTAL DATA, OBTAINED 1o ="
0.5 DURING MSRE RUN NO.8 P ot

7
20
/ /——_
/ /
C.3 -
: Iy . . . . o
0.2 / 2 ¥ 23 —100
/ ] A—r.-.-fa"'fi"“
0.1 / / e L 2
. / / .
-.‘ib ° ¢
fi’-.‘.
O .
0 4 8 12 16 20 24 28 32 36 40

TIME AFTER INCREASE IN REACTOR POWER LEVEL (hr}

Fig. 5.6. Effect of Bubble Stripping Efficiency on Transient Buildup of
135
Xe

Reactivity. Step increase in power level from 0 to 5.7 Mw; vol % circu-

lating bubbles, 0.10.

ORNL-DWG 67-1081

' EXPERIMENTAL. DATA, OBSERVED
DURING MSRE RUN NO. 8

0.5 | . AN

a,, VOLUME PERCENT —
\ \cmcuu\flNG BUBBLES

'-——.'-.-0-.* \ |
P00 - 0.0
0.2 . \-& \ . .
‘ | \ \N o5 \
. \ R
7 EO.SO
0
0 4 -8 12 16 20 24 28 - 32 36 40

TIME AFTER DECREASE IN REACTOR POWER LEVEL ({hr)

Fig. 5.7. Effect of Volume of Circulating Gas on Transient Decay of 135xe

Reactivity. Step decrease in power level from 5.7 Mw to 0; bubble stripping effi-

ciency, 10%.
ORNL—DWG 67-—14082

0.7

e EXPERIMENTAL DATA, OBTAINED
DURING MSRE RUN NO. 8

\
)

o
3

| \\
\\
\

)
/

BUBBLE STRIPPING

REACTIVITY MAGNITUDE (%o84/k)
[®]
D

Eb'
¢ eolee \ EFFICIENCY (%)
——| \
02 T 0. >\\ \\\1o~..
l\\\ \20 \
g
0.1 ; \--.._\\ 50 |~
S
0 : :
0 4 8 12 16 20 24 28 32 36 40

TIME AFTER DECREASE IN REACTOR POWER LEVEL (hr)

Fig. 5.8. Effect of Bubble Stripping Efficiency on Transient Decay of 135ye

Reactivity. Step decrease in power level from 5.7 Mw to 0; vol % circulating bub-

bles, 0.10.

A

ORNL-DWG 67 -1083
06 .

0.5 / AT

g e EXPERIMENTAL DATA, OBTAINED
< DURING MSRE RUN NO.8
2 04 ™
& N
Z 03 ~ ‘
2 \ .
= ? ®ececoe \ .
> '“_‘:JLLA. \ \ &, BUBBLE STRIPPING
= 0.2 L . EFFICIENCY (%)
5 —Z. L
= P ————
‘g L \\"‘\ \ : 10\
I&J o \-‘_\\ 20\\
] 150(2)____----____
o |
0 a 8 §2 16 20 24 28 32 36 . 40

TIME AFTER DECREASE IN REACTOR POWER LEVEL (hr)

Fig. 5.9. Effect of Bubble Stripping Efficiency on Transient Decay of 135y,

Reactivity. Step decrease in power level from 5.7 Mw to 0; vol % circulating bub-

bles, 0.15.
ORNL-DWG 67-1084
1 1 1

Y
iy

® EXPERIMENTAL DATA, OBTAINED DURING

/ \\ MSRE RUN NO. 9
0.6 \\
I 7 B AN AN \
Y -\\ \
~.
~
P

T T~

' €,, BUBBLE
STRIPPING

REACTIVITY MAGNITUDE (% 84/4)

™

N

AN
N
™~
N
S~
~J
\

.3 — EFFICENCY |
) e [ ] o o 4 A
0.2 . = ® =~ \10\
P
' 20
® \ \\
\\
0
0 q 8 12 {6 20 24 28 32 36 40

TIME AFTER DECREASE IN REACTOR POWER LEVEL (hr)
Fig. 5.10. Effect of Bubble Stripping Efficiency on Transient Decay of
135)(e Reactivity. Step decrease in power level from 7.4 Mw to 0; vol % circu-

lating bubbles, 0.10.

ORNL DWG 67-1085

07
- 06 /’—-“\ _ -
~g
N \ .
- 05 N e EXPERIMENTAL DATA, OBTAINED
3 \ DURING MSRE RUN NO. 9
(71 ]
2 ~ ~ T
, § \ \ _
= 03 : \\ ¢, sBUBBLE STRIPPING
> LB
SR S ol N : \ EFFICIENCY (%)
— .";-._ \ N
o S e o
\\
0.1 -
\\___
100 T
0
0 4 8 12 16 20 24 28 32 36 40

TIME AFTER D.ECREASE IN REACTOR POWER LEVEL (hr)

Fig. 5.11. Effect of Bubble Stripping Efficiency on Transient Decay of
135)(e Reactivity. Step Decrease in power level from 7.4 Mw to 0; vol % circu-
lating bubbles, 0.15.
ORNL-DWG 67-1086

0.8
/
0.7 ™S ..
/ \ o EXPERIMENTAL DATA,OBTAINED
\ DURING MSRE RUN NO. {0
x 0.6
3 N\
> ) \
w 0.5 | =~
Q \
jum
=
4 N
é ’ \ N
fi
- T~ N | ¢,» BUBBLE STRIPPING
= . \ . EFFICIENCY (070)
- [ L Y \ 10
Q * ] . » [ J P \ \
I?J ™ |\ i \
x 0.2 L B ™~ 20
’ AR \\ N
* “.\ 50 \\
® ¢ T~100
o L M . .-..,__'___ — \
' M e
®
0
o 4 8 12 16 20 24 28 32 36 40

135

TIME AFTER DECREASE IN REACTOR POWER LEVEL (hr)” "

Fig. 5.12. Effect of Bubble Stripping Efficiency on Transient Decay of |

Xe Reactivity. Step decrease in power level from 7.4 Mw to 0; vo!l % circu-
lating bubbles, 0.10.

ORNL-DWG- 67-1087

0.7
< 06 A ® EXPERIMENTAL DATA ,OBTAINED — |
= / \ DURING MSRE RUN NO. 10
@ ,
£ os AN
W \
=
5 04 //__‘.\'"“‘ \
2 ' \ N .
> \ V .BUBBLE STRIPPING
S 03 N EFFICIENCY (%)
5 IS i Bl T \
< e L 2 - \ N0
T 02 S e
. / '*....‘ \ ~
\ [ ) Iy \
(o} ) !:n,..__‘:}(g L
100 e
oL |
0 4 8 12 16 20 - 24 28 32 36 40

TIME AFTER DECREASE IN REACTOR POWER LEVEL (hr)

Fig. 5.13. Effect of Bubble Stripping Efficiency on Transient Decay of
]35Xe Reactivity. Step decrease in power leve! from 7.4 Mw to 0; vol % circu-

lating bubbles, 0.15.
Part 2. Materials Studies

6. Molten-Salt Reactor Program Materials

6.1 MSRE SURVEILLANCE PROGRAM — HASTELLOY N
W. H. Cook H. E. McCoy

Several stringers of Hastelloy N test specimens were located in an axial position near the
center line of the MSRE for sur;/eillance purposes. These specimens were from heats 5085
(cy‘lindrical v_esrsel) and 5081 (miscellaneous parts)., They were removed after 4800 hr at 645 +
10°C, during which the reactor had operated 7612 Mwhr. The peak thermal dose was 1.3 x 10%°
nvt, and the peak fast dose (>1.2 Mev) was 3 x 10'° nvt,! The peak-to-minimum thermal flux
over the length of the test specimen array was a factor of 5, but this was not found to be a signif-
icant variable in this evaluation. The details of the removal of the specimens have been de-
scribed previously,? but we want to mention again that there was some bowing at various points
along the surveillance assembly. We used an optical comparator to eliminate those specimens
bowed by more than 0.001 in. over the gage length.

The surveillance control specimens were exposed to MSRE-type fuel salt and duplicated the
thermal history of the in-reactor specimens. They were also bowed, and the ones bowed By more

than 0,001 in, were not used.
‘ Samples of both reactor and control specimens were metallographically examined. While the
réactor samples had structures that were dirtier than the controls, no major .changes in structure
were found. No evidence of attack or deposition was found within the gage lefigths' of'any of the
samples. However, all test specimens from heat 5085 (both irradiated and control) were found to
have a surface layer on the 14 -in.-diam portion of the samples on the sides which were in con-
tact or near contact with the graphite. Such layers were not found on the smaller l/s-in.-diar'n.

gage length of the specimens. Figure 6.1 shows the nature of the surface layer on these samples.

1 private communication with H. B. Piper. .
2MSR Program Semiann. Progr. Rept, Aug. 31, 1966, ORNL-4037, p. 97.

—

95
= =

V.00 INLHES
T5500%

Fig. 6.
loy N in near contact with graphite.

Hastelloy N Surface from Exposed MSRE Surveillance Samples. Surface deposit from Hastel-

Note also the extensive carbide precipitation that has occurred along the grain and twin bound-
aries during the exposure. A similar surface reaction layer was noted on samples from heat 5081
but did not occur as frequently. The 20-mil-thick Hastelloy N straps (heat 5055) that bound the
graphite specimens also had such a reaction product on the sides that contacted the graphite.
These straps showed limited ductility when bent at room temperature. They tended to break
intergranularly in the reaction region when bent sharply. Although we have not made a positive
identification of the surface layer, we assume that it is a carbide produced by a reaction be-
tween Hastelloy N and graphite.

Both tensile and creep-rupture tests have been conducted on the surveillance specimens.
The creep-rupture testing is not yet complete, so only the results of the tensile tests will be
reported.

! is shown as a

The total elongation at fracture when deformed at a strain rate of 0.05 min™
function of temperature in Fig. 6.2. Both heats show some reduction in ductility in the irradiated
condition. Heat 5085 exhibited a slight reduction in ductility when tested unirradiated at room

temperature, while the irradiated specimen of heat 5081 exhibited the same effect. With the ex-
L\

97

ception of these values, the ductilities remained essentially constant up to temperatures of
500°C. At temperatures above 500°C, the ductility of the irradiated and the control material
decreases with increasing temperature, with the irfadiatéd material showing a greater loss in
ductility. At temperatures above 650 to 700°C, the control material exhibits improved ductility,
whereas the ductility of the irradiated material continues to decrease.

Figure 6.3 compares the properties of the irradiated and control specimens at a lower strain
rate, 0.002 min~—'. Qualitatively, the behavior is very similar to that noted at a strain rate of
0.05 min~!. However, the loss in ductility is magnified at the lower strain rate. The variations
in the properties of the two heats of material have been reduced to where the differences are
minimal. ) ] \

Figures 6.4 and 6.5 show how the ratio of the irradiated to the unirradiated tensile property
varies for heats 5081 and 5085, respectively, as a function of temperature. For both heats, the
yield strength is unchanged by irradiation. The ultimate tensile stress is reduced about 8% up
to a temperature of about 500°C, where the reduction becomes considerably greater. The reduc-
tion in ductility at room temperature and at elevated temperatures is also clearly demonstrated.

We have compared the irradiated and control surveillance specimens. Let us now look at how
the properties of the control specimens have changed during their 4800 hr of thermal exposure to

molten salt. Table 6.1 shows representative properties of heat 5085 after several different heat

treatments. The first group of tests was run with the material in the as-received condition (mill

annealed 1 hr at 1177°C). Annealing for 600 hr at 650°C had no appreciable effects on the

properties. The MSRE surveillance specimens were given a 2-hr anneal at 900°C before insertion

ORNL-DWG 67-2452A
70

60

w

o

\|.
>

9 A PSS = ‘ \\‘\ /
[l ‘_,:/ \ \ ] /
z - . 4
E 40 =I/ \ y Vi £
40 8 —— AT
% —_— . . \ e
o [ ] \ Al\ ‘//
d \ z|\\ -
30 N \ 7
. VAN
Z CONTROL . IRRADIATED N
o N
o 8 o 5081 N\
s o 5085 Y 1
¢=0.05 min™! \\“
~
10 \\\,,\
0
0 100 200 300 400 500 600 70O 800 900 000

TEST TEMPERATURE (°C)

Fig. 6.2. Comparative Tensile Ductilities of MSRE Surveillance Specimens

and Their Controls ot a Strain Rate of 0.05 min-l.
98

ORNL-DWG 67-2453A

50

P, -
>

Y
o
/

\
o \ "1"
\\ , / .
CONTROL IRRADIATED \\ //
l 1Y A
s o 508 N = '
A e 5085 \ * _.-/

&= 0002 min' , ‘\’\M

-

N
(@]

TOTAL ELONGATION (%)
\ w
o)

N
. [
10 AN
\;\g\____;o_
0
0 {00 200 300 400 500 e00 700 800 ‘ 900 100C

TEST TEMPERATURE (°C)

Fig. 6.3. tomparative Tensile Ductilities of MSRE Surveillance Specimens

and Their Controls at a Strain Rate of 0.002 min-I.

ORNL-DWG 67-2458A

1.2
1 O s s m—_fy mm & [ ]
—T T AT
e . S S~gT —4,
> I>—- e - NA Sa
ElE 0.8 g \‘\ b,
&S ' \ ||\" --..__:
ola Ve
la ] \ A
o ’ s \
o
olu 0.6 []
== :
alz \
al2 e YIELD STRESS \
EE oa A ULTIMATE STRESS M
> ; w TOTAL ELONGATION - \
&=0.05 min | N
0.2 :\
N\
~
~N
-
O S
0 100 200 300 400 500 600 700 BOO 900 1000

TEST TEMPERATURE (°C)

Fig. 6.4. Comparative Tensile Properties of lrradiated and Unirradiated
MSRE Surveillance Specimens, Heat 5081.

into the reactor, and the properties of the material in this state are also indicated in Table 6.1.
The properties at 25°C were unchanged, and slight reductions in tensile strength and ductility
were noted at a test temperature of 650°C. After exposure to salt for 4800 hr at 650°C, there

is a further reduction of 10% in the tensile strength and the ductility’. Table 6.2 shows that heat
5081 underwent similar changes at elevated temperatures but did not suffer the reduction in

ductility at 25°C noted for heat 5085.
]

‘.

99

ORNL—DWG €7-2459A

1.2

-4—

0.8 = —y=———]

- iy NP B

06

IRRADIATED PROPERTY
UNIRRADIATED PROPERTY

04

e YIELD STRESS
4 ULTIMATE STRESS
B TOTAL ELONGATION

é=0.05 min '

0.2

0 100 200 300

400

500

600 700

TEST TEMPERATURE (°C)

800

900

1000

Fig. 6.5. Comparative Tensile Properties of Irradiated and Unirradiated
MSRE Surveillance Specimens, Heat 5085,

Table 6.1. Tensile Properties of Hastelloy N — Heat 5085

Test Strain Yield Ultimate Uniform Total Reduction

Specimen Heat Treatment Temp Rate Stress Stress Elongation Elongation in Area
No. °C) (min~!) (psi)  (psi) (%) @) %)
.76 As received 25 0.05 52,200 116,400 51.3 52.5 ' 56.3
77 As received 427 0.05 30,700 102,900 57.6 59.4 49.9
284 As received 600 0.02 32,200 85,100  46.6 47.6 40.4
78 As received 650 0.05 28,700 80,700 35.5 36.7 33.2
285 As received 650 0.02 31,900 65,000 25.8 31.8 27.2
283 As received 650 0.002 30,500 64,300 22.6 24.1 28.8
79 As received 760 0.05 32,100 61,500 24.7 27.0 31.9
80 As received 871  0.05 30,700, 42,300 9.0 31.8 33.4
81 As received 882 0.05 23,100 23,100 1.8 40,2 43.9
1339 A.R. + 600 hr at 650°C 25 0.05 41,100 115,700 52.5 52.8 41.7
. 1340 A.R. + 600 hr at 650°C 650 0.05 33,800 76,000 36.8 37.5 36.4
4295 A.R. + 2 hr at 900°C 25 0.05 41,500 120,800 52.3 53.1 42.2
4298 A.R. + 2 hr at 900°C 500 0.05 32,600 94,800 51.2 54.1 40,9
4299 A.R. + 2 hr at 900°C 500 0.002 33,500 100,200 52.0 53.3 41.7
4296 A.R. + 2 hr at 900°C 650 0.05 29,600 75,800 31.7 33.7 34.6
FC-3 , 25 0.05 45,500 111,200 46.8 . 46.8 31.5
DC-19 A.R. + 2 hr at 900°C + 500 0.05 33,600 94,300 48.8 . 49.3 41.5
DC-26 4800 hr in 500 0.002 33,400 91,700  43.3  44.3 37.8
DC-14 MSRE salt at 650°C 650 0.05 31,800 70,100  25.8 26.8 30.0
DC-25 650 0.002 31,500 62,500 22.8 24.3 27.2

100

Table 6.2, Tensile Properties of Hastelloy N — Heat 5081

-

. Test Strain Yield Ultimate Uniform Total "Reduction
Spe:;men Heat Treatment Temp Rate Stress Stress Elongation Elongation in Area

o (Cc) (min~!) (psi)  (psi) (%) (%) (%)
4300 A.R. + 2 hr at 900°C 25  0.05 52,600 125,300 56.7 59.5 50.5
4303 A.R. +2 hrat 900°C - 500 0.05 32,000 100,300 57.8 60.7 44.4
4301 A.R. + 2 hr at 900°C 650 0.05 32,200 81,800 31.7 33.9 29.9
4302 A.R. + 2 hr at 900°C 650 0.002 32,900 74,900  29.0 29.5 31.8
AC-8 25  0.05 47,700 118,700  55.9 57.6 48.8
AC-19 A.R. + 2 hr at 900°C 500 0.05 35,800 97,800 53.6 56.6 46.2
BC-9 + 4800 hr in 500 0.002 36,200 95,300 46.2 47.0 38.1
AC-27  |MSRE saltat 650°C | 650 0.05 32,400 68,400  23.8 24.6 23.1
AC-17 - 650 0.002 33,600 66,700 22.8 23.2 21.6

Some of the test specimens have been e};amined metallographically. Figure 6.6 shows the
microstructure of control specimen No. AC-8 from heat 5081, which was tested at 25°C. This
specimen exhibited good ductility, and this is reflected in the intragranular, shear-type fracture
and the lack of intergranular cracking. Figure 6.7 shows the irradiated specimen from the same
heat that was tested at 25°C. The fracture is largely intergranular, and there are numerous
intergrarular cracks. The control and irradiatéd specimens from heat 5085 tested at 25°C ex-
hibited the same geheral characteristics as those illustrated in Fig. 6.7. The other specimens
examined were tested at 650°C, The failures were all intergranular, with the irradiated speci-
mens exhibiting the typical characteristics of a high-temperature, low-ductility, intergranular
fracture. ' |

We compared the ductilities of the surveillance specimens with those for specimens irradiated
in other -experiments without salt present. Heat 5081 had been irradiated previously in the ORR.
The ORR experiment was run at 700°C to a thermal dose of 9 x 10*° nvt, and the matetial was
in the as-received condition. The MSRE surveillar-lcer specimens weré run at 650°C to a thermal
dose of 1.3 x 1029 nvt, and the preirradiation anneal was different. However, none of these dif-
ferences are thought to be particularly significant, and the results compare rather well. Figure
6.8 shows that the postirradiation ductilities of heat 5081 after both experiments are very similar.

The most important question to be answered concerning these data is how they apply to the
operation of the MSRE.

The surveillance specimens were exposed to a thermal dose of 1.3 x 10?°% nvt. (The MSRE
vessel will reach this dose after about 150,000 Mwhr of operation.) This burned out about 30%
of the '°B and produced a helium content of about 10~ 3 atom fraction in both heats. The high-
temperature tensile properties are exactly what we would expect for this dose. Our in<ell and
postirradiation creép tests, on several MSRE heats, indicate that the .creep ductility will be in

the range of 0.5 to 3%, with the higher values being noted for lower stresses (or longer rupture

3W. R. Martin and J. R. Weir, Nucl. Appl. 1/2, 160 (April 1965).
101

S INCHES

T> 100X

0.0

Fig. 6.6. Microstructure of Hastelloy N (Heat 5081) Exposed to Salt for 4800 hr at 650°C. Tested at
25°C at a strain rate of 0.05 min™ | (specimen No. AC-8). Etchant: glyceria regia. (a) Fracture. (b) Edge

about Y in. from fracture.

102

0.035 INCHES
> 100X

o

T 100X

T

Fig. 6.7. Microstructure of Hastelloy N (Heat 5081) Exposed for 4800 hr in the MSRE at 650°C. Tested
at 25°C ot a strain rate of 0.05 min™! (specimen No. D-16). Etchant: glyceria regia. (a) Fracture. (b) Edge

about Y in. from fracture.
L)

103

ORNL—DOWG 67—4779

1.0
[ ]
<>\
0.8 l

o ¢ = 0.002 min~"

g &2

5

= N\ : O THIS STUDY, b, =1.3X10%° nvt, T=650°C
2 06 ® ORNL —TM ~ (005, ¢ _=9x102% avf, —]
r 7= 700°C

w

-

(=

5

= ]
ol®
W= ok
Lo B8
alz
o|d
sle
x|z ®

s }
— 0.2 \

\ |
‘N\”
-0

400 - 500 600 700 800 900 ' {000
T TEST. TEMPERATURE (°C)

Fig. 6.8. Comparative Effects of Irradiation in MSRE and ORR on the Duc.
tility of Hastelloy N, Heat 5081,

lives).* We are reasonably confident that our tests on the .surveillance specimens will indicate
a similar behavior. Our work also seems to indicate a saturation in the degree of radiation
damage at a helium atom fraction of about 107, and we feel that the properties of the material
will not deteriorate further. The low-temperature ductility reduction was not expected. It is
probably a result of grain-boundary precipitates forming due to the long thermal exposure. Ir-
radiation play;s some role in this process that is yet undefined. The low-temperature properties
are not ‘‘brittle’’ by any standards but will be monitored closely when future sets of surveil-

lance specimens are removed.

6.2 'MECHANlCA_L PROPERTIES OF HASTELLOY N
H. E. McCoy | -

Since the surveillance assembly holder was deformed, it was not possible to remove only
part of the specimens and return the remainder to the‘MSRE; therefore, all 162 specimens were
removed. The number of specimens required for actual surveillance purposes was rather small,
and we used the others for learning‘ more about the general radiation damage éharacteristics of

Hastelloy N. These data are included along with other tests in the following discussion.

‘*MSR Program Semiann. Progr. Rept. Aug. 31, 1966, ORNL.-4037, pp. 118 and 121.
104

Table 6.3 shows the postirradiation tensile properties of several heats of Hastelloy N. Heat
5085 shows pr0pérties that are very similar to those observed for the surveillance specimens.
Using a test condition of 650°C and a strain rate of 0.002 min~ ', heat 5065 has a total elonga-
tion of 10 to 14% if irradiated cold and about 6% if irradiated hot. This is slightly less than
observed for heats 5085 and 5067. Note that heat 5085 is the only heat irradiated at an elevated
temperature that shows a significant reduction in ductility at 25°C.

Several of the si)ecimens from heat 5081 were tested at various strain rates and temperatures
to determine how the properties varied with these parameters. The variation of the total elonga-
tion with strain rate for the control specimens is shown in Fig. 6-9; The ductilities at 400,

500, 760, and 850°C are reasonably independent of strain rate. At 650°C, the ductility is very
dependent on strain rate down to rates of about 0.05 min~!, Figure 6.10 shows that the ductility
of the irradiated material is only slightly strain-rate-dependent at 400°C. At 500°C (Fig. 6.11),
the ductility decreases markedly with decreasing strain rates below values of about 0.05 min~1,
This probably corresponds to the transition from transgranular to intergranular fracture. At
650°C (Fig. 6.12), the ductility depends very heavily upon strain rate, showing a rapid decrease

at rates of 2 min~! and not having reached a constant value at a strain rate of 0.002 min~—!. At

760°C (Fig. 6.13), the ductility is a strong function of strain rate down to rates of 0.005-min~1!,
where the dependence on strain rate becomes much less. At 850°C (Fig. 6.14), the ductility is
strongly dependent on strain rate down to rates below 0.05 min™!, where it beco’mes fairly con-
stant. This family of curves will allow one to predict the strain-rate sensitivity of the ductility

of Hastelloy N at a given temperature.

ORNL DWG 67-2455A

70
60 -
__a300°Ce 3 A s A
< ~500°C ! ! :
2 50 yam po =T
— N S
pra §850°C, : V P ,,” /
9 I _L I z"f
:: 40 (— 0760°Co : [e) ,i
) e
% ' : //
o >0 P
. LA .
1 w650°Ca | 8 HEAT 5081
= . INERl
o 20 L TEST TEMPERATURE
e 400°C ®650°C
10 A 500°C © 760°C
v 850°C
i U 11
103 2 5 402 2 5 0! 2 5 10° 2 5 10"

STRAIN RATE (min~ ")

Fig. 6.9. Influence of Strain Rate on the Ductility of MSRE Surveillance
Control Specimens, Heat 5081.
Table 6.3. Postirrodiation Tensile Properties of Several Heats of MSRE Hastelloy N After Irradiation in a Helium Environment

Boron Irradiation nvt, Test Yield Ultimate Uniform Total Reduction
Heat Condition Level Experiment Temp Thermal Temp E Stress Stress Elongation Elongation in Area - Specimen
No. (ppm) No. ©c) Dose  (°C) (psi)  (psi) %) %) ) No..
x 1020

5065 A.R. 20 ORR-149 43 8.5 - 25 0.05 102,900 135,100 31.5 35.5 52.0 570
5065 AR. 20, ORR-149 43 8.5 200 0.05 82,300 119,600 36.1 39.2 54.8 571
5065 A.R 20 ORR-149 43 8.5 650 0.05 44,600 76,900 26.3 27.3 24.5 572
5065 A.R. 20 ORR-149 43 8.5 650 0.002 40,800 57,400 12.2 13.1 21.9 574
5065 A.R. 20 ORR-149 43 8.5 871 0.05 33,500 36,400 1.7 1.8 3.25 573
5065 A.R. 20 ORR-149 43 8.5 871 0.002 23,200 23,400 0.9 1.8 6.07 575
5065 20 ORR-149 43 8.5 25 0.05 101,500 132,200 30.0 33.6 55.2 581
5065 8 hr 20 ORR-149 43 8.5 200 0.05 78,600 118,300 37.3 39.8 45.9 582
5065 > at 20 ORR-149 43 8.5 650 0.05 35,100 77,700 25.3 25.8 22.6 583
5065 8710Cfi 20 ORR-149 43 8.5 650 0.002 38,100 66,200 13.7 14.3 17.3 585
5065 20 ORR-149 43 8.5 871 0.05 37,100 38,500 1.7 1.9 2.94 584
5065 20 ORR-149 43 8.5 871 0.002 25,900 25,900 0.8 1.6 5.26 586
5067 A.R. 20 ORR-155 500-700 1.4 25 0.05 59,000 123,700 .49.4 51.2 45.4 2289
5067 A.R. 20 . ORR-155 500-700 1.4 650 0.05 38,400 66,800 14.0 14,0 20.5 2290
5067 A.R. 20  ORR-155 500-700 1.4 650  0.002 36,400 59,900 9.6 9.6 16.0 2201
-5085 A.RI. 38 ORR-}SS - 500-700 - 1.4 25 0.05 46,300 110,600 41.1 41.2 40.3 2285
5085 A.R. 38 ORR-155 500-700 1.4 650 0.05 30,800 64,400 18.0 21.2 19.1 2286
5085 AR. 38 ORR-155 500-700 1.4 650 0,002 30,200 51,800 9.3 9.9 15.4 2287
5065  A.R. 20  ORR-155 500-700 1.4 25  0.05 50,100 117,000 54.4 56.1 53.2 1857
5065 A.R. 20 ORR-155" 500-700 1.4 . 650 0.05 37,400 62,100 12.3 12-.4 17.2 1858
5065 A.R. 20 ORR-155 500-~-700 1.4 650 0.002 34,800 49;1 00 6.4 6.5 15.3 1859
5065 A.R. 20 ETR-41-31 600 £ 100 3.5 550 0.002 . 49,100 68,600 9.1 9.4 1273
5065 A.R. 20 ETR-41-31 600 100 3.5 600 0.002 42,200 56,300 8.2 8.5 1276
5065 AR. 20 ETR-41-31 600 £ 100 3.5 650 0.05 41,200 59,000 10.8 11.3 1270
5065 A.R. 20 ETR-41-31 .600+ 100 3.5 650 0.002 42,600 51,800 5.9 6.1 1271
5065 A.R 20 ETR-41-31 600 £ 100 3.5 760 0.002 41,900 46,000 2.8 2.8 1274
5065 A.R 20 ETR-41-30 <150 5 650 0.05 46,700 76,200 21.6 22.2 28.8 383
5065 A.R. 20 ETR-41-30 <150 5 650 0.002 37,700 53,300 9.3 10.0 10.1 380
5065 A.R. 20 ETR-41-30 <150 5 650 0.002 40,000 57,500 il.4 11.6 17.5 384

SOT1
106

ORNL—DWG 67— 2449A

60
50 L * o | 1| b
— 7] .
40
[ ] | ]
- | |
& ! (] | s
=z [ ]
Z 30
=
» HEAT 508 400° C
20 4 UNIFORM ELONGATION
e TOTAL ELONGATION
o = REDUCTION IN AREA
O ‘
10° 2 5 0?2 2 5 10! 2 5 1° 2 5 1o

STRAIN (%)

STRAIN RATE (min ')

Fig. 6.10. Influence of Strain Rate on the Ducfi'lity of MSRE Surveillance

Specimens at 400°C, Line based on total elongation values.

ORNL-DWG 67— 2504A

60

o
o
\

o]
lo]

A
[T AY

N
(@)

HEAT 5081 500°C

e TOTAL ELONGATION
10 . 4 UNIFORM ELONGATION
’ REDUCTION IN AREA

a
0 . L

2 2 5 o' 2 5 1o°

STRAIN RATE (min ")

2 5 10’

Fig. 6.11. Influence of Strain Rate on the Ductility of MSRE Surveillance

Specimens at 500°C. Line based on total elongation values.
107

ORNL-DWG 67-2450A

24
| ] /
|
4 A
. y
, 20 A
A
'] /
9 e L A .
-~ | 1"’fl"
= //
= 2
E 12 [ ] [
n 1;,/"' A
’ pE
“/ .
g =" HEAT 5081 650°C
A UNIFORM ELONGATION
. e TOTAL ELONGATION
4 n REDUCTION IN AREA
0 . ' — ‘
03 2. 5 402 2 5 410! 2 5 10° 2 5 10!

STRAIN. RATE (min™")

Fig. 6.12. Influence of Strain Rate on the Ductility of MSRE Surveillance

Specimens at 650°C, Line based on total elongation values.

ORNL-DWG 67-2451A

24
’ [ ]
20 HEAT 5081 760°C /
A
4 UNIFORM ELONGATION ‘ //
® TOTAL ELONGATION ' /)
£ 187 » REDUCT!ON IN AREA p
= ]
z 1
a ] o
o :
/
8 affil
e T
s
1
N ST 1
' [ ]
[y
0 -2 —4 0 |
103 2 5 1072 2 5 10 2 5 10 2 5 10

STRAIN RATE {min~!)

Fig. 6.13. Influence of Strain Rate on the Ductility of MSRE Surveillance

Specimens at 760°C. Line based on total elongation values.
108

ORNL DWG 67-2500A

12

1

{0

HEAT 5081 850°C
9
4 UNIFORM ELONGATION ;:

. 8 ¢ TOTAL ELONGATION — /
§ 7 ® REDUCTION IN AREA 7
z )i
g © /
E 5 |//
«n s i

4 [ ] i

3
3 gl
| eTT]
2 a A a
A
i ? :l ] ,_
o .

103 2 5 102 2 5 40

2 5 40
STRAIN RATE (min™%)

Fig. 6.14. . Influence of Strain Rate on the Ductility of MSRE Surveillance

Specimens at 850°C. Line based on total elongation values.

6.3 PRECURSORS OF MSBR GRAPHITE
W. H. Cook

Basically, the isotropic graphite sought for molten-salt breeder reactors (MSBR) should have

these properties:

Permeability to helium 1077 t6 1073 cm? /sec
Pore entrance diameter ' None larger than 1 {4
Electrical resistivity <1000 microhms cm? cm™ !
Coefficient of thermal expansion Approx 4.5 X 1078 (DC)—1
Ash content ' <150 ppm

Boron ‘ ' <1 ppm

Bulk density >1.86 g/cm3

In addition, it should be a well-crystallized graphite without fillers such as lampblack or carbon
black. Other parameters, such as mechanical properties, probably would follow satisfactorily if
the preceding requirements were met.

Our immediate needs are modest quantities (50 1b) for general evaluations and for the specific
and important radiation damage studies.

We have made initial examinations on seven experimental grades of graphite. None satisfy
all the basic requirements listed above; however, one shows a good pore spectrum and four ap-

pear to have potential. Two do not show promise,

'y
109

The Chemical Engineering Development Department of the Y-12 Plant® has made some ex-
perimental MSBR-type graphite that has a microstructure as shown in Fig. 6.15. This has been
fired to 3000°C and has not had any impregnations to decrease its pore sizes. Except for some
minor flaws, the structure is unusually tight. This is also shown by the pore spectrum shown
in Fig. 6.16, in which the major portion of the pore entrance diameters are less than 0.3 . The
larger pore entrances shown may be the result of the small cracks and voids visible in Fig. 6.15.
It was fabricated as a small piece, 1.6 in. diam x 0.8 in. long, and there are some problems in
fabrication, purity, and high electrical resistivity, but it does look encouraging in these early
stages of development. It is the only one of the seven different grades that was fabricated
specifically for MSBR requirements.

The other grades are materials that were originally developed for other purposes. As one
might expect, the major problems with these are high gas permeabilities and pores that are too
large. All these had pore entrance diameters as large as 3 y, whether they had uniform or only

surface graphitic impregnations. Four of the grades, which had uniform impregnations through-

Sy-12 Plant, operated by the Union Carbide Corporation for the Atomic Energy Commission, Post
Office Box Y, Oak Ridge, Tenn.

Y—-76831

=

0.023 INCHES
T~ 100X

Fig. 6.15. An Experimental Isotropic Graphi:
As polished. 100x.

Grade B-5-1, Showing Low Porosity and a Few Flaws.

110

ORNL-DWG 67-4780

BULK VOLUME PENETRATED (%)
o

1.0

20.9
12.4
1A
9.60
8.10
6.80
5.60
4.30
3.15
2.10
0.92
0.8
0.7
0.6
0.50
0.40
0.30
0.20
0.100
0.090
0.080
0.070
0.060
0.050
0.040
0.030
" 0.020
0.018
0.014

PORE DIAMETER {u)
Fig. 6.16. A Pore Spectrum Plot of an Experimental lsotropic Graphite, Grade B-5-1.

out the structure, had sufficient properties to warrant further investigation. Two are being
purchased in 50-1b quantities for irradiat\ioh studies and general evaluations:

An interesting feature of six of the grades of graphite above that had pore entrance diameters
as large as 3 pu was the_xt the impregnations appeared to decrease the accessible void volumes
but did not lower the range of the pore entrance diameters.

Decreasing the accessible void volume is helpful, but having the pore entrance diameters
less than 1 pu is of greater importance. Both MSBR and impregnatibn requirements appear to
dictate that the base stock must have the major amount of its accessible voids with pore en-

trance diameters approximately 1 p.

6.4 GRAPHITE IRRADIATIONS

C. R. Kennedy

by

Irradiation experiments to demonstrate the ability of graphite to sustain massive neutron ex-
posures have been designed and are being fabricated or are in progress. Graphite of MSBR
quality will be irradiated in the DFR, HFIR, EBR-II, and ORR. The major irradiations will be
obtained from capsules placed in the HFIR, The specimens will be located in two rod assemb-lies,
which will replace two of the californium production rod assemblies. All the 32 graphite ring
specimens in the rod assembly are designed to be irradiated at 700 + 25°C. After a one-reactor-
cycle experimént to verify the design temperature, the specimens will be removed, examined, and
recycled alternately every six months until a total irradiation time of 21/2 years is obtained. At
this time, the total irradiation exposure will range from 5 to 10 x 10?2 vt (E > 0.18 Mev). The
specimens will be examined for dimensional stability, gas permeability, and mechanical in-

tegrity.
111

Irradiations in the DFR are essentially backup experiments to the HFIR irradiations. These
irradiations will be used primarily to obtain relative comparisons of experimental grades to more
standard grades; The exposures in these irradiations will not exceed 5 x 102! avt (E > 0.18 Mev).

Irradiations in the EBR-II will be restrained growth experiments to confirm the ability of the
MSBR graphite to sustain plastic creep deformation under irradiation. Again, the exposure ob-
tained will be much less than that obtained in HFIR, with a maximum of about 1022 nyt (E >
0.18 Mev). The maximum tensile strains obtainable from these experiments will be about 3%.

The experimental prb,gram in the ORR consists of very closely controlled 'creep experiments
to obtain quantitative creep coefficients. These data, although obtained at a very low exposure
rate and thus low exposures, are essentially for a‘ stress analysis of the graphite bodies. These
experiments will be very limited in scope in view of the base of exrstmg information available
on the creep behavior of graphite. The current experrment has been constructed and mstalled

in the ORR.

6.5 BRAZING OF GRAPHITE

W. J. Werner

Studies were continued to develop methods for brazing large graphite pipes to Hastelloy N.
Our current activities consist of work in the following areas: (1) development of a corrosion-
resistant alldy which will readily wet and flow on graphite but which does not suffer from the
transmutation problem associated with gold-containing alloys and (2) devising techniques for
the manufacture of graphite-to-Hastelloy-N assemblies of the size and configuration required for

loop experiments,

Large Graphite-to-Hastelloy-N Assemblies

A vacuum or inert-atmosphere induction brazing furnace for brazing graphite-to-metal as-
semblies ufi to 4 in. in diameter by 12 in. long has been designed and is currently under con-
struetion. Several pieces of graphite, molybdenum, and Hastelloy N have been prepared for
brazing, using the tapered-joint de51gr1 reported previously.® Figure 6.17 shows the size and
configuration of the components. The graphrte is AT] grade due to the una@arlabrhty of MSRE-

grade material in the desired size and configuration.

6.6 CORROSION RESISTANCE OF GRAPHITE-TO-METAL BRAZED JOINTS

W. H. Cook -

It was reported that the braze joining grade CGB graphite to molybdenum did not have any
microscopically visible attack after a 5000-hr ex posure to LiF-BeF ,-ZiF 4-ThF 4-UF, salts at

®MSR Program Semiann. Progr. Rept. Feb. 28, 1966, ORNL-3936, p. 140.
112

Y-78516

Fig. 6.17. As-Machined Components for Graphite-to-Molybdenum-to-Hastelloy-N Tapered Joint.

1300°F (705°C) contained in Hastelloy N.” The brazing alloy was 60 Pd—35 Ni-5 Cr (wt %).
The chemical analyses of the test salt did not show any significant changes from the analyses
of the as-received salt.

The microstructure of the brazed joint is shown again in Fig. 6.18 for reference purposes.
Electron-probe microanalyses show that molybdenum diffused into the brazing alloy, as was
deduced by Jones and Werner from its microstructure.® There was some migration of nickel and
palladium into the molybdenum, along with a slight amount of chromium.

The speckled precipitates in the brazing alloy are primarily Mo, Ni, Cr, and C in descending
quantities. The long, acicular crystals located toward the graphite sides of the brazed joint
are primarily Mo, Cr, Ni, and C in descending quantities. These are believed to be essentially
mixed carbides of molybdenum and chromium.

The microstructures suggest that this is a well-formed joint that has not been appreciably
altered by the corrosion exposure. A deposit is present on the surface of the braze metal. The
deposit has been identified as pure palladium except for some nickel in the region adjacent to
the brazing alloy. While it extends slightly over the graphite, it seems to be attached only to

the braze metal. Figure 6.19 shows another view of the deposit with a rough surface and a spotty

"MSR Program Semiann. Progr. Rept. Aug. 31, 1966, ORNL-4037, pp. 115—17.
B1bid., p. 113.

PHOTO 87447

PALLADIUM COATING

BRAZING ALLOY

LYBDENUM

Fig. 6.18. Microstructures of the 60 Pd—35 Ni—~5 Cr (wt %) Brazing Alloy Used to Join Grade CGB Graphite to Molybdenum. (a) After 5000-hr ex-
posure to molten fluoride salts at 1300°F (705°C). Etchant: 10% oxalic acid. 100x. (b) Enlarged microstructure of palladium coating located pri-
marily on the brazing alloy. Etchant: oxalic acid. 500x.

€11
ettt at

MOLYBDENUM

Fig. 6.19. External Appearance of Palladium Coating on Brazing Alloy (60 Pd—35 Ni~5 Cr, wt %) of the Grade CGB Graphite Brazed to Molybdenum
After 5000-hr Exposure to Molten Fluoride Salts at 1300°F (705°C). (a) Elevation view. (b) High magnification (100x) of typical appearance.

1498
115

deposit on the molybdenum. No evidence of palladium was found in the fluoride salts or on the
walls of the Hastelloy N container. While the palladium must have been leached from the braze
metal and then redeposited, the mechanism for the transfer is not clear. The thickness of the
coating appears to be dependent on the test time. Howe.verr,-since it was only 0.001 in. thick
after the 5000-hr exposure to the molten fluoride salts at 1300°F (705°C) and seems to be present
only on the braze, it seems valid to say that the braze was essentialiy unattacked for this period.
An amorphous-appearing, metallic-like coating on the graphite® surfaces exposed to the
molten salts was found to be Cr,C, by x-ray diffraction. Electron-probe microanalyses indicated
that some vanadium is also present, There is no explanation for the presence of vanadium un-
less it came from the salts, because the graphite and Hastelloy N normally contain vanadium in

average quantities of 0.0009 and 0.5% respectively.

6.7 THERMAL CONVECTION LOOPS

°

A. P. Litman

We are continuing to study the compatibility of structural materials with fuels and coolants
of interest to the Molten-Salt Reactor Program. Natural-circulation loops of the type described

10,11 ,re used as the standard test in these studies.

previously
Three loops are now in operation, and details of their service are shown in Table 6.4. The

long-term loops, Nos. 1255 and 1258, fabricated from Haistelloy N and type 304 stainless steel,

Ivid., p. 117.

10G. M. Adamson, Jr., et al., Interim Report on Corrosion by Zirconium Base Fluorides, ORNL-2338
(Jan. 3, 1961). '

11 MSR Program Semiann. Progr. Rept. Aug. 31, 1965, ORNL-3872, pp. 81-87.

Table 6.4. Thermal Convection Loop Operation Through February 28, 1967

Maximum ‘
Loop Loop Material Hot-Leg Specimens Heat-Transfer Medium Temp . gT " Hours
No. CF) ("F) _Operated
1255 Hastelloy N Hastelloy N + 2% Nb LiF-Ber-ZrF4-UF4- 1300 160 43,024
{permanent) ThF4 (70-23-5-1-1
' mole %)
1258 Type 304 L stainless steel Type 304 L stainless LiF-Ber-ZrF4-UF4- 1250 180 31,749
steel (removable) ThF, (70-23-5-1-1
mole %)
10 Hastelloy N None Nal“-KF-BF3 - 1125 265 6,734
(48-3-49 mole %) :
9 Type 446 stainless-steel- None Lil"-Ber-ZrF4-UF4 1400 300 5,255%

clad Nb—1% Zr ) (65-29.1-5-0.9 mole %)

“Loop plugged on 10-1-66.
116

respectively, and circulating MSRE-type fuel salt plus approximately 1 mole % ThF, continue
to operate without incident,

Recently, we installed specimens in the hot leg of the stainless steel circuit so as to
generaté additional data on that system. It is of interest to compare the compatibility of the
salt which has now reached maturity in the loop with earlier results. To date, the specimen in
the hottest portion of the loop, 1250°F, has experienced weight losses as shown in Table 6.5,

2

Interpolation of these results indicates a rate loss of approximately 3.6 mg cm™ month™ !,

This is lower than the rate revealed in the previous test.

Table 6.5. Effect of Time on Weight Change of
Type 304 L Stainless Steel Specimen in
Contact with LiF-BeF,-ZrF ;-UF ;-ThF

(70-23-5-1-1 Mole %) at 1250°F

, Time ‘ Weight Change
(ht) , (mg/cmz)
25 -1.6.

115 -2.1
450 -2.95
1125 —4.4

Loop No. 10, fabricated from Hastelloy N and circulating a fluoroborate mixture, is now
scheduled to operate for one year, after which time it will be dismantled and examined. It

continues to circulate without difficulty.

6.8 EVALUATION OF MSRE RADIATOR TUBING CONTAMINATED WITH ALUMINUM

D. A. Canonico D. M. Haseltine

The failure of the aluminum blower blades at the MSRE site and the resultant damage were
discussed in the last semiannual report.!? It was concluded that no damage was visually ob-
served; however, the\possibility that some undetected aluminum still might be in intimate con-
tact with the Hastelloy N tubes did exist. An experiment was conducted to determine the effect
of a prolonged exposure to aluminum at 1200°F.

Aluminum pieces from the blower blades and a Hastelloy N tube similar to those used in the
heat exchanger were placed in intimate contact and held for times up to 1000 hr. The results
of the 1000-hr exposure are shown inm Fig. 6.20a. For comparative purposes, the 5-hr exposure is
shown in Fig. 6.205. It is evident that the penetration (approx 0.010 in.) is similar in both

-

12MsR Program Semiann. Progr. Rept. Aug. 31, 1966, ORNL-4037, pp. 103-7.
117

Fig. 6.20. Metallographic Sections Through Aluminum Cone and Hastelloy N Tube. Samples were held
ot 1200°F for (a) 1000 hr and (b) 5 hr.

photomicrographs. The Hastelloy N tubing has a wall thickness of approximately % , in. The
penetration after 5 hr was about 16%, and after the 1000-hr exposure it had not increased.

The microstructure seen in the specimen held for 1000 hr is considerably different from that
in the 5-hr sample. The extended exposure has allowed the various phases to grow and has re-
sulted in a diffusion couple of aluminum-nickel complicated somewhat by the presence of zinc

and other minor elements.

This work supports the conclusions reported previously, that the radiator system is satis-

factory for further operation.
7. Chemistry

7.1 CHEMISTRY OF THE MSRE

Fuel Salt Composition and Purity

R. E Thoma

More than half the total power generated by the MSRE (21,464 Mwhr) was produced during the " .
current report period. Heat balance.and nuclear calculations indicate that a total of 1.083 kg of
235y, 0.473% of the original uranium inventory, has now been consumed by fission. Table 7.1 .
summarizes the results of fuel composition and purity analyses for the MSRE fuel in each of the
power runs, including the three conducted in the last six months, Nos. 8 to 10, and the current
run, No. 11. These data show that the uranium concentration of the fuel salt has decreased
appreciably since power operation began. It should not be inferred, however, that burnup is
evident in the results of chemical analysis for runs 4 to 10, for virtually all the decrease noted
in Table 7.1 is the result of dilutions of the fuel salt by flush salt. Figure 7.1 shows a cfim_—
parison of MSRE inventory values ! with the results of chemical analyses ? for runs 6 to 10.
Analytical values shown here have been adjusted to compensate for changes in isotopic com-
position of uranium and for those periodic variations in analytical bias as determined by the
Analytical Quality Control Group.? -

Step decreases in the inventory values reflect the dilution of the fuel salt by residues of
flush salt remaifiing in the reactor fuel circuit after flushing operations are completed. Fuel
salt is- also removed from the fuel circuit by sampling procedures and by transfer of fuel to flush
salt. From the-results of analyses of flush salt specimens we have deducéd that each drain-
flush-fill sequence results in a net transfer of 7.1 kg of uranium from the fuel to flush salt. The
computation of this value has involved a number of assumptions, such as those concerned with
the precision of uranium analyses in the 100-to-800-ppm concentration range, as well as the
configuration and dimensions of MSRE fuel circuit components where salt residues may reside.

Figure 7.1 also shows values which should have been obtained in the absence of dilution-transfer

1H. B. Piper, personal communication.

2Chemical analyses were performed under the supervision of C. E. Lamb, ORNL Analytical Chemistry
Divisien.

3G. R. Wilson, ORNL Analytical Chemistry Division.

118
Table 7.1. Summary of MSRE Fuel Salt Analyses

Number . U® Inventory \ ' -

Run o Concentration (wt %) Values (wt %) Concentration (mole %) Concentration {(ppm)
No. Samples 7Li Be Zr u® R TLiF BeF, ZrF, UF ] Fe Cr Ni

4 22 10.51 £0.137 6.55 T 0.161 11.14 £0.205 4.642 £0.028 4.622 4.622 63.36 £ 0.567 30.65 +0.583 5.15 £0.116 0.83 £t 0.011 131 t65 48 t7 40 L20
5 1 10.65 6.53 11.45 4.625 4,622 4.622  63.63 30.30 5.25 0.816 68 51 15

6 13 ©10.51 10,289 6.54 £0.200 11.31 £0.231 4.630 £0.027 4.622 4.619 63.35 £1.072 30.60 10.946 5.23 £0.145 0.825 10.013 111 T 44 50 +8 56 £24
7 11 10.55 £0.054 6.67 ¥0.174 11.34 £0.215 4.640 £0.017 4.619 4.614 63.04 £0.495 30.9510.580 5.20%0.118 0.819 £0.009 88132 4816 48 t16
4-7 . 47 10.523 £ 0.178 6.572 0.179 11.239 £0.271 4.638 £0.025 4.622 4.614 63.290£0.722 30.698 £0.696 5.188 £0.126 0.824 £0.011 114 55 49 +7 46 +21
8 8 11.78 £1.406 6.53 1 0.199 11.16 £0.193 4.632 1 0.011 4.601 4.599 65.84 +2.486 28.57 £2.123 4.82 £0.347 0.771 £0.058 122 45 64 27 61 £36
9 4 10.99 £ 0.099 6.63 £0.068 11.15210.370 4.603 £0.031 4.587 4.586 64.17 £0.006 30.04 £0.147 4.99 £ 0.185 0.794 £0.011 150 £17 61 £5 52 20
4-9 59 10.528 £ 0.159 6.570 £0.175 11.222 £0.266 4.635 £0.026 4.622 4.586 63.312 £0.679 30.685 £0.665 5.179 +0.125 0.824 £ 0.011 118 +52 51 9 48 T 24
10 10 11.14 £ 0.079 6.58 £0.188 11.0510.152 4.609 £0.020 4.575 4.569 64.65 £0.450 29.64 £0.480 4.92 1 0.064 0.791 £0.010 150 £30 6024 74 35
410 69 10.785 £ 0.641 6.571 £0.179 11.197 £0.260 4.631 £0.026 4.622 4.586 63.835 £1.342 30.258 £1.163 5.095 £ 0.212 0.811 $10.029 122 51 539 52 +27
8-10 22 11.345 £0.883 6.569 £0.173 11.106 £0.212 4.616 £0.022 4.601 4.569 64.998 £1.614 29.320 £1.402 4.898 £ 0.226 0.784 0,036 140136 6215 6533

2Corrected to compensate for isotopic composition. x.denotes beginning of run; y denotes

end of run.

611
120

ORNL-DWG 67-4782

4.66 -
AN
4.64 RUN 7 -
RUN 6 . ' /
— 1 RUN 8 L ‘

4.62 i S TT - + RUN 9
"o 1 T/ T
E 1 . : . P —— \-__
% L !
% 460 : RUN 10
3 S
2 . ) L~q
2 458 !
5 e . T iy

456 CHEMICAL ANALYSES CORRECTED FOR ANALYTICAL BIAS

— —— BOOK VALUE ADJUSTED FCR DILUTION BY FLUSH SALT -
— — BOOK VALUE WITHOUT ABJUSTMENT FOR DILUTION
454 — . BY FLUSH SALT : -
4.52 -
0] 2 4 6 8 10 12 14 16 _ 18 20
' Mwhr (X 10°)
Fig. 7.1. Weight Percent of Uranium in MSRE Fuel Salt During Runs
6 to 10. y ‘

losses; these values indicate that while chemical analyses were heretofore not sensitive enough
to reflect burnup losses, it may be possible that such losses will be reflected in subsequent

operations with the present fuel salt.

MSRE Fuel Circuit Corrosion
R. E. Thoma

Since oxidative corrosion of the MSRE fuel circuit results in the formation of chromous fluo-
ride, the concentration of chromium in the fuel salt serves as the principal indicator of the extent
of generalized corrosion. The chromium content of the fuel salt has remained very low throughout:
the operating history of the MSRE. In the current report period the chromium concentration of the
salt has remained at 62 £ 5 ppm (Table 7.1), corresponding to a uniform removal of chromium from
the walls of the reactor circuit from a maximum depth of about 0.1 mil.

Prior to the removal and replacement of the original metal and graphite surveillance speci-
mens in August 1966, * the average chromium concentration of the fuel salt was 48 +7 ppm (Tablé
7.1), a value wl-mich was attained during the first power operations with the reactor. A sample
taken early in October 1966, which followed the change of the surveillance specimens, showed
the chromium content had reached about 62 ppm. All subsequent specimens of fuel salt analyzed

‘after that time have shown the presence of approximately 62 ppm of chromium, indicating the

4MSR Program Semiann. Progr. Rept. Aug. 31', 1966, ORNL-4037, p. 97.
y

121

introduction of about 85 g of chromium into the fuel since October 1966. The increase in chro-
mium concentration from 48 to 62 ppm may possibly be assignablé to corrosion of the new sur-

veillance assembly. If so, the corrosion sustained by the assembly must be greater than that

- previously experienced by the entire fuel circuit from December 1965 up to the present time.

The metal surface exposed to the salt in the surveillance assembly is about 1000 in. 2.

To pro-
duce 85 g of chromium would require the'leaching of chromium from this surface, if uniform, to an
implausible depth of about 10 mils.

The MSRE fuel salt contained a very low concentration of U3* at the outset of power oper-
ations, with possibly a maximum of 0.16% of the uranium (366 g, 1.54 gram atoms) in the usdt
form (cf. section entitled ‘‘Extent of UF , Reduction During MSRE Fuel Pieparati'on”). At the
termination of MSRE run 7, fuel burnup had consumed 1.66 gram atoms of uranium. Just how much
oxidizing capacity is produced by fuel burnup is uncertain because it involves unverified infer-
ences as to the final chemical identity of many fission products, but between 0.6 and 1.0 equiva-
lent of oxidizing capacity should result for each gram atom of 23°U fissioned. If 1.0 equivalent
was produced, the fuel salt would have become slightly oxidizing by the end of run 7. No such
conclusion is justified, however, from the results of chromium analyses. These results indicate
rather that oxidation corrosion ceased before the MSRE had generated a total of 1 Mwhr of power.
Absence of corrosion iri runs 4 to 7 should probably be attributed at the beginning to the presence
of U3+ and perhaps subsequently to the deposition of noble-metal fission product films. At the
termination of run 7 (7800 Mwhr) a uniform film formed by the deposition of all the Mo, Nb, and
Ru produced would have a thickness of approximately 150 A.

The current experiment with the MSRE is scheduled to be terminated at the end of 30,000
Mwhr of operation. At that time the present surveillance specimens willl be removed for inspection
and testing. It will be of considerable value tolle‘arn whether the present assetfibly has sustained

the amount of corrosion which has been observed in the fuel samples since last October.

Extent of UF ; Reduction During MSRE Fuel Preparation
B. F. Hitch C. F. Baes, ]Jr.

Uranium was added to the barren fuel salt of the MSRE as a binary mixture of 27 mole % UF,
in ’LiF. This fuel concentrate had first been purified by the usual sparging with an HF-H,
mixture to remove oxide, followed by sparging with hydrogen alone to complete the reduction of
structural metal fluorides such as NiF , and FeF ,.5® During this final reduction step, a small
portion of the UF , should also have been reduced, the amount depending upon the duration of the

treatment and the equilibrium constant for the reaction

1 _ 7
UF ,(d) +5H2(g) = UF,(d) + HF(g) .

5_]. H. Shaffer et al., Reactor Chem. Div. Ann. Progr. Rept. Jan. 31, 1965, ORNL-3789, pp. 99—109. '
6_]. H. Shaffer, MSR Program Semiann. Progr. Rept. July 31, 1964, ORNL-3708, pp. 288—303.
122

The exact amount of UF , thus introduced into the MSRE fuel has become a matter of special
interest with continued operation of the MSRE owing to evidence that significant amounts of

some fission products are far more oxidized than would seem compatible with the presence of
significant amounts of UF ; in the MSRE fuel. Consequéntly,.the data collected by Shaffer et al.’
during the purification of the fuel salt concentrate at the production facility recently has been

. examined in detail in an attempt to determine the equilibrium quotient for the above reaction:

PHF XUF3

T PLTX
- Tm, TUF,

’

and to determine the extent of UF4 reduction in the LiF-UF4 mixture.
For small amounts of reduction, the UF 3,/UF4 ratio may be related to Q and the volume (V) of

H, passed per mole of UF, (nU) by®
(1'.1UF3/nU)2 - ZQPI:]/: (V/nyRT) + (n{)JFs/nU)2 ,

provided equilibrium conditions are maintained during sparging. The last term on the right is the

initial "UFs/nU ratio. Replacing ”UFs/”U by QP;]/:/PHF ,

1 2 ( V) 1
= + .

2 1/2 0 2
B Pir QPH2 nyRT (PHF)
In accord with this equation, plots of l/PI?IF vs V, based on data collected at 700°C during the
purification of the various batches of fuel concentrate, were found to be linear. All plots could
be fitted reasonably well with lines of slopes corresponding to ¢ ~ 0.9 x 10~ % atm 172, From
the final value of Py, knowing Q and Py , the average amount of uranium reduction at the end

2
of the hydrogen treatment was estimated to be 0.16%. _
In an attempt to confirm this estimate of Q and the amount of reduced uranium present initially

in the MSRE fuel, an 11.4-kg portion of unused fuel concentrate was studied further in the lab-

7Unpub1ished data, supplied by J. H. Shaffer. - |
8Combination of

PHF
RT

dnUF3= dv

and
nyry;  Pyp

1/2
-1 p

Q:

n

to e_liminaie PHF’ followed by integration, gives

|4 1
=——2[r+ln 1-nl+cC,
n,RT  QPY

H,

where r = nUFs/nU' For small values of r this simplifies to the equation in the text.

123

oratory. >Hydrogen sparging was initiated at 5S10°C. At this relatively low temperature, no sig-
nificant reduction of U** to U3* should occur; however, HF evolution was detected immediately
and continued at a significant level until 250 liters of H, had passed and 0.0019 mole of HF per
mole of uranium had been evolved. This indicated that inadvertent exposure of the salt to oxi-
dizing impurities such as water or oxygen had occurred during prior storage, during transfer of
the sample to the reaction'vessel, dr in later handling. Since the presence of HF at this tempera-
ture in the amounts seen should have quickly oxidized the UF , present, it was not possible to
confirm the amount of UF , initially present in the fuel concentrate. In two subsequent H,
sparging runs at 700°C, however, data were obtained which permitted improved estimates of Q

from plots of 1/ PfiIF vs V. The resulting values of Q are about twice that estimated from the

H2 F|ow
Temperature (ml min_l kg_l) Q (ofml/2)
. Run1 707 53 1.74 x 107°
Run2 , 705 35 1.85%10~°

salt production data. It is not reasonable to attribute this discrepancy entirely to the differences.
in temperature since, judging from Long’s measurements of the temperature dependence of Q in
LiF-BeF melts,® more than a 30°C difference would be requiréd. It seems more likely that the
discrepancy is due partly to nonequilibrium sparging conditions in the production treatment. The
present value of Q = 1.8 x 10~ ¢ atm!/2 determined for the fuel concentrate is somewhat lower
than the value ~4 x 10~ ¢ atm'/?2 which may be'estimated for the MSRE fuel salt at 700°C from
Long’s measurements. This indiéates that UF, is not as easily reduced in the fuel concentrate
as in the fuel salt.

Even though equili_brium conditions might not have prevailed during purification of the fuel
concentrate, 0.16% reduction of UF4 remains a_yalid estimate since, in effect, it is based upon
_ the integrated amount of HF evolved by reduction which, in turn, is related by material balance

to the amount of UF3 formed.

Adjustment of the UF, Concenfl;ution in the MSRE Fuel Salt

W. R. Grimes R. E. Thoma

The fission product isotopes of molybdenum, niobium, fechnetium, rufhenium, and tellurium
were expected to appear principally in their elemental forms in the MSRE system. While some
might be carried as suspended metal or even in solution as moderately unstable fluorides of low
valence state, they were expected to precipitate, in large part, on the metallic portions of the
reactor. Although this suggested behavior has indeed taken place in the MSRE, appreciable quan-

tities of molybdenum, ruthenium, and tellurium (and probably technetium and niobium) have also

-QG. Long, Reactor Chem. Div. Ann. Progr. Rept. July 31, 1965, ORNL-3789, pp. 68—72.
124

been observed in the cover gas in the MSRE pump bowl. Substantial fractions of the fission prod-
: uct niobium, molybdenum, tellurium, and ruthenium have been found on or in the MSRE moderator
graphite. 10 The presence of these materials as gas-phase species suggested that the fuel salt
contained, at the outset and until this year, much less uranium trifluoride than intended and very
much less than is tolerable. A

The fuel as charged into the MSRE for start of the power opération prpbably had NUF4 at very

near 9 x 10~ 2 mole fraction and, at most, Ny at1.4x 10~ ° mole fraction, which corresponds to
0.16% of the uranium being UF .. The UF content of the MSRE fuel was determined!! after ap-
proximately 11,000 Mwhr of operation by study of the equilibrium corresponding to

1 ' ) )
UF4(d) +—2— Hz(g) = UFs(d) + HF(g) .

The result showed that the concentration of UF_ corresponded to less than 0.05% of the total
uranium and probably to less than 0.02%. The MSRE fuel salt was considered to be far more oxi-
dizing than was necessary or desirable and certain to become more so as additional power was
:‘produced unless adjustment was made in the UF, concentration. A program was therefore ini-
tiated to reduce 1% of the 228.5-kg inventory of U**, or 9.64 gram atoms, to U3™ by the addition
of small quantities of beryllium metal to the circulating salt. Initially, 4 g of beryllium was intro-
duced into the quel sa;lt by melting a mixture of 7LiF-BeF2 carrier salt and powdered beryllium in
"the MSRE pump bowl sampler cage. Subsequently, three adflitions have been made by suspending
specimens of 3/8-in. beryllium rods in the salt in the pump bowl. The capsules used for adding
beryllium were similar in size and construction to those used for sampling for oxide analysis but
were penetrated with numerous holes to permit reasonable flow of fuel salt. The beryllium rods
have reacted with the ffiel salt at a steady rate, dissolving at approximately 1.5 g/hr. To date,
27.94 g of'beryi1i1‘1m has been introduced into the MSRE fuel salt, a quantity correspdnding to the

conversion of 0.65% of the U** to U3™,

7.2 FISSION PRODUCT BEHAVIOR IN THE MSRE

S. 8. Kirslis F. F. Blankenship

"The initial results of tests on the chemical behavior of fission products in the MSRE were
reported previously, ! with descriptions of the experimental facilities used and a discussion of
the objectives of this work. Most fission products behaved as expected, with the exception of
the noble metals, which showed an unexpected tendency to volatilize and to deposit on metal sur-
faces and in graphite. This obsetvation, implying that the fuei salt was more oxidizing than was

desirable, contributed to the decision to reduce the fuel b‘y repeated small additions of beryllium

10S. S. Kirslis, MSR Program Semiann. Progr. Rept. Aug. 31, 1966, ORNL-4037, p. 165.
11A. 5. Meyer, ‘‘Hydrogen Reduction}of MSRE Fuel,”’ Intra-Laboratory Correspondence, Jan. 3, 1967.
125

metal. The special pump bowl tests in this report period were mainly directed toward following
the effect of beryllium additions on the volatilization and deposition behavior of the noble metals.
Also completed in the period were the radiochemical analyses on the first set of long-term surveil-

lance specimens of graphite and Hastelloy N exposed in the MSRE core.

Long-Term Surveillance Specimens

The bulk of the radiochemical analyses were reported previouslry12 on the graphite and Hastel-
loy N specimens exposed in the MSRE core for 7800 Mwhr of power operation. Further analyses
were made on selected graphite samples for *5Zr, 25Nb, '4!Ce, !*%Ce, and '37Cs to provide a
more complete pictute of the behavior of these isotopes. A few samples were also analyzed for
147Nd and 91'Y because of the interest in rare-earth-type isotopes themselves, as well as in their
rare-gas precursors. The ne\'a-v data are tabulated, along with previous results on these elements,
in Tables 7.2—7.4. In a few cases, the previous resuits were slightly corrected on final evalu-
ation of the counting data. (Similar corrections for the other previously reported fission products
seldom exceeded the 10% analytical error; the revised values will not be reported here.)

The new data generally followed the indiéations from previous results. The flhat profile of
gasZr in the interior of the middle graphite bar, at a level 100 times that of the blanks, suggests
a slight volatility of zirconium, since ?5Zr has no long-lived gaseous precursor. Some of the
95Nb in the interior may have arisen from the decay of 95Z1, but this correction is negligibl\e for
the first two surface samples. Even in the interior, the concentration of °3Nb in a given layer
was always higher than that of 95Zr, whereas the reverse would have been true if most of the
95Nb had been formed in place from °5Zr. Thus the observed high concentrations of 5Nb in the
graphite may not be ascribed to precursor behavior.

The additional data on '4!Ce, 14%Ce, 'Y, and *’Nd confirmed the previous conclusion that
the distribution of the rare earths and alkaline earths in graphite reflects the diffusion behavior
of the precursor rare gas. The species with short-lived gaseous precursors showed steep concen-
tration gradients, whereas those whose gaseous precursors had half-lives of several minutes
showed relatively flat interior concentration profiles. A diffusion model has been developed'?
which satisfactorily accounts for the observed distributions of the fission products with rare-gas
precursors. |

The very flat interior concentration profile of 137Cs in graphite was confirmed by the new
results in Table 7.2. A distribution similar to that of 8°Sr was expected on the basis of precursor
behavior. The data suggest a mobility of '37Cs itself, in accord with the known volatility of ele-
mental cesium over cesium carbide. !4

Since 99Tc is the worst neutron poison after ?>Mo in the noble-metal fission product group,

radiochemical analysis for it was attempted on several graphite samples. It proved extremely

1201SR Program Semiann. Progr. Rept. Aug. 31, 1966, ORNL-4037, pp. 165—91..
13R. J. Kedl, unpublished communication.
19MSR Program Semiann. Progr. Rept.. Aug. 31, 1966, ORNL-4037, pp. 165-91.
126

Table 7.2. Radiochemical Analyses of Middle Graphite Bar

Weight Depth Disintegfations pef Minute per Gram of Graphite
Sample of Cut
() 95 95 144 141 137 147 91
(mils) Zr Nb Ce Ce Cs Nd 1y
Wide Face Exposed to Circulating Fuel
1 0.8463 6.02 1.12x10'° 6.93x10'! 3.08x10° 3.17x10° 1.53x10° 571x10® 1.90x10"°
4 1.2737 9.7 1.37x10% 0.16x10° 8.43x107 6.53x10° 2.02x107 <1x10? 1.83x 108
7 0.0814 7.50 1.16x10° 7.76x10%® 3.80x107 1.43x10° 1.98x107 <4x10® 3.12x10°
11 0.9145 6.94 8.20x107 2.04x10® 2.22x107 7.33x10%® 2.17x107 8.9x10° 2.07 x 10°
14 08962 6.98 1.51x10° 4.28x10% 4.07x107 555x10° 1.86x107 5.1x107 3.33x10’
17  1.0372 8.25 3.16 x 107  2.46 x 107 3.90x 10% 1.96 x 107 4.26 x 107 2.66 x 10’
23 0.8176 6.64 1.09x10°
Side Face Exposed to Circulating Fuel
2 0.7583 7.68 8.33x 1011 2.48x10% 3.38x10%% 1.27x107 3.82x10% 2.46x10'°
6  0.9720 10.10 1.52x 102% 2.41 x 107 7.26 x10° 5.12 x 10°
8  0.5139 5.43 4.87 x10%  3.00x 107 2.50%10°
12 0.3395 3.65
15  0.6976 7.61 2.54x 107 8.39 x 103
18  0.3737 4.15 ¢
Other Side Face Exposed to Circulating Fuel
3 06135 6.21 6.36 x 1011 2.45 x'10° 3.35x 10! 8.58x 107 2.59x10° 1.60x 10"
5  0.7543 7.84 2.80 % 101% 1.08x 10%® 7.76 x 10° 1.99 x 107
9  0.6198 6.55 2.36x10° 2.18 x 107 3.01 x 10°
13 1.1652 12.53 '
16  0.8469 9.24 2.04x107 3.77 x 108
19 0.9406 10.45
Face in Contact with Graphite
20  1.1381 9.23 6.68x10° 1.04x10'! 1.56x10° 2.15x 10'°? 4.8%10°
Unexposed Graphite Blanks
10 4.48 X105  2.46 x 10°  1.07x10° 1.41x10° 2.90x10*' <1x10® 1.97x10°
24 131 x 10° 3.28x10° 8.06x10° 6.38x10° <8.3x10* <3x10% 1.41x10°
Notes: 1. The samples are arranged in order of successive cuts on each face (see Fig. 7.22 of Ref. 25, p. 179).
2. The sample weights given here have been corrected for the average 4.5% loss during milling.
3. The depths of cut were calculated from the sample weights and areas and the known graphite density.
g
4.

The activities tabulated are corrected to the time of shutdown, 11:00 AM, July 17, 1966.
127

Table 7.3. Radiochemical Analyses of Top Graphite Bar

Weight Depth Disintegrations per Minute per Gram of Graphite
Sample .y of f-:m 95, 95p 144, . g, 137 147y 4 91y
(mils) .
Wide Face Exposed to Circulating Fuel
25 0.3602 6.23 1.28x10° 1.56x10'" 4.06 x10® 8.60x10° 2.24x10" <1x10° 3.59x10"
20 -0.4355 7.93 1.06 x 10° S 1.22x107 T <2x107 5.72x 10°
58  0.5260 9,94 9.5 x10® 4.73x10%° 1.87x10%® 1.57x 107
60  0.2916 5.51 1.49 x 107
62  1.0783 20.38 1.11 x10° 3.54x 107 1.49x10’
Side Face Exposed to Circulating Fuel
26 0.4615 11.41 5.97 x10'°% 2.38 x 10® 4.94 x 10°
31 © 0.4564 11.88 2.50 x 10%  5.22:x10% 2.13x 10°
ther-Side Face E;tposed to Circfilofing Fuel
28  0.6703 17.11 5.20% 10" 2.46x 10® 5.38x 10°
33" 0.5404 14.43 8.62 x 107
Wide i’oce in Con‘focf with Groplhife
27 0.6422 11.32 1.24x10° 5.59x 10" 2.57x 10® 4.03x10° 2.50x 107 1.85x 10® 1.55x 10°
32 0.5375 9.95 2.28 x 108 - | 1.0x 107 <4x10° 2.79% 10"
59 0.3154  5.96 6.25x10° 1.05x 107 2.56 x 10® 1,87 x 107
61  0.5835 11.03 |
63  0.7310 13.82 2.04x 10° 2.17x 107 9.93 x 10°
Notes: The samples are arranged in order of successi\;e cuts on each face (see Fig. 7.22 of Ref. 25, p. 179).

The sample weights given have been corrected for the average 18.9% weight loss during milling.

1
2
3. The depths of cut were calculated from the sample weights and areas and the known graphite density.
4. The activities tabulated are corrected to the time of shutdown, 11:00 AM, July 17, 1966.

difficult to purify ??Tc from confaminating activities sufficiently well to provide an accurate

count of this low-specific-activity isotope. The results indicated qualitatively amounts of °°Tc

to be expected from the decay of the determined amounts of its parent *°Mo. In view of this test

and the chemical similarity of Tc and Mo, it is fairly safe to calculate 99T¢ concentrations from -

‘observed ??Mo concentrations on the assumption that °9Tc will remain where the ?°Mo was found.

The nature of the distribution of various fission products in the graphite surveillance speci-

mens is shown graphically in Fig. 7.2. The lines are rougil averages obtained by visually drawing

smooth curves through individual points, often widely scattered, representing all of the available -

data. The steepness of the concentration gradients of the noble metals and the very different gra-

dients of species with noble-gas precursors ('*%Ba and #9Sr) are strikingly apparent.
128

Table 7.4. Radiochemical Analyses of Bottom Graphite Bar

Weight Depth Disintegrations per Minute per Gram of Graphite
Sample ’ of Cut
() ) 95, 95Np 144~ 141~ 137 g 1,47Nd 91y
. (mils)
Wide Face Exposed to Circulating Fuel
9 11 8 9 : 7 8 8
34 0.8032 15.04 2.40x10 3.34 X 10 6.05x 10" 8.11 x10 2.80x10 _ <3x10 1.20xX 10
38  0.5979 11.64 2.49 X 10° 1.09 x 107
64  0.2323 4.68 3.12x10% 8.40x10° 6.92 x 108
66 0.3120 6.28
69  0.7183 14.49 7.63x 10° 1.84 x 10°
Side Face Exposed to Ciréuloiing Fuel
. 11 8 ' 10 7 9
35 0.3904 10.70 2.63x10 7.89x10% 1.04 %10 7.27 x10 3.69x 10
39 0.4480 12.98 9.00x10® 1.09x 107 2.29x10°
Other Side Face Exposed to Circulating Fuel
11 8 9
37 0.5480 - 15.39 3.05x10 5.81 x 10" 9.19xX10
41 0.3520 9.62 4.49x10% 3.05x10% 7.25x10°
Wide Face in Contact with Graphite
36 0.4810 9.12 3.47x10° 4.20x10'! 9.05x10%® 1.07 x 10'° 1.12x10% s5.10%10°
40 0.5936 11.77- 2.50x 109 <3 X 108 7.69 x 107
65  0.4756  9.58 3.07x10% 2.72x10° 1.10x10®
67  0.4025 8.10
68 0.6260 12.61 4.09 X 105 1.03 x 108
Notes: 1. The samples are arranged in order of successive cuts on each face (see Fig. 7.22 of Ref. 25, p. 179).
2. The sample weights given have been corrected for the average 9.1% weight loss during milling.
3. The depths of cut were calculated from the sample weights and areas and the known graphite density.
4. The activities tabulated are corrected to the time of shutdown, 11:00 AM, July 17, 1966.

-

Uranium. Analyses of Graphite Specimens

Since appreciable quantities of uranium had been found in graphite from previous in-pile tests

in which the fuel was allowed to cool and radiolyze, a number, of the milled graphite samples were

analyzed for uranium by a chemical fluorometric method and by delayed-neutron counting. The

chemical method on dissolved samples showed only that the uranium concentrations were less

than 30 ppm by weight. The very sensitive delayed-neutron counting method gave the results
shown in Table 7.5.

The very low surface and volume concentrations of 235U shown in Table 7.5, corresponding

to about 1 g in the complete graphite core, could have no discemible nuclear or chemical effect
129

13 ORNL-DWG 67-774

-

<
.
]

) Y
1
Wi —t
L W
T
e | I |
g
. L\
o
<§t 14
50 BN
foed l\ ‘\ N, —-.\\ -
W | W VAN —~ a5
o \ NN o — Sr
s VN NN -
] . P —
z AN \\
=
Y \ \ ™~
a 10 L ¥ 11 Y - 432
Ly L} N e Te
) A1 L ~
& ‘\ \\ \\
e T\ ~ —\40gq
x \ A SN
S 99
© N 20 N
z =
%2} \
o
409 .Y
Y ™~
\ -~
AN \\ 103
\\ RU
\\
e 134
\ I
10
0] 10 20 30 40 50

DISTANCE FROM SURFACE OF GRAPHITE (mils)

Fig. 7.2. Distribution Profile of Fission Products in

lGruphife from MSRE coré. )

on the operation of the reactor. The interior uranium concentrations in the top and bottom graphite
were near the blank graphite value, but those in the middle graphite bar were distinctly higher and
exhibited no concentration gradient. .

It is interesting to compare the depos1t10n of uranium and molybdenum in the first layer of
graphite. The avérage value for uranium from Table 7.5 is 0.072 ug of 235U or 0.22 pg of total
uranium per square centimeter. The average amount of °?Mo on the graphite surfaces at the time
of reactor shutdown was 0.039 pg of ®°Mo per square centimeter. Assuming that 99Mo indicates
the deposition behavior of the stable molybdenum fission products there must have been nearly
2 pg of total molybdenum per square centimeter of graphite after 7800 Mwhr of reactor operation.
Thus the weight of total molybdenum depositing in the first layer of graphite was nearly ten times
that of total uranium. _ |

The amounts of seven fission products deposited on top, middle, and bottom samples of Hastel-

loy N from the surveillance assembly were previously fepbrted. 15 In addition, the following val-

151bid., p. 53. : .
Table 7.5. Uranium in Graphite Surveillance Specimens

Micrograms i Micrograms
Sample Milled of 235y of 235y
No. Graphite Bar Layer per Gram per Square
of Graphite Centimetef
25 Top 1 2.72° 0.080
29 2 " 0.15
58 3 0.16
60 4 0.36
1 Middle 1 3.56° 0.090
4 2 1.26
7 3 1.07
11 , 4 0.87
23 6 1.18
34 Bottom 1 0.66° - 0.047
38 2 0.06
64 3 0.07
66 4 0.11
Graphite . <0.082
blanic B

aAverage of dup.licate samples which agreed within about 10%,.

ues were obtained for 25Nb deposition: top, 2.68 x 1019 dis min—! cm~2; middle, 2.74 x 101° dis
min~! cm~?%; and bottom, 3.79 x 10'° dis min~* cm~ 2. If the niobium were distributed uniformly
over the 1.2 x 108 cm? of Hastelloy surface in the MSRE, these values correspond to 42, 43, and
90%, respectively, of the calculatefl total 5Nb present at reactor shutdown in the reactor system.
Thus, on the average, about half the *Nb produced was deposited 'on the reactor metal walls.
A similar average for the deposition of 9°Nb on graphite was nearly half the total present. Cor-
respéndingly, analyses for ?5Nb in recent fuel salt samples, after correction for °5Zr decay
since sampling, indicated that _only a low fraction of the total present remained in the salt.

The Hastelloy N samples were also analyzed for 6°Co, °*Mn, Co, and Fe. The thermal
and fast fluxes calculated from these values agreed satisfactorily with values from the analysis
of dosimeter wires included in the surveillance package. For ?5Nb, as for the fission products
previously reported, there was no correlation between flux and the amount deposited on Hastel-

loy N.

Fuel Salt Samples

Six additional 10-g samples of circulating molten fuel, taken primarily in connection with

recent pump bowl tests to assess the effect of beryllium metal additions on the volatilization
131

1

and plating behavior.of noble-fnefal fission products, were analyzed radiochemically for the 13
isotopes listed in Table 7.6. The data from the last of the previous five samples, FP7-12, are

included in Table 7.6 for comparison. The noble-metal activities are plotted in Fig. 7.3.

Effect of Operating Conditions

In the period of reactor operation covered by the seven tabulated samples, there were four
reactor drains and a number.of_ shutdowns, the longer of which lasted 83.5 days, 14.5 days, and
- 14.1 days. .Sample FP8-5 was taken to test the effect of reactor drain and long shutdown (83.5
days) on the concentrations of fission products. The activities were calculated back to the time
of shutdown (July 17, 1966) and should thus be comparable with the results of sample FP7-12,
taken a few days before the shutdown. It is seen that the concentrations of alkaline earths, rare -
earths, and ?5Zr were not significantly altered. The small rise in '3!I concentration is attributed
to a slight contamination of the sample, which would have a large apparent effect when 'multiplied
by the large correction factor for decay for more than ten half-lives. " There were, however, sig-
nificant decreases by a factor of 4 'in the concentrations of °3Ru, !°®Ru, and !?°Te, suggesting
a slow deposition of these species on the metal walls of the drain tank. Similarly, the noble-
metal concentrations were lower in sample FP11-22, taken 3.2 days after shutdown, than in the
' previous sample l‘i‘P11-12, taken during power operation. The other species increased in concen-

tration due to the 16 days of power operation between the two samples.

Effect of Beryllium Additions

It is difficult to conclude from the data in Tablé 7.6 that there was a significant effect of
fuel reduction on the noble-metal concentrations in the fuel. The c-oncentrations frequently rose
rather than fell, as expected, after adding beryllium. It was interesting that the ?°Mo, !%3Ru,
106Ry, and !3?Te showed parallel rises and falls. The °°Mo results were impossibly high for
samples FP11-8 and FP11-12. The high values were checked by reruns on fresh samples. If all
the °°Mo produced by fission remginedruniformly distributed in the fuel, the calculated concen-

—1. A simple calculation shows that if all the %Mo

tration would be 1.4 x 101! dis 1;1in‘1 g
produced by neutron activation of the ®®Mo in the first 0.1-mm thickness of the Hastelloy N
- reactor containment vessel diffused instantaneously into the fuel melt, the increase in ?°Mo
conc‘en_tration would be only about 10° dis/min per gram of fuel. It thus appears that a mechanism
is required which either concentrates fission-produced ?9Mo and other noble metals in the pump
bowl or resuits in 1arge temporal and spatial variations in their concentrations. Dissolved ?°Mo
would undoubtedly be uniformly distributed. If the noble metals circulated as a shspension of
insoluble metal particles, it is conceivable that they might concentrate in the pump bowl or vary
in concentration with pump bowl level, cover gas pressure, and other operating variables. Since
clean metal surfaces are not wet by the fuel salt, there might also be a tendency for metal par-

ticles to collect around helium bubbles, which are probably most numerous in the pump bowl. The

froth or flotation hypothesis can be checked experimentally in several ways.
Table 7.6. Analyses of Fuel Salt Samples

Experiment FP7-12 FP8.5 FP10-12 F P10-20 FP11-8 FP11-12 FP11-22
Sampling date 7-13-66 10-8-66 12-28-66 1-9-67 241367 2.21-67 3.9.67
Operating time, days® 4.1 off, 11.9 on 11.9 on, 83.5 off 14.5 off, 13.2 on 14.5 off, 25.4 on 14 off, 16.1 on 14 off, 24.1 on 40 on, 3.2 off
Nominal power, Mw 7.2 0 - 7.4 7.4 7.4 7.4 0
Accumulated Mwhr 7200 7800 13,800 15,800 19,000 20,400 22,400
Be addition, g 5.5 10.65 11.65
Isotope Fission Disintegrations per Minute per Gram of Sa 1t?
Yield (%) c '

9.6 hr ?1s¢ 5.81 1.32 x 10!! 1.28 x 1011 1‘.31><1011 1.33x 10! 1.53 x 1011

51 day %%r 4.79 3.96 x 1010 3.70x 1010 3.83 x 1010 4.70 x 1010 4.78 x 1010 6.50 x 1019 7,93 x 10!°

33 day !*lce 6.0 6.88 x 1010 7.20 x 1010 4.63 % 1010 4.11 x 1010 9.23 x 1019 1.10x 101!

285 day !*%Ce 6.0 1.91 x 1010 1.79 % 1010 2.43 x 1010 9,68 x 1010

66 hr Mo 6.06 3.15 x 1010 3.56 x 1010 4.77%x 10 3.20x10''>  2.54x10!'!? 9.09x10!°

39.7 day !°3Ru 3.0 7.14 x 10° 1.42 x 10° 8.02 x 108 6.08 x 108 5.22 x 10% 5.33 x 10° 3.66 x 10°

1.01 year '9%ry 0.9 -2.13x 108> 5.00x 107 ~4.0x 107 2,78 x 1072 1.62 x 108 ~1.9%x 108 1.46 x 108

y -

77 hr 132%Te 4.7 3.81x 1010 1.56 x 1010 2.04 x 1010 5.50x 1010 3.76 x 1010 2.80x 1010

33 day 129m71e 0.35 4,94 x 108 1.38 x 108 1.49 x 108 ~3.1x 108

8.05 day 131 3.1 5.36 x 10'° 7.94 x 1010 5.46 x 1010 7.16 % 1010 4.96 x 1010 7.55x 101  8.30x 101°

35 day 25Nb 6.2 (2.45x 10112 4,77 x 1019 6.73 x 10° 2.36 x 10192~ 1.14 x 10° 1.21 x 1010

65 day °°Zr 6.2 6.55 % 101° 6.01 x 1010 5.42 x 101° 5.21 x 10190 9.34 x 1010 9.25x 101% 1,12 x10!!

12.8 day !*°Ba 6.32 1.42 x 1011 9.00x 10! 9,97x10!° 1.30x 101! 2,74 x 1011>

30 year '37Cs 6.0 3.13 x 108 4.04 x 108

“Duration of previous shutdown and of continuous operating time just before sample was tak

shutdown,

bCalculated to the time of sampling or to the previous shutdown.

en; vice versa for the two samples taken during

c€l
133

ORNL-DWG &7-4783

41os jo woab uad ajnuiw sad suolypibajuistp

[+;] -] ~
22-hdd = NMOgLNHS Aop-¢ T T xkll T T Y T T T T T A |DCI T 17T — ]
/| /| 3
A_”_.r_
2h=thdd M-t —1 —] h\%l — gttt | —t— — |~ —H T+ 4+ — — ]
‘28 699t} [Tlpm / 7 | =3
mnzn_.._..ll|llmlllllllIIA.|-||-|IIllloifl I T % — 8 ? - |— — | — —
/ N
I.I.l.. / 4 .y
02-0kdd FF—TF—T— —F — — —— T+ — — =i = 0 — 4 — — re— S — — ——
eg 6590} - / / \\ o
2l-okdd H H — — +— 4+ — 14 IJ. L e — | A L IV- —
_ / N
/
NOILIQaY - _
ag bg'g
§-8dd H —=——F +— +— —F — = | —+V— TV — — — 1117 T Tl e 1t = —t — |Ih|P||r||-
et o -
NwIN&n_ S Oy — TNMOQJLNHS hOUlm.meI! T Ill!n..‘ Mwu| — — Tt -n._ Hll]A.Ullfl‘lh"l 1 EI.I“I R R
..H_I = 3R <4 =
» 8 o g ’
&
o n o st n ‘2 o n o ) [Te] o~ @
Q Q o o e}

,:om JO woib

Jad ajnuiw J1ad suoipibajuisip

16 20 24

12
Mwhr (x103)

F'ig. 7.3. Nobel-Metal Activities in Fuel Samples.

Calculated back to the

time of sampling or to the previous shutdown.
134

The 9“_‘Nb concentrati,\ons in Table 7.6 varied erratically and did not parallel the behavior of
the  other noble metals. This is ascribed to analytical difficulties, which are being further in-
vestigated. An unavoidable difficulty is that a large correction for ®*Zr decay must be made for
each salt analysis. .

The other fission products (°!Sr, 8%Sr, 14!Ce, 14%Ce, 131, ?5Z1, and !*°Ba) generally
showed approximately the expected yields in the salt phase within analytical error. It was re-
ported previously *® that reactor power levels calculated from the observed concentrations of
elements like strontium and cerium, which remain in the fuel i)hase, were consistently about 20%
lower than the power levels calculated from coolant heat balances. This discrepancy has per-
sisted in the analyses of the new fuel salt samples. To date there have been several dozen

radiochemical analyses, all of which indicate lower than nominal power levels.

Pump Bowl Volatilization and Plating Tests

Six pump bowl tests were run in this report period in which metal specimens were expdsed to
the gas phase and the fuel phase of the pump bowl for 10 min. ‘The previous technique 7 was
modified slightly in that the coils of silver and Hastelloy N were attached beside the capsule
cables rather than being wound on them. This facilitated sample disassembly in the hot cell
and furnished an additional specimen of stainless steel cable in the gas phase. Each of the
four gas-phase specimens (Hastelloy N, silver, nickel-plated key, and stainless steel cable) and
the single fuel-immersed specimen (stainless steel cable) were leached or dissolved, and the
solutions were analyzed for at least nine fission products and for #°°U.

The gross features of fission product behavior in these runs were similar to those reported
previously, with heavy deposition of the noble-metal fissiofi products on all specimens and light
contamination by °5Zr, rare earths, and alkaline earths. All the analytical data obtained will
not be presented since the mass of detail would be confusing. However, to make clear the effect
of the three beryllifim additions, the depositions on the Hastelloy N spécimens in the gas phase
and the stainless steel samples in the fuel phase are presented in detail in Tables 7.7 and 7.8.
The results on the other gas-phase specimens were generally qualitatively similar to those on the
Hastelloy N samples. | | _ ,

It is seen from Table 7.7 that noble-metal volatilization usually increased after the first and
third beryllium additions and decreased as expected only after the second addition. Neverthe-
less, tfie average deposition on the gas-phase specimens was slightly lower after the three ad-~
ditions than it had been previously. Much larger decreases in volatilization followed reactor,
shutdown. Experiment FP8-5 was run just before power operation was resumed after an 83.5-day
shutdown, and FP11-12 after a 3.2-day shutdowh. In each case, activities were calculated back

to the time of shutdown. The deposition on the gas-phase specimens of elements with stable

1%1bid., p. 168.
17rbid., p. 69.
Table 7.7. Fission Produet Deposition on Hastelloy N

Experiment FP7-12 FP8.5 FP10.12 FP10-20 FP11-8 FP11-12 FP11-22
Sampling date 7-13-66 10-8-66 12-28-66 . 1-9-67 2.13-67 2+21-67 3.967
Operating time, days?® 4.1 off, 11.9 on 11.9 on, 83.5 off 14.5 off, 13.2 on 14.5 off, 25.4 on 14 off, 16.1 on 14 off, 24.1 on 40 on, 3.2 off
Nominal power, Mw 7.2 -0 7.4 7.4 7.4 7.4 ' 0
Accumulated Mwhr 7200 7800 13,800 15,800 19,000 20,400 22,400
Be addition, g 5.5 10.65 11.66
Isotope Fission Disintegrations per Minute on Total Specimenb
Yield (%)
66 hr 2o 6.06 3.35 x 1019 4.00 x 10!° 1.36 x 101! 5.61 x 1010 1.51 x 10!! 2.49x 1010
39.7 day '°%Ru 3.0 2.14 x 10° 1.51 x 108 3.60 x 108 1.03 x 10° 4.71 x 108 2.25 x 10° 1.14 x 107
1.01 year 9°Ry 0.9 ~7.2 x 107 4.86 x 106 ~7.0x 108 3.63 x 107 1.5x 107 7.1 x 107
77 hr '32Te 4.7 3.67 x 101! 1,55 x 101! 3.36 x 101! 1.20 x 1011 1.56 x 10! 2.74 x 1010
33 day 1%977¢ 0.35 2.40 x 108 8.79 x 108 2.5 x 10? ‘
8.05 day !31I 3.1 2.71 x 1010 Y 1.77 x 1010 1.01 x 1010 5.77 x 107 7.58 x 10° 9.68 x 108
35 day 9°Nb 6.2 3.91 x 10° 1.1 x 107
65 day 257¢ 6.2 7.49 x 106 <3.6 x 10° <4.0x 108 $4.5x10° £3.2x 108  ~6.54x% 10°
12.8 day '*%Ba 6.32 1.84 x 108 - 2.73 x 108 1.72 x 108 1.35x 108 7.40 x 107
235y, ug total 2.63 © 0.64 . 091 0.423 1.88 1.79

“Duration of previous shutdown .and of continuous operating time just before sample was taken; vice versa for the two samples taken during

shutdown.

PCalculated to the time of sampling or of previous shutdown.

SET
Table 7.8. Fission Product Deposition from Fuel on Stainless Steel

Experiment FP7-12 FP8-5 FP10-12 FP10-20 FP11-8 FP11-12 FP11-22
Sampling date 7-13-66 10-8-66 12-28-66 1-9-67 2-13-67 2.21-67 3.9-67
Operating time, days? 4.1 off, 11.9 on 11.9 on, 83.5 off 14.5 off, 13.2 on 14.5 off, 25.4 on 14 off, 16.1 on 14 off, 24.1 on 40 on, 3.2 off
Nominal power, Mw 7.2 7.4 7.4 7.4 7.4 0
Accumulated Mwhr 7200 7800 13,800 15,800 19,000 20,400 22,400
Be addition, g 5.5 10.65 11.66

Isotope Fission Disintegrations per Minute per Specimenb

Yield (%)

66 hr ?9Mo 6.06 1.30x 10! 2.23 x 10! 3.13 x 101! 7.28 x 10! 4.20x 10'1 7,13 x10°
39.7 day !%3Ru 3.0 3.10x 10° 5.90x 108 2,24 x 10° 9.50 x 10° 1.88 x 1087 1.38 x'10°
1.01 year 196Ry 0.9 6.60 x 107 1.77 x 107 ~9.2 % 107 2.96 x 108 5.6 x 1057  4.04 x 107
77 hr 321e 4.7 2.21 x 101! 3.59 x 1011 1.96 x 10! 3.84x 1019 3,67 x10!° 2.17x10!°
33 day 129mTe 0.35 3.17 x 108 1.09 x 108 1.34 x 10° |
8.05 day 31 3.1 3.85x 101° 1.23x 10° 1.71 x 1010 1.19% 101% . 3.38x10° 2.15 x 10° 2.63 x 10°
35 day ?5Nb 6.2 7.87x 1010 1.7 %107
65 day 95Zr 6.2 2.84 x 107 <6.5 x 109 <1.2 x 107 S1x107 ~6.3x108 7.2 x 108
12.8 day !*%Ba 6.32 4.81 x 108 1.71 x 108 4.34 x 107 3.25 x 107 3.61 x 107
235y, pg total 24 17.6 1.65 3.57

“Duration of previous shutdown and of continuous operating time just before sample was taken; vice versa for the two samples taken during

shutdown,

PCalculated to the time of sampling or of previous shutdown.

o1
137

fluorides was not affected either by fuel reduction or reactor shutdown. The fact that noble-
metal fission products volatilized to a considerable degree after reactor shutdown indicated that

the process of fission was not directly related to the vo-latilizing process.

As in previous runs, it was observed that deposition on the four different metal specimens
was of similar magnitude for all species. Less reaction of noble-metal fluorides with silver in
particular was expected since AgF is relatively unstable. In ekperiment FP11-22 a gold wire
coil was substituted for the silver one, and it picked up as much of the noble metals as the other
metal specimens. This lack of chemical discrimination is puzzling and k:_asts some doubt on the
supposition that the noble-metal fission products leave the fuel phase as high-valent fluorides.
To'eliminate the possibility that the observed deposition was due to reaction with oxide films on
the metal specimens, experiments are planned in which the specimens will be reduced before in-
sertion into the pump bowl. These observations of deposition indicate the value of laboratory
experiments to identify the chemical nature of the noble-metal species that volatilize from radio-
active fuel salt. Such experiments appearto be feasible using either mass spectrometric or gas

chromatographic techniques.

The data in Table 7.8 show that the average deposition of noble metals was heavier on the
stainless steel cable immersed in fuel than on Hastelloy N in the gas phase, although contami-
nation by salt, as indicated by °5Zr, 14°Ba, etc., was no greater. It was interesting that the
’plating behavior generally paralleled the volatilization behavior. Thus on the average, the amount
of noble metals plated decreased after the second beryllium addition but increased after the first
and third additions, as did deposition on the gas-phase specimens. Also, the amount plated was
distinctly lower in the two runs made after reactor shutdown. This parallelisrfi is consistent with
the theory that high-valent noble-metal species in the fuel take part in both the plating and vol-
atilization processes. More information on the nature of the high-valent species in the fuel may
be obtained by immersing specimens of other metals besides stainless steel in future pump bowl

tests.

Becaqse. of the practical interest in the behavior of fhe molybdenum and %%Tc fission p;od-
ucts, the results of these pump bowl tests for 99%o are given in detail in Table 7.9. The ques-
tionable effect of fuel reduction and the definite effect of reactor shutdown are shown clearly.

It is seen that deposition on the nickel-plated key was. relatively heavy, as qbser_ved previously,
while depositions on the other ‘specimens were of similar lower magnitudes. The larger size of
the nickel specimen probably accounts for the larger total amount deposited. The heavy depqsit
on the key suggests the truly gaseous nature of the ?*Mo Spe.cies, since the key is surrounded
by the latch and only its bottom tip projects into the pump bowl spéce. The qualitative nature of
these tests is shown by the data in Table 7.9. It might be concluded from Table 7.9 that depo-
sition of ??Mo on silver was uniformly higher than on Hastelloy N. This is not borne out by
previous data, nor by the behavior of the other noble metals. Individual values are probably
reproducible only within a factor of about 4. On this basis, only sizable changes in deposition,

such as between the last two runs, are to be considered significant.
138 :

Table 7.9. Deposition of %Mo on Pump Bowl Specimens

Experiment FP7-12 FP10-12 FP10-20 . FP11-8 FP11-12 FP11-22
Sampling date 7-13-66 12-28-66 1-9-67 2-13-67 2-21-67 3-9-67
Operating time, days® 11.9/4.1 13.2/14.5 25.4/14.5 16.1/14 24.1/14 40/3.2
Nominal power, Mw 7.2 7.4 7.4 7.4 7.4 0 ’ =
Accumulated Mwhr 7200 13,800 15,800 19,000 20,400 22,400
Be addition, g 5.5 10.65 11.66
Sample ‘ Disintegrations per Minute per Specimenb
x 1010 x 1010 x 1010 x 1010 x 1010 x 100
Hastelloy N 3.35. 4.00 13.6 5.61 151 2.49
Silver 12.7 67.0? 16.2 18.2 29.3 2.82 (Au)
SSG° ' : 9.57 12.7 8.90 20.2 1.83
Nickel? 17.6 141 105 26.7 54.6 14.4
SSL° 13.0 22.3 31.3 . 72.8 42.0 0.71 :
1 —-1f -

Fuel, dis min~ " g 3.15 3.56 4,77 32.0? 25.4? 9.09

®The slash separates the durations of the previous shutdown and of continuous operating time ]ust before the
sample was taken; vice versa for sample FP11-22, taken during shutdown.

Calculated to the time of sampling or of previous shutdown.
°Stainless steel cablé in gas phase.
dNickel-plated key.

®Stainless steel cable immersed in fuel phase.

mgMo activity in fuel sample taken simultaneously.

Uranium on Pump Bowl Metal Specimens

Also recorded in Tables 7.7 and 7.8 are the quanfities of 235U determined by delayed-neutron
counting in the leaches of the metal specimens exposed in the MSRE pump bowl. The reported !
results varied from 0.4 to 24 pg of *>°U per sample. Of the total of 25 samples run, only five
exceeded 5 ug of 235U, with the remainder averaging 2 pg of ?33U. No correlation was apparent R
.between the deposition of urériium and that of fission products'; thus the uranium found did not
merely represent contamination of the samples by fuel salt. It is not surprising that uranium -
shows different chemical behavior from the noble metals. What is difficult to explain chemically
is that it volatilizes or plates at all. Fortunately the extent of uranium volatilization and plating
is small (as shown more conclusively in the examination of the long-term surveillance specimens)

and is of negligible practical concern.

Freeze Valve Capsule Experiments

Although valuable qualitative information was derived from the tests in which metal specimens
were exposed in the pump bowl, these tests suffered from the drawback that the results could not
be interpreted quantitatively. Therefore; a capsule was designed to take a pump bowl gas sample

of known volume from whose analysis the gaseous concentrations of fission product species
139

could be calculated. It was required that the sampling device be small enough to pass freely
through the bends of the 1 1/2-in.-diam sampling pipe and that it should operate éutomatically when
it reached the pump bowl.

The device shown in Fig. 7.4 operated satisfactorily to furnish 20-cc samples of pump bowl
gas. The capsule is evacuated and heated to 600°C, then cooled under vacuum to allow the
Li BeF, in the seal to freeze. The double seal prevents loss of Li,BeF , from the capsule
duting sampling. The weighed, evacuated capsule is lowered into the pump bowl through the
salf—sampling pipe and positioned with the bottom of the capsule 1 in. above the fuel salt level
for 10 min. The freeze seal melts at the 600°C pump bowl temperature, and pump bowl gas fills
the 20-cc volume. The capsule is then withdrawn to 2 ft above the pump bowl, and the capsule
is allowed to cool. Thete is a slight ‘‘breathing’’ action of the capsule when the initial sample
is partially exhaled as the capsule warms from the Li ,BeF , liquidus temperature (457°C) to
600°C. As the capsule cools from 600 to 457°C in the sampling pipe, additional gas is drawn

into the capsule. The double freeze seal prevents loss of the salt seal during the exhalation.

ORNL-DWG 67-4784

STAINLESS STEEL

CABLE
3/ _: )
VOLUME | 3/—in. OD NICKEL
20
e
LizBeF, NICKEL CAPILLARY
2Befy -

Fig. 7.4. Freeze Valve Capsule. .
140

It isrthou‘ght that the reactive fission product gases first inhaled react rapidly with the interior
metal (nickel or stainless steel) walls of the capsule, so that correction need not be made for
loss during the exhalation. The final inhalation is 2 ft up the pipe, where the atmosphere should
be relatively pure helium.

The cooled resealed capsule is withdrawn from the sampling pipe and transported in a carrier
to the analytical hot cells. Here a Teflon plug is placed over the protruding capillary and the
exterior of the capsule is thoroughly leached free of fission product activities. The top of the
capsule is cut off, and the interior metal surface is leached with basic and acid solutions for 1 hr.
The bottom of the capsule is then cut at two levels to expose the bottom chamber of the capsule.
The four capsule pieces are placed in a beaker and thoroughly leached with 8 ¥ HNO, until the
remaining activity is less than 0 1% of the original activity. The three leach solutions are radio-
chemically analyzed for the activities shown in Table 7.10 and for 235U by delayed-neutron
counting.

Table 7.10 gives the total activities in the leach solutions for the freeze valve capsule runs
made before and after the first 5.5-g beryllium add1t1on to the fuel. The runs made after the next
two beryllium additions, failed since the capsules were lowered a little two far into the pump bowl
-and withdrew salt samples. In the two successful runs, the final leach of the segmented capsule
contained about nine-tenths of the total of each of the activities. The activities either were
| trapped by the sealing salt or reacted very rapidly with the walls of the bottom chamber or the
capillaries.

The results of Table 7.10 generally confirmed the qualitative indications from the pump bowl
tests. For comparison, Table 7.10 includes the ana1y51s of fuel salt sample FP10-20, which was
taken between the two freeze valve capsule runs. It is seen that the 20-cc gas samples con-
tained more **Mo, '°*Ru, '°Ru, and !*?Te than were contained per gram of fuel salt. The
'slight decrease in most of these activities after adding 5.5 g of beryllium is probably within
analytical error. The °°Nb activities were relatively low and may reflect the analytical diffi-
culties with this isotope. On the other hand, the 95Zr and 1*%Ba activities were very low,
indicating that less than 0.5 mg of fuel salt had entered the capsules. The small amounts
of these activities found may have been due to the slight volatility of Z1F , and to the 16-sec -
140xe precursor of 14°Ba. The quantities of 235U found were low, but higher than expected.

Although 0.5 mg of fuel salt would more than account for the amount of ?*°U found, this ex-
planation is difficult to accept since the values for ®5Zr and !*°Ba were lower in the sample

for which the 235U value was higher. Tentatively, the 23°U values are taken as representing
volatilization of uranium. The sharp drop in the uranium value after adding 5.5 g of beryllium
metal was not reflected in the amounts of uranium found on the metal specimens of the corre-
sponding pump bowl tests (see Table 7.7). This correspondence will be examined in future tests.

The known volume of the gas samples permits us to calculate quantitatively the molar gas-
eous concentrations of the noble metals and of uranium in the helium cover gas, also given in
Table 7.10. It is seen that the gaseous concentrations represent significant partiél pressures

of uranium, molybdenum, ruthenium, and tellurium. The concentrations for 99Mo multiplied by the
141

Table 7.10. Freeze VYalve Capsule Results
Experiment FP10-11 FP10-22 FP10-20
Sampling date 12-27-66 1-11-67 1-9-67
Operating time., daysa 14.5 off, 12.6 on 14.5 off, 27.7 on 14.5 off, 25.4 on
Nominal power, Mw 7.4 7.4 7.4
Accumulated Mwhr 13,600 16,200 15,800
Be addition, g J 5.5
Fission Before Be After Be
Isotope Yield & ) Fuel Salt b
(%) Dis/min Total Ppm Dis/min Total® Ppm (dis min~} g_l)

66 hr ?9Mo 6.06 2.04 x 101 5.4 1.36 x 10*! 3.6 4.77 x 1010
39.7 day '?3Ru 3.0 3.80x 10° 1.4 2.63 x 10° 1.0 " 6.08 x 10°
1.01 year '%%Ru 0.9° ~6.7x10' ~0.24 7.74 %107 0.27 2.78 X 10"
77 tr 13%1e 4.7 5.73 x 1010 1.7 5.08 x 10'° 1.6 2.04 x 101°
8.05 day ‘3'1 3.1 9.75 x 10° 0.75 2.03 x 107 0.18 7.16x101°
35 day ?5Nb #6.2 $3.26 x 107 2.09 x 108 ‘ 2.36 x 101°
65 day °°Zr . 6.2 <2.9x10° <0.0018 $2.2x 107 £0.012 5.21x 10'°
12,8 day '*°Ba 6.32 2.75 x 10° 0.034 3.48 x 10° 0.042 9.00 x 10'°
. . ,

Sy 3.86 ug 46 0.55 pg 6.5 14,000 ug/g

“Duration of previous shutdown and of continuous operating time just before sample was taken.

bDisintegrations per minute calculated to the time of sampling or of previous shutdown.

total cover gas flow (5000 liters/day) correspond to more than half the total ?°Mo produced per

day. Since previous material balances indicated that the bulk of the ®°Mo should deposit on the

Hastelloy N surfaces or remain in the salt, it is suspected that the gas concentration in the

sampling volume surrounded bg} the mist shield is higher than in the remainder of the pump bowl

gas phase. A similar calculation for the gas sample before beryllium addition indicates that

more than 2 g of uranium per day would be swept out by the cover gas. The next gas sample

indicated much lower uranium loss. Clearly, more analyses of freeze valve capsule samples

are needed to determine more accurately the volatilization of uranium.

Special Pump Bowl Tests

A special capsule has been designed (Fig. 7.5) to determine the concentrations of hg}dro-

carbons and volatile fluorides in the pump bowl gas. A known volume of the sample gas is

made to pass through a weighable stainless steel cloth container containing CuO. From the

. weight change and the fluoride ‘analysis of the CuO, the gaseous concentrations of hydrocarbons

and volatile fluorides can be calculated. A single test of this type with flush salt in the reactor
has been successfully run and showed detectable amounts of hydrocarbons and HF. Quantitative

results will be reported when results are available on several tests.
142

ORNL-DWG 67-4785

TOP OF
PUMP BOWL

[ CONE SEAT

— ]

MIST SHIELD —— o]

T———Cu0 IN STAINLESS
STEEL CLOTH BASKET

[T~ FUEL SALT __ |l —

Fig. 7.5. Hydrocarbon Analyzer.

The above test can be run only with the.reactor fuel pump turned off. A variation is being
developed which can be run with the reactor at power and the fuel circulating.

Another special capsule has been designed for exposing graphite specimens to both the fuel
phase and the gas phase of the MSRE pump bowl. This test will permit the evaluation of noble-
metal deposition on graphite under various specific short-term operating conditions.

-

General Discussion of Fission Product Behavior

Additional qualitative pump bowl tests and new quantitative freeze valve capsule tests have
confirmed the previously reported tendency of the noble-metal fission products t\o leave the fuel
by volatilization and by plating on Hastelloy N. The observed behavior was difficult to interpret
thermodynamically, as discussed previously. These difficulties were compounded by the currently
reported observations that uranium also volatilized and that the behavior of noble metals was not

significantly affected by the addition of 27.8 g of beryllium metal to the fuel, sufficient to reduce
143

0.65% of the U** content to U3*. Analyses of the reduced salt for U3+ by a recently developed
method indicated that the U4*/U3" ratio was far too low to produce the observed partial pressures
of volatile high-valent fluorides of noble metals afid uranium.

In order to retain the volatile fluoride hypothesis, recourse must be had to kinetic explana-

. tions. It may be postulated that sizable steady-state stoichiometrically equivalent concentrations
of US* and U3* are produced radiolytically in the fuel melt. ‘This postulate implies that the oxi-
dized and reduced species do not undergo back reaction fast enough for their concentrations to
fall to very low values. In the radiolysis of water at high temperatures, steady-state concentra-
tions of H, and O, are produced and maintained analogously. U3* and U* are logical choices
for the surviving oxidized and reduced species, since U*" is the bulk fuel constituent that is
most easily reduced and most easily oxidized. Now the US* can oxidize the noble-metal fission
products to high-valent volatile states. Bubbles of helium known to be circulating with the salt
can act as kinetic traps to preserve the oxidized s;;ecies from reaction with U3*. The probability
of reduction is much poorer in the dilute gas phase. The oxidized fluorides may in this way be
delivered to the pump bowl and swept out.' '

A difficulty with this theory, aside from its several assumptions, is that noble-metal volatili-
zation was nearly as great three days and 83.5 days after reactor shutdown (Table 7.7) as during
reactor-operation, whereas the radioactivity of the fuel melt and consequently its rate of radioly-
sis had changed by several orders of magnitude. A very strong dependence of the rate of back
reaction of U* and U5 on their concentrations would be required to reconcile this fact to the
- theory.

An alternate theory postulates that the noble metals circulate as metal sols in the fuel melt.
Since the fuel melt does not wet clean metals, the sol would tend to accumulate at the interfaces
of the fuel with the gas phase, such as helium bubbles. When the helium bubbles go through the J
spray ring and burst, some of the metal sol may be sprayed into the gas phase as a gaseous sus-
pension, which would act like a volatile species.

Difficult to explain by this theory are the volatilization of uranium and the appreciable pene-
tration of noble metals into the surfaces of the graphite surveillance specimens.

It is clear that information is needed on the nature of the volatilizing uranium and noble-metal
species before a credible account of their behavior can be given. Experiments in this direction
are planned. b '

The reduction of the fuel melt by adding beryllium metal will be continued until at least 1%
of the U** has been converted to U3*. This additiqnal reduction may have some effect on noble-
metal volatilization. The addition of a little hydrogen to the helium cover gas may be effective
if volatile fluorides are involved. The effectiveness of hydrogen in reducing noble-metal volatili-
zation may be tested in the hot cell by measuring the volatilization from a‘large sample of molten

MSRE salt during sparging with helium and with hydrogen.
~ ' 144

7.3 PHYSICAL CHEMISTRY OF FLUORIDE MELTS
The Oxide Chemistry of LiF-BeF,-ZrF, Mixtures
B. F. Hitch C. F. Baes, ]r.'
Previously described measurements®®+19 of the solubility of BeO in LiF-BeF, mixtures and of
the solubility of Z1O, in simulated MSRE fuel-salt—flush-salt mixtures have been completed. The
results, which have been reported more fully elsewhere,?® may be summarized by the following

expressions:

In LiF-BeF, saturated with BeO at 500 to 700°C,

log X ,_ = —0.901 + 1.547Xp o — 2625/T .
0] 2

In (2LiF-BeF ). + ZiF , satutated with Z10,, at 500 to 700°C,
a

— 3/2
X s b(XZrF4) ’

O : 1/2
(XZrF4)

whereifl
log a = ~1.530 — 2970/T ,
log b=—1.195 — 2055/T .
In these expressions, the mole Afraction is defined
n,

X’ _ I
i - B
o it "Ban + "sz4

Z0, replaces BeO as the least soluble oxide when X, - exceeds 0,0008. This corresponds to
4 .

1.6 wt % fuel salt in flush salt. With further increases in the amount of fuel salt, the qxi.de

tolerance decreases at a given temperature, passes through a minimum at X, - approximately
o : 4

equal to 0.01, corresponding to approximately 20 wt % fuel salt, and then increases (Fig. 7.6).

Solubilities of SmF; and NdF in Molten LiF-BeF, (66-34 Mole %)
F. A. Doss F. F. Blankenship J. H. Shaffer
Measurements of rare-earth fluoride solubilities in molten LiF-BeF2 mixtures have been re-

sumed to supplement earlier data on rare-earth trifluoride solubility behavior in molten fluoride

_ systems.?! This experimental program will examine the behavior of those rare earths which are

;ac. F. Baes, Jr., and B. F. Hitch, Reactor Chem. Div. Ann. Progr. Rept. Dec. 31, 1965, ORNL-3913,
p. 20.

19MSR Program Semiann. Progr. Rept. Feb. 28, 1966, ORNL-3936, p. 133.

20B. F. Hitch and C. F. Baes, Jr., Reactor Chem. Div. Ann. Progr. Rept. Dec. 31, 1966, ORNL-4076,
p. 19. . :

21'I'\’ezulctor Chem. Div. Ann. Progr. Rept. Jan. 31, 1960, ORNL-2931, p. 77.
145

6R NL-DWG 67-769

7(°C)
700 600 500
1000 T T ]
Ry
/440(
47‘50 .
14
e N4y

Je ‘S
200 P o, 29

OXIDE CONCENTRATION (ppm)}

)
®
N <o ' L
100 o
—Us - N~
e
)
N ~
50 , AN
(2)
-
20
1.0 14 1.2 13
1000/7 (ok)

Fig. 7.6. Estimated Oxide Tolerance in MSRE Salt
Mixtures. (1) Flush salt saturated with BeO, (2) flush-
salt—fuel-salt mixture of minimum oxide tolerance, and

(3) fuel selt.

of interest to fuel reprocessing studies for the reference design MSBR and whose solubilities in
the proposed MSBR fuel solvent have not been previously measured. The solfibility of SmF, in
LiF-BeF, (66-34 mole %) was determined for comparison with previous data in a similar fluoride
solvent and with anticipated measurements of SmF, solubilities in the same solvent.

The experimental method essentially duplicated that of the earlier investigation. Filtered
samples of the saturated solvent were withdrawn at 50° temperature intervals while heating and
cooling the molten mixture between 550 and 800°C. Rare-earth concentrations in the filtered sam-
ples were determined by radiochemical techniques. In the current study rare earths weré labeled
in situ with an appropriate radioisotope at temperatures in excess of that required for complete
dissolution of the added rare-earth fluoride, Earlier studies incorporated the radiotracer during the
preparation of the rare-earth trifluoride. '

Experimental values of the solubility of'SmF3 in LiF-BeF , (66-34 mole %) at 600, 700, and
800°C were 0.013, 0.024, and 0.040 mole fraction respectively. Solubilities of NdF3 in the same
solvent were 0.010, 0.0‘1‘9, and 0.035 mole fraction at 600, 700, and 800°C respectively. Heats of
solution of 10.5 kcal/mole for SmF, and 11.5 kcal/mole for NdFa- were calculated from the linear
dependence of the logarithm of the solubility values on the reciprocal of the absolute temperature,

as shown in Fig, 7.7,
146

ORNL-DWG 67— 4786

5.0

4.0

ol TN

2.0 |— \

RARE EARTH FOUND IN SOLUTION (mole fraction x10%)
L
i

8 | 9 10 14 BERY: 3
10‘OOO/T(°K)

Fig. 7.7. Temperature Dependence of the Solubility of SmF3 and NdF3 in
Molten LiF-BeF2 (66-34 Mole %). ,

Current solubility values for SmF coincide with those obtained previously in the solvent LiF-
BeFZ-UF4 (62.8-36.4-0.8 mole % respectively) only at 650°C. The heat of solution of SmF by
the previous data is about 13.9 kcal/mole. Although no discrepancies in the two experimental

procedures are apparent, further studies would be required to resolve possible solvent effects.

Possible MSBR Blanket-Salt Mixtures |

S. Cantor R. E. Thoma

-

The blanket salt proposed in the MSBR reference design repott is an ’LiF-BeF ,~ThF  mixture
(71-2-27 mole %), whose liquidus temperature is 1040°F (560°C). Alternate compositions of

7LiF‘-Ber-ThF4 may appear to be attractive as blanket salts if the advantages of lower liquidus
147

temperatures are not offset by the attendant reduction of thorium concentration. (The concentra-
tions of BeF are not sufficient to affect the viscosity adversely.) Mixtures of 7Li‘F-BeF2-ThF4

whose compositions lie along the even-reaction boundary curve

L == 3LiF - ThF (ss) + 7LiF - 6ThF

appear to qualify as the best blanket salts from phase behavior considerations. The ThF4'con-
centration of these mixtures varies from 6.5 to 29 mole % as LiF concentrations change from 63 to
71 mole %. Liquidus temperatures for this range of compositions vary from 448°C (838°F) to
568°C (1044°F). In the molten state these mixtures contain thorium concentrations ranging'from
850 to 2868 g/liter at approximately 600°C. These data are summarized in Figs. 7.8 and 7.9 and
in Table 7.11. Optimization of thorium corncentration and liquidus temperatures may be made with
the aid of Fig. 7.8; if 2000 g/liter of thorium is adequate to achieve good breeding gain, the
composition containing 17.5 mole % ThF, may be us'ed, and the liquidus temperature of this mix-.

ture is 516°C (960°F),

ORNL-LR-DWG 37420R2

ThF4 1411
wo
TEMPERATURE N °C
1080 COMPOSITION IN mole 7
LiF-4ThE,
7LiF-6ThF, o
e LiF-2ThE, *00
3LiF-ThF4 sS
LiF-2The, ‘
950
P 897
2LIF'BEF2 900
7LiF‘GThl—; .
P 762 850
P 597 Sog
£568
3LIF-Th, Ao, >
£ 565° s 50
700
[+ 850
"3, O 033 5

Sy 2 &\ 50 £ 526

LF AVARN BeF

2

845 2LiF- BeFa/sToofatso 400 400 450 500 545

P 465 £ 370

Fig. 7.8. Phase Diagram of the System LiF-Ber-ThF4. |
100 —

1050 [—

1000 —

TEMPERATURE (°F)

850 —

800 ——

750 —/——

950 ——

900 ——

TEMPERATURE {°C)

593

566

538

510

454

426

148

ORNL-DWG 67— 4787

3400
/////,! 3000
////// 2600
LIQUIDUS TEMPERATURE ///,
,///////// 2200
/2/, ’ 1800
//// THORIUM

7 /// 1400
////// 1000
600

10 15 20 25 30

Tth (mole %)

Fig. 7.9. Thorium Content of LiF—Ber-ThF4 Compositions on the Even-
Reaction Boundary Curve L — 3LiF . Th F4(ss) +7LiF+6Th F4.

Toble 7.11.

MSBR Blanket Compositions

THORIUM CONCENTRATION (g / liter at 800 °C)

Thorium Concentration

'

‘Liquidus Temperature

Composition (mole %)

(g/liter) Degrees C Degrees F 'I‘hF4 LiF Bos:l:‘2
863 450 842 7 63 30
1207 469 876 10 66 24
1828 500 932 16 69 15
2419 550 1022 22.5 71 6.5
2868 568 1054 29 71

149

7.4 SEPARATIONS IN MOLTEN FLUORIDES

Removal of Rare Earths from Molten Fluorides by Precipitation on Solid UF,

F. A. Doss H. F. McDuffie J. H. Shaffer

Studies of the precipitation of rare-earth trifluorides from solution in LiF-BeF, (66-34 mole %)
on solid UF, have continued.2? This program is concerned with the retention of rare-earth fission
products on a bed of solid UF, as a possible method for reprocessing the fuel solvent of the
reference design MSBR. Experiments conducted thus far have examined the removal of selected
rare earths from the simulated fuel solvent upon addition of solid UF,. Results obtained fbr the
removal of NdF, are illustrated in Fig. 7.10. Material balance calculations of the system indicate
that the composition of the solid phase varied from 0.54 to 0.22 mole %_NdF3 in UF, as UF, was
added to the salt mixture. Subsequent changes in the concentration of NdF3 and UF, in the salt
mixture, as the temperature was varied from 550 to 800°C, corresponded to a constant solid

composition of 0.22 mole % NdF3 in UF, over the temperature region.

\

22MSR Program Semiann. Progr. Rept. Aug. 31, 1966, ORNL-4037, p. 145.

ORNL-DWG 67— 4788

N

N .

NEODYMIUM FOUND IN SOLUTION (mole fraction X10%)

'\
T
\.
d
0.4 : o —
®
WEIGHT OF SALT.MIXTURE: 1.{{kg

0.2

o]

0 0.2 0.4 0.6 0.8 1.0 1.2~

UFy ADDED (moles)

Fig. 7.10. Removal of NdF3 from Li|'=-BeF2 (66-34 Mole %) at 550°C by
Precipitation on-Solid UF3.
150

Extraction of Protactinium from Molten Fluorides into Molten Metals

D. M. Moulton W. R. Grimes
F. F. Blankenship J. H. Shaffer

~ Studies of the extraction of protactinium from the MSBR blanket salt (73LiF-2BeF -25ThF )
by reduction with thorium-bismuth amalgams have continued. In an experiment at 650°C in which
the salt originally contained about 600 ppm UF and a trace of PaF , thoriu:fi was added to the
bismuth in several successive small increments, and after equilibration the concentrations of Li,
U, Th, and Pa-in the bismuth and U and Pa in the salt were measured (Li and Th in the salt
stayed constant).

Letting Xm and Xs denote mole fractions in the metal and salt phases and with activity"

coefficients of unity, we can define a half-cell voltage for each metal:

RT X_
8:83—-r§lnxm, (D | -

8

where 8(1) is a constant to be evaluated later. Plotting In (X_ /X ) for each of the four metals vs
the amount of added thorium gave four parallel lines. This was taken to mean that the metals were
indeed in equilibrium; therefore, all the £’s were set equal to each other, requiring that the 8; be-
adjusted to fit. ‘ ‘ -

The stahdard half-cell potential 80,°f a metal is nomally defined for the reaction M = M?* + ne—,
where M is the pfire metal in the physical state appropriate to the temperature and M” " is the ion in
solution referred to some standard state. Baes?3 has tabulated 80 for a number of elements, using
as the standard state of the ion a hypothetical unit mole fraction solution in the MSBR fuel solvent
(extrapolated from dilute solution)l, except for Li and Be, for which the standard state is the Li BeF
mixture. These values are for the LizBeF4 fuel salt but are taken here as a reasonable approxi-
mation of the Li-Th-Be blanket salt. Any variation in the concentration of the ion in the salt can
be accounted for by writing '

RT 1

where Ve 18 the activity coefficient in the salt; thus 80 = & at unit ion activity ()/SXS = 1), and Ye =
1 at low concentrations, so 80 is, as stated, the voltage at an extrapolated hypothetical unit mole
fraction solution.

Highly electropositive metals dissolved in bismuth do not form ideal solutions at all, even at
low concentration; these are intermetallic compounds formed with substantial free energies. Never-

theless, preserving the form of Eq. (2), we can write

RT y X
= _ — |pnmm 3
o " 75 "X )

23¢. F. Baes, Jr., Thermodynamics, pp. 409—33, IAEA, Vienna, 1966.
151

Here y <<1 even in very dilute solutions.

At this point an alternative choice of standard states is useful. In the extraction proceés we
do not deal with pure metals but with amalgams exclusively, and we are concerned not with ac-
tivities but with concentrations. If we write (3) in the form of (1) having ,Xm/XS , the measfifed

quantity, as the independent variable, we find

E=F —-—ln—-———ln-X—m. (39

Since we are dealing with solutions of constant composition (Li and Th in the salt) or dilute (all
the others), all of the y’s may be taken as constant with composition. Then we can define an
alternate standard potential 83: '

RT Ym

SR CR— P ,
0 0 e Y,

f

and suBstituting this into (3 ’)V we obtain (1), We see at once that 8; is the voltage of the half-
cell when X = X_ . Here the standard states or hypothetical unit mole fraction solutions have th‘e_
properties cortesponding to infinite dilution. To evaluate the 8(1)’5 we need know 80 and y_/y, for
one of the four metals and just the ratio /Xm/XS'for the others. This will enable us to construct an
electromotive series for the system amalgam-salt which will ificlude protactinium, for which we do
not know y or 80 . It should be streséed that . and 8; are alternate ways of accounting for the
same phenomenon, the formation of intermetallic c.om“pourids with bismuth. |
For lithium we have y = 9.8 x 10-5.24 Assuming that this salt acts like LizBeF4, we say
that y = 2.2 [ extrapolating from (1)], which gives 8(1) (Li) = 1.803. Now we may use this to find

the other Sé’s, which are shown in the accompanying table.

‘ Orry & 1b c
M ~& (923 SR -&l ¥,
Li 2.601. 1.80 9.8 x 1073
Th ' 1.772 1.47 4% 10”7
\
Pa _ 1.32
~ U 1.368 ©1.28 4x10°

8—80 (v) at 923°K is half-cell potential for pure metal > ion at hypothetical unit mole fraction in
Li, BeF, (except to 0.67 mole fraction for Li).

b—gé (v) at 9239K is half-cell potential for amalgam - ion in LiZBeF4 at Xm = Xs .
Cym is activity coefficient used to describe activity when referred to standard state described by 80.
. - :
Values of y_ can be estimated for Th and U. The salt concentration of uranium is low, so y,

may be set equal to 1. A rather crude approximation to y_ for Th is an extrapolation of y_ for U in

z

24Slightly changed from ORNL internal memorandum of D. G. Hill to W. R. Grimes, June 29, 1966.
152

LizBeF“',23 giving y_ = 7.5, Using these and the tabulated values of 80 , we find y_ as shown.
Interpolation of values obtained at Brookhaven?® gives y_ for Th and U as 10" and 1.7 x 10~*
respectively. An error of an order of magnitude in ym/ys for Li gives four orders of error in y for
Th and U, so considering the approximations in the calculations, the agreement is reasonable. On
the other hand, 8(1) changes by the same amount for all valences, so that the difference in 8(1), which
is the quantity most directly related to the extraction efficiency, is not affected by errors in y_ .
Therefore, although the absolute values should not be taken too seriously, they indicate fairly

well the relative ease of extraction of the four elements into bismuth.

Extraction of Rare Earths from Molten Fluorides into Molten Metals

D. M. Moulton - W. R. Grimes
F. F. Blankenship J. H. Shaffer

As part of the program of removing lanthanide fission products from MSBR fuel solvent by re-
ductive extraction into liquid metal, the study of the activity of lithium in bismuth amalgams has
been continued. The method used has been the equilibration of amalgam-salt sys.tems with metallic
beryllium, which is insoluble in bismuth and thus has a known unit activity. An analysis of'déta
reported earlier?® reveals that, regardless of whether the amalgam was prepared by adding Li or
IBe to the bismuth-salt system, only about two-thirds of the reductant metal appeared as amalgé-
mated lithium. Because of the difficulty in accurately measuring very small changes in salt com-
position, mass balances were made on the basis of the metal phase only, so that it was not pos-
sible to tell whether the missing one-third was due to some systematic experimental error or
represented an actual reduction process. | '

The use of barium as a reductant has allowed both salt and metal phases to be analyzed ac-
curately. In this experiment Ba was added on top of the salt, through which it sank and dissolved
in the bismuth. A similar procedure had been used for Be (which floats, however), while Li had
been added through a dip leg directly to the bismuth. Here again it was found that only about 60%
of the Ba could be accounted for by Ba and Li in the bismuth. Now, however, the lost barium all
appeared in the salt phase, with the mass balance generally 95 to 105%. It is clear, therefore,
that some unexpected reduction process was going on.

The equilibrium constant for the reaction
Ba(m) + 2Li* = Ba?" + 2Li(m)

is found to be 0.016, corresponding to a AF® of +7.14 kcal/mole [ assuming y(Li+) - 1.4,
y(BaZ+) =1]. 8; for the reaction Ba(m) = Bal?t + 2e~ is —1.63 v, using 8(1) of Li as —1.80.

All of these values are based on the amalgamated metal at unit mole fraction as the standard and

25Brookhaven National Laboratory Nuclear Engineering Progress Report, July—September 1958, January—
April 1959; Annual Report, July 1, 1956. -

26J. H. Shaffer et al., MSR Program Semiann. Progr. Rept. Aug. 31, 1966, ORNL-4037, p. 142.
153

not on the pure element. By comparison, 80 for Be is —1.72,%3 and 80 for Ba is probably about
—-2.2 to —2.3. '

The only'redugible material present in sufficient quantities to explain this behavior is beryl-
lium ion. Excluding for the present the possibilities of Be* and Be® dissolved in the salt, metal-
lic beryllium can be present only at unit activity. Thus, at equilibrium a lithium afialgam mole
fraction of about 0.18 can be reached before any beryllium is formed, while Ba amalgam should not
reduce Be2* at all. However, one can imagine the formation of Be either by the metallic Ba sink-
ing through the salt or by Li amalgam at locally high concentration before it is thoroughly mixe;d.
If, then, any Be so formed, or Be introduced as such, should be physically disposed in such a way
that it no longer can be made to contact the bismuth'by agitation, it will never come to a true
equilibrium. For instance, since Be floats, it may splash onto the container walls and not drain
back. It is intended to test this explanation by trying to observe beryllium on the container walls,
but this had not been done at the time of this writing. At least, it can be shown that the amount
of lost reductant is roughly proportionJal to the amount added, which would be expected if this is
the case. Unfortunately, some other old work, in which, when salt was contacted with an already

equilibrated amalgam, more Li disappeared,?® is still not explained.

Protactinium Studies in the High-Alpha Mol ten-Salt Laboratory
C. J. Barton H. H. Stone

Data given in the previous progress report?’ showed that a large fraction of the protactinium
dissolved in molten LiF-ThF, (73-27 mole %) is converted by exposure to solid thorium to a form
that does not bass through a sintered copper filter. However, more than half the reduced protac-
tinium remained suspended in the molten salt. It was also réported that efforts to convert the dis-
solved protactinium into a more readily recoverable form by electrolytic reduction had been un-'
successfu}. Further studies of the reduction process and other potential recovery methods are

briefly summarized here.

Electrolytic Reduction. — Three experiments were performed in an effort to transfer protactinium
from molten LiF-ThF4‘ (73-27 mole %) to liquid bismuth by electrolysis. The fraction of the prot-
actinium reduced to a form that would not pass through a sintered copper filter varied from 0.5 to
0.95. In all three cases, -analysis of samples of the bismuth layer that had passed through sintered
stainless steel filters indicated that only a few pé_rcent of the original protactinium content of the
salt phase was dissolved in the molten metal. Because of lack of encouraging results, we have

temporarily discontinued the electrolytic approach to protactinium recovery.

Thorium Reduction in @ Bismuth-Containing Melt. — An experiment was performed to determine
whether transfer of reduced.protactinium from the salt phase to bismuth would be improved by

carrying out the reduction in the presence of dissolved bismuth. This experiment was carried out

27C. J. Barton and H. H. Stone, MSR Program Semiann. Progr. Rept. Aug. 31, 1966, ORNL-4037‘, p. 156,
154

in a niobium-lined nickel pot equipped with a niobium dip leg. The 23!Pa, 233Pa.(in irradiated
LiF-ThF4) , and 0.55 g of E’.i:!O3 were m@xed with LiF-ThF4 (73-27 mole %) and given a 30-min
treatment with mixed HF and H,. After adding bismuth metal to the molten salt, the salt phase
was exposed to a freshly cleaned thorium rod, approximately 1/4 in. in diameter, for three consec-
utive 45-min periods. After the third exposure the protactinium concentration had been reduced to
about 12% of the initial value. _

The crusts collected from the three thorium rods used in this experiment (13.0 g total) con-
tained approximately 6.5 g of bismuth, far more than could be accounted for by the change in
bismuth content of the filtered samples. The biSmuth. recovered as a separate phase weighed 51 g,
as compared with the calculated value of 89 g. Further experimentation will be required to provide
a satisfactory explanation of these observations.

Thorium Reduction in the Presence of Iron Surface (Brillo Process). — Shaffer reported?8® that
when a tracer quantity of *3*Pa precipitated from molten LiF-BeF -ThF , (73-2-25 mole %) in the
presence of a large amount of iron surface in the form of steel wool (Brillo), most of the 233pga
activity remained with the steel wool when the salt drained away. We have performed a series of
experiments that confirm Shaffer’s finding with 2*'Pa concentrations in the approximate range 30
to 100 ppm of 231Pa.

The experiments were performed in three steps. A weighed quantity of previously purified
LiF-ThF, (73-27 mole %) was mixed with 231Pa and 233Pa in a nickel vessel, treated with mixed
HF and H, to remove oxides, and then treated with H2 to reduce NiF2 to metal and PaF5 to PaF4.
A weighed amount of grade 00 steel wool (0.068 m?/g surface area) was placed in a mild steel
liner which was enclosed in a welded nickel vessel equipped with a stainless steel dip leg. A
stream of purified hydrogen flowed through the vessel while it was heated to 800 to 820°C for
about 4 hr and then cooled in flowing helium. The dip legs of the two vessels were then con-
nected by a short transfer line and heated to about 650‘5C. Helium pressure applied to the salt -
vessel transferred the salt to the steel-lined vessel, and helium continued to flow through both
vessels to provide mixing of the molten salt. A thorium rod, attached to a l/g'ifl- nickel rod, was ' .
inserted in the salt with its lower end about 1/4 in. above the bed of steel wool. After the thorium
was exposed in the salt for the desired length of time, it was withdrawn, and a filtered sample of _ -
the salt was obtained with a sintered copper filter stick. This procedure was repeated, usually
with a clean thorium rod, until a gross gamma count of 233Pa in the filtered sample showed that
most of the protactinium had been removed from solution. The vessel connections were then re-
versed to transfer the salt back to the nickel vessel. Both vessels were cooled with helium flowing'
through them.

After it had reached room temperature, the steel-lined vessel was cut through with a small saw
and the contents of the liner, consisting primarily of a ball of salt enclosing the steel wool, were

removed and weighed. This material was crushed and, in some experiments, separated into mag-

28_]. H. Shaffer, ORNL intemal memorandum to W. R. Grimes, Nov. 15, 1965.
155

netic and nonmagnetic fractions by use of a small magnet. A sample of the unfiltered salt in the
nickel vessel was also removed for analysis, Alpha pulse-height analysis was used to determine
the 2°!Pa content of all samples, ,

We found that thorium exposure of 1.5 to 2.2 hr reduced 95 to 99% of the protactinium to a form
that would not pass through the sintered copper filter medium. From 93 to 98% of the protactinium
remained with the steel wool, along Iwith about 20% of the salt initially present in the experiment.
A surprisingly large amount of iron was found in the filtered salt samples before the salt was ex-
poséd to the steel wool. The source of this iron has not been determined, but analysis of the
filtered samples indicated that, in general, the iron content of the salt decreased as the protactin-
ium was reduced by exposure to thorium. This suggested the possibility that iton coprecipitated
with the reduced protactinium was carrying the protactinium to the surface of the steel wool. We
performed a pair of experiments to test this theory in which we attempted to hold all- the variables
constant except the iron/protactinium ratio. The results were not conclusive because répeated
analyses of the salt samples have shown such a large scatter in the iron values that it was dif-
ficult to calculate reliable iron/protactinium ratios. We are currently conducting experiments with
S9Fe tracer in an effort to establish the role of dissolved or colloidal iron in the retention of re-
duced protactinium by steel wool and to test the reliability of the currently used analytical pro-
cedure for the detemmination of iron in fluoride,samples. Additional experiments will be required
to establish the important variables in the Brillo protactinium recovery process and to explore
modifications that would make it more readily adaptéble to a large-scale process.

Conclusions. — The results of our efforts to transfer protactinium from a molten breeder blanket
mixture to liquid bismuth, either by electrolytic reduction or by exposure of the salt to thorium
metal, have not been favorable in the small-scale glove-box experiments conducted to date,
Nevertheles-s, the encofiraging results obtained in tracer-level experiments with protactinium and

rare earths suggest that continued studies are desirable.

Preliminary Study of the System LiF-ThF ,-PaF,

C. J. Barton . H. H. Stone G. D. Brunton

In the protactinium recovery studies described in the preceding section of this report, it has
been generally assumed that protactinium is present in the LiF—ThF; (73 mole % LiF) melts as
PaF4. This assumption has seemed plausible since the melts have, in every case, received a
treatment with H, at temperatures near 650°C, and H, is known to reduce pure PaF _ to PaF  at
much lower temperatures, (The Pa*" state seems to be the lowest known in fluoride systems.)
We have, however, conducted a few preliminary experiments to test this assumption and to see if
the phase behavior of PaF , is similar to that of ThF in mixtures with LiF.

~ About 100 mg of .231PaF4 was prepared by evaporating a measured portion of purified stock
solution (9 M in HF) to dryness in platinum and heating the residue to 600°C in flowing HF-H,
mixture. Conversion to PaF, was confirmed by .weight and by the brown color of the material. A

portion of this material was mixed with LiF.-ThF4 (73 mole % LiF) to yield a mix with 68 mole %

-
156

LiF and 32 mole % (Th,Pa) F,. Another portion was mixed with LiF and the LiF-ThF , mixture to
yield a mix with 73 mole % LiF and 27 mole % (Th,Pa) F4. Both mixtures were admixed with
ammonium bifluoride (whose decomposition products on heating help to minimize possible
hydrolysis), heated to 650°C, and cooled slowly. |

Examination o-f the slowly cooled melts showed that segregation of PaF -rich phases from the
bulk of the LiF-ThF material occurred in both cases. Material from the mixture with 68 mole %
LiF is believed to be a solid solution of LiPan in LiThFS. One of the phases from the sample
with 73 mole % LiF is believed, because of its similarity to the analogous uranium compound, to
be Li4PaF8.‘ The PaF does not appear isomorphous with ThF ; the LiF-PaF, system may, in
fact, be more like the LiF-UF , than the LiF-ThF , system. Itis obvious that study of the binary
LiF-PaF  system is needed before attempting further deductions conceming phase relations in
the ternary system LiF-ThF4-PaF4.

‘A portion of the LiF-ThF -PaF mixture with 73 mole % LiF was transferred to a small thorium
crucible and heated to 650°C in a helium atmosphere. Examination of the material with the polar-
izing microscope revealed some Li , ThF_, but a large part of the mixture was in the form of opaque
angular fragments, which are probably protactinium metal. X-ray examination will be required to
confirm this conclusion. '

~

7.5 DEVELOPMENT AND EVALUATION OF ANALYTICAL METHODS
FOR MOLTEN-SALT REACTORS

Determinations of Oxide in MSRE Salts

R. F. Apple J. M. Dale A. S. Meyer

The analyses of oxide in radioactive salt samples from the MSRE are summarized in Table

7.12.

Table 7.12. Oxide Cencentrations of MSRE Fuel-Salt Samples

Oxide Concentration
Sample Date Received '

(ppm)
, FP-6-1 47-66 ' 49
FP-6-4 ‘ 4-14-66 | 55
FP-6-12 . 5-10-66 | 50
FP-6-18 5-25-66 47
FP-7-5 6-22-66 66
FP-7-9 7-4-66 59 }
FP-7-13 7-15-66 ' 66
FP-7-16 7-22-66 56
FP-8-7 : 10-14-66 44

FP-9-2 11-9-66 44

157

The abrupt decrease in concentration between samples FP-7-16 and FP-8-7 coincided with the
change from copper to nickel sampling ladles and méy reflect an oxide blank of about 15 ppm. The
overall standard deviation of these results, 8 ppm absolute, includes any variation in the oxide
content of the fuel in the reactor and in oxide contamination during sampling and transfer. The
results of standardization with 5nO, samples?® performed during this period indicate that the pre-
cision of measurement of the water evolved by hydrofluorination is about 3 to 4 ppm absolute.

During the last week of December the oxide apparatus became inoperative. Because of hy-
drogen reduction experiments being pe‘rformed in the same hot cell, it was not possible to enter the
cell and investigate the equipment until early February. It was found that the electrolytic moisture-
monitor cell was not functioning, and it was removed from the apparatus. With the current MSRE
sampling schedules it should be possible to install a new cell with an improved coating®? ‘and re-
sume oxide determinations of radioactive salt samples during the month of March.

This replacement will be the first major maintenance required for the hot-cell apparatus since
its installation in February 1966. The moisture-monitor cell had been in operation for several
additional months during laboratory tests of the apparatus. This one-year operation represents a
reasonable service life for a moisture-monitor cell, even under normal operating conditions in a
nonradioactive environment.

During the period when the hot-cell equipment was inoperable, the oxide development apparatus
in Building 4500S was réactivated. Several samples of nonradioactive fuel and solvent salt from
the second ORR molten-salt loop were analyzed for oxide content. The results of these analyses
are summarized in Table 7.13.

It is probable that the high oxide levels found in the third and fourth samples were in part due
to the contamination during the brief period of exposure when the crushed samples were transferred
to the hydrofluorinator. When possible, future samples from this source will be melted into a nickel

ladle in a dry box.

294SR Program Semiann. Progr. Rept. Aug. 31, 1966, ORNL-4037, p. 192.
30, . ’ . o
Ibid., p. 193. -

Table 7.13. Oxide Concentrations of Fuel and Solvent Salt Samples from the
Secand ORR Molten-Salt Loop

Salt Condition Sample Weight Oxide Concentration
Sample Designation as Received (g) (ppm)
19-29-66 Pellets ' 10.4 200
10+17-66 Pellets 10.1 l 220
11-21-66 Crushed : 18.8 420
F-194 Crushed ' 18.4 820
Solvent salt batch No. 2 Fused into ladle 48.4 115

Solvent salt batch No. 17 Fused into ladle - 13.1 520

- 158

The increase in oxide between the last two samples represehts the oxide pickup when the flush

salt was circulated in the loop.

Determination of U3*/U** Ratios in Radiocactive Fuel by a Hydrogen Reduction Method

J. M. Dale - R. F. Apple A. S. Meyer

A proposed explanation of the unexpected distribution of certain fission products in the MSRE
system is that the fuel had become sufficiently oxidizing to produce significant partial pressures
of volatile fission product fluorides such as MoF , TeF, and RuF which then migrated into the
graphite and the blanket gas. Because the accumulation of fission products in graphite is vitally
important to breeder reactors, a method for the determination of U? fruet ratios in the radioactive
fuel samples was needed.

The possibility that a significant fraction of the iron and nickel is present in the fuel as
colloidal metal.particles3! precluded any adaptation of the hydrogen evolution method for U’ *
because these metallic components would also yield hydrogen on acidic dissolution. An alternate
approach, suggested by C. F. Baes, is based on a transpiration method in which a sample of the

molten fuel is sparged with hydrogen to reduce oxidized species according to the reaction

n—um

ME_  + H,— MFr.n + (n — m)HF ,
in which MF  may be UF,, NiF, FeF ,, CrF,, or UF4 in order of their observed reduction po-
tentials, The rate of production of HF is a function of the ratio of oxidized to reduced species in
the melt. |

Some components of the oxide apparatus3? were adapted for the transpiration measurement on

' is connected to a source of

radioactive fuel samples. The inlet of a modified ‘‘hydrofluorinator’
thoroughly dried hydrogen and helium (see Fig. 7.11), and the outlet is connected to a heated
manifold fitted with six liquid-nitrogen-cooled NaF traps. This arrangement permits the collection
of the HF from four successive reduction s’teps with hydrogen and two blank spargings with helium
on the same 50-g fuel sample. When the transpiration run is completed, the traps are disconnected
from the manifold and removed from the hot cell for desorption and titration of the collected HF.
Interpretation of these titrations in terms of U3*/U* ™ ratios in the samples is made by com-
paring the HF yields for the reduction steps with those calculated for hypjothetica.l salt compo-
sitions. If U*" is the only reducible species in the melt, a derivation based on a material

balance and the hydrogen reduction equilibrium yields the rather simple relationship for U3 ¥

as a function of hydrogen exposure:

1 v 3+
Uln -IT—:E';'; - U =Bt: . (1)1

31MSR Program Semiann. Progr. Rept. Feb. 28, 1966, ORNL-3936, p. 162.
321bid., pp. 154—62.
in which

U, udt

159

ORNL-DWG. 67-226

HYDROGENATOR

LIID
He NITROGEN
TRAPS

ACCESS AREA S HOT CELL
N s © o
N B HEATED MANIFOLD
§ (ORI O OINONONONO!
N
\ U U UUUU
N NoF TRAPS FOR COLLECTING HF Mgfigugénongnr
N
\

Fig. 7.11. Diagram of the Hydrogenation Apparatus for Determination of
u3*/u4* Rotios in MSRE Fuel.

’

are concentrations of total and trivalent uranium expressed as mole fractions

H

. . . + , .
t is the reduction time from ““0’’ U®" concentration, min,

1/2
) KP-H2 Vo
SRT
where
3+
U PH atm 1/2
U4 + P1/2 ’ !
H)
P!_l = partial pressure of H2, atm,
PH = partial pressure of HF, atm,
Vs = purge flow, liters/min,
S = quantity of fuel sample, moles,
R = gas constant, liters atm (°K) ~! mole—1,
T = temperature at which the sparge flow is measured, °K.
When corrosion products (M].2 *) are present, the relationship assumes a different form as each

corrosion product ion successively undergoes reduction to the metal:

~

in which

a. =
1

a4 U-a U j=n K2 1 U
anU-—U + Uln W + 4 E K2 —_— =Btn', (2)

3+
a
=1 n U .

1
1+ (Ky /K) M2T1/2
J

)
160

sz * = concentration, mole fraction, of the jth corrosion product in order of reduction, at
t =0,
n

t., min, is measured from the instant reduction of the nth corrosion product starts,

K PHF

M. =
i M.2+ 1/2 p1/2
(M) Py

The yield of HF from any reduction step is calculated from the initial and final concentrations

{
of U_3+ and corrosion products as follows:

. j=n ,
‘micromoles of HF ={ (U2F - U3") +2 ¥, [(M].”)i — M2M)| S x10°. ) ;
i=1

The relationship of Egs. (1) and (2) is illustrated in Fig. 7.12', in which the yield of HF is
plotted as a function of Bt for an oxidized fuel containing U* * and Fe2" in concentrations ap-
proximating the elemental analyses of the MSRE fuel. On a logarithmic plot the accumulated HF
yield follows a near-linear relationship [ Eq. (1)] derived for an iron-free melt until a critical
U3t/U*? 1atio is reached and reduction of Fe2” starts. The reduction of Fezf’ then predominates,
with the rate of change of U? * reduced in accordance with Eq. (2) until substantially all of the

iron is reduced and the HF yield approaches that derived from uranium reduction alone.

A computer program is being developed by the Mathematics Division to calculate HF yields
for any preselected reduction steps on any melt composition. This will permit optimization of
reduction steps for best distinction of U*/U*? ratios. In the absence of such a program it was
possible to select four reduction steps on the criteria that the first two steps at 496°C would re-
" duce at least 99.9% of the Fe2' and Ni2 ¥ without materially affecting the U3* concentration and
that the last two steps at 596°C would reduce U™ without reducing Cr? *. The latter reductions
thus serve as an indication of the ‘‘oxidation state’’ of the fuel. Hydrogen fluoride yields'frorr‘x
sample FP-9-4 and from various hypothetical initial fuel compositions are plotted in Fig. 7.13.
Broken lines connect calculated yield points, and a solid line connects the experimentally measured
HF yields from the sample. The relatively low yields of HF in the first two steps are in agreement
with earlier experimental evidence®! that only a small fraction of the iron and nickel is present in
the ionic state. The high yields in the last two steps indicate that the fuel is in oxidizing or near-
oxidizing condition. On the basis of the HF yield of step III and in consideration of estimated ex-
perimental errors, a ratio of U3t/U*? of 0.000 to 0.001 was aséigned. Sample FP-10-25, taken after
the addition of sufficient beryllium xfietal to increase the U3* /U*" ratio by 0.0037, was analyzed
similarly, and the HF yield at step III corresponded to an applroximate U3*t/Uu*" ratio of 0.005.
This represents. an increase of 0.004 to 0.005 over sample FP-9-4, in reasonable agreement with
that calculated from the beryllium reduction. Hydrogen fluoride yields from sample FP-11-5, taken
after a possible eiposure of the fuel to air, agreed within estimated experimental error with sample

FP-10-25. |
161

ORNL - DWG. 67-122

1000 3
F INITIAL .
_ CONDITION .
B % 143 o +2 . T
- No(f,oL(J)), %Fe2)  MSRE FUEL SALT, 1.22 moles i
] : \ X5y =80x1073 1
\\ | Xspe=11x 1074
100 \ E
2 \ ]
C \ N
- \ :
\
- \ T
w
he SAMPLE ;
2 10} f E
s [ B(0,0) \ s X 3.
X I c(o.oz,o.i)\:;_—:_,::_‘ / % ]
n \‘1:':.._," / ! 7
- G0N0 — 4
/
= I'// 3
‘E // ;
b " N
- D(O.B,<O.1)—-—'_(/ -
R — o )
: 2 @ 2 8
< < o s
u ~ ~
01! g - m v
I il I T

STEP DESIGNATION (sparge time, min atm"z; melt temperature)
Ho Flow, 200 cc/min at 27°C, tatm

Fig. 7.12. Calculated Yields of HF from the Hydro-
genation of a 50-g Fuel Sample at 1000°K.

ORNL-DWG. 67-224

T L] lll!lll L) T T'l]‘]' ] ¥ ‘|l|‘l|i ¥ 1 ] llllll T LB A
1000 - Hypothetical Initial C_osnditions —
: Xs*4 = 8.03x 10 3
i XEU+3 =0 //:
. -4
- XZFe+2 =115 x 10 -
g 100 i
E s Total HF :
A i 4
N - HF from U*4if 1
< 5 no Fe*2 were present
10E- , E
- HF from U*% :
[ / ]
1 1 o lllllll 1 1 illlll‘ 1 1 Illllll 1 1 lllllll 1 J b i1)1
0.4 1 . 10 100 1000 10,000

(Bt +1.4)x 107, minutes

t

Fig. 7.13. Calculated and Observed Yields of HF from MSRE Fuel
Somple FP-9-4,
' 162

A sample (IPSL-21) taken from the second ORR molten-salt loop has also been analyzed and
found to contain about 1% of the uranium as U®*, Because the composition of this sample is in
the region in which Cr? *is also reduced, a more accurate estimation of U®*/U* * ratios must be
deferred until the computer program is available. _

Although tentative analyses by this technique are béing reported, it should be recognized-that
the method is still in the developmental state, and adjustment of the reported values may be nec-
essary as experience is gained. The agreement in the observed increase in U? */U** ratios with
that predicted from the beryllium addition supports the validity of the method. Pregliminary calcu-
lations indicate that it may be possible to apply the method to continuous in-line analysis of the
fuel for the determination of both U3 */U* ¥ ratios and significant ionic concentrations of corrosion

products.

EMF Meoasurements on the Nickel-Nickel(ll} Couple in Molten Fluorides

D. L. Manning H. W. Jenkins33 Gleb Mamantov34

This study was initiated to evaluate, by potentiometric measurements on concentration-type‘
~cells, several metal-metal-ion couples for use as possible reference electrodes in molten fluo-
rides. The approach is similar in principle to the utilization by Laitinen and Lui35 of the Pt/Pt2*
couple as a reference electrode in molten chlorides.

The apparatus used to contain the melt is a small -vacuum dry box (24 in. long, 20 in. deep,
and 15 in. high) which is outfitted with a furnace, vacuum and controlled atmosphere facilities,
and a moisture monitor, Utmost care is exercised to protect the molten fluorides from moisture
contamination because of the affinity of the molten fluorides to moisture and the resultant pre-
cipitation of oxides. The melt (~400 ml) is contained in a graphite cell; a spiral nickel electrode,
platinum stirrer, Pt—10% Rh thermocouple, and the inner electrode compartment are positibned in
the melt. The inner compartment of the cell consists of a thin-walled boron nitride compartment
which contains the same fluoride melt.and a fixed concentration of dissolved NiF,. A nickel
electrode is inserted into this cohpartment. Hot-pressed boron nitride is an insulator in molten
fluorides but is slowly penetrated by them. As a result of this effect, BN can be utilized to
separate the two half-cells and yet achieve electrical contact, because such contact is made when
the BN becomes wetted by the melt. For molten LiF-NaF-KF, benetration of the BN (1/3 ,min. wall
thickness) occurs within a day or so; however, it is much slower for LiF-BeF ,-ZrF, melts. Ap-
proximately a week is required for the resistance through the BN to drop to ~ 1000 ohms or less.

With stirring, weighed portions of NiF, were added to the melt. After each addition, emf
measurements were made and a sample withdrawn for nickel analysis. From the equation AE =

(RT/2.3nF') log (X'l/Xz) and plots of AE vs log X, , it was demonstrated that the nickel system

3?'University of Tennessee, Knoxville.
34 . . .
Consultant, Department of Chemistry, University of Tennessee, Knoxville,
35 i ’
H. A. Laitinen and C. H. Lui, J. Am. Chem. Soc. 80, 1015 (1958).
163

exhibits reasonable Nermstian behavior at 500°C. The concentration 6f nickel varied from approxi-
mately 10=% to 103 mokle fraction. From the standpoint of Nemstian reversibility, therefore, the
nickel couple appears to be a good.choice for a reference electrode in molten fluorides.

Stability studies conducted so far are encouragihg. For a run in molten LiF-NaF-KF, the emf
remained constant to within 2 mv over a two-week period. These studies are continuing, and

longer stability tests are planned.

“

Studies of the Anodic Uranium Wave in Molten LiF-Ber-ZrF4

D. L. Manning

Additional observations were made of the anodic wave at ~ +1.4 v vs a platinum quasi-
reference electrode in fluoride melts that contain U**, It is tentatively postulated that the wave
is due to the anodic oxidation of U** to U*. The reduction of U** to U3* is observed at ap-
proximately — 1.2 v. For simple charge-transfer processes, fhe wave heights for the two electrode
reactions should be thé same for the same rate of voltage scan. The anodic wave, however, is
larger than the cathodic wave when a platinum—10% rhodium electrode is used. This suggests
that the anodic wave may be a charge transfer followed by a catalytic reaction, that is, dis-
proportionation of U5 ™. The observations, however, were not the same when the waves were
recorded at a Pt—40% Rh electrode. At this electrode, the anodic and cafl;odic wave heights were
about the same. It would thus appear that the electrode material itself may be influencing the
magnitude of the anodic wave, that is, acting as a catalyst for the disproportionation of us®.
Also, the possibility that the products of the electrode reaction (U® +,_U6+, etc.) are attacking the
electrode should not be overlooked. These studies are continuing, and it is planned to test other

electrode materials,

Spectrophotometric Studies of Molten-Salt Reactor Fuels

J. P. Young

The determination of changes in the ppm concentration level of U3" in molten fluoride fuel
salts which contain U*” is one of the expected applications of the proposed in-line spectropho-
tometer3® for the MSBR. In cooperation with F. L. Whiting®*?® and Gleb Mamantov,3* the volta-
metric generation and simultaneous spectrophotometric observation of ppm concentrations of U3*
are being investigated in LiF-BeF ,-ZrF , melts which contain 2% w/w U**. Such studies are
carried out in a nickel captive-liquid cell*’ which was modified to include a three-electrode
voltametric system. The cell is designed so that spectrophotometric measurements can be made
in the immediate region of the indicator electrode during tfie time that a redox reaction is occurring,

The cell holds about 1 ml of molten fluoride salt solution. A preliminary description of the cell

36MSR Program Semiann. Progr. Rept. Aug. 31, 1965, ORNL-3872, p. 145.
37y, P. Young, Anal. Chem. 36, 390 (1964).
164

and its operation has been published.3® The results obtained have demonstrated that changes of
50 ppm in the U3* concentration could be easily seen in a 1-cm path length of molten salt which
containé 2% w/w U** and.up to ca. 500 ppm U3.+. This does not imply that the method would not
work at higher concentration levels of U3 *. rather, this is the highest concentration level that was
studied experimentally.

Several spectrophotometric determinations of U* *in LiF-BeF, and LiF-BeF, containing ca.
4% w/w ZrF, or ThF, have been carried out using the very sensitive U* * absotption peak at 235
nm.3? Transparency of fluoride solvents in the region around 235 nm is a necessary requirement
for this type of determination; therefore, the optical behavior of fluoride solvents of interest is
under study. Moderate absorbance has been observed in the region of 235 nm in LiF-BeF, melts
which contain ca. 10% w/w ThF4.' It is believed, however, that this absorbance is caused by a
contaminant in the ThF , that was added.

The study of the spectra of solute species which might interfere with the proposed spectropho-
tometric determination of U:"'_+ in molten fluoride salts was continued. The spectra of several
trivalent rare-earth ions in molten LiF-BeF , at 550°C were recorded. Although Er** and Sm®”* ‘
exhibit absorption peaks at 377 and 400 nm respectively, in the vicinity of the U3 * absorption at
360 nm, the molar absorptivity of these absorptions is under 10 and no interference would be ex-
pected. The spectrum of Ho®" in LiF-BeF , was also obtained. This solute species likewise °
exhibits relatively weak absorbance; the molar absorptivity of its most intense absorption peak,
452 nm, is less than 10. All of the absorption peaks observed for these three rare-earth ions,
over the wavelength range of 250 to 2000 nm, are transitions within- the f orbitals and would be

expected to be relatively weak.

7.6 ANALYTICAL CHEMISTRY ANALYSES OF RADIOACTIVE MSRE FUELS
| F. K. Heacker C. E. Lamb L. T. Corbin
The remote apparatus for determining the oxide content of MSRE salt samples was modified to

pemit measurement of the oxidation state of 50-g fuel samples. The method is based on the

equation:

n—m
MF. + > H2—>MFm+(n—m)HF,

n

where MF may be UF_, NiF , FeF , or UF .- The reaction is driven to the right by sparging the
molten sample with hydrogen. The liberated HF is collected on NaF traps cooled by liquid nitro-
gen. After the HF from the hydrogenation steps is collected, the traps are removed from the hot

cell, and the HF is desorbed at 300°C and titrated. The oxidation-state measurements were made
in conjunction with beryllium additions to the MSRE fuel salt. To date, three samples have been

analyzed.

38]. P. Young, Gleb Mamantov, and F. L. Whiting, J. Phys. Chem. 71, in press.
39MSR Program Semiann. Progr. Rept. Aug. 31, 1966, ORNL-4037, p. 195.
165

‘Sample Analyses.

In addition to the oxidation-state éamples, several other special samples were analyzed. These
included MSRE off-gas samples, beryllium addition samples, and six highly purified LiF - BeF,
samples.

For the past two years, beryllium determinations made on radioactive MSRE fuel-salt samples
have exhibited a 2% positive bias from book values. In an attempt to locate the source of the
bias, six highly purified LiF - BeF, samples were submitted to the High-Radiation-Level Analytical
Laboratory. Three of the éamples were fused at 700°C to simulate the conditions of the fuel salt. -

The results obtained on these samples are shown in Table 7.14.

Table 7.14. Results Obtained on Analysis of Highly Purified Lithium-Beryllium
Fluoride Samples

Weight Percent Beryllium

Sample - ‘ Difference (%)
True ’ Found?®
LB 1 3.05 3.07 0.6
LB2 6.98 7.01 | +0.43
LB 3 10.19 10.11 —-0.79
RT 17 13.76 13.74 ~0.15
RT 2° ¢ 15.13 15.10 —~0.20
RT 3° 18.06 17.95 —0.61

@yalues listed as found are average values calculated from approximately 20 determinations.
bsamples fused at 700°C.

Although the individual determinations exhibited a +2% spread, no significant bias is evident
from the data shown in Table 7.14. In addition to the above results, determinations made on syn-
Vth-etic solutions similar to dissolved MSRE fuel-salt samples show no appreciable bias in the
method.

From July 1, 1966, through December 31, 1966, 48 routine salt samples were analyzed as shown
in Table 7.15. Several of the MSRE fuel-salt samples were submitted with silver and INOR-8 wires
coiled around the stainless steel cable between the latch and ladle. The latch, wires, and cable

were separated and prepared for radiochemical analyses.

Quality Control Program

A surhmary of the MSRE control results for the third and fourth quarters of 1966 is shown in
Tables 7.16 and 7.17. The values shown in Tables 7.16 and 7.17 are a composite of the values

obtained by four different groups of shift personnel.
166

Table 7.15. Analysis of Radioactive MSRE Fuel Samples from July 1, 1966,
Through December 31, 1966

Number of Analyses Made

RCA% Msa®?
U Zr Cr Be F Fe. Ni Mo Li Prep Prep Oxide
38 38 38 38 38 38 38 2 38 38 8 6

4Radiochemical analysis,
bMasS spectrographic analysis.

CTwo 50-g oxide samples are being held for future analysis, and two were lost due to a malfunction in
the oxide apparatus moisture monitor. ’

Table 7.16. Summary of Control Results for July, August, and September 1966

28 (%)

Determination and Determinations . Bias
Method Made Fixed " Found (%)
Be photoneutron 21 5.0 2.75 +0.43
Cr amperometric 35 15.0 - 14.69
Fe spectrophotometric 33 15.0 6.65
Ni spectrophotometric 33 15.0 8.89 +5.21
U coulometric 24 1.0 1.90
Zr amperometric 26 5.0 4.47
Table 7.17. Summary of Control Results for October, November, and December 1966
Determination and Determinations 25 (%) l Bias
Method Made Fixed Found (%)
Be photoneutron 146 5.0 1.77 —0.28
Cr amperometric 86 : 15.0 8.06 - —4.33
Fe spectrophotometric 60 15.0 5.43 +2.28
Ni spectrophotometric 54 15.0 5.71 +1.80°
U coulometric 172 \ 1.0 0.66 +0.36

Zr amperometric - 77 5.0 ’ 3.21 +1.35

8. Molten—Salt Convection Loops. in the ORR

.H. C. Savage E. L. Compere
J. M. Baker M. J. Kelly
E. G. Bohlmann

Irradiation of the first molten-salt convection loop experiment in the Qak Ridge Research Re-
actor .beam hole HN-1 was terminated on August 8, 1966, after development of 1.1 x 10'8 fissions/cc
(0.27% 235U bumup) in the 7LiF-].3eF2-ZrF4-UF4 (65.16-28.57-4.90-1.36 mole %) fuel. Average
fuel power densities up to 105 w per cc of salt were attained in the fuel channels of the core of
MSRE-grade graphite. . | : [

Successful operation of the major heating, cooling, tempetature control, and sampling systems
was demonstrated; however, leaks developed in two of the four cooling systems. The experiment
was terminated after radioactivity was detected in the secondary containment s3lzstem as a result of
fuel leakage from a break in the sample line near the loop.

Irradiation of a second loop, modified to eliminate causes of failures encountered in the first,
was begun in January 1967. Fueled operation began January 30, 1967. Operation is continuing at

an average core fuel power density estimated at 160 w per cc of salt.

8.1 OBJECTIVES AND DESCRIPTION

The loop is designed to irradiate a representative molten-salt fuel circulating in contact with
graphite and Hastelloy N at typical temperature differences and at a core power density of 200
w/cc. In patticular, it allows us to study the interaction of fission products with graphite, metal,
and fuel and gas phases, and the chemistry of the fuel salt at high levels of burnup. Provisions
for sampling and replacement of both gas and salt pemit conditions in the loop to be determined

and to be altered during operation.! =4

1 MSR Program Semiann. Progr. Rept. Aug. 31, 1965, ORNI1.-3872, pp. 106—10.
" 2MSR Program Semiann, Progr. Rept. Feb, 28, 1966, 0RNL-3§36, pp. 152-54,

3Reactor Chem. Div. Ann. Progr. Rept., Dec, 31, 1965, ORNL-3913, pp. 34—35.

4I'\’eactor Chem. Div. Ann. Progr, Rept, Jan. 31, 1965, ORNL-3789, pp. 4548,

167
168

8.2 FIRST LOOP EXPERIMENT

In-Pile Irradiation Assembly

The core of the first loop consisted of a 2-in.-diam by 6-in.-long cylinder of graphite obtained
from MSRE stock. Eight vertical 1/4-in. holes for salt flow were bored through the core in an octag-
onal pattern with centers 5/8 in. from the graphite center line. A horizontal gas separation tank con-
nected the top of the core through a return line to the core bottom, completing the loop. These, and
the core shell, were fabricated of Hastel.loy N. The heaters, and the cooling tubes in the core and
return line, were embedded in $prayed-on nickel, as was the 12-ft sample tube leading from the

loop to the sample station in the equipment chamber at the ORR shield face.

Operations

The loop was operated with MSRE solvent salt for 187 hr at Y-12, and several salt sa_mplés
were taken. It was inserted in beam hole HN-1 of the ORR on June 9, 1966, and operated 1100 hr
with solvent salt; during this period the equipment was calibrated and tested, and its performancé
was evaluated. The loop was advanced to the position nearest the reactor lattice on July 21, and
water injection into the air streams to the tubular core coolers and the jacket around the gas
separation tank was tested. One of the two core coolers leaked and was plugged off. Water in-
jection was discontinued until after uranium was added.

On July 27, after sampling, eutectic 7LiF-UF4 (93% enriched) fuel salt was added to develop a
uranium concentration of 1.36 mole %. At this time a capillary tube in the sample removal system
broke, precluding further sampling. An associated in-leakage of air impelled solvent salt to a cold
spot in the gas sample line, thereby plugging it.

During subsequent operation the fission heating rate was determined. Water was released into
the loop container during this period from what proved to be a leak in the cooling jacket around the
gas separation tank. After a short reactor shutdown on August 8 to permit removal of water accu-
-mulated in the container, the irradiation was resumed. That evening, release of substantial radio-
activity into the loop container indicated fuel leakage. The lloop temperature was lowered to freeze
the salt, and the loop was retracted to 2% flux. It was removed to hot cells for disassembly and

examination on August 11, 1966.

Chemical Analysis of Salt

Samples of solvent salt taken prior to irradiation and after 1100 hr in-pile, and of irradiated
fueled salt obtained after dismantling, were analyzed chemically and radiochemically. A sample
of salt found between the metal core shell and the graphite was also analyzed. Results are given

in Table 8.1 and discussed below.
169

Corrosion

The level of corrosion products, particularly chromium and nickel, in the salt increased in the
successive samples. This was possibly due to uptake of moisture by the solvent salt prior to
loading, with consequent corrosion of the Hastelloy N. The corrosion appears to have occurred in
the addition tank, since a sample which was taken directly from the addition tank, without entering

the loop, showed similar levels of corrosion products.

Fission Products

Fission products were counted in a fuel sample after 110 days‘ of cooling, and concentrations
are given below as a percentage of the amount produced, calculated on the basis of observed fis-
sion heat (4.8 x 10!7 fissions/g). . ,

Cerium-144 and -141 (77, 64%) and 8°Zr (65%) were somewhat below the calculated production.
Cesium-137 (41%) and ®°Sr (42%), both with noble-gas precursors of ~ 3 min half-life, were still
lower and could thereby have been lost to the gas space or graphite voids. Tellurium-127 was
present in 10% of the amount produced in the salt. Ruthenium-103 and -106, which were expected -

to deposit on Hastelloy N surfaces, were not detected (<0.03%) in the salt.

Nuclear Heat and Neutron Flux

Nuclear heat was measured at various loop insertion positions by comparing electrical heat
requirements under similar conditions with the reactor at zero and at full power. Reactor gamma
heat with the loop fully inserted was 2900 w (with unfueled salt). With fuel containing 1.36 mole %
uranium (93% enriched), fission heat in the fully inserted position was 5800 w. The corresponding
overall average fission heat density was 80 w per cc of salt at 650°C, and in the graph'ite core the
average fission heat density was 105 w per cc of fuel salt.

The average effective thermal neutron flux in the salt was estimated independently from the
nuclear heat, from activation of solvent salt zirconium, from cobalt monitors on the loop e);terior, _
and by neutron transport calculation. The results agreed well, ranging between 0.9 and 1.2 x 1013

-2

neutrons cm~? sec™ 1.

Hot-Cell Examination of Components

After separation from other parts of the package, the loop proper was kept for approximately
three months in a furnace at 300°C to prevent fluorine evolution by fission product radiolysis of the
salt. At this time it was removed for detailed examination. ‘

The tubular core cooler, type 304 stainless steel, was found to have broken entirely loose with-
out ductility at its outlet end as it left the core near a tack weld to the core shell. Intergranular

cracks originated on the outer circumference of the coiled tube.
- Table 8.1. Loading and Samples from First In-Pile Molten-Salt Convection Loop

(analyses as mg/g or mole %)

"Li Be Zr vé F Cr Fe Ni Mo
Composition as loaded
Solvent salt
. Composition as manufactured (mole %) b (64.78) (30.06) (5.16)
(mg/g) 114.5 68.4 ° 118.9 698.2 ~™~0.020 ~0.,020 ~™0.100
Fuel, eutectic (mole %) (72.46) (27.54)
(mg/z) 48.5 619.6 331.9
Fueled loop mixture (calcd) (mole %) (65.16) (28.57) (4.90) (1.36)
(mg/E) 106.5 60.2 104.5 74.8 654.0
Hastelloy N — representative analysisc (mg/) 70.400 46.000 696.200 161.000
Analyses of loop samples (mg/g)
Hours Hours of Sample
Molten Radiation No,
120 0 1 (solvent salt) 98.0 66.5 119.5 683 . 0.310 - 0.275 0.137 n.d.
166 0 3 (solvent salt) 97.5 66.1 122.0 673 0.355 0.285 0.455 n.d.
1260 208 6 (solvent salt) 116.5 © 69.1 120.0 704 0.670d 0.092 0.540 n.d.
After shutdown,
1578 329 9 (fueled mixture) 113 58.5 99.4 71.5 0.780 0.258 0.555 <0.015
Graphite—Hastelloy N annulus specimen S-1 105 56.1 101.5 71.7 3.250 3.800 25,300 4,740

292.96% 235y,

bj. H. Shaffer, MSR Program Semiann. Progr. Rept. Feb, 28, 1965, ORNL-3812, pp. 150--52. .

“Heat SP-19 for comparison,

Sler activation gave a Cr concentration of 0,990 mg per g of salt.

0L1
171

The cooling jacket on the gas separation tank leaked at a weld. The fuel leak resulted from a
nonductile break in the Hastelloy. N sample line tubing near the attachment to the core bottom. The
sprayed nickel was also cracked in this region. '

Fuel salt in the form of a scale a few mils thick was found on the interior of the core shell,
between it and the closely fitting graphite core. The analysis in Table 8.1 indicates that the scale
is a mixture of fuel salt and Hastelloy N (probably metal debris from cutup operation).

Hot-cell metallurgical examination of the interior surfaces of the Hastelloy N comprising the
core bottom and core shell wall revealed no evidence of any interaction with salt or carbon, or other

change,

8.3 EVALUATION OF SYSTEM PERFORMANCE
Heaters

_The molten-salt loop package used 21 continuous or intermittent heaters, all 1/8--in.-OD Inconel
sheathed, MgO insulated, with Nichrome V elements designed for continuous operation at 870°C.

No failures occurred. -

Coolers

The heat removal rate of the loop coolers was entirely adequate to remove the 8.8 kw of fis-
sion and gamma heat, even after the loss (described earlier) of one of the two cooling coils around
the loop core section. The air plus water-injection technique appears adequate and resporfsive.
The use of water injection was not necessarily the cause of failure of the two cooling units, but

only made the failures evident.

Temperature Control

The response of the heating and cooling systems to rapid changes in the nuclear heat could be
tested only under full fission conditions in-pileé. Since this was regarded as important, reactor set-
back tests were conducted. Temperature control system response was adequate to maintain the salt
molten during a reactor setback with resultant loss of 8.8 kw of nuclear heat, and to return the loop

to nomal operating condition during a rapid (11-min) return to full power,

Sampling and Additic‘m

The sampling and addition system and operating procedures were adequate to permit several
~additions and removals of molten salt while operating the loop in-pile, and to transport shielded
samples under inert gas atmosphere to the analytical laboratory. A broken capillary connecting

tube prevented additional sampling.
172

Salt Circulation

Convective sal-t citculation, at rates of 5 to 10 cc/min, was achieved by causing the return line
to operate at temperatures below the core temperature. Flow stoppagés occurred from time to time.
These were attributed to bubble formation resulting from different solubility of the argon.cover gas
at the varied temperatures around the lo,op'. Salt flow was reestablished by evacuation ~a'nd readdi-

tion of cover gas. Loss of flow had no adverse effect on loop operation.

8.4 SECOND IN-PILE IRRADIATION ASSEMBLY

A second in-pile molten-salt convection loop,l essentially identical to the first convection loop
experiment,® was constructed, and in-pile irradiation began eatly in January 1967. Problems en-
countered in the first convection loop experiment and information from subsequent postirradiation
hot-cell examinations led to modifications to the second.loop to eliminate these problems.

The coolant tubes embedded in sprayed-on nickel around the core section are now 1/“—-in.-OD X
0.035-in.-wall Inconel tubing instead of the 14—in.-0D x 0.035-in.-wall type 304 stainless 'steel used
on the first loop. The stainless steel tubing should have been entirely adequate for the service,
and no reason for the failure observed has been uncovered; but Inconel is the preferred material for
exposure to the high-ten;perature steam-air mixture (~400°C) generated when air-water mixtures are
used as coolant. Since the rupture of one of the core coolant tubes occfirred adjacent to a point
where the tube was tack welded to the core wall, the tack weld waé eliminated in favor of a mechan-
ical strap attachment. An expansion loop to relieve stresses has also been included in each of the
coolant outlet lines. A mockup of the modified cooling coil was operated at temperature with air-
water mixtures for more than 400 hr, ir;cluding 120 thermal shock cycles (600 to 310°C), with no
sign of difficulty. Thermal cycling occurs during a reactor setback and startup, and it is estimated
that no more than about 20 such thermal cycles will occur during a year of op‘eration.

The two failures which occurred in the capillary tubing (0.100 in. OD x 0.050 in. ID) used in
the salt transfer system appear to have resulted from excessive mechanical stress. Consequently,
the wall thickness of this line has Been increased to 0.050 in., and additional mechan}cal support
hqs been added such that there is now no part of the salt sample line which is unsupported — con-
trary to the case in the first loop assembly.

The cooling jacket of 1/16-in.-thick stainless steel surrounding the reservoir tank has been re-
placed by an Inconel tube wrapped around the outside of the tank and attached by means of sprayed-
on nickel metal, as is done on the core section and cold leg. Also, provisions for uée of an air-
water mixture as coolant have been added, since it was found that air alone did not provide suffi-
cient cooling in the first experiment. - _

Continuous salt circulation by thermal convection was not maintained in the first experiment.

It was concluded that loss of circulation was caused by gas‘acc‘umulation in the top 6f the core
section. Accordingly, the salt flow channels at the top and bottom of the eight 1/4-in. holes .for salt

flow in the graphite core were redesigned to provide better flow conditions at the inlets and exits of
173

-

the vertical holes. Further, the top and bottom of the core section, horizontally oriented on the first

loop, were inclined at 5° to minimize trapping of gas.

Operation

The loop was placed under irradiation in beam hole HN-1 of the ORR on January 16, 1967, after
325 hr of satisfactory preirradiation test operation. - Preirradiationv.operat‘i'on included COmpié"te re-
moval and replacement of an original flush charge of solvent salt, providing a good demonstration
of the operability of the sampling and addition éystem

The loop operated satisfactorily under irradiation at 650°C with MSRE solvent salt (’LiF- BeF -
ZiF,, 64.8, 30.1, 5.1 mole %) during checkout and calibration measurements,

On _]anuary 27, 1967, samples of irradiated solvent salt were taken and submitted for chemical
and radiochemical analyses. ' . _

On January 30, 1967, 7LiF—UF4 (63-27 mole %) eutectic fuel (93% 23°U) was added, along with
additional solvent salt, resulting in a fuel composition of 7LiF~BeF2-ZrF4-UF4 of 65.26-28.17-
4.84-1.73 mole %. Irradiation began January 30, 1967, with the experiment in a relatively low flux
position. Convective circulation was established a't an estimated rate of 30 to 50 cc/min (™~ 2 min
circuit time), and the loop was operated in several flux positions. Controls wete tested and ad-
justed, and nuclear heat was determined as a function of position. Samples of fuel salt were with-
drawn February 6, 1967.

Operation in the fully inserted, highest flux position was achieved February 21, 1967, and has
since been maintained. The average flux effective in all the salt is estiméted to be slightly above
1 x 1013, The average power density in the fuel salt in the core is estimated to be about 160 w/cc.
The total fission heat genera{ion is about 9 kw; with a gamma heat of about 4 kw, the system heat
generation is about 13 kw. The loop is operating satisfactorily.

* Sampling and addition systems have operated reliably and without difficulty in all cases. It is
anticipated that additional samples will be taken and fuel and solvent salt additions made from time

to time during the course of the experiment.
Part 3. Breeder Reactor Design Studies

9. Molten-Salt Breeder Reactor Design Studies

E. S. Bettis
C. E. Bettis D. A. Dyslin  G. H. Llewellyn  W. Terry
R. J. Braatz H. F. Kerr T. W. Pickel L. V. Wilson
9.1 GENERAL

The Breeder Reactor designnl"3 has been considerably narrowed in scope during this period.
A modular concept has been adopted as the basic design, and our effort has been concentrated on
more clearly defining the features of one 250-Mw (electrical) module. No work has been done on
overall plant design during this period, and no attention has been given to the steam system.

* With minor exceptions, all design effort has been concemed with components and layout
within the reactor cell. . A layofit was made of the drain tanks with interconnecting piping for all
three salt circuits. Figure 9.1 is a plan view of the general arrangement. This preliminary lay-
out was made to gef an idea of the magnitude of the ‘‘tank farm’’ problem, and it appears that a
satisfactory solution t§ this question will not be overly. difficult.

Again, in order to get a preliminary evaluation of the nature of the communication problem
between reactor plant and chemical processing plant, one method of effecting this intercon-
nection was developed. Basically this plan involves a batch transfer between piants, the
batches being moved between the two plants by gas pressure. Transfer to and from.the chem-
ical ‘plant is in control of the reactor plant operator. Although the system of communication
between reactor and chemical plant has not been worked out in any detail, it appears that the
communication scheme can be done with certainty and safety.

We believe that the optimization of the reactor is now accurate enough to warrant serious

detailing of the désign. .A basic concept of fuel cell, blanket, plenum chambers, and reflector

IMSR Program Semiann. Progr. Rept. Aug. 31, 1966, ORNL-4037, p. 207.
2Design Studies of 1000-Mw (e) Molten-Salt Breeder Reactors, ORNL-3996. I
3MSR Program Semiann. Progr. Rept. Feb. 28, 1966, ORNL-3936, p. 172.

174

REHEATERS

STEAM GENERATORS—|——

34 ft

INSULATION

COOLANT PUMP

DUMP TANKS
2nd SALT ———

BLANKET
FUEL~«\:::::::::

o)

43

L2

201t
K
o7 i1
21
BLANKET
HEAT EXCHANGER
AND PUMP — |
PRIMARY
HEAT EXCHANGER
AND PUMP ———— 7 4t
8 fi

STEAM PIPING

1133

AXA3
I W)

ORNL-DWG 67-3082A

HX = HEAT EXCHANGER

Fig. 9.1.

Reactor and Steam Cell Plan.

OFF-GAS
34 ft
T e, ST e e e e eeror oy
LUSH SALT STORAGE 4
{ DECONTAMINATIO
________ AND 16 ft &in.
STORAGE
: b
i = o 211
Y P I -
AL T sy AN R FUEL
D g P S S PROCESSING !
L@ | DECONTAMINATION
i 1 s | .
TN ol 2 T ¢ A 16 f1 &in.
RIS =t ST STORAGE
AR
: |: ’ (I A :I 1
1.1 [ LI —_—
i )
: . REACTOR ¢ !
| ‘\ ," I
: S : 14 ft
}:—kwA419ft ‘ 18 ft ‘
ot 8 ff et 22— 4 ft —= L*—4ft —*J 4
' 157 ft - ' I
nlAll ' "B".‘__

SL1
176

thickness for a reactor with an average power density of approximately 40 kw/liter has reached
a firm status. It is this module which will receive detailed attention. For comparison/purposes
a module of equal power but having twice the average powerdensity was calculated. Table 9.1
shows the comparison between these two reactors regarding performance and economy.

The blanket and fuel heat exchangers are also considered to be in a sufficiently firm design
to justify detailing of these components. Both heat exchangers have undergone significant

changes in physical arrangement and will be discussed in some detail later.

Table 9.1. Comparison of the Effects of Core Power Density on the Characteristics
of a 250-Mw (Electrical) Molten-Salt Breeder Reactor

Average core power density, kw/liter . 78 : 59
Power, Mw 556 _ 556
Vessel diameter, ft | ' 11.4 12
Vessel height, ft \ ~12 17
Core diameter, ft 6.34 8
Core height, ft 8 10
Core volume, ft> 252 503
F-raction of fuel in core - 0.164 0.165
Fralction of blanket in core 0.05 0.06
Fraction of graphite in core 0.786 0.775
] Blanket thickness, ft 2 . 1.5
Fraction of salt in blanket 0.65 - 0.60
Breeding ratio ' 1.06 1.07
Fuel yield, %/year 6.79 6.02
Fuel cycle cost, miils/kwhr .0.42 0.43
Fissile inventory, kg - 175 218
Fertile inventory, 1000 kg | 41 ' 43
Specific power, Mw (themal)/kg 3.18 2.55
Number of core elements 336 336
Velocity of fuel in core, fps 9.7 6
Average flux, e§ 100 kev ‘ . 2.9 x 1014 1.5 x 1014
Fuel volume, ft3
Reactor core - 41.3 83.0
" Plena 10.0 24.0
Entrance nipples 5.6 8.5
Heat exchangers and piping ' 102.0 105.0
Processing 10.5 8.8
Total : 169.4 231.3,
Radial peak average flux ratio 1.58

Axial peak average flux ratio 1.51

177

The arrangement of components in the reactor cell as previously repbrted posed some very
difficult stress problems. A new arrangement of these components abpears to have removed or
greatly lessened these problems. Analyses are currently in progress which will indicate whether

or not further changes are required.

Operational expériencé with the MSRE has indicated clearly that the handling of fission
gases from a molten-salt reactor requires very careffil design. The gas system for the breeder
reactor is a most important part of the plant, and a start has been made in the design of this
system. A flowsheet which seems to describe this system has been developed. Also, analyt-
ical studies of gas stripping efficiencies and parameters which determine or define the nature

of the gas sparging system have been initiated.

~ Associated with the gas handling prob\lem is the matter of draining the reactor in the event
of a pump stoppage. Modifications in the primary heat exchanger indicated by the sparge gas
handling and core drainage influenced us to put an overflow line from the sumpA tank of the pump
into a fuel dump tank. These c;hanges in the primary heat exchanger have been rather extensive,

although the basic heat exchanger remains very much as the original concept.

9.2 FLOWSHEET

Since the reference design is based on the modular reactor concept, the flowsheet of Fig.
9.2 is for one module. The number of the various components in this module is shown. The
steam system is shown very sketchily for the 1000-Mw (electrical) plant, and an indication of

a steam header fed by the four modules is shown.

It will be noted that the maximum coolant-salt pressure is somewhat higher than has been
reported previously. This pressure occurs at the discharge of the coolant pump, and, of the
260 psi at this point, 110 psi comes from a gas overpressure at the suction side of the coolant

pump.

~ We believe it to be imperative that, in the event of a failure in any part of the salt systems,
the pressures should be such that blanket salt would flow into fuel salt, and coolant salt would
flow into blanket and fuel salt, depending on the location of the failure. This made it necessary
to put a high bias pressure on the coolant salt. The pressures shown on the flow diagram repre-
sent design-point pressures and take %nto account static heads and dynamic pressure drops in

all cases.

In previous flowsheets a separate coolant pump circulated salt through the reheater. We now
bypass the discharge of the main coolant pump through a salt-flow regulating valve to the re-
heater. This valve is little more than a variable orifice, and we believe the use of such a valve

is justified.
PROCE SSING

FUEL SALT

HEAT EXCHANGER

AND PUMP

FUEL

4 f1t¥day

REACTCR

J

L ———

BLANKET SALT
HEAT EXCHANGER
AND PUMP

COOLANT
BLANKET SALT
SALT PUMP
PROCESSING

|
| ———
—(

FUEL SALT

DRAIN TANKS

POINT ft3/sec

AR O0ROEOEE

FUEL
psig TEMP (°F)

250 13
250 9
250 137
250 110
25.0 50
25.0 31
250 18
BLANKET
43 125
43 8.0
43 1110
43 200
4.3 145

1300
1300
1300

1000
1000
1000

1250
1250
1250
1150
1130

' t
(% E \BOH_ER SUPERHEATER

POINT f1¥/sec

BlEBEHEEBGEE

COOLANT
psig  TEMP [°F)

375
375
37.5
5.0
5.0
8.1
8.1
37.5
375
375
37.5

BLANKET SALT COOLANT SALT

DRAIN TANK DRAIN TANK

122
1o
260
208
212
252
194
203
198
161

138

1125
1125
1125
125
850
1125
850
850
850
1111
111

PRARDBIRIPODIRDD

REHEATERS

STEAM

2.52

10.07

7.15
2.92
2.92
513
5.13
1.28
1.28
513
5.00
4.30
7.16
1007
1007
2.52

Fig. 9.2, Module Flowsheet.

3600
3600
3515

3515

3500
600
570
570
540
540
V72

15inHg

3500
3475
3800
3800

REHEAT STEAM

POINT 10% Ib/hr psia TEMP (F)

1000
1000
1000
1000
866
552
650
650
1000
1000
706
92
551
695
700
700

ORNL-DWG 67-3081A

FEED WATER
SYSTEMS

PERFORMANCE DATA
1000 Mwle} GENERATION PLANT

NET OUTPUT 1000 Mwle)
GROSS GENERATION 1034.9 Mw(e)
BOILER FEED PUMPS 29.4 Mw
BOOSTER PUMPS 9.2 Mw(e)
STATION AUXILIARY 25.7 Mw(e)
REACTOR HEAT INPUT 2225  Mwlt)
NET HEAT RATE 7601 Btu/kwhr
NET EFFICIENCY 449 %

LEGEND
FUEL _—
BLANKET =
COOLANT
STEAM @~ ————— —n——
WATER —
FREEZE VALVE

DATA POINT X

8L1
179

9.3 REACTOR CELL COMPONENT ARRANGEMENT

Because of the relatively high melting point of the blanket salt, it is desirable to connect
the coolant circuits of the fuel heat exchafiger and the blanket heat exchanger in series. Coolant
salt leaving the fuel heat exchanger at a temperature of 1111°F enters the blanket heat exchanger,
where it picks up 14°F to attain the design-point temperature of 1125°F. The coolant-salt line,
carrying 37.5 ft3/sec, is a large pipe (™20 in. in diameter) an.d therefore rigidly connects these
two heat exchangers. Since the blanket heat exchanger is located high in the reactor cell while
the fuel heat exchanger must be located below the bottom of the reactor vessel, there is a shift
in the center of gravity of the coupled heat exchangers when they are filled with salt.

In the original layout of the reactor cell which had the reactor, fuel heat exchanger, and
blanket heat exchanger in a triangular array, a twisting moment resulted about the axis of the
fuel connection between the reactor and the fuel heat exchanéer. Also, the method of suspen-
sion mounting of the heat exchangers from constant load cell s could not provide accommodation
for the differential thermal expansion of the system under all conditions. ,

A new layout of the reactor cell shown in Fig. 9.3 was made. In this configuration the three
components, reactor, fuel heat exchanger, and blanket heat exchanger, are mounted in line. The
coolant lines connecting the fuel and the blanket exchanger are made concentric, and the shortest
possible spacing between exchangers is used. The connection from fuel heat e€xchanger to re-
actor is as short as possible, using concentric piping for this as we have been doing previously.

The connection from the blanket heat exchanger to the reactor vessel has been changed. In-

. stead of the short concentric pipe used previously, the inlet and outlet lines from reactor vessel
to blanket exchanger are run independently. These lines are purposely made long and flexible
by running them through four 90°bends. Thus we intend to decouple the interconnection between
the reactor and the blanket heat exchanger by use of flexible piping.

The three components, reactor, fuel exchanger, and blanket exchanger, are rigidly inter-
connected then by only one link for each; the reactor connects by the concentric fuel line to the
fuel exchanger, which connects by one concentric coolant line to the blanket exchanger. At de-
sign point these connecting lines are very near the ambient cell temperature of 1000°F. The fuel
line is actually at 1000°F, and the coolant line is very near 850°F. With the three cell compo-
nents in line, a different plan of sfipport is now proposed. Let us point out that this arrange-
ment has not yet been analyzed for stresses, so we cannot be sure it is satisfactory in its
present conception. '

The plan now involves mounting a stainless steel bed plate in the reactor cell in such a way
as to be tied at one end but free to expand with cell temperature. This bed plate is located
horizontally below the bottom of the reactor vessel and is mounted through necessary insulated
supports to the load-bearing concrete of the cell. The reactor vessel and both heat exchahgers
are mounted rigidly to the plate as shown in Fig, 9.4. We believe that this general arrangement
will work and that any overstressing found by analysis of the system can be-corrected without

departing drastically from the concept.
ORNL-DWG 67-3083A

BLANKET HEAT EXCHANGER

FUEL AND BLANKET

PUMP DRIVE MOTORS COOLANT SALT PUMP

DUMP TANKS STEAM GENERATORS

8ft Qin.

REHEATERS

STEAM
PIPING

58ft Oin.

HEATER

Y

PRIMARY HEAT
EXCHANGER AND PUMP

Fig. 9.3. Reactor and Steam Cell Elevation.

081
a [}
ORNL-DWG 67-33644
' 3714t OQin. | ' 22 f1 Qin. '
12+ Oin. in. 6t Oin. ~=— 15t 10in.
’ N— A : N——"

! N ' -
\ BL. HEAT

REACTOR EXCHANGER

4%t Qin.
(TYP)

PRIMARY
HEAT
EXCHANGER

ditatihl :l\‘h-“ L'}

I

FIXED SUPPORT

Il

—]_———INSULATION

101t 6in.
141t &in.

Fig. 9.4. Reactor Cell Mounting.

181
182

Vertical growth of all components can take place independently, and the only accommodation
needed is a low-temperature sealing bellows at each pump-drive penetration through the cell
membrane. Any control-rod drive mechanism would involve a similar bellows seal.

The treatment of the reactor cell has not received any attention, but it is certain that some
formof radiation shield must cover the structural concrete and this shield must have cooling.
Inside the radiation shield, between it and the furnace of the cell, there must be some thermal

insulation. -

9.4 COMPONENT DESIGN
Reactor Yessel

The reactor is shown in the elevation drawing of Fig. 9.5. The reactor vessel has a cy-
lindrical body 12 ft in diameter and approximately 17 ft high with dished heads on top and
bottom. A heavy supporting ring essentially the diameter of the core carries legs which weld
to the bed plate and provide the mechanical mounting for the vessel. From the center of the
bottom head a concentric fuel line communicates with the fuel heat exchanger. This fuel line
has an outside diameter of 24 in. with a concentric retumn liné 16 in. in outside diameter, Fuel

flows in these lines at a rate of approximately 18 fps.-

Inside the reactor vessel are dished heads forming plenum chambers for distribution of the
fuel cells which rise frfom these heads. These plenum chambers are removable from the reactor
vessel, being held in place by a flanged mounting ring equipped with clamps. This flange has
not been completely designed, but we are considering the use of flat graphite gaskets to pre-
vent bonding of the flange surfaces under the high temperature and rather,high loading pressure
obtained at desigfi conditions. The inner head communicates with the inner concentric line
through a slip joinf which permits movement of the head relative to the pipe. Such movement
results from thermal expansion and must be accommodated. In addition, the slip joint allows

removal of the core when the flange seal is broken.

The core of the reactor is made up of 336 cylindrical graphite fuel cells mounted as close
together as tolerances permit to form essentially a cylindrical array approximately 8.3 ft in
diameter. The graphite cells are extruded cylinders with center holes 11/2 in, in diameter,
surrounded at 120° angles with three 7/a—irl.-diam holes. At the top of each cell there is a
graphite cap méchined to provide a smooth communication between the four holes of the

cell. Figure 9.6 shows the arrangement of one of these cells.

At the bottom of a cell, the graphite cylinder is joined to a transition nipple. This join-
ing is done by a combination of a short threaded engagement for mechanical attachment and a
rather longer brazed joint for leak-tightness. " The connection to this nipple is a matter of
development, and work is proceeding to establish the best practice for effecting this con-
nection. The lower end of the nipple is threaded to screw into tapped holes in the upper

curved head.
183

ORNL-DWG 67-3078A

— 12 ft Oin. DIAM.

; CORE: 8 ft Oin. DIAM x 10 ft Oin. HIGH
336 FUEL CELLS

-CONTROL TUBE PURGE

]
PLENUM
f T0
WOy BLANKET
. ‘ X . HEAT
1 AL EXCHANGER
10/9¢0
:.\fl
N \. -1
3
. N 4.)
. \ g
4-in.-DIAM U H
GRAPHITE ®
SPHERES—-__i% ,
| (=
T\
2
7% z
6-in. .0 £k &
GRAPHITE % o2 o
REFLECTOR—4{// =3 =
& ©
.REACTOR
VESSEL
_ r 1 1 l /
\ ! =Ty | :
V. o
Dy M
......... ’

FUEL SALT/ i
PLENUMS \

TO FUEL HEAT EXCHANGER

-Fig. 9.5. The 250-Mw (Electrical) Reactor.
184

{ft 6in.
—-‘_fi"——\_——_—_r-_‘
VAN
rd LSS Sy ’
L ,(S’/ //// o’ o // //§
\ / " £ oS
.7 # ‘ PR
~ // /\
\\\ /—\‘ ,{-\ '\\
N ~
NN EAN
N\
N R
AL R
101t Qin.
(REACTOR)
5.00Qin, DIAM
fi’
N
N N
'\ M
§ "
\ ~
N,
N X \
~, N \
NIR N N
N
1ft &in
’L,L’f///;\’
1ft Qin. ’I’
d N
\ By

CRNL-DWG 67-267TA

2 ?/3 in.

1Y in. DIAM

3-HOLES,
Yg in. DIAM

GRAPHITE TO
METAL BRAZE

1%gin, OD X 1¥4in, 1D

1% in. OD X 1¥%gin. 1D

FUEL INLET PLENUM—/

N

NN NN

LD

TAPERED THREAD

1
-/ “m———— FUEL QUTLET PLENUM

Fig. 9.6. Fuel Cell Element.
185

From smaller holes in the inner head, concentric with the upper one, smaller nipples ex-
tend through the upper head. These nipples are also attached to the inner head by a pipe

thread, each one being screwed into place before attaching the fuel cells above.

As a fuel cell is put into place, the 11/2-in. center hole is slipped over the inner nipple.’
This engagement is a slip joint and allows for screwing the cell into place, and in operation
it permits the necessary differential expansion between the cell and the inner nipple. After
all cells are in pléce, the entire core can be leak-tested, the graphite-to-metal transition joints
having previously been tested individually. When the assembled core has been tested, it can

be lowered into the reactor vessel and the flanged. joint can be clamped.

Just inside the reactor vessel there is a graphite reflector 6 in. thick. This reflector is
only on the vertical walls; there is none below or above the core. It is made of rectangular-
shaped pieces with mortised joints to form a self-supporting wall resting on a ledge at the
bottom of the vessel. The volume between the reflector and the vessel provides a flow
passage for the blanket éalt as it discharges through holes in the bottom of the distribution

ring located near the top of the vessel.

We are considering the use of 4-in.-diam graphite balls in the region between the core and
the reflector. These balls have holes in them so that graphite occupi.es 40% of the volume in
this ll/z-ft—thick region. A retainer of perforated metal prevents the balls, which float in the
blanket salt, from crowding into the top head of the reactor vessel. Instead, by floating up
against this retainer, a measure of support is afforded the core structure, the floating balls

tending to exert a considerable force on the fuel cells.

With this core configuration a two-fluid reactor is achieved. Fuel salt enters the outer
concentric pipe at the bottom of the reactor. It flows out to the outer radius of the core and
in between the upper inner dished heads, feeding salt into the nipples emerging from the upper
head. Fuel passes up through the three 7/8-in. holes of each fuel cell and down through the
center 11/2—in_. hole through the inner pipe nipple into the collecting plenum. From there it
flows through the slip joint into the inner concentric fuel line to the pump suction at the

center of the fuel heat exchanger.

The blanket salt flows into a distributor at the top of the reactor, and, emerging through
holes in the bottom of this header, it flows behind the graphite reflector to the bottom of the
reactor. There it flows up through the balls in the blanket annulus and out of the suction

line near the top of the reactor.

Some of the blanket salt fills the interstices betweén the fuel nipples and the fuel cells.
Circulation of the blanket salt is governed by thermal convection. Heat generation and tem-

perature calculations are in progress to determine whether this flow is adequate.

-

The center location for a fuel cell is left vacant. This is for a control rod. In Fig. 9.6
this control rod is represen\ted as a hollow graphite cylinder 5 in. in outside diameter, 4 in.

in inside diameter, and equipped with a gas inlet at the top. By controlling the gas pressure,
186

blanket salt can be positioned at any height within this cylinder, and this blanket volume con-
trols the reactivity of the core. Calculations indicate a change of 1.8% in reactivity between
the full and empty condition of this cylinder.

It may be that, instead of a hollow cylinder, a graphite rod 5 in. or less in diameter will
‘be used to displace blanket salt from this position in the core. Effective control can be ob-
tained by driving this rod in a more conventional way. In the event of failure of the drive, |
the rod would be lifted out of the core by the more dense blanket, thereby reducing the re-
activity. ‘ |

The reactor has been optimized as far as nuclear performance is concemed. “There is no
very sensitive parame;er, and rather large changes in all dimensions can be tolerated without
changing the characteristics materially. Table 9.1 gives the calculated performance of this
reactor. This core volume is 503 ft3, with a fuel volume in the core of 83 ft3. This gives an
average power density in the core of 39 kw/liter. _

For comparison, the performance of a reactor having a power density of 78 kw/liter is given
in Table 9.2. This reactor -has half the volume in the core and would give somewhat better
i)erformance. At present we do not know just what the limit is on graphite radiation tolerance.
The smaller reactor, whatever the graphite limit, would be expected to have approximately half
the core life of the larger one, which is used as the reference design.

Some general comments can be made about the reactor as it is currently conceived. As we
stated above, the nuclear petformance is so insensitive to parameters that the final reactor
design is likely to be close to the present concept. We would prefer to have the core removable
in modules rather than having to remove i(t in one piece. This is particularly desirable if we
have to maintain the lower power density and if it becomes desirable to make reactor modules
larger than the 556-Mw (thefinal) size with which we are concerned at present. We do not yet
have what we consider an acceptable way of making these core submodules, but work is con-
tinuing to try to develop one.

The thickfiess of the heads has not received rigorous attention. At present the top or load-
bearing head is 1 in. thick. This is thicker than the load seems to require but is used to give
an adequate thickness for tapping for the nipples. The analysis of all stresses in these heads
is being made but has not yet been completed. The same is true for the pressure drops.

The top of the reactor as shown has a welded joint which can be ground off for removal and
rewelded when a new core has been installed. Probably this access joint will Be moved up
outside at least one layer of shielding in the final core design. In this case some of the shield-
ing blocks would be supported by the top of the reactor vessel. The closure point for the re-

actor would, by this means, be more accessible.

Fuel Heat Exchanger

The heat transfer work which had been done for the Breeder Reactor study was reevaluated.

This was made necessary by the results obtained from the MSRE.. It was found that the salt

187

Table 9.2. Fuel Heat Exchanger Data Sheet

Fuel-coolant-salt ?xchanger, counterflow, two-shell pass, two-tube pass with donut-type baffles

Number required -
Rate of heat transfer, Mw
Rate of heat transfer, Btu/hr

Shell side
Hot fluid or cold fluid
Entrance temperature, °F
Exit temperature, °F
Entrance pressure, psi
Exit pressure, psi
AP across exchanger, psi
Mass flow rate, lb/hr

Tube side
Hot fluid or cold fluid
Entrance temperature, °F
Exit temperature, F
Entrance pressure, psi
Exit pressure, psi
AP across exchanger, psi
Mass flow rate, lb/hr

Tube material -
Tube OD, in.
Tube thickness, in.

Tube length, tube sheet to tube sheet, ft

Shell material

Shell thickness, in.
lShell ID, in.

Tube sheet material

Tube sheet thickness, in.

Number of tubes

Pitch of tubes, in.

Type of baffle
Number of baffles

Il

1
528.5
1.8046 x 10°

Cold (coolant salt)
850 '

1111

198

161

37

1,685 x 107

Hot (fuel salt)
1300

1000

137

50

87 ‘
1.093 x 107

Hastelloy N
0.375
0.035

Inner annulus, 13.53
Outer annulus, 14.50

‘Hastelloy N

1
66.7

Hastelloy N

" Top outer annulus, 11/2

Top inner ah.nulus, 2578

Floating head, 33/4

Inner annulus, 4347
Outer annulus, 3794

Inner annulus, 0.600 radial;
0.673 circumferential

Outer annulus, 0.625 triangular
Donut

Inner annulus, 4 -
Outer annulus, 10

188

thermal conductivity was less by about a factor of 3 than the value used in the original heat
transfer calculations. This difference resulted in degradation of the overall heat transfer co-
efficient, necessitating an increase in heat transfer surface of the heat exchanger of approxi-
mately 20%. A new heat exchanger design with a larger number of tubes and a slight change
in the baffle arrangement was developed. The parameters of this new heat exchanger are given
in Table 9.2. B

While the new heat exchanger has a higher fuel holdup by reason of the additional tubes, a
more favorable arrangement of the lheader and other reductions in parasitic volumes compen-
sated for this larger number of tubes, so that the net fuel inventory in the heat exchanger—pump
complex did not change significantly from that required in the first heat exchanger used in the
reactor study.

A significant change in the heat exchanger design concerns the method of flanging the heat
exchanger into the coolant circuit piping. The coolant flows up around the heat exchanger and
is discharged symmetrically across the outer bank of fuel tubes. In this arrangement the toroidal
coolant header is replaced by a jacket around the tubes. Figure 9.7 shows the arrangement of
this new heat exchanger. | .

After traversing the outside of the fuel tubes in a countercurrent flow pattem, the coolant
salt leaves the heat exchanger through a central discharge pipe which connects to the center
of the concentric coolant-salt line through a slip joint. The entire heat exchanger can be re-
moved from the circuit by opening the large flange, éutting the large fuel pipe communicating
with the reactor, and disengaging the center fuel pipe by slipping it out of the slip joint pro-
-vided for this purpose. The drain and overflow lines and auxiliary pump lines must also be
cut to remove the exchanger.

Removing and replacing a primary heat exchanger is a major repair. However, with the de-
‘sign as shown, the difficulties appear to be less formidable than if both coolant lines had to
be cut and welded to perform the task. Adequate remote cutting, aligning, welding, and in-
specting equipment and methods must be demonstrated for this application.

Because of the expansion of the fuel tubes, the toroidal header at the bottom of the ex-
changer will move relative to the vessel enclosing the exchanger. For this reason and also
to permit removal of the exchanger, a drain line cgnnot be connected to this header. Instead,

a dip line for draining fuel extends into this header from the top of the exchanger. This line
connects through a freeze valve to the fuel drain tanks for the reactor. When the drain valvé

is thawed and the drain tank is vented, the fuel salt in the heat exchanger flows through this

line into the drain tank. The gas pressure in the pump bowl is sufficient to effect a complete
drain of the fuel salt from the system. | _

In case of a pump stoppage with the attendant necessity to drain the reactor, the following
sequence takes place. Upon pfimp stoppage the reactor automatically drains into the pump sump
tank and through the 5-in. overflow line; the salt goes to the drain tank. A gas flow into the re-
actor inlet plenum is furnished by the gas separator lines from the pump sump cover gas. This

automatic drainage takes place in the order of a few seconds. However, the concentric lines
189

ORNL-DWG 67-3080A

UPPER BEARING =Y B
N W7
SEAL 1
! GAS CONNECTION
/ i // :
SHIELDING A )
L4 v /
/—* — 3£t Oin.
START-UP LEVEL
//_
M| /
&-ft O-in ODTANK —___|
att Qin _
M.S. BEARING \ | | HieH 0PR. LEVEL
| l—Low OPR. LEVEL
IMPELLER % | — L
, Sia | 15 in.
FUEL TO 7 N \
REACTOR | ~ ~—T""F———1——FILL AND DUMP
. 3 \ ) 32% in.
________ l |
4ul mll' |
i !H”H[h A‘\\;\ - \IHIJ -
| ! dlll"|l|!l" A r”"hih"fltl il M il
|
—A I
f X 1l
| m‘ ~48t1 Oin.
|
OUTER TUBES L L
3794 AT ¥in, oo M I | N
| R
INNER TUBES ———_ [l |l | @t ein

4378 AT 3gin.OD

6ft 6%in. ODx ! g
Kij

FUEL DRAIN

~‘
~
r

P

l

i
mm!!!!fllm\

\' AVl

Fig. 9.7. Primary Heat Exchanger and Pump.
190

to the reactor and the head space above the fuel tubes in the heat exchanger remain full of salt |
and will overheat if not cooled. '

When the pump stops, the pressure in the lower header.of the heat exchanger drops to the
static pressure of the salt, about 14 psi. If now the gas valves in the line are opened, the
stored gas will flow into the bottom of the heat exchanger, displacing the fuel into the pump
sump from which it overflows into the drain tank. By proper sizing of this tank, enough fuel
can be displaced f-rom the heat exchanger to bring the level at equilibriumvbelow the top tube
sheet of the exchanger. The lines and volumes above the tube sheets will be empty of salt,
and no overheating will result. ' i '

We have calculated the cooling required for fuel in the tubes of the heat exchanger under .
the stagnant conditions. Because of the geometry of the coolant salt system, there ig con-
vective flow of coolant salt even when no coolant pump is running. This convective flow of
coolant is adequate to remove aftetheat from the stagnant fuel in the tubes.

To provide cooling but to avoid overcooling requires some modification of the steam circuit
of a module. We have not studied the modifications in sufficient detaill to justify a discussion
at this time. ' | _ i

The large sump tank on top of the heat exchanger is not intended to hold the drained fuel
from the reactor. Because of the undesirability of installing the afterheat cooling system in
the reactor cell and also in order to be able to drain the lines to the reactor, an overflow pipe
was installed from the sump tank to a separate dump tank. This tank is equipped with a cool-
ing system a{dequate for afterheat removal.

The sump tank on top of the heat exchanger appears to be larger than necessary inasmuch
as excess fuel flows through the 5-in. drain line to the drai;'l tank. The size of the tank is
dictated by the fact that it must hold the reactor volume (83 ft3) upon starting the reactor.
Fuel cannot be forced by gas pressure out of the drain-tank fast enough to keep the fuel pump
primed on startup. The fuel must be displaced into the sump tank and held there by gas over-

ptessure in the drain tank until the reactor is filled by the action of the pump.

Not shown in the drawing of the heat exchanger is the gas separator and associated ap-
paratus for handling the sparge gas for the fuel system. The plans for handling this gas are
being worked out, and drawings of the equipment are not yet ready. The criteria and flowsheet

are discussed in Sect. 4. ,

Blanket Heat Exchanger

The blanket heat exchanger is shown in Fig. 9.8. This heat exchanger is similar to the fuel
heat exchanger but considerably smaller and less. complicated. There is no need for the gas
sparge system, and there is no afterheat 'pro'blem that has to be provided forhere. In addition,
the coolant flow is less complicated since counterflow is unnecessary. The femperature\dif-
ference is modest between blanket salt and coolant, and the total rise in temperature is only
11°F as the coolant passes through. Table 9.3 gives the characteristics of the blanket ex-

changer.
191

_ UPPER BEARING

ORNL-DWG 67-3079A

—- i St :

i %AN:Z SHIELDING
%8
!'/

%

SEAL

il . y
24in. OD

3t Oin. ' / BAFFLE
; | OPERATING LEVEL

e /
p //— MOLTEN SALT BEARING
| == BLANKET "PROCESSING
— [— DRAIN
—— — \—-. . IMPELLER-20-in. PITCH DIAM
. TS '
1f1 7 in, |s/ ‘\g\ Al —— 4 - BLANKET TO REACTOR
7 R ,&‘\‘\A\‘ \% e
. (ZET AL L L 77 S
23 ft 7in. ‘ - FLANGE
Vi
: GAS CONNECTION
GAS SKIRT
| OUTER TUBES-827 AT ¥ in. OD
9ft Oin.
L INNER TUBES-814 AT % in. OD
-4 ft 5in. 0D
6in
+
3ft 6in.
R R
N /—'——
s ) i
\ k%
. \ f’ COOLANT SALT
A

Fig. 9.8. Blanket Heat Exchanger and Pump.
192

Table 9.3. Blanket Heat Exchanger Data Sheet

Blanket-coolant-salt exchanger, one-shell pass, two-tube pass with disk- and

donut-type baffles

Number required
Rate of heat transfer, Mw
Rate of heat transfer, Btu/hr

Shell side
Hot fluid or cold fluid
Entrance temperature, °F
Exit temperature, °F
Entrance pressure, psi®-
Exit pressure, psi®
AP across exchanger, psi®
Mass flow rate, 1b/hr

Tube side
Hot fluid or cold fluid
Entrance temperature,. °F
Exit temperature, °F
Entrance pressure, psi'Ei
Exit pressure, psi®
AP across exchanger, psib
Mass flow rate, 1b/hr
Velocity, fps

Tube material

Tube OD, in.

’i‘ube thickness, in.

Tube length, tube sheet to tube sheet, ft
Shell material

Shell thickness, in.

_Shell ID, in.

Tube sheet material

Tube sheet thickness, in.

Number of tubes

Pitch_of tubes, in.

Total heat transfer area, t'tA2
Basis for area calculation
Type of baffle

Number of baffles

Baffle spacing, in.

Disk OD, in.

Donut ID, in.

Overall heat transfer coefficient, U,
Btu hr— ! ft—2

4
27.75
9.471 % 107

Cold (coolant salt)
1111

1125

138

129

15

1.685 x 107 »

Hot (blanket salt)

1250

1150

111

20

91

4.3 x 10°
10.5

Hastelloy N
0.375

0.035

8.3
Hastelloy N
L

40:78
Hastelloy N
1

Center section, 810

Annular section, 810
0.8125

1318

Tube OD

Disk and donut

“Includes pressure due to gravity head.

bPress‘.ure loss due to friction only.
193

9.5 REACTOR PHYSICS
O. L. Smith

In addition to optimization studies describe;i above, work on MSBR reactor physics in-
cluded (1) a series of cell calculations performed to examine the sensitivity of the MSBR
cross sections and reactivity to various changes in cell structure and composition, and (2)
several two-dimensional calculations of the entire reactor. All of the calculations were based
upon the most accurate description of the system that was available as of January 1, 1967.
The graphite-moderated portion of the core was 10 ft in length and 8 ft in diameter, contained
~0.2 mole % 233U, 27 mole % 232Th, and had a fuel volume fraction of 16.48% and a fertile
volume fraction of 5.85%. The reactor had a 1.75-ft-thick radial blanket consisting of 60%
fertile salt and 40% graphite, surrounded by a graphite reflector 6 in. thick. The top axial
blanket was 1.5 ft thick and contained 60% fertile salt and 40% graphite. The bottom axial
blanket was 1 ft thick and contained 3.18% Hastelloy N, 16.48% fuel salt, and 80.34% fertile
salt. A number of structural details below the lower blanket were included in the two-dimen-
sional calculations. |

Figure 9.9 shows the geometry of a cell in the graphite-moderated core. Graphite dowels
of appropriate size are located at the six corners of the cell to yield the desired fertile salt
volume fraction. The cell calculations were performed with the code TONG and involved
varying (1) cell diameter, (2) fuel distribution (i.e., fuel separation distance, s), (3) 233y
concentration, (4) 232Th concentration, (5) fuel volume fraction, and (6) fertile volume frac-
tion. Each of these parameters was varied separately while holding the others constant.
Table 9.4 and Fig. 9.10 show the effect on reactivity of varying the parameters. The varia-
tions are shown relative to a reference cell which had a diameter (flat to flat) of 3.156 in.

and a fuel separation distance, s, of 1/8 in.

ORNL-DWG 67-4789

~3\
COS N
\ @ ./ ‘ »

Fig. 9.9. Cell Geometry.
¢

194

Table 9.4. Effect on Reactivity of Yarying

Cell Properties

Percent
Case Variation eff
1 Reference cgll 0
2 Cell diameter —36.7 - 0,0271
3 Cell diameter +58.4 +0.0317
4 Fuel separation, s —~100 ~—0.0011
5 Fuel séparation, s + 200 +0.0003
6 233y concentration —25 —0.1031
7 233y concentration +25 +0.0763
8 232, concentrat'ion —25 +0.0936
9 232Th concentration + 25 —0.0768
10 Fuel volume fraction - 25 —0.0851
11 Fuel velume fraction + 25 +\0.0512
12 Fertile volume fraction —~25 +0.1048
13 Fertile volume fraction + 25 —0.0845
) ORNL-DWG 67—4790
PERCENT FUEL SEPARATION CHANGE
-100 0 100 200
0.15
) [ | |
S
0.0 e\p)/(é\ -
N\, 55
e o
. (&)
= N% L DIANETER |
3 . FUEL SEPARATION s |
.,t —y
.-—-—"--- I
< // N edé’
<© NH
- N N,
< vy Y
& \/P
\1\3 \ CFy
O/ O
-010 "—QQ@\’I‘f 7
-40 -20 0 20 > 40 60
- PERCENT CHANGE

Fig. 9.10. Effect on Reactivity of Changing Cell Properties.
195

The results of these calculations (compare cases 1 and 3) indicate that there may be a
. reactivity advantage (attributable to increased self-shielding of the ?32Th resonances) to
.using a cell somewhat larger than the 3. 156-in. reference cell. If the conversion ratio is’
not a'dversely affected, use of a 5-in. cell may, for example, pemit reduction of the 233U
inventory. Pending further study, case 3 (which differs from case 1 only in cell size) is
considered to represent the current cell dimensions used in the design.

Table 9.5 and Fig. 9.11 show information about the flux distribution for case 3. Table
9.5 shows the ratio of the average flux in the fuel to thé cell average flux, the ratio of the
a\llerage flux in the graphite to the cell average flux, and the ratio of the average flux in the
fertile salt to the cell average flux for the epithermal and fast flux ranges. Figure 9.11 shows
the thermal flux distribution in the cell. The results in Fig. 9.11 are based upon an annular
approximation to the cell of Fig. 9.9. _ .

The two-dimensional calculations used the 5-in.-diam cell composition and cross sections

for the core, and they were performed with the diffusion theory code EXTERMINATOR-2. Nine

Table 9.5, Flux Ratios in Epithermal and
Fast Energy Ranges

Energy Range Fuel Graphite ' Fertile

~ 0.821-10 Mev 1.226 0.929 0.878

0.0318-0.821 Mev  1.090  0.984  0.958

1.234--31.82 kev 1.014 0.998 0.991
0.0479~1.234 kev 1.0 1.0 1.0
1.86—47.9 ev 1.0 1.0 1.0

ORNL—DOWG 67-4731

2.0
w 1.5 CELL AVERAGE FLUX
g 7 ~~
- — — -t e
S )
5 10 ‘
§ ——  FUEL _"I"“/ ‘—l' \ii GRAPHITE —-‘ i-¢—~
>:<) GRAPHITE FUEL ?\_FERT;LE
—J
z 0.5

o]
o) i 2 3 4 5 6 : 7 8 9

RADIAL POSITION {cm)

Fig. 9.11. Thermal Flux Distribution in Cell (E <1.86 ev).
196

energy groups were used and are shown in Table 9.6. Figure 9.12 shows the radial flux dis-
tribution in 'the core midplane for neutron energy groups 1 and 6 separately, and Fig. 9.13 shows
the total flux distribution. The radial peak-to-average flux ratio in the graphite-moderated part
ofthe core is 1.58. Figure 9.14 shows the axial flux distribution for groups 1 and 6 at a radial
distance 18 in. from the axis of the core, and Fig. 9.15 shows the total flux distribution. The

. axial peak-to-average flux ratio in the graphite-moderated core region is 1.51. Thus the total
peak-to-average flux ratio is 2.39, the peak occurring at the geometric center of the graphite-
moderated core. ‘

The central cell of the reactor is intended for control purposes and consists of a graphite
tube 5 in. in outside diameter and 4 in, in inside diameter. It is envisioned that control will
be achieved by regulating the height of the fertile salt in the tube. If the completely empty
tube is filled with fertile salt, .the change in reactivity is &k/k = ~0.018%. If the empty
tube is filled with graphite, the reactivity change is dk/k = +0.0012%. Thus there appéars
to be a substantial amount of reactivity control available by varying the height of the fertile

column in the tube.

Table 9.6, Neutron Energy Groups Used in

Two-Dimensional Calculations

Group Energy Range

Pk

0.821-10 Mev
0.0318—~0.821 Mev
1,234-31.82 kev
0.0479-1.234 kev
1.86—47.9 ev
0.776~1.86 ev
0.18—0.776 ev
0.06--0.18 ev

O 0 3 O o s W N

0.01-0.06 ev

197

ORNL-DWG 67-4792

{x10'%

2.0

1.6

.4 CORE ~mfe—BLANKE T— 7|

-2

GROUP 1

/

FLUX (neutrons cm

o
®

GROUP 6
0.6 \ \

0.4

0.2

0 { 2 3 4 5 6
RADIAL POSITION (ft)

Fig. 9.12. Radial Distribution of Neutron Flux in Groups
1 and 6 at Core Midplane.
x10'%

-2

FLUX (neutrons cm

198

ORNL-DWG 67-4793

(x10'3)

2.8 \

N
IS
L~

)

CORE BLANKET

sec
M
Q

[o)]
=

n

- FLUX (neutrens em™2

o
®

0.4 : \\ ,
—

0 1 2 3 4 5 6 7

RADIAL POSITION (ft)

Fig. 9.13. Radial Distribution of Total Neutron Flux
at Core Midplane,

ORNL-DWG 67— 4794

14 -
/BOTTOM OF / _ TOP OF

CORE / CORE —
12 N [~
10 [ BLANKET / - \ BLANKET
08 / T e GROUP 6 ‘

Zj: N
Y

0] 1 2 3 4 5 6 7 8 9 10 1" 12 13 iq 15
' AXIAL DISTANCE FROM BCTTOM OF REACTOR (ft)

Fig. 9.14. Axial Distribution of Neutron Flux in Energy Groups 1 and 6 at a Distance 18 in. from Core Axis.
{x10'3)

FLUX (neutrons cm™2 sec !y .

0.8

199

ORNL-DWG 67- 4795

BLANRET-\\‘ 1 / / | A TOPOi—'CORE
i/ NEEEER

0 1 2 3 4 5 6 . 7 8 9 10 " 12 13 14 15
‘AXIAL DISTANCE FROM BOTTOM OF REACTOR (ft) '

Fig. 9.15. Axial Distribution of Total Neutron Flux at a Distance 18 in. from Core Axis.

9.6 MSBR GAS HANDLING SYSTEM

Dunlap Scott  A. N. Smith  R. J. Kedl.

The gas handling system for the MSBR serves several functions in the operation of the

plant. These include: supplying helium for purging and pressurization of the gas spaces;

rapidly removing !?5Xe from the salt; transporting, removing, and storing the radioactive

fission product gases; and processing the helium for recycle into the gas supply‘ or disposal

to the atmosphere. A scheme for accomplishing these functions is described below.

Xenon Removal

Preliminary studies of the migration of '*°Xe to the graphite in molten-salt breeder re-

‘actors indicated that it would be possible to reduce the '35Xe poison fraction to an acceptable

value of 1/2% if a gas stripping system could be devised which would process the entire reactor

" fuel-salt inventory in about 30 sec. Since the reactot systém salt circuit time is about 9 sec,

it is required that the entire inventory be processed every 3 to 4 passes around the circuit.

The required xenon processing time could be extended by processing the fuel for removal

of the precursor, 1*°l. However, the advantages of '*°I processing are limited by the fact that

20% of the total !35Xe produced in the fissioning of 2*3U appears directly, and very high proc-

essing rates would be required to obtain even modest gains. For example, if the entire fuel in-

ventory were processed for 13°I removal once every hour, the xenon removal process rate would

only be extended to about 80 sec, and a very high !l processing rate would increase the xenon
200

processing time to only about 110 sec. Therefore; processing for iodine is not, at this time, being
considered as a method of removing !%°Xe. '

The method proposed for removal of 3%Xe involves stripping of xenon-enriched helium bubbles
from the fuel stream by means of a gas separator located at the heat exchanger outlet. The bub-
bles will be'generated by injecting helium into the salt at the fuel pump suction-, and transfer of

xenon from the salt to the gas will be effected during passage through the heat exchanger.

. Mechanical Design

The flowsheet for the MSBR dff-gés system is given in Fig. 9.16. Helium is injected into
the suction of the fuel salt pump and is removed by centrifugal separation at the heat exchanger
outlet. The liquid-gas mixture from the separator is then féd into a cyclone separator where the -
entrained liquid is removed and returned to the pump bowl.. The gas from the pump shaft purge
and the instrument lines is combined with that.from the cyclone separator and fed into the 48-
hr xenon holdup system. A-small portion of the flow from this system is fed into the long-term
xenon holdup and then through the noble-gas separator to a recycle system for supplying clean
helium to the pump shaft purge and the instrument gas lines. The remainder of the gas from the
48-hr xenon holdup system is fed into the gas injector system at the fuel pump suction. It is
the salt-powered gas injector that serves as the prime mover for the high-gas-flow recycle

system. Some of the design criteria of the critical components are described below.

Gas Injector System

1. The gas addition rate shall be sufficient to produce about 1% void volume in the salt
at the pump suction,

2. The location and manner of gas injection shall provide a balanced distribution of bubbles
in the liquid entering the pump impeller. This consideration affects the bubble distribution in
the heat exchanger and the bubble size, as well as the pump denamic balance.

3. There shall be operator control over the gas addition rate into the gas injector.

4. The supply for the gas injector shall be taken from the outlet line from the 48-hr xenon
holdup helium recirculation system. There shall be a backflow preventer arrangement at this
gas supply point to prévent salt from getting back into the charcoal beds as a result of a-
sudden pressure transient in either the salt .or gas system. | ‘ /

5. The method of injecting bubbles into the pump suction shall be a fuel-salt-powered jet
pump taking. ité. salt supply from the heat exchanger discharge.

The suction pressure for the salt-powered jet pump shall be less than 10 psia at a flow rate

of 6.5 scfm of helium.
iy

BLANKET

PUMP

CYCLONE
SEPARATOR

MISCELLANEOUS PURGE GAS

‘\

. 12
25 To mopu

3 4

sl
>

MODULE
HEADER

1234

CW= COOLING WATER

POINT  ft3/sec

) - o

25
25
25

Q@@

psia
277

'64.7

457
327

" HX = HEAT EXCHANGER

TEMP (°F)
1000

1000
1000
1000

LES

VOLUME
HOLDUP
AND FISSION
PRODUCTS

HELIUM
SUPPLY COOLANT PUMP PURGE ——=
DRAIN TANK VENT ——]
ACCUMULATOR
&
FILTERS
AND
PRE - TREAT
93 HI-FLOW
AUXILIARY
SURGE CHARCOAL
TANK BED
X

CHEMICAL PROCESS PURGE ———=

ORNL-DWG 67-3077A

HX CW HX cw of
. 90-day
clflfi%%i\l_ 2G
@ . - BEDS
[ { ? | X
] CW
e 13 113 {
3
{
gl [ 3 EITT13
. 91).
£ 3 3 3 s
T T
13 3 & 3
Kr -85 AND
T TRITIUM REMOVAL
I

PRI

psia TEMP (°F)

247
24.0
23.0
18.0

70
247
247

1000
1000
230
230
1000

DATA

_U
o
Z

BORO®O®B®

MAINTENANCE
CONTAINMENT
VENTILATION

psia TEMP (°F)

scfm
200 180
200 165
200 150
2.00° 447
200  44.7.
025 247
50 MAX 17.7 MAX
50 MAX 17.2 MAX
SO MAX 147

Fig. 9.16. MSBR Gas Flow Diagram,

100
70
70
70

CIOD1NE
TRAP AND
ABSOLUTE
FILTERS

NOTE :

TOTAL HEAT LOAD ON 48-hr
. DELAY SYSTEM IS ~ 5 Mw

102
202

Bubble Sepnrofor System l _

1. The bubble éeparator shall be installed within the outer annulus of the fuel-salt line from
the heat exchanger to the reactor. ‘

2. The separator, including the entrance region, shall be kept as short as is reasonably pos-
sible, and the separator shall be as close to the heat exchanget as good design pemits. The
purpose of this restriction is to pemit removal of the separator along with the heat exchanger
and at the same time permit locating the heat exchanger close to the ;eactor.

3. The design of the system shall include assurance that the separator will not admit gas
into the salt line to the reactor except when the fuel pump stops. Check valves or liquid sub-
mersion of the discharge will not be used to prevent backflow, since the reactor drain scheme
will use the separator outlet as a gas source. |

4. The bubble fraction of the salt as it enters the reactor shall be less than 0.1% during
steady-state operation.

5. The fuel line outside the gas separator shall be free of protrusions all the way to the
heat exchanger outlet. Equipment for cutting and welding the 24-in.-diam fuel line will occupy
this space during some maintenance operations. |

6. The gas outlet from the bubble separator and the gas line to the pump ‘bowl shall stay
within the salt lirie if possible. This will reduce the number of penetrations through the pipe
wall.

7. A modified gas cyclone separator shall be an integral paft of the bubble separator dis-
chérge stream. The liquid would be discharged to the pump bowl, and the gaseous discharge

would go to the xenon holdup system.

- Yolume Holdup System

The first stage of the high-flow 48-hr xenon holdup is a gas volume holdup whose several
functions are listed below.

1. The first stage shall be a cooled volume for the decay of the very short-lived gaseous
fission products. A reentrant tube system similar to the one in the fuel drain tank of the MSRE
* will be investigated for providing the cooling.

2. There shall be a provision for a final demisting of the gas as it comes from the cyclone
separator. ‘

3. Since there will be a total of more than a kilogram of fission products produced each day
in the reactor complex, and since a sizable fraction is involved in the short-lived gaseous
products, a method of collecting, cooling, and conveying to a disposal area must be included.
The collecting device should pass only gas which has resided in the volume holdup for at

least the specified minimum transit time. | ‘

4. The volume holdup system shall be sized to reduce significantly the heat load and
particulate loading rate from short-half-life fission products in the first stage of the charcoél

" bed, which is immediately downstream.
203

Noncritical Components .

Many of the remaining components in the off-gas handling system ate reasonably well
understood, and, while they must be designed and ultimately tested, it is believed that they

will offer no critical problems. Some of these are listed below:

1. High-gas-flow charcoal bed design.
Biological charcoal bed design (low flow).

3. Noble-gas separator and disposal system. This includes 85Kr and tritium removal from the
helium which is to be returned to the recycle system. .

4. Helium compressor for recycle of clean helium.

Gas sampling system for purity control and for surveillance of the chemical condition of the
salt. ‘

6. Instrumentation and controls for operating the gas system.
10. Molten-Salt Reactor Processing Studies
M. E. Whatley

A close-coupled facility for processing the fuel and fertile streams of a molten-salt breeder
reactor (MSBR) will be an integral part of the reactor system. Studies are in progress for obtaining
data relevant to the engineering design of such a processing facility. A

The fuel processing plant will operate on a side stream withdrawn from the fuel stream which
circulates through the reactor core and primary heat exchanger. For a 1000-Mw (electrical) MSBR,
approximately 14.1 ft® of salt will be processed per day, which will result in a fuel-salt cycle
time of approximately 40 days. The presently envisioned process has been described previously; !
the significant process steps are recovery of the uranium by continuous fluorination, recovery of
the carrier fuel salt by vacuum distillation, and recombination of the i)urified UF _ and barren car-

rier salt.

10.1 CONTINUOUS FLUORINATION OF AMOLTEN SALT
L. E. McNeese B. A._ Hannaford -

Uranium present in the fuel stream of an MSBR must be removed prior to the distillation step,
since UF | present in the still would not be completely volatilized and would in part be discharged
to waste in material rejected from the distillation system. Equipment is being developed for the
continuous removal of UF, from the fuel stream of an MSBR by contacting the salt with F, in a
salt-phase-continuous system. The equipment will be protected from corrosion by freezing a layer
of salt on the vessel wall; the heat necessary for maintaining molten salt adjacent to frozen salt
will be provided by the decay of fission products in the fuel stream. Present development work
consists of two parts: (1) studies in a continuous fluorinator not protected by a frozen wall and
(2) study of a frozen-wall system which is suitable for continuous fluorination, but with an inert
gas substituted for fluorine in the experiments. Work on the two systems is described in the fol-

lowing sections.

Nonprotected System

The experimental system has been described previously! and consists of a 1-in.-diam fluori-

nator 72 in. long and auxiliary equipment which allows the countercurrent contact of a molten salt

IMSR Program Semiann. Progr. Rept. Aug. 31, 1966, ORNL-4037, p. 232.

204
3

205

with F . It is intended to demonstrate the effectiveness of this type of system for uranium recov-
ery and to obtain engineering design data on backmixing and F , utilization. The present work
has used salt containing no beryllium, but LiF-BeF2 mixtures will be used when suitable facil-
ities for handling beryllium are made available.

This system also allows development of techniques and auxiliaries required for continuous
fluorinators. These include means for controlling and measuring molten-salt flow rates and
methods for sampling and analyzing molten-salt streams and gas streams. A gas chromatograph

is used for analysis of the fluorinator off-gas for UF _, F,, and N,. Uranium concentration in

the salt dufing fluorination is determined from salt sampli:s taken at 15-min intervals.

Data obtained from this system have been reported previously;! three additional experiments
have been made at 600°C using an NaF-LiF-Z:F , mixture containing 0.37 to 1.16 wt % UF, and
having a melting point of ~.475°C. Salt feed rates of 9.8 to 30 cm?®/min and F, rates of 235 to
335 cm3/min (STP) have been used with molten-salt depths of ~.48 in. Uranium removal during
one pass through the flu’orinator varied from 99.36 to 99.89%, as determined from salt samples.

During the best run' to date, a molten-salt feed rate of 10.2 cm3/min gnd an F, feed rate of
235 cm?®/min (STP) were maintained for a 2.5-hr period. At steady state, 99.89% of the uranium
was removed by the fluorinator, as determined from inlet and exit concentrations of uranium in the
salt. Later in the run, a molten-salt feed rate of 22.5 cm3/min and a/n-’F2 rate of 270 cm?/min

(STP) were maintained for a 2.5-hr period. With these conditions, the uranium removal at steady

- state was 99.62%. The uranium concentration in the feed salt was 1.16 wt %, and fluorine utiliza-

tions were 15.6 and 30% respectively. The equipment operated smoothly during the run, salt and
gas feed rates were constant, and the system was operated at steady state for 2 hr at each of the
above conditions. |

The available design parameters include the ratio of salt and fluorine flows and the height of
the tower. Any recovery can be attained, and an economic optimum will probably fall above

99.9%.

Protected System

Components for the protected system have been fabricated and are being installed. This
system will be operated with a frozen layer of salt on the fluorinator wall and will allow the.
countercurrent contact of an inert gas with molten salt in equipment suitable for continuous
fluorination.

The fluorinator is constructed from 5-in.-diam sched-40 nickel pipe and will provide a section
protected by a frozen wall ~6 ft high. Internal heat generation is provided by Calrod heaters
blaced inside a 3/4-in.-diam nickel tube éldng the center line of the vessel. The frozen-wall thick-
ness will be dependent on the radial heat flux and can be varied from 3/8 to 11/2 in. Frozen-wall
thickness will be determined from temperature gradients measured by two sets of four internal
thermocouples located at different radii with respect to the vessel center line. A feed tank
and receiver vessel will allow the feeding of a salt volume equivalent to approximately ten

fluorinator volumes through the system.
206

The system will be operated with a salt mixture (66 mole % LiF—34 mole % Z:F ,) having a
phase diagram similar to that of the LiF-BeF , fuel salt for the MSBR and should point out problems

associated with the operation of a frozen-wall fluorinator.

10.2 MOLTEN-SALT DISTILLATION STUDIES

J. R. Hightower L. E. McNeese

Relative volatilities of several rare-earth fluorides with respect to lithium fluoride have been
measured in an equilibrium still operated at 1000°C and (.50 mm Hg pressure. Recent work has
refined previously reported values for CeF ,, LaF ,, and NdF ; and has added the relative volatility
for SmF ,. The rate of vaporization of LiF has been measured at 1000°C at several pressures;
these data are use;ful in estimating the probable error in measured relative volatilities and for
predicting vaporization rates in eqfiipment suitable for MSBR processing. The results of a study
on the buildup of materials of low volatility at a surface where vaporization is occurring are also

given.

Relative Volatility Measurement

A distillation step will be used in the MSBR processing plant to remove rare-earth fission
products from the fuel stream. To design the still it is necessary to know relative volatilities
of the rare-earth fluorides with respect to LiF, the major constituent of the still pot. The equi-
librium still used for these measuremenfs and the operating procedure have been previously de-
scribed.? Relative volatilities of four of the rare-earth trifluorides with respect to LiF at 1000°C

and 0.50 mm Hg have been obtained from recent experiments and are listed below.

Rare-Earth Fluoride Mole Fraction Relative Volatility with Respect to LiF

CeF, " 0.02 | 3x 103

LaF, 0.02 3x107* -
NdF, 0.05 _ 6x10* _

SmF 0.05 2% 10~4 | b

These relative volatilities allow the required rare-earth fluoride (REF) removal efficiencies in a

still of simple design without rectification.

Vaporization Rate Studies

Data on the variation of vaporization rate with total pressure are necessary to assess the
error in relative volatilities measured in the recirculating equilibrium still and to predict vaporiza-

tion rates in equipment suitable for MSBR processing. A diagram of the equipment used for the

21bid., p. 229.

o
[ Y

x

207-

measurement of vaporization rates. is shown in Fig. 10.1. The distillation unit was made from 1-
in. nickel tubing bent into an inverted U. The salt was vaporized from a graphite crucible in the
left leg of the still, and the condensate was collected in a similar crucible in the right leg.

For operation of the still, a mixture oflLiF and a rare-earth trifluoride having a known com-
position was placed in a crucible in the vaporizing section of the still, and a second crucible
was placed below the condenser. The system was purged with argon while being heated to the
desired temperature. When the temperature in the vaporizing section reached 1000°C, the still
pressure was decreased to the desited value, and the condenser cooling air was turned on. The
condenser temperature of about 500°C caused the salt vapor to solidify on the condenser walls.
After a given length of time the St‘ili was pressurized with argon and fhe cooling air turfled off;
this allowed the condensate to melt and drain into the crucible below the condenser. Vaporization
rates were determined from the decrease in weight of the salt in the vaporizing section of the still,
since only part of the salt vaporized was collected in the other crucible. Results of these tests
are given in Table 10.1.

At pressures near the v;apor pressure of LiF, the LiF vaporization rate increased slowly as
the pressure was lowered. At a pressure significantly lower t.han the LiF vapor press‘ure, the rate

increased by an order of magnitude. When the total pressure is higher than the vapor pressure of

ORNL—-DWG 67 - 15074

1—in. NIGKEL
TUBING —=

faN

¢ ANANANANAN
T

CONDENSER

. N . ¥

GOOLING AIR

~=—— FUNNEL

f'\’
THERMOWELL—e{T]
VAPORIZ!NG/‘ :H
SECTION
CRUCIBLE — CONDENSING
SECTION
CRUCIBLE
: CRUGIBLE
. =~ SUPPORTS
ARGON ~ Y '
NLET —=C— —— vacu

SWAGELOCK VACUUM
EEH/ FITTINGS \Efé PUMP

Fig. 10.1. Appuratus for Vaporization Rate Measurement.
208 .

Table 10.1. Variation of LiF Vaporization Rate
with Total Pressure at 1000°C

bondenser Pressure®’ Vaporization Rate
(mm Hg) (g cm ™2 sec“l)
1.0 7.8 x 107°
0.50 3.3x 1073
0.35 48X 1077
0.1 ' 2.4 x 107*

“The vapor pressure of LiF at 1000°C is
about 0,53 mm Hg.

the salt, the rate of vaporization should be controlied by the rate of diffusion of LiF dand REF
through the argon present in the system; the measured rate .at 1.0 mm Hg was comparable with the
rate calculated by assuming the rate to be diffusion controlled. Since these data indicate that the
recirculation rate in the equilibrium still is controlled by diffusion of the salt vapor through argon,
an error in the measured relative volatilities could arise because of differences in the rates of dif- .
fusion of LiF and REF vapor. Calculations indicate that the error in the rel\ative volatility from .
this source is only about 1% and that other effects such as a nonuniform concentration gradient

in the liquid are much more important.

Buildup of Nonvolatiles at a Yaporizing Surface

" During vaporization of a multicomponent mixture, materials less volatile than the bulk of the
mixture tend to remain in the liquid phase and are removed from the liquid surface by the processes
of convection and molecular diffusion. Low-pressure vaporization does not generate deeply sub-
merged bubbles and therefore provides little convective mixing in the liquid. An appreciable
variation in the concentration of materials of low volatility may occur if these materials are
removed by diffusion only. 7

Consider as an exemplary case a continuous still of the type showr} in Fig. 10.2. Fuel carrier
salt (LiF-BeF 2) containing fission product fluorides is fed to the bottom of the system contin-
uously. Most of the LiF-BeF , fed to the system is vaporized, and a salt stream containing most
of the nonvolatile materials is withdrawn continuously. The positive x direction will be taken as
vertically upward, and the liquid withdrawal point and the liquid éurface will be located at x = 0
and x = [ respectively. Assume that above the liquid withdrawal point, molten LiF containing
REF flows upward at a constant velocity V. At the surface, a fraction v/ V of the LiF vaporizes,
and the remaining LiF is returned to the bottom of the still.

Above the withdrawal point, the concentration of REF satisfies the relation

d?’C _ dC - |
D— -V _—_=0, . (1)
dX2 dX_ : .
209

ORNL-DWG 67-294A

PRODUCT SALT
v,aCy

i

RECIRCULATING
SALT
V-v, Cg

—‘~ x=£

]
|
| x
|
|

| DISCARD SALT
WITHDRAWAL F-v,C,
POINT

FEED SALT
F, Gy

Fig. 10.2. Continuous Still Having External Circulation and a Nonuniform

Liquid Phase Rare-Earth Fluoride Concentration Gradient.

and the .boundary conditions are, at x =1,
dC
-D = |x=1‘+ VC =vaC +(V - v)C_,
and at x = 0,
dC ‘ '
-D E|x=O+VCO=(V—v)CS+FCf—-(F -v)C,,

where
D = diffusivity of REF in molten salt of still pot concentration, cm?/sec, -
C = concentration of REF in molten salt at position x, moles of REF per cm? of salt,
x = position in molten salt measured from liquid withdrawal point, cm, i |
V = velocity of molten salt with respect to liquid surface, cm/sec,

C . = concentration of REF at x = I, moles of REF per cm? of salt,

(2)

3)
210

-»
‘ : cm? LiF (liquid)
F = LiF feed rate, s
cm? vaporizing surface - sec
em?® LiF (liquid)
v = LiF vaporization rate, ,
cm? vaporizing surface - sec
a= relative volatility of REF referred to LiF,
C, = concentration of REF at x = 0, moles of REF per cm? of salt,
C = concentration of REF in feed salt, moles of REF per cm? of salt.
Equation (1) has the solution
FCifl — (w/V) (1 = a) [1 — exp (-V{ - x)/D)]}
Cx) = . 4 .
va+(F =il - /YA - a) [l —exp (-VI/D)]}
The fraction of the REF removed by the still is" ' .
F - v)C,
fraction REF removed = —
FC; '
F—v 5)
" Fevyva/[l = (w/V)] (1 = @) [1 - exp (-VI/D)}
The fractional removal of REF for a continuous still having a perfectly mixed liquid phase
is '
1
fraction REF removed = . (6)
. 1+ [va/(F = w)]
The ratio of the fractional removal of REF in a system having a nonuniform concentration to that
in a still having a uniform concentration will be denoted as ¢ and can be obtained by dividing
Eq. (5) by Eq. (6). Thus '
1+ [va/(F — v)] '
¢ @

1+ ve/F - I/ - /M A = @) [1 — exp (-VI/D))}

' Values of ¢ calculated for a still in which 99.5% of the LiF fed to the still is vaporized
[v/(F-v) =199] and in which the relative volatility of REF is 5 x 10~ * are given in Fig, 10.3.

The following two effects should be noted:

1. The value of ¢ is essentially unity for VI/D < 0.1 for any value of v/V (fraction of LiF
vaporized per circulation through still). Within this region, a near-uniform REF concentration
is maintained by diffusion of REF within the liquid, and mixing by liquid circulation is not re-
quired.

2. The value of ¢ is strongly dependent on v/V for VI/D > 1. Within this region, a near-

uniform REF concentration can be maintained only if liquid circulation is provided. For VI/D =

1]
Lo

&

211
ORNL-DWG 67-30A
LO | | I 1Y II 'i' i i
! |
SN N —
| ~— =0.5 v To.1
0.5 N -
. (.
\, \.«;—209
0.2
0.4 \
\
0.05 \
0.02 \
0.01 : \
a=5xi0""
V_ =499 i
0.005 Fv 0 \
\L=,|
Y
0.002
0.00{ :
0.1 0.2 0.5 1 2 5 10 20 50 100
Ve

Fig. 10.3. Ratio of Fraction of Rare-Earth Fluoride Removed in Still Having

Nonuniform Concentration to That in Still Having Uniform Concentration.

100, ¢ has a value of 0.0055 with no liquid circulation and a value of 0.99 if 90% of the LiF is
returned to the bottom of the still. . ‘ ‘ '

An actual still would probably operate in the region VI/D > 1; so the importance of liquid
circulation cannot be overem[;hasized. Liquid phase mixing by circulation is believed to be an

essential feature of an effective distillation system.

10.3 VACUUM DISTILLATION EXPERIMENT WITH MSRE FUEL SALT

W. L. Carter.

Experimental equipment has been designed for an engineering-scale demonstraflon of vacuum
distillation of molten-salt reactor fuel. The distillation will be carried out at about 1000°C and 1
mm Hg pressure to separate LiF-BeF, carrier salt from less volatile fission products, primarily
the rare earths. Uranium tetrafluoride will not be present during distillation, having been previ-

ously removed by fluorination.
212

This experiment is pért of a program to develop all unit operations (see Fig. 10.4) in the
processing of a molten-salt breeder reactor fuel. Vacuum distillation is the key step in.the proc-
ess because it recovers the bulk of the valuable LiF-BeF  carrier, decontaminated from fission
products, for recycle to the reactor. Feasibility of distilling fluoride salts was established in
batch laboratory experiments by Kelly;? the present experiment will demonstrate the operation
on an engineering scale and furnish data on the relative volatilities of the components of the mix-
ture. \ |

In the interest of simplicity, fabrication time, and economy, no attempt is being made in this
experiment to reprodu'ce actual MSBR operating conditions, such as high internal heat generation
rate in the still volume or the design of a still that cafi serve in a processing plant for a breeder.
Such advances are the next logical step after an engineering-scale demonstration. However, it
is the purpose of this experiment to show that molten salt containing fission products can be fed
continuously to the still at the same rate at which it is being distilled, with the simultaneous
accumulation of fission products in the bottoms. The still can also be operated batchwise to

concentrate fission products in some small fraction of the original charge.

3u. J. Kelly, ““Removal of Rare Earth Fission Products from Molten Salt Reactor Fuels by Distilla-
tion,’” a talk presented at the 11th annual meeting of the American Nuclear Society, Gatlinburg, Tenn.,
June 21-24, 1965,

ORNL-OWG 66 — B2RA

- F, RECYCLE

NaF
SORBER or
REACTOR 100 - 400°C gF, .
coLD
i | SORBER Y REP
BLANKET UFs+ Py
LiF-BeF,-UF, : ’ -
FLUORINATOR :
~550°C
_jet— F2
LiF—BeF,—UF,
LiF + BeF, SPENT
) 2 NaF + MgF, , .
OFF - AND VOLATILE FP’s
GAS
poill)—
STILL
= MAKE-UP - ~{000°C ' UFg
LiF -BeF,-UF, '
H REDUCTION - -
2 WASTE |
FUEL ‘ ' *
MAKE - UP : .
DISCARD WASTE
FORZr RARE EARTHS
REMOVAL INLiF

Fig. 10.4. Principal Steps in Processing Irradiated Fuel from a Molten-Salt Reactor.

i

.

Wl

213

- The experimental program is in two parts: About 90% of the time will be devoted to nonradio-
active operation, and the remaining 10% to radioactive operation in distilling a small quantity .
of fuel from the MSRE. The first phase is expected to log 500 to 1000 hr of operation. The same
equipment is to be used in the radioactive experiment after being thoroughly inspected at the
conclusion of nonradioactive operation. Radioactive runs will be carried out in an MSRE cell.

Components of the experiment are a feed tank (48 liters), still (12 liters), condenser, conden-
sate receiver (48 liters), associated temperature and pressure instrumentation, and vacuum sys-
tem. The still, condenser, and receiver are fabricated as a unit, The vessels are mounted in an
angle-iron frame, which is 3 x 6 x 7 ft high, allowing transport of the entire facility as a unit
once instrumentation, power, and service lines have been disconnected. Each process piece is
surrounded on all sides by shell-type electric heaters; these in tfirn are enclosed in 4 to 8 in. of
thermal insulation. A diagram of the assembly, which shows the feed and sampling mechanisms,

is shown in Fig. 10.5.

An experiment is carried out by charging a molten mixture of carrier salt and fission product

~fluorides into the feed tank, which is held at a temperature slightly above the melting point. In

most cases this is 500 to 550°C. Concurrently the still, condenser, and condensate receiver are
heated and evacuated, and the space above the liquid in the feed tank is evacuated. The final

pressure is adjusted to about 0.5 atm in the feed tank and 0.5 mm Hg in the rest of the equipment.

ORNL-DWG 66—10952 A

VENT —— o
CoRTESL STILL VENT
N J‘p;’: __________ VOL =12 liters
P={mmHg

r=1000°C

! : ]
Np INLET -

i o

VACUUM PUMP

LIQUID N,
COLD TRAP

CONDENSER
10 in. DIAM X 51 in.

CONTAMINATED FEED
LiF =65.6 mole %
BeF,=29.4 mole %
Zrfa = 5.0 moie %
FISSION PRODUCTS

SAMPLER

H
FEED TANK =1
VOL = 48 liters RN CONDENSATE TANK
F =6 psia : VOL =48 liters
7 = 500°C ) ; F=0.5mmHg

T'= 500°C

Fig. 10.5. Vacuum Distillation of LiF-Ber-Zde-

DECONTAMINATED
LiF -BeF,-ZrF,
214

When the still temperature reaches that of the feed liquid, a 12-liter charge is forced into the
still through a heated 1iné, and the temperature is raised to 950 to 1000°C, the temperature range
in which diétillation begins.

The still pot is the highest point of the system and is an annular volume surrounding the top
of the condenser. Salt vapors flow into the top of the condenser and condense along its length by
losing heat to tlile surroundings. Freezing is prevented by supplying the necessary external heat
to keep the éoqdenser surface above the liquidus temperature. In leaving the condenser the
distillate passes through a small cup from which samples can be removed for analyses.

Liquid-level instrumentation in the still allows control of feed rate to correspond to the dis-
tillation rate, which is estimated to be 400 to 500 cm?® of distillate per hour. Determining the
actual distillation rate for molten-salt systems is an important part of this experiment.

All vessels and lines that contact molten salt are fabricated of Hastelloy N. In the region of
the still and upper section of the condenser, the normal use temperature for this alloy may be
exceeded by as much as 200°C. Consequently, the vessels are to be examined thoroughiy by
dimensional, radiographic, and ultrasonic methods before and after nonradioactive operation.

Provision is made for hanging test specimens of candidate metals of ‘construction in the still.

g

215

OAK RIDGE NATIONAL LABORATORY
MOLTEN-SALT REACTOR PROGRAM

FEBRUARY 28, 1967

M. W. ROSENTHAL, DIRECTOR
R. B. BRIGGS, ASSOCIATE DIRECTOR
P_R. KASTEN, ASSOCIATE DIRECTOR

a9on

MSBR DESIGN STUDIES
E.S. BETTIS
. BETTIS*
. BRAATZ™
. CARTER**
. KERR
. LLEWELLYN"
. NELMS*
. PICKEL™
. POLY
. STODDARD*

EEIATIQIEDD
O EP I -Aram

MSRE OPERATIONS

COMPONENT AND SYSTEMS DEVELOPMENT

1&C DESIGN AND DEVELOPMENT

R. L. MOORE 1&C
D. G. DAVIS* 1&C
P. G. HERNDON 1&C
J. W. KREWSON* 1&C
T. M. CATE* 18C
- B. J. JONES 1&C
NUCLEAR ANALYSIS

p—{ B.E.PRINCE R
T.W. KERLIN uv

- FUEL PROCESSING DEVELOPMENT
|| M. E.WHATLEY cT
W. L. CARTER** cT
L. M. FERRIS cT
B. A. HANNAFORD cT
J. R. HIGHTOWER cT
R. B. LINDAUER cT
L. E. McNEESE cT
E. L. NICHOLSON C1
C. E. SCHILLING cT
F. J.SMITH cT
J. BEAMS cT
J. F.LAND cT
C. T. THOMPSON cT

i]
REACTOR CHEMISTRY

EIOMOED

TOMTMC< - —-MJP»P»ITX

nogoro

METALLURGY

ADAMSON*
CooK*
CANONICO"
FRANCO-FERREIRA
KENNEDY
MeCOY*
WERNER”

L. GRIFFITH*
. LANE™
LAWRENCE"
SHOOSTER*
THOMAS®
WADDELL™
WILLIAMS

P. N. HAUBENREICH DUNLAP SCOTT R
NUCLEAR AND MECHANICAL ANALYSIS
COORDINATION ENGINEERING ANALYSIS
C. H. GABBARD™** R 4. R. ENGEL R R. J. KEDL R
o : C. H. GABBARD** R L
H. B. PIPER R
J A WATTS R
OPERATION INSTRUMENTATION AND CONTROLS
R. H. GUYMON R || REMOTE MAINTENANCE
J. H. CROWLEY R | J.R. TALLACKSON 1&C R. BLUMBERG R
P H. HARLEY R R. K. ADAMS” 1&C J. R. SHUGART R
T. L. HUDSON R C.D. MARTIN 1&C
A. |. KRAKQVIAK R C. E. MATHEWS* 1&C
H. R. PAYNE R J. L. REDFQRD 1&C
M. RICHARDSON R R.W. TUCKER 1&C
H. C. ROLLER R S.E.ELLIS 1&C
R. C. STEFFY R C. E. KIRKWOOD 1&C
T. ARNWINE R J.C. LeTELLIER 1&C MSRE COMPONENTS
W. H. DUCKWORTH R T. SAMUELSEN 1&C A. M. SMITH R
J. D. EMCH R - R. B. GALLAMER R
E. H. GUINN R R. E. CARNES 'R
T.G. HILL R C. A GIFFORD R
C.C.HURTT R
1. C. JORDAN R
H. J. MILLER R
L. E. PENTON "R CHEMISTRY
W. E. RAMSEY R L -
R. SMITH, JR. R R. E. THOMA RC
J. L. STEPP R '
5. R WEST R
J. E. WOLFE R
5. 1. DAVIS R
PUMP DESIGK AND DEVELOPMENT
CHEMICAL PROCESSING L] A.G.GRINDELL** R
— - P.G. SMITH R
_ R.B. LINDAUER cT L v WILSON® 5
DESIGN LIAISON D. D. QWENS R
p—
C. K. McGLOTHLAN R
MATERIALS
MAINTENANCE ¥. H. COOK Mec
B. H. WEBSTER \_‘
A.F.GILLEN
L. P. PUGH
AC ANALYTICAL CHEMISTRY DIVISION
CT CHEMICAL TECHNOLOGY DIVISION
D DIRECTORS DIVISION
GEAC GENERAL ENGINEERING AND CONSTRUCTION DIVISION
1&C INSTRUMENTATION AND CONTROLS DIVISION
MAC METALS AND CERAMICS DIVISION
R REACTOR DIVISION .
RC REACTOR CHEMISTRY DIVISION
UT UNIVERSITY OF TENNESSEE

"PART TIME ON MSRP

DUAL CAPACITY

W. R. GRIMES* RC
F. F. BLANKENSHIP ' RC
E. G. BOHLMANN' RC
H. F. McDUFFIE* RC
MSRE ON-SITE CHEMISTRY
R. E. THOMA** ' RC
5. 5. KIRSLIS RC
P. 0. NEUMANN® , RC
IRRADIATION PROGRAM
J. M. BAKER : RC
E. L. COMPERE RC
H. C. SAVAGE ] RC
J. E.SAVOLAINEN |
J. R.HART ] RC
B. L. JOHNSON RC
J. W. MYERS ’ RC
R. M. WALLER | RC
r
PHYSICAL AND INORGANIC CHEMISTRY
€. F. BAES, JR.* ! RC
C.J. BARTON* ‘ RC
3. BRAUNSTEIN®, RC
G. D. BRUNTON* RC
S. CANTOR® : RC
J. H. SHAFFER* [ RC
R. E. THOMA** ‘ RC
C. F. WEAVER* ! RC
C.E.L. BAMBERGER" RC
F. A. DOSS RC
H. F. FRIEDMAN* RC
L. 0. GILPATRICK* RC
D. M. MOULTON : RC
J. D. REDMAN® RC
K. A. ROMBERGER* RC
* D. R. SEARS* . RC
H.H. STONE* RC
W.K.R. FINNELL . RC
B. F. HITCH RC
C. €. ROBERTS RC
W. P. TEICHERT RC
|
I
b
ANALYTICAL CHEMISTRY
L. T. CORBIN* AC
1. C. WHITE* AC
A.S. MEYER* . AC
R. F. APPLE ‘ AC
C. M. BOYD* AC
J. M. DALE AC
F.K. HEACKER AC
C. E. LAMB* AC
W. L. MADDOX* AC
G. MAMANTOY utT
D. L. MANNING® AC
J. P,

|
!
!
YOUNG* I AC
}
|
i

———— .

e

N

!lf
-

VENOUL AN~

O ErE MO EFUVUOONMEIAMECOOMP I NITUMIVIETON OO OO

. Adams

. Adamson
Affel

. Alexander
. Apple
Baes
Baker
Ball -

. Barthold
.. Bauman

. Beall

MmMUo-TNMOoOTX

. Bender -
. S. Bettis
. 5. Billington

E. Blanco

. F. Blankenship

O. Blomeke

. Blumberg

. L. Boch

. G. Bohlmann
. J. Borkowski
. E. Boyd

Braunstein

. Bredig

. Breeding
. Briggs .
. Bronstein
. Browning
. Bruce

. Brunton-

. Burger

. Canonico
Cantor

PTOAOMAODS- >

. W. Cardwell
. L. Carter

|. Cathers -
M. Chandler

. L. Compere

A. Conlin

. H. Cook

T. Corbin

. B. Cottrell
. A, Cristy

217

INTERNAL DISTRIBUTION

PAVLSIME-ATINOVOADVDIANPEINIZEOSIPIDEME-EZAN =0T -

ORNL-4119

UC-80 — Reactor Technology

NMEDVDOETONAFTMOPZVEOIP>PIIOAINMOPAIIP>OEZTMA DI OMOS-I=Ir T

. Crowley
. Culler

Dale

. Davis
. Davis

DeVan
Ditto

. Donnelly
. Dunwoody
. Eatherly

Engel
Epler
Ergen
Ferguson

. Ferris

Fraas

. Friedman

Frye, Jr.

. Gabbard

Gall

. Gallaher-
. Gilliland
. Goeller:

. Grimes
.‘Grindell

. Guymon

. Hammond
. Hannaford

Harley

. Harman

Harrill

. Haubenreich
. Heddleson
. Herndon

. Hibbs (Y-12)

Hightower

. Hill

Hise

. Hoffman -
. Holt

. Holz

. Hor'ron

. Householder
<r

101.
102,
103.
104-108.
109.
110.
111,
112,
113.
114.
115.
116.
117.
118.
119.
120.
121.
122,
123.
124.
125.
126.
127.
128.
129,
130.
131.
132.
133.
134.
135.
136.
137.
138.
139.
140.
g 141.
142.
143.
144,
145.
146.
147,
148.
149,
150.
151.
152.
153.
154-155.
156.
157.

-—

VA ErMEC-ADEOMPIOONITIAOEONAIANIACP>AAMOLCOCP>PO0NDIIAONZTEOVET

TTMEBPAOCrIONrFrA>OAYME-EXTMMEMBDOOMOZ- IO Pr-mMPmMEI-YTTAEZEV-A-20I

. L

. Hudson

. Inouye

. Jordan

. Kasten
Kedl

. Kelley
Kelly

. Kennedy
. Kerlin
. Kerr

. Kimball
Kirslis
Knowles

Krakoviak

. Krewson

. Lamb
Lane

. Larson
Lawrence
. Lincoln

. Lindauer
. Litman

. Liverman
. Livingston

Lundin

Lyon

. MacPherson
. MacPherson
. Maienschein
. Martin

. Martin

. Mathews

. McClung

. McCoy

. McDuffie

. McGlothlan
McHargue

. McNeese
Meyer

. Miller

. Mills

. 'Mixon

. Moore

. Morgan

. Moyers

Nichols

. Nicholson -
. Oadkes

. Osborn

. Parker

. Parsly
atriarca

158.
159.
160.
161.
162.
163.
164.
165.
166.
- 167.
168-244.
245.
246.
247.
248.
249.
250.
251.
252.
253.
254.
255.
256.
257.
258.
259.
260.
261.
262.
263.
264.
265.
266.
267.
268.
269.
270.
271.
272.
273.
274.
275.
276.
277.
278.
279.
280,
281.
282.
283.
284,
285.

“CITOO=EP>IFTINIV-WIE>I

ZARE-PPO-OUEIIOODEME-0OION

—

U

C
]
C
. W
F
E
nl
E
H

—E>TVOOMPOETM

. R. Payne
. M. Perry
. B. Pike

. B. Piper
. E. Prince

L. Redford

. M. Reyling
. Richardson

@)

. Robertson
. Roller

. Rosenthal
. Savage

. Savolainen
. Schaffer

. Schilling
ap Scott

. Seagren

. Shaffer

. D. Shipley

. J. Skinner

. M. Slaughter
. N. Smith

. J. Smith

P. Smith
L. Smith
G. Smith

. H. Sneli

. F. Spencer
Spiewak.

. C. Steffy

. E. Stevenson

H. Stone

. A. Sundberg

R. Tallackson

B. Trauger
W. Tucker
C. Ulrich
C. Watkin
M. Watson
S. Watson
F. Weaver

. H. Webster
. M. Weinberg

R. Weir

. J. Werner
. W. West
. E. Whatley

C. White

)
&_v‘

290

341,
342,
343.
344.
345.
346.
347.
348.
349.
350.
351.
352.
353-354.
355.
356.
357.
358.
359.
360.
361.
362.
363.
- 364,
365.
366.
367-368.
369.
370-647.

286.
287.
288.
289,
-294.

T-D-CMFEMETIITAETNE-ETPO-0D--

—
—
R

| \
G. C. Williams . 295-297. Central Research Library
L. V. Wilson 298-339. Laboratory Records Department

G. J. Young . "~ 340. Laboratory Records, ORNL R.C.
Biology Library .
ORNL —.Y-12 Technical Library

Document Reference Section

EXTERNAL DISTRIBUTION

. C. Bowman, Union Carbide Technical Center, 12900 Snow Road, Parma, Ohio 44130

. H. Brannan, Carbon Products Division, 270 Park Avenue, New York, New York 10017
. A. Charpie, UCC Electronics Division, 270 Park Avenue, New York, New York 10017
. F. Cope, Atomic Energy Commission, RDT Site Office (ORNL)

. W. Crawford, Atomic Energy Commission, Washington 20545

. A. Douglas, Materials Systems Division, UCC, Kokomo, Indiana 46901

. Giambusso, Atomic Energy Commission, Washington 20545

. J. Larkin, Atomic Energy Commission, ORO

A. Lieberman, Atomic Energy Commission, Washington

. D. Manly, Material Systems Division, UCC, 270 Park Avenue, New York, N.Y. ]00]7
. L. Matthews, Atomic Energy Commission, RDT Site Office (ORNL)

B. McDonq|d Battelle-Pacific Northwest Laboratory, Hanford, Washington
W. McIntosh, Atomic Energy Commission, Washington 20545

. A. Rosen, Atomic Energy Commission, Washington 20545

M. Roth, Atomic Energy Commission, ORO
Shaw, Atomic Energy Commission, Washington 20545

. E. Sinclair, Atomic Energy Commission, Washington 20545
. L. Smalley, Atomic Energy Commission, ORO
. D. Stoughton, UCC, P. O. Box 500, Lowrenceburg, Tennessee

A. Swartout, UCC, 270 Park Avenue, New York, New York 10017

. F. Sweek, Atomic Energy Commission, Washington 20545
. W. Ullmann, UCC, P. O. Box 278, Tarrytown, New York 10591
. J. Whitman, Atomic Energy Commission, Washington 20545

James Wright, Westinghouse Electric, P. O. Box 355, Pittsburgh, Pa. 15230
Research and Development Division, AEC, ORO
Reactor Division, ORO

W.
Given distribution as shown in TID-4500 under Reactor Technology category (25 copies — CFSTI)

W. Grigorieff, Assistant to the Executive Director, Oack Ridge Associated Universities